rf-photography.ca

Creating impressionistic is something I find very satisfying and enjoyable about photography. There are several ways to create these types of images to generate a painterly, artistic result.

The making of impressionistic photos can be done with multiple exposures or with a single exposure and with either a film or digital camera. The guideline that follows will show how to do it both with film and digital as well as with both multiple and single exposures.

Film

If you’re using an automatic film camera, it has to have a multiple exposure feature built into it. All of Canon’s EOS SLRs have this feature. Other camera manufacturers included it in some of their models but not all. In some models you can only make 2 shots on the same film frame but in other models you can do up to 9. Canon’s EOS SLRs allow up to 9 – but you can do more and I’ll explain how below.

I use Canon cameras so I’m going to explain the technique as it would be done with an EOS SLR.

The first thing you need to do is compensate for the build up of exposure on the film frame. If you put 9 shots on a frame of film and don’t make any exposure adjustment, you’re going to end up with a very overexposed photo – not what we want. Assuming 100 ISO film, the easiest way to compensate is to adjust the ISO setting in the camera to equal the number of shots you want to make. So if you’re going to take 9, you’d set your film ISO to either 800 or 1000. You’ll need to take the camera out of auto-ISO mode to do this.

Once you’ve got the exposure compensation set, you go into the Multiple Exposure feature (how you access it through the menus is different depending on the camera but the manual will show you) and set the number of shots you want to take. The camera won’t advance the film until the last shot in the sequence has been taken. With the exposure compensation and number of frames set, you’re ready to make your multiple exposure, impressionistic photo. Compose as you would normally. Take the first shot then move the camera slightly in any direction – not much, just a small amount. Make your next exposure and move the camera again. Continue this until the last shot has been taken. It does take practice to figure out how to move the camera and how much to get the result you want. You can also with with a zoom lens and zoom the lens between shots. If you try this, start at the wider end of the zoom and work toward the narrower end. Starting at the wider end allows you to make sure there’s nothing unwanted to distracting in the frame because you’re only going to go narrower from there. With time, you’ll soon get a good feeling for how to go about it.

Remember I said you could do more than 9 shots on a frame with a Canon SLR? The way to do it is to set the number of exposures to 9 like we did above. After you take 8 shots, you can reset the counter and the exposures will continue on the same frame. So, if you set it to 9, take 8 then reset it to 9 you can get 17 shots on a frame. You can do this as many times as you want (within the limits of the ISO setting of the camera, of course). You need to make the appropriate exposure compensation before you start too. If you wanted 17 shots, you’d set your ISO to 1600.

The shot below left is a 17 shot multiple of a small maple tree that had turned colour for autumn. The shot on the right is a 26 exposure multiple of a plot of lupines.

If you’re using a manual film SLR you can still do multiple exposures, it’s just a little more work.

You still need to make the same exposure compensation as with an automatic SLR.

Getting the camera set to make the multiple exposures is where the little extra work comes in.  After each exposure, you’ll need to press the film rewind release button.  This disengages the internal clutch for the film advance mechanism.  While holding the rewind release button, cock the film advance lever.  This will reset the shutter but won’t advance the film.  After the shutter is reset, let go of the rewind release button and you’re ready to take your next shot.  It’s a good idea to cock the film advance lever after letting go of the rewind release button.  This ensures the film is properly tensioned in the camera and that it stays flat.

Digital

The process, with most digital cameras, is a fair bit different.  With the exception of a couple Pentaxes and Nikons, digital SLRs don’t have a multiple exposure feature.  It’s interesting that while Canon was known for including the feature on all of its EOS film SLRs, it has yet to include the capability into any of its DSLRs.

Since, for the most part, you can’t do it in camera, you have to do it with editing software – the Digital Darkroom.  Using a digital camera to generate your working originals is simple.  Shooting in RAW is best because it gives you the most flexibility when editing.  You can either make as many exposures as  you want in camera or you can take one in camera and make multiple copies after the fact.  I prefer to make as many as I want in camera because it takes less time than making copies later.  You don’t even have to move the camera between exposures necessarily because you can position the layers manually later in the Digital Darkroom.

The one thing you will need to make digital impressionistic images is computing power.  RAW files from my 5D, for example, are about 73MB when converted to 16 bit TIFF files.  Layering 16 or 17 of those on top of each other makes for some pretty big working files.  A good computer with at least 2GB of RAM and a sizable amount of free hard drive space to use as a scratch disk will make the work easier and faster.

I’ll show you two ways of creating multiple exposure, impressionistic images digitally.  I encourage you to take these ideas, play with them, change them and see what you come up with.  Please feel free to write me with your methods and results if you feel like sharing them.

The shot on the left is the original.  I made 15 in camera exposures.  The middle photo is done in Photoshop CS3 using layers.  The one on the right is done using HDR software.

To create the effect in Photoshop using layers, open all the images in Adobe Camera RAW (or your RAW editor of choice).  If you make changes to any of the RAW files before converting, you should synchronise the changes to all of the images so you’ve got consistency in your starting exposures.  Convert the RAW files to 16 bit TIFFs and open them in Photoshop.  Other image editing programs may be able to be used as well.

Choose one of the exposures as the base layer – it doesn’t matter which (they’re all the same, remember).  Copy each of the others and paste them onto the base layer (Select>All, Edit>Copy, Edit>Paste or CTRL+A, CTRL+C, CTRL+V – the equivalent to the CRTL key on a Mac is the Option key).  Change the Blend Mode of each layer above the Background to Lighten.  This is the equivalent of a 1/2 stop increase in exposure and it allows the layers below to show through.  You can try changing the Blend Mode of one of the layers (again it doesn’t matter which one) to Soft Light.  This will darken the overall result slightly and give some additional shadow contrast.

Once all the layers have been copied, select the Move tool (keyboard shortcut V) and begin moving each layer where you want it in relation to the others.  You can do this with your mouse or using the arrow keys on your keyboard.  When all the layers are where you want them, either Merge or Flatten all the layers.  You need to do this step in order to do any Cloning or Healing.  If you don’t, then these edits won’t apply properly.  With the layers merged or flattened, make whatever adjustments or edits you feel are necessary to come up with a result you’re happy with.

Another way to do this with layers is by adjusting the layer Opacity.  I don’t personally like this method as well, but have a go with it because you may like what it does for you.

The general rule of thumb is that each layer above the Background should have the Opacity set to 100% divided by the layer number.  So the first layer above the Background (2nd total layer) would be set to 50%, the second layer above the Background to 33% and so on.  When you get 15 or 16 layers stacked on top of each other the opacities of the layers at the top of the stack are pretty low but that’s OK.
A third way to create these digitally is with HDR software.  HDR stands for High Dynamic Range.  The software is typically used to automatically blend images with different exposure settings to create a result that has a greater brightness range than the sensor can capture naturally.

I decided to try using the software to create impressionistic images because these programs are purpose-built to merge a number of exposures together.  For my own purposes, I like a software called Dynamic Photo HDR from a company called MediaChance (they’re a Canadian company too so supporting a ‘local’ business is an added bonus).  The reason I prefer it is that it allows you to manually adjust the placement of each layer before the merge is done.  Other HDR software doesn’t give you this option.  Photoshop CS2 and CS3 have HDR capabilities built into them but the functionality isn’t useful for non-aligned image layers.

Using DPHDR, I load my camera exposures into the software.  The software will take the exposure settings from the EXIF data.  Because the software is designed to work with images at different exposure settings, you have to set the exposure for each image manually.  I typically set each at .01.  With that done, the next step is alignment.  The software will attempt to aling the layers automatically and after doing so you have the opportunity to adjust the alignment manually.  This is where I move each layer around to place all of them where I want them.  The software only allows you to see one layer in relation to the base layer at a time so it takes some trial and error to get the alignment of all the layers where you want it.  After doing a few, you’ll get a sense of where the layers are in relation to each other.  Once that’s complete, you can let the software do the merge and output the HDR file.

HDR files are 32 bit.  To make these useable for other purposes they have to go through a process called tonemapping which redistributes and compresses the 32 bit brightness and colour information into a 16 or 8 bit space (whichever you choose).  DPHDR has tonemapping capabilities.  What I normally do is save the generated HDR file, open it in Photoshop and do the tonemapping there.  I just like the result I get doing it this way but there’s nothing wrong with the tonemapping functions of DPHDR, so try both and see what you like best.

With the HDR file converted to a 16 bit TIFF, I can then do whatever other editing and adjusting necessary to create the final result.

Single Exposures

Impressionistic photos can also be created with film or digital using single exposures.

With a single exposure you’ll need longish shutter speeds.  The longer shutter speeds allow you to move the camera while the shutter is open to create the impressionistic effect.  A number of the shots in the Galleries have been created using this effect.

Shutter speeds in the 1/3 second to 1.5 second range work well.  There are any number of ways you can move the camera.  Longer, smoother, more sweeping movements create softer, more flowing images.  Shorter, more compact movements create a more staggered, less smooth look.  Moving the camera in longer, vertical motions with trees, for example, has the effect of elongating the scene elements.  If you’re using a zoom lens, try rotating the camera while holding the zoom ring during the exposure.  This can create some very pleasing swirl effects.  Also try rotating the camera and lens.  This creates a different type of swirl effect that also can create some interesting photos.

As I experiment with this technique more, I’m really beginning to like it.  It’s the closest I’ve been able to come so far to replicating what I could do on film with multiple exposures.

New Life for Old Film Scans

Several years ago, I was commissioned to do a series of still life images of Calla Lilies.  I still have the film scans on my hard drive and recently decided to see what I could come up with using HDR software and multiple copies of the scan files.  Two of the originals are below.  The impressionistic images are further down.

I made 15 copies of each original (16 in all), loaded the copies into DPHDR, set the exposures, manually moved the layers and then created the HDR file.  Next I opened the HDR file in Photoshop and tonemapped it to a 16 bit TIFF.  After doing this, the edges of the various misaligned layers were visble as you can see in the intermediate shot below.

I created a Gaussian Blur layer mask set to Hide All then painted the blur layer back in along the edges of the flowers.  The two photos below re the end results of the two originals above.

I like the effect of the result.  It’s a softer, more painterly look.  The softness and added volume around the edges gives an enhanced sense of dimensionality and there seems to be a greater sense of flow than in the originals.

This is a follow up to the previous intro to timelapse photography and will discuss using HDR images as the input. I started a group on Vimeo for HDR timelapse and there’s some absolutely stunning work in the collection from some truly talented artists. I’d suggest checking out if you’re looking for some creative inspiration.  HDR timelapse is a very cool sub-genre of timelapse that I was first introduced to a little over a year ago by Walter Arnold of The Digital Mirage.  I haven’t decided yet whether to thank him or curse him for it.    I would really like to see him do some more of this kind of thing though because the clips I saw that got me interested in the technique were very good.

HDR timelapse isn’t any different from regular timelapse once the shooting and HDR merging/tonemapping is done. Where it ups the complexity level is in the shooting. In addition the regular intervalometer, you need an interval timer that will combine with bracketing to get your input images for HDR merging. Shooting Nikon, I’m fortunate that some Nikons have that functionality built in. Other camera makers don’t and getting it can be a bit tricky. The Promote Control mentioned in the previous tutorial will allow you to combine the two. If you’re a Nikon or Canon shooter, Chris Breeze of Breezebrowser fame has developed some software for shooting tethered to a laptop that allows you to combine interval timing and bracketing. The downside is, of course, that you have to be tethered which limits the usefulness of the application in some situations. Canon’s EOS Utility doesn’t offer the combined functionality, nor does shooting tethered via Lightroom v3.x. It’d be very cool if a smartphone app could be developed that would allow you to control the camera via an iPhone or Blackberry (or other smartphone) but still capture to the onboard memory card (sometimes I really do wish I could do programming…… well, no, not really).  When I used to shoot Canon I tried a couple third party intervalometers and while I could get them to combine with in camera AEB, it was kludgy at best.  Some of them top out at 99 shots before having to be reset so if you’re doing a 3 shot bracket, you only get 33 brackets before having to stop to reset the intervalometer.  I don’t believe something like the Canon TC 80N3 can be used to combined interval timing and AEB.

What mode do I shoot in? This is a matter of a fair bit of debate in timelapse circles.  It’s also a bone of contention in non-HDR timelapse but seems to come more into play when HDR is added to the mix.  There are those who believe full manual is the only way to shoot timelapse and particular HDR timelapse.  Their reasoning is that they don’t trust the ‘automatic’ part of modes like Av and feel there’s too much inconsistency in shutter speeds to be reliable.  Personally, I don’t put a lot of stock in that.  I recognise there can be variations but don’t feel they’re significant enough to cause problems.  In addition, I want the shutter speed to be variable to account for changing light conditions.  If I start shooting in cloudy conditions and over time the sun comes out, if I’m shooting in manual I’m going to be overexposing in the brighter conditions.  Shooting in Av, the shutter speed will adjust automatically and help keep my exposures consistent.  The unwanted side effect of shooting in anything but manual, according to some is a thing called ‘flicker’.  Flicker is a visible variation in light levels in the video clip that, in extreme cases can have an almost stroboscopic effect.  While I recognise flicker can and does happen, I’m not convinced the major cause is shooting in a programmed mode like Av.  Light levels change naturally over time.  In timelapse those changes in light level are sped up and can appear as flicker.  In addition, I think the added post-capture processing of merging and tonemapping HDR sequences can have an impact on flicker.  Tonemapping algorithms, even if the settings are the same from set to set, are going to process each set differently.  There are reasons for this.  Again, changing light levels from set to set will cause the tonemapping operators to be applied differently.  In addition, HDR is still a relatively new technology in photography and tonemapping operators are still being improved and refined.  The newness of the technology I think helps contribute to some of the inconsistencies in tonemapping from set to set.  So while I don’t completely dismiss the concept that shooting in manual will help minimise flicker, I’m not convinced that it’s the major contributor some feel it is.  Bottom line, shoot manual or Av, whichever you’re comfortable with.  In addition, if you want to try to capture what the hardcore timelapsers call the ‘Holy Grail’ – the transition from sunrise to daytime or daytime to sunset – you really have to shoot Av because the light levels are changing so much and so quickly.

What format do I shoot in? The same concepts apply here as with regular timelapse.  RAW gives you more flexibilty but takes up more card space.  JPEG takes up less card space but gives you less flexibility in editing so you need to be more careful in setting up the shoot to be sure you’re not cooking in something like a wonky white balance that will affect the rest of the processing later on.  Keep in mind that HDR merging will exaggerate colours and colour casts in source images so if you’re going to shoot JPEG you want to take extra care in setting WB and in setting up the camera processing settings to give you JPEGs that are tamer in terms of saturation and contrast than if you were doing single shots.

What should the spacing between shots be? This can get a little tricky with HDR timelapse.  Because you’re capturing a bracketed set, you need to make sure that the time between sequences is long enough that you can capture the full bracketed sequence before the counter gets to zero for beginning the next sequence.  The reason for this is that most interval timers start counting down automatically when the shutter is opened the first time; not when the shutter closes the last time.  There are some that don’t start counting till the single shot is done or the seqence of shots is done but most don’t.  You need to build your total shutter speed for the sequence into your spacing between sequences as a result.  So, if the total shutter speed for a 5 shot sequence is 8 seconds, you can’t have the interval timer set for 6 seconds.  It needs to be 9 or longer.  The other thing to consider is if you are going for the Holy Grail and going from light to dark (daylight to sunset) your shutter speeds are going to get longer over time.  You want to build some slack into your timing so that you’re not having to reset the timer all the time.  Use an aperture that’s a stop or two wider than you might normally or an ISO setting that’s higher than you normally would to give you that working time.  Depending on how dark you’re going to go, you may need to stop the camera and reset the sequence but you just don’t want to be doing it all the time.  When you touch the camera to reset the sequence (not a problem if you’re using a remote), there will be some jostling and pronounced shake.  You can use this to your advantage.  You can use it to adjust the focal length or camera position include a transition or two in the sequence which can add visual interest.  That’s what I did in this particular sunset clip.

How long to I need to shoot for? Same methodology applies as in regular timelapse.

How many shots do I need to take? The same math applies as in regular timelapse with one exception.  You need to multiply the number of shots by the number you’re capturing in your bracket.  So if I need 240 shots for a 10 second clip and I’m taking a 5 shot bracket, I need to actually shoot 1200 images.  Card space becomes vital when doing HDR timelapse because of the huge numbers of shots being captured.

OK, I’ve got my shots, now what do I do? The first step, obviously, is to process and tonemap your HDR images.  Ideally you’ll want to use an HDR software application with batching capability.  Also ideally you want to use an application that will allow you to save a preset and apply that preset to tonemap the images in batch mode.  Take one sequence, merge and tonemap it manually.  When you’ve got the settings where you want them for the look you want in the resulting video clip, save those settings as a preset.  In batch mode, have the software save the intermediate 32 bit images – this way if you don’t like the tonemapping results you can come back and reprocess them with different settings – and have it apply your tonemapping preset and save the tonemapped images.  Some HDR apps will let you choose a custom name followed by a numerical suffix.  This will let you get your sequencing done at the same time as you process your HDR images.

From there, the steps are the same as with regular timelapse.  Import the sequence into your software of choice and export the video clip.

In the next instalment, I’ll work through a sequence from start to finish in a free piece of software called Virtualdub (Windows only).  Vdub has some nice plugins that can help you deal with flicker as well as camera movement such as wind buffeting.

This is the last instalment in the series of HDR software reviews. If a new application comes along in the future, I may add it to the set and I’m in the process of looking at some new beta versions of a couple previous applications so will likely update those reviews in the future with new information; but for all intents and purposes this is it. What started out as a list of about 5 applications has grown to a 10 part series. From a purely personal standpoint, if I never see the images that have been used in all these reviews again it’ll be too soon. I hope the thoughts I’ve laid out have been useful for some people and perhaps given some insight from the standpoint of a casual user.

This last instalment will look at Picturenaut. I was initially reluctant to include it because unlike all the other applications in the series, Picturenaut isn’t a commercial product. My feeling is that if someone’s offering a piece of software for sale, it’s open to be reviewed and critiqued but if someone’s offering up a piece of software for free it’s a different matter.

Picturenaut can be downloaded from the HDRLabs website which, as many of you probably know, is owned and maintained by Christian Bloch. Christian is well known in HDR circles and is the author of The HDRI Handbook which is highly recommended as one of the two seminal reads for users of HDR both new and advanced along with Practical HDRI.

In a comment to another one of the reviews in the series, it was mentioned that I should include Picturenaut. I contacted Christian to see if he’d mind (essentially get his permission). He had no problem with it with one caveat. He asked that I relate the story behind Picturenaut. So here goes.

Picturenaut ‘fell out of’ The HDRI Handbook. The book is essentially Christian’s diploma thesis. In the early stages of trying to find a publisher, he was turned down so put the thesis onto the web for anyone to download (it was downloaded over 20,000 times). This is back in 2003 when HDRI wasn’t being used anywhere in photography really but only in the motion picture industry (it’s been used there for several years in CGI). A programmer named Marc Mehl contacted Christian with an interest in trying to develop a program for HDRI. Working together, with Christian helping Marc find and refine the process and the ‘under the hood’ algorithms, Picturenaut was produced. Marc continues to update Picturenaut today. Donations for maintenance and development of the program can be made through the HDRLabs website and all donations go to Marc for continued development. Download it, give it a try and if you like it (you probably will), make a donation (and yes, I have). The other thing that’s important to note about Picturenaut is that it’s got an Open SDK which means that anyone with a knowledge of C++ can develop plugins for it. If you know C++ and you want to help improve the software, have at it.  An Open SDK is what allows users to create their own actions and scripts in Photoshop.

Now, with the story behind the software known to all who read, let’s take a look at the application itself. I have to say I’m more than a little nervous about this one because I know Christian’s going to be reading it at some point and no doubt my mistakes will be pointed out when he comments (go easy on my Christian, I’m just a dumb photographer).

The GUI for Picturenaut is about as clean as they come.  A menu bar at the top with both dropdowns and icons to launch actions.

There are no preferences that can be set up so you can pretty much start using the program straight away.  It will open RAW files, it also supports Radiance, TIFF and OpenEXR 32 bit files.  You load TIFF or JPEG images as your source images as well.  There is no batch functionality; which would be a nice future development to see.

To load a bracketed set of images, either click on the icon of the images with the + sign or go to File>Generate HDRI.  A dialogue box opens to select the images.

You’ve got a few options here to consider.  There are the usual Align and Deghost.  Exposure correction is an option that will allow the software to adjust the exposure of the image from the camera as may be necessary to properly compute a gamma curve before averaging the pixel values together.  For example, sometimes you’ll look at the EXIF of your images and your -3 shot might show as -2.97.  Picturenaut will make the necessary correction to properly average the pixel values together if the Exposure Correction option is checked.  The result should be more accurate blending of the source images.  Color Balancing is another option to consider.  Based on the information from the EXIF data, slight differences in colour can occur due to an inexactness in the exposure information as noted above.  Color Balancing will try to correct those errors and match colours in the 32 bit image to the colours in the source files (i.e., prevent false colour or dramatic colour shifts).  If Exposure Compensation is left checked, Color Balancing shouldn’t be necessary.  Next you have a dropdown for Weighting.  This determines how the pixel values in the source images will be used.  Standard uses all the pixel values in each image.  If particular areas of a source image are noisy (i.e., underexposed shadows), noise in the final image could be enhanced.  Mid Emphasis uses primarily the middle-toned values in each image.  This can work well if you’ve fully exposed a bracketed series and have your darkest shadows brought up to a middle tone and your brightest highlights brought down to a middle tone.  If your shadows are darker or highlights brighter, using Mid Emphasis may sacrifice some shadow or highlight detail in the merged image.  The last choice is for Curve.  This is the gamma curve of the source images.  If you’re using RAW files, selecting Linear from the dropdown should be the way to go.  If you’re using TIFF or JPEG images as your source, selecting Computed would be best which will allow the software to reverse engineer the gamma curve and properly correct for it to linearize the data for merging.  Selecting Standard Curve or User Defined should only be used if you know exactly what gamma encoding is used in the source images.

Once you load the source images, you don’t have to wait long to get your merged image on screen.  This is the fastest application of all those tested thus far without question.  To get into the tonemapping, either click the icon at the top with the little sliders on it or go to Image>Tonemapping.  The tonemapping screen is fairly simply laid out as well.

You can make the preview image full screen and drag the tonemap operators to a second display if you have one.  Moving to a second display is not ‘sticky’ so has to be done each time.  More fulsome dual monitor support would be nice.  In the preview window, scrolling with the mouse wheel will zoom in and out within the image.

There are 4 tonemap operators to choose from.  The one that will likely get used the most is Adaptive Logarithmic.  The other that may get used some is Photoreceptor.  The other two won’t be of use to most folks.  Adaptive Logarithmic preserves colour well and applies a soft shoulder and foot to the contrast curve (film photographers will understand the concept).  For more digitally based photographers, think of the PS Curves adjustment with a short, soft curve at the top and bottom with a long straight section in the middle, or with a very shallow S-shape in the middle.  That top and bottom curve are the shoulder and foot.  Photoreceptor applies a more dramatic straight cut off at the top and bottom which will give more contrast and ‘punch’ to shadows and highlights but doesn’t do anything for the midtones which will tend to be flatter.  Colour preservation isn’t as good and I found I’d get a fairly pronounced green/cyan colour cast to images using Photoreceptor.

Sliders for the various adjustments are live (big thumbs up).  In addition to the adjustment sliders, there are options for adjusting the histogram via a Levels type adjustment and the gamma.  Unchecking Automatic Luminance gives you control of the highlight and shadow points on the histogram.  You can adjust overall Luminance or each channel individually.  I’d recomment leaving Automatic Contrast unchecked.  In the Gamma tab, you can adjust the gamma curve of the tonemap function to better suit your needs, your desired look or your source images.  To do this, you can adjust the sliders, select from the dropdown menu or, if you have gamma curves loaded on your system, use one of those (not many users are going to have gamma curve files of their own).  Generally you’ll start with the default 2.2 and adjust from there or stick with the Computed option in the dropdown.  You don’t really want to use Linear here because your output image is no longer longer comprised of linear luminance values like your RAW input files would be.  You want to apply a gamma correction to your output LDR file.  The Rec 709 checkbox is a setting you won’t find in other HDR software.  This one was a mystery to me so I did a little digging.  As far as I can find, it’s a gamma correction function for HDTV that is different from the still image gamma curves we’re used to using.  In general, checking it darkens the image overall, increases colour saturatoin and boots local contrast.  Check it and uncheck it, see if you like what it gives you.  You can choose the bit depth of your output image as 8, 16 or 32.  You can save tonemapping settings as presets so they can be recalled in the future (nice!).  If you switch between tonemap operators, the last settings you used are retained when you return to a previous operator.  The big thing I’d like to see added is the ability to tag the output files with a colour space.  If you use JPEG or TIFF images with a colour space embedded, Picturenaut will honour that colour space through the merging, tonemapping and saving stages.  If you use RAW files; however, your output LDR image will be untagged.  This can result in fairly dramatic differences in colour from what you see in the tonemap preview when the image is opened in an image editor for further post-tonemap tweaking.

How’s it all work?  Let’s take a look.  The first thing to note is that you’re not going to get the really whacked out, hyper-real results with Picturenaut that you can get with other HDR software.  Picturenaut is intended for real world use where realistic results are expected and desired.  Given that, there won’t be any of the surreal looks shown with previous parts in the series.

The image below is the result of tonemapping the default existing 32 bit file.  Overall contrast is good.  Highlights are very well controlled and there’s good shadow contrast.  Picturenaut crunches dynamic range as well as or better than the other applications reviewed previously.  What should be noted is the blue in the water.  This wasn’t present in the tonemapping preview in Picturenaut and is a result of not being able to tag the output LDR with a colour space.  The blue can certainly be removed with further editing and if TIFF files are used as the inputs, the water retains the white colour it had in the tonemapping preview.  It would take very little additional work to get this to where I wanted it.

Let’s take a look at the merge within Picturenaut.

It’s a tad dark but I could have produced a bit lighter version during tonemapping if I’d wanted.  The blue in the windows on the right isn’t as strong as with some other software but that can be enhanced with further editing.  The colour in the floor and walls is good.  The brightness in the windows on the left is well controlled.  Again, the software has crunched the dynamic range very well.  With a little more work in PS and/or LR, a very useable result could be achieved.

Picturenaut also has a deghosting function.  Turning it on doesn’t slow down processing very much at all.  As can be seen from the image below, it works fairly well.  Not as well as CS5 HDR Pro; which at this point has the best deghosting of anything out there, but probably pretty close to Photomatix.

Overall, Picturenaut is a terrific program.  That much better considering the cost.  It’s fast, it compresses dynamic range well, it can generate extremely good, realistic output.  On a commercial interior gig I had earlier this summer, Picturenaut saved my bacon.  I wasn’t getting results I liked with any of the other software at my disposal so decided to try Picturenaut.  Bingo!  Images I could make use of and the client was happy.

Picturenaut definitely goes into the recommended column.  There are a few things that could make it that much better but it’s pretty damn good as it is right now.  As I noted above, if you download it, use it and like it then make a donation which will help keep development going.  We all know there are plenty of photo-related software programs out there that are being sold commercially that are buggy, crash prone and difficult to use that should still be in Alpha development, never mind even Beta.  Picturenaut is the opposite.  It’s a well thought out, well developed piece of software that’s free and should be offered commercially.

As with the others, if you see any glaring or non-glaring errors, let me know and I’ll make the necessary corrections.

I received an email from the folks behind Photo Engine a few weeks ago telling me they were going to be launching a new software package that did HDR and a whole lot more and asking if I’d be interested in being part of the beta group. Sure, I said.

I’ve been playing with the software now for a couple weeks. What follows is essentially a review like others in the series but keep in mind that things could change since this is only the first beta of the program.

I’ll say upfront that Photo Engine is a very complex piece of software. It takes a bit of time to get a feel for it. There’s some terrifically complex coding in the background related to the HDR Relight feature to be sure. In general, while what I’ve seen so far is looking like it’s going to be a very good program, there some complexities and features that are superfluous and attention could have been paid elsewhere. More on that as we progress.  I also think that some of what’s in this software is geared toward CG artists and may not be a lot of use to photographers.

There’s a particular feature of Photo Engine that I haven’t tried yet and that won’t be covered in this preview.  It’s the HDR Relight feature.  I’ll state upfront that I’m a bit skeptical as to the applicability of this for photographers – it may be one of those aspects of the software that CG artists will use more – but will keep an open mind and reserve final judgement till I’ve had the chance to try it out.

On opening the software, you’ll see a GUI that has a screen with a lot going on.  There are three basic components of the app. as shows in the upper right corner – Browse, Edit and Help.  Browse is the default window.  This is where you select your files to work on.  A screenshot is below.

For file types the application can make use of, you’ll see a picture thumbnail.  For others, you’ll get a text thumbnail.  Down in the bottom section are a list of recent Photo Engine projects (using the proprietary .rcd file type) which have been worked on.

You’ll see in the screenshot above that .hdr files don’t have a picture thumbnail.  While the program can read Radiance file types, it can’t display the thumbnail.  OpenEXR files won’t be listed at all because the program can’t work with those.  32 bit TIFF will have a thumbnail with an exclamation point because while it can read and open ‘normal’ 8 or 16 bit TIFFs, it can’t work with 32 bit TIFFs.  Nor can it work with PSD files at all.  I’d suggest the lack of support for a wider range of 32 bit files is a negative for Photo Engine.  While we’re at it, while Photo Engine can read 32 bit Radiance files, it can only write 32 bit files in its proprietary .rcd format.  This makes Photo Engine essentially incompatible with other HDR or image editing applications on the market.  While I understand that the folks at Oloneo are trying to produce a ‘one stop shop’ software application, not building in cross-platform functionality is a big negative.

Opening a single file to work on is as simple as double clicking.  Opening a bracketed series to merge and tonemap is a bit different.  First you select the respective files in the browser using Shift + click or CTRL + click.  Next, in the window on the upper right titled Project Image Selection you click Add.  This moves all the selected files into the project window where you can work with them further.  To begin creating a high dynamic range document, move down into the HDR Tonemap window, select Auto Align or not, then click Create HDR Tonemap Project.  If you’ve added only the images you need for a single merge, there’s no need to highlight the images in the Project window.  If you’ve added images for more than one merge, you’ll need to highlight the ones you want to include before clicking the Create HDR Tonemap Project button otherwise all the images in the Project window will be used.

You’ll notice in the top of the Browse screen in the middle is the typical … icon to open a location and select files.  You can also use the dropdown menu and select by file or by folder.  As you open and work with various folder locations, these will be stored in this dropdown menu and you can select from a recent location quickly without having to go through the folder hierarchy again.  If you’re working with RAW files, the program will give you thumbnail previews, but these do take a bit of time to appear.

Once the files begin to merge, you’re taken to the Edit window.  You can switch back and forth between Edit and Browse without affecting the current merge.  In the Edit window, your tonemapping controls are on the right and your edit history is on the left as seen in the screenshot below.

There is no multiple monitor support.  While the image on screen is fairly large if you’re using a larger monitor, it would become relatively smaller on a smaller monitor.  Adding multiple monitor support so that the edit and history windows could be moved to a second screen would be beneficial.  Similarly, in the Browse screen, having all the supporting windows on a second monitor would be a good idea.

Based on my work thus far with Photo Engine, I’d suggest the Advanced Local Tonemapper is the way to go.  The difference between it and the regular Local Tonemapper is the addition of the Detail controls which can prove useful.  There are also Auto Tonemapper and Globabl Tonemapper options but these provide little user control and less than pleasing results generally.

The variety of controls available in the Edit window is quite extensive.  At the top are the usual tonemapping controls.  Moving down there are LDR adjustments, a white balance panel, print toning and below that very fulsom colour controls.

In addition to the white balance panel on the right, at the top there is an eyedropper which can also be used to select white balance.  I find the eyedropper to be quick and effective.  Activate it then click in the image on a white, black or neutral tone to set your WB.

The Photographic Print Toning panel is one of those aspects of the software that, to me, seems superfluous.  While I understand Oloneo is trying to give users a one-stop shop for image editing, I think it highly unlikely that advanced users are going to do print toning in this application.  Particularly when there’s no print module in the app.  If I want to tone an image for printing, I’m going to do it in Photoshop where I have significantly more control over the final outcome.

Below the print toning, there’s an Advanced section.  Here are where the colour controls are.  Two curve adjustments are available – Brightness and Saturation.  If you right click on the curve in either panel you’ll be presented with 3 options.  Bezier Spline, Catmull-Rom Spline and Linear.  What the ……?!?!  Here again, is where the developers may have gone a bit overboard.  None of these, on first blush, looks like the curve adjustment we’re used to in PS.  Ignore Linear.  It produces a non-smooth curve that is nothing like what we’re used to.  Catmull-Rom looks more like the standard PS curve adjustment but the interpolation between points is different.  That leaves us with Bezier Spline.  And this is the one that will work like the PS curve photographers are used to.  You can add points by clicking on the line and dragging.  What about the tangents on the curve?  These will adjust the slope of the curve locally.  Might be useful from time to time.  If all you want to do is get the ‘standard’ S-curve, click on the tangent arrow of the top point and drag it up.  Next click on the tangent arrow of the lower point and drag it down.  Voila, your standard S-curve contrast adjustment.  If you click to add points on the upper and lower sections of the curve and drag these, you’ll get something similar to the standard PS curve but not quite the same.  In this case you’ll probably want to move the upper and lower tangent points as well to make the curve smoother.  Lengthening or shortening the tangent lines will change the inflection point of the curve.

As noted above, this level of adjustability is more than most photographers are going to want.  Keeping in mind my earlier thought that this software is meant for both photographers and CG artists and doing some research into these various curve types, it does seem that these curves will be more familiar to those working in the CG world.

The Brightness curve works like the PS curve in the Luminance blend mode; affecting brightness without colour.  The Saturation curve below that begins making colour adjustments.  The bottom section of the curve works on areas of lower saturation while the upper section of the curve works on areas of higher saturation.  Dragging down or up will reduce or increase relative saturation levels respectively.

Below that are individual adjustments for Hue, Saturation and Brightness (Luminance).  The spectrum for each is split with a line for each colour/hue.  Clicking and dragging the point on the line adjusts saturation, brightness or hue for that particular colour, isolated from the rest.  Right clicking on a point and selecting ‘Free Mode’ allows you to move the position of the colour line left or right along the spectrum, effectively changing the relative relationships between the various colours.  Perhaps another bit of superfluousness.  While an interesting adjustment to have available, I’m not sure how much photographers are going to use it.  The adjustments that these controls make are precise; however, so the controls are effective.

The last thing to mention is the history panel on the left.  Like the history panel in PS or LR, a record is kept of each adjustment you make.  You can undo one thing at a time or several.  Like in LR (unlike in PS), if you save the file in the proprietary .rcd format, the edit history is stored as well so when you open the file in the future, you have access to everything you did previously.  This is only true if you save in the .rcd file format; however. Right clicking on a history point will allow you to add a comment or edit an existing comment.  Perhaps useful if you want to recall why you used a certain setting in the past.  You can also create an edit version (similar to the LR Snapshot) which you can come back to in the future.  If you create an edit version, then back up in the history to change something, you lose your version.  This makes the version feature less useful.  Versions should be retained so they can be brought back at any time (like the LR Snapshot).  The Play button steps through all the edit history from start to finish automatically.  This would be useful for creating tutorials.  You can have the playback stop each time a comment is found as well so the comments can be used to explain certain steps or processes.

Photo Engine is a colour managed application.  When you save a file you’ll be presented with a dialogue to tag it with a colour space (sRGB, AdobeRGB or ProPhotoRGB), a bit depth (if saving as a TIFF), a compression type for TIFF and quality level for JPEG and a resolution.  If you want, you can also add copyright information into the file. The program defaults to the .rcd file type for saving.

There are no text help files installed with the program, nor are there any on the Oloneo website.  The help comes in the form of video tutorials.  While a nice supplementary form of help, a text help file where people could search and get quick assistance would be beneficial.

There are no preferences to be set in Photo Engine.  Rather than having to choose what file type, bit depth, colour profile, etc, each time a file is saved, it would be preferable to have these established as user defaults which can be overridden if desired.  This would speed the workflow process.  In addition, there is no possibility; that I can find, to save tonemap settings as presets.  Particularly given the varied and complex adjustments available, the ability to establish presets would be beneficial.

How’s it all work?  Examples are below but generally, pretty well.  Overall, speed is good and comparable with the other top programs on the market.

First the realistic result of the existing HDR image.

Colour and contrast are good.  The blue in the water is well controlled.  Overall, a very good result and one that wouldn’t need a lot of additional work.

Next the surreal result on the existing HDR image.

It’s clear that Photo Engine can go from mild to wild.  This is actually one of the more appealing results on this image of all the ones tried thus far.

Now, how does Photo Engine do with files merged inside the application?

The realistic version.

The blue colour in the windows on the right is good.  Overall colour is good.  The result is a bit dark overall but this can be corrected with some more work in PS post-tonemap.

Next the surreal result.

Once again, we’ve gone from mild to wild.  Unlike in some other applications, the blue in the right side windows has been retained to a large degree.  The windows on the left are blown out, as in other software but in a different way and a good deal of the dirt on the windows has been retained.

In both cases, Photo Engine has handled the dynamic range in the images very well.  The surreal/grunge results are different from others and in a lot of respects better.

Overall, Photo Engine shows a great deal of promise.  There are some user-functionality issues that would make it better and easier to use.  The fact that, I think, they’re targeting both photographers and CG artists means there are aspects of the software that aren’t as relevant to photographers but that can be dealth with.  There is no version, at this point, for the Mac OS; however, Oloneo states that the software works well with Parallels Desktop 5 (which I believe is the current version at the time of writing).  There’s no deghosting function which is why that wasn’t tested.  The one other thing that Photo Engine doesn’t have that would be nice to see is a batch function.

Once I try the HDR Relight functionality, I’ll come back and provide an update.  If  you’ve been trying out Photo Engine and find an area where I’ve made a mistake in my commentary, please let me know, I’ll take another look and upate as required.

HDR Darkroom is another relative newcomer to the HDR software scene. I say relative newcomer because it came to market after many of the well known apps like Photomatix, Dynamic Photo HDR and FDRTools.  I only became aware of it in the last couple months and decided to add it to the growing list of software apps in this series.

I should state upfront that after sending some questions to the folks who produce HDR Darkroom, they’ve provided me with a full version (sans watermark) and in return, I’m going to provide them with some samples to use on their web gallery at no cost.

Unlike many of the other applications coming out on the market, HDRDR is ‘just’ an HDR application.  It’s not trying to be all things to all people, which is a plus in my book.  They’re concentrating on one thing with HDRDR and concentrating on doing it well.  32 bit files can be saved either as Radiance (.hdr) or OpenEXR (.exr).  When saving JPEG files, there are no quality options.  The file is saved at full size and full quality.  Personally, I prefer this.

On opening the program, the GUI is clean and simple with a menu bar at the top and menu icons down the left.  This allows you to use whatever method you’re comfortable with to work.  Positioning your mouse over an icon brings up a description of what it is so you’re not working blind if you use the icons.

There are no preferences available to be set so you can simply start working.  There is a batch function (thumbs up) but no colour management implementation (thumbs down).  Under the Process menu is an interesting item labeled ‘Export to Photoshop’.  Upon further inspection, it’s not quite as interesting as first thought as it only will export a BMP version of the file you’re working on.  If it could be reprogrammed to export a JPEG or TIFF, that’d be cool.  As it is, the standard workflow of tonemapping, saving the tonemapped image out as a JPEG or TIFF then opening that file in PS for any further work is the way to go.  There is no deghosting function so that won’t be tested.

How’s it all work?  Opening an existing 32 bit file is pretty quick.  The program goes through an initial ‘tonemapping’ step in order to display the image on screen for a first visual representation.  If you like what you see, you can save out the file from here.  Additional work is probably going to be needed.  A screen capture of the intial preview of our default existing image is below.

It’s actually not too bad, but it does need more work.  Into the tonemapping functions we go.  Going into the tonemapping functions opens up a new screen, but it’s not a double screen so you can’t flip back and forth – there’s no need to really so it doesn’t matter than you can’t.  In the tonemap screen, the tonemapping controls are on the right side of the screen and the image in the middle.  You can adjust the size of the preview between Small, Medium and Large.  The default is Medium.  Even selecting Large, the entire image fits in screen so you can see everything that’s happeningn with the tonemapping.  There is no multiple monitor support so you can’t pull the tonemapping bar to a second screen and get a bigger image on the primary.  Unless you’re working with a very small screen this should be alright because the tonemap bar doesn’t take up a lot of screen real estate and the workspace is clean otherwise so there’s nothing else to keep you from getting a good idea of what’s happening as you adjust the tonemapping operators.  Multiple monitor support is always a nice thing, but in this case – unlike with some of the other apps we’ve looked at – it’s not vital.

There are three tonemapping operators:  Local Tone Balancer (LTB); Local Tone Enhancer (LTE); and Fast Tone Compressor (FTC).  The first two are local tonemappers, the last is a global operator.  Based on what I’ve experienced in using the software, LTE is going to be the operator of choice in most cases.  It has the most flexibility and produces the best results on a consistent basis.  Unless noted otherwise, LTE is the one I’m using in this commentary.  There is also the ability to save tonemapping presets via the dropdown menu at the top.  A screen shot of the tonemapping GUI is below.

The sliders don’t offer a live preview but the update speed is quick.  Switching from one tonemapping operator to another, then back brings you back to the last settings you used in that operator – nice.  Keeping the Strength setting in the range under 50 is generally where you’re going to want to be for a more naturalistic/realistic look.  Fill Light is one that’ll be used sparingly in most cases.  It does do a good job though of working on the darker/shadow areas yet not brightening the lighter/highlight areas.  The Brightness slider does a nice job of making adjustments to the overall exposure and helping retain/regain shadow/highlight contrast.  Below the Tonemapping Parameters is a section for Post Processing tools.  There are two tabs – Color Balance and B/W Point Clipping.  The Color Balance sliders act like the colour balance tool in Photoshop and can be useful for adjusting the white balance in the image.  The B/W Point Clipping sliders on the second tab are where you’re going to set your white and black points.  Both work well and the B/W sliders provide enough fine control that you can get the overall contrast where you want it.  Eyedroppers for setting white balance as well as white/black points would be nice additions to make adjusting those settings slightly quicker.  Once done, click OK to apply the settings and you’re taken back to the original screen after the settings are applied which is pretty quick.  Tonemapped files can be saved as 8 bit JPEGs, 8 bit TIFFs or 16 bit TIFFs (noted in the dropdown as 48 bit – 16 per channel).  It can also save as PNG and BMP but these will be less useful for photographers.  The realistic version of our default existing file is below.

This is pretty darn good as it stands right now.  The blue in the water is a little too intense but that’s easily fixed after tonemapping.  Without question this is on par with the best apps out there for generating a realistic result on this particular image.  Getting to this result took very little time and required not a lot of playing with the sliders.  There are Undo and Redo icons on the left side of the screen.

An interesting part of the way HDRDR works is that after you tonemap and go back to the main screen, you see the results of the tonemapping on screen but your 32 bit image is still there in the background.  When you save the file out as a JPEG or TIFF, it saves a copy so what you see on screen is still the 32 bit image.  If you then go back into the tonemapping operators, it opens the original 32 bit image back up with the previous tonemapping settings applied but clicking Reset takes you back to the original.  This is a nice workflow process if you want to try different tonemapping adjustments.

Now let’s look at the other end of the tonemapping spectrum.  Two files are below.  The first is a surreal/grunge look with the LTE operator, the second is a surreal/grunge look with the LTB operator.  They’re different.  In both cases the Strength slider is pushed well up past 50, which is where you’re going to want to go if you’re working for the less realistic look and in the LTB version, the Local Lighting slider is also pushed up toward the top end.  My only quibble with the LTB version is what it’s done to the waterfall pools.

HDRDR can swim in both ends of the tonemapping pool.  Getting from a realistic result to a less realistic is as simple as moving the Strength slider.  You’ll likely want to tweak some of the other settings as well but the Strength slider really does control a lot in this software.  Simple is good.  And the speed is very good througout.  As fast or faster than the other applications tested thus far.  Another important thing to note is that the program has handled the full drange of this scene very well.  Recall from the introductory article that I chose this particular scene because of the range of brightness it contained.  There are no presets other than the default settings but creating your own is easily done and with the simplicity of making adjustments to create different looks/versions, creating different presets takes little time.

We’ll move on to see how HDRDR does merging and tonemapping a set of files.  The screen for setting up the merge is very simple.  You can choose to align or not and you can choose to have the exposure information taken from the EXIF data or by guessing.  If you were using film scans or other images that didn’t have exposure information in the EXIF, you could use Guess as a starting point.  Once you choose your images and set the software to work the loading, aligning, merging and initial preview tonemapping happens in a pretty snappy fashion.  Perhaps not quite as fast in this case as Photomatix or CS5 HDR Pro but still quite quickly and faster than many of the other apps. tested thus far.

Once again, the intial preview is a very good starting point as seen in the image below.

The colour in the windows on the right is good.  The brightness in the windows on the left is well controlled.  Not quite the best we’ve seen so far but still very good.  Colours throughout are accurate, if a tad cool which is easily addressed.

Making the tonemapping adjustments and applying the tonemapping settings was a slightly slower with this 32 bit file than the existing one from above.  I’m not sure why that might be.  It could simply be a combination of the settings needed to get the result required more time to process than previously.  The Strength slider for this was a bit higher than on the first image so which would mean more processing behind the scenes so that could explain the slight time difference.  The difference was small enough that it could also have been due to other processes going on in my system at the same time.  More testing would be needed to know for sure.  Either way, it’s still quite quick.  Below is the natural/realistic version.

Here again, the colour in the windows on the right is good.  Brightness in the windows on the left is better than in the preview version.  Good shadow/highlight detail is retained throughout.  The colour from the dirt in the left windows shows up well.  It’s a very nice result and one that I’d only have to do a little extra tweaking to after the fact via a Curves/Levels type adjustment and/or perhaps a slight Shadow/Highlight adjustment.

The grunge/surrealistic version is below.

As with the realistic version, colour is good, detail is good in the highlights.  What’s interesting about this version compared to surrealistic versions from the other applications is the way HDRDR has rendered the highlight areas in the room itself.  The highlights areound the ticket windows, the old newsstand on the left and on the large pillar in the back corner add enhanced visual interest to this version that the others haven’t had to the same extent.  The mapping of the light in HDRDR, in this case anyway, has created a very appealing result.  As with the other images, the Strength slider is the one that largely controls the realistic or surrealistic look and in this case it was all the way at 100%.  In some other apps., moving the equivalent of the HDRDR Strength slider to 100% would cause the windows to become completely blown out but highlights are still well controlled here at the same time an interesting ‘grunge’ result has been generated.

After getting the fully unlocked version, I have to admit that I cheated and used it for more than just this review.  I wanted to see how it would handle some night scenes with wildly varying light levels and light colours.  I used it to shoot a timelapse segment in Toronto.  The last 5 seconds of this clip are images merged and tonemapped in HDRDR.  Tried the same files in Photomatix and liked the HDRDR results better so used them.

A Summer Night in Dundas Square from Robert Fisher on Vimeo.

Support for HDRDR is decent.  There are no Help files installed with the program.  There are some video tutorials on the HDR Darkroom website as well as some FAQ sections.  I emailed some questions to their general support email address and got responses within a business day.  The HDRDR offices are in Europe and I’m in North America so considering the time difference, that response turnaround is very good.  The responses I got were helpful.  My understanding from trading emails with them is that they’re working on revamping the website and I expect it’ll have enhanced information on it when the new site is brought onstream.

I did run into a few bugs when working with the software.  Two are minor and not worth going into detail on.  The third involves a difference in tonemapping when working in Batch mode vs. doing individual merge/tonemap operations.  Again, based on emails I’ve traded with their support group, they have a handle on the problem and expect to have a fix in the near future.

To discuss the Batch mode a bit more for a second, like everything else with HDRDR, the interface is simple and the process is quick.  Running the same set of files through both HDRDR and Photomatix, HDRDR completed the batch process slightly faster than PM.  You can choose to save either the 32 bit file, a tonemapped JPEG file or both.  The 32 bit file is saved in the Radiance format.  Since the software can save individually merged files in both Radiance and OpenEXR formats, giving that flexibility in Batch mode would be preferable.  Being able to save the tonemapped file as either an 8 or 16 bit TIFF file would also be preferable to just offering JPEG.  You have access to the same tonemapping operators in Batch as in standard mode.  You cannot; however, invoke a preset in Batch mode.  If you test out a merge to get your tonemap settings  figured out, you’ll have to write down the settings  and input them manually in Batch.  Being able to use presets here would also be a good enhancement.  You can merge up to 10 images in Batch mode which should be enough.  Nikon offers the widest bracketing at up to 9 in some of its models so unless you’re doing manual bracketing at 1/3 or 1/2 stops and capturing huge numbers of images per sequence (which really isn’t necessary), the ability to merge up to 10 images should suffice.

There is a noise reduction feature in the software which I didn’t try.

All in all, HDRDR is a very nice piece of software that does one thing and does it quite well.  The software is on sale right now for $79.  There are both Mac and Windows versions.  At that price it’s definitely a good deal.  Even at the full price of $99 it’s the same price as Photomatix Pro and in terms of the quality of the software and support, it’s pretty much on par with PM.

As with previous reviews, if you see any errors please let me know and I’ll work to make the corrections.  Thanks for reading and I’m happy to recieve your feedback.

The latest beta of PM 4, b2, came out a few days ago. It’s got some new features and a few tweaks so I’m going to give a bit of a preview-review of some of the new functionality. The folks at HDRSoft know I’m doing this and are fine with it. They’ve asked me not to show any screen shots of the GUI and I won’t. That’s unfortunate though because some of the biggest changes involve the GUI. I will try to describe some of them as best I can.

Part II of this review series discussed the current version of Photomatix.  As I noted there, it’s a terrific piece of software and, until PS CS5 HDR Pro was introduced, has been arguably the ‘gold standard’ of HDR software apps.  With the introduction of the new features in CS5 HDR Pro, there’s (arguably) a bit of a tussle for that title of ‘gold standard’.  Does any of that change with what’s coming in the new version of Photomatix?  Let’s take a look.

There are quite a few changes in the works for the new version vs. v3.x.  Some of the key ones are:

- semi-manual deghosting feature by drawing around area to be deghosted

- showing thumbnails for tonemapping presets at the bottom of the screen when in tonemapping mode

- combined panel for tonemapping and exposure fusion (no need to select one or the other when loading images, no multiple tabs for Details Enhancer & Tone Compressor)

- ability to zoom via the mouse scroll wheel when the cursor is inside the image in tonemapping mode

- improved multi-threading support for RAW demosaicing and fusion/adjustment

- ability to adjust tonemapping sliders with the mouse scroll wheel

- transfer of image metadata to merged/tonemapped image

- improved functionality of alignment based on ‘matching features’

These aren’t all of the proposed changes.  There may be new ones added, some may be discarded as the beta process moves forward to a final release of v4.  Many of the changes are cosmetic.  Some deal with background processing which should improve results and improve speed.

The new preview thumbnails for tonemapping presets show up in a sort of film strip type of panel at the bottom of the screen.  This strip can be moved around but it always seems to be ‘in the way’.  In order to have the image on screen large enough to get a sense of what’s happening with the tonemapping, this preset strip blocks part of the image.  The thumbnails are ‘live’ in that clicking on one will invoke that preset on the main image.  There’s a tab for Built In presets and a tab for My Presets so if you have any favourite tonemapping settings saved, these will be available as well.  The thumbnails are small.  You can get a bit of a sense of what the result will be not a really good sense.  You can also close this strip.  It’ll re-open the next time you open an image for tonemapping.  You can drag it to a second monitor but it’ll come back to the main monitor next time.  I can see how this would be positive for some users.  It strikes me as a less useful feature addition.  Other aspects of the GUI can be turned on or off in Preferences (e.g., Workflow Shortcuts).  This should also be able to be turned on or off as the user wishes.

Perhaps the biggest change to the front end is the addition of the semi-manual deghosting feature.  When loading a set of bracketed files, the Processing Options give you two choices for removing ghosts, the old, fully automated method or the new semi-manual method.  Choosing semi-manual then brings up another screen called Ghosting Selection where you determine the area to be deghosted.  The instructions indicate to drag your mouse over the area to be deghosted.  Almost sounds like it creates a box or selection automatically, right?  It doesn’t though.  It’s like the Lasso tool in Photoshop.  You drag your mouse around the area you want selected.  You don’t have to join the ends.  If you only go part way around, the ends will be joined automatically.  Next you right-click in the area and have the option to select for deghosting or remove selection.  Remove selection is greyed out if you haven’t got the selection finalised yet.  You can make multiple selections using this process.  Once the areas are selected you can click on Preview Deghosting or simply click OK.  There’s a slider at the bottom of the Ghosting Selection screen that has no description.  This adjusts the brightness inside the image to help seeing ghosted areas.  Right-clicking also allows you to select a different image from the bracketed set to use for making the selection.  Choosing the 0, or +1 probably makes the most sense from the standpoint of ability to view areas within the image in most cases.  Make your selection based on your own image sets.  Click OK and wait.  Wait a long time.  It took so long, I wondered whether the program had frozen.  This semi-manual deghosting feature is very slow.  Significantly slower than the fully automated deghosting process.  How well does it work?  See the image below for a comparison of the new semi-manual method, the PM fully automated method from the current version and HDR Pro.

It works pretty well.  There’s a small bit at the edge of the left side that didn’t get taken care of but that did in the automated process of the current version.  Other than that, it’s about on par with the old method.  Given the time it takes to process using this new method I’d suggest that there’s not much of a benefit to it.

The ability to adjust the sliders with the mouse scroll wheel is a nice feature.  Click on the adjustment point of any of the tonemapping operators then scroll with the mouse and the adjustment is effected.  There’s a downside though.  With all of the tonemapping operators open, the tonemapping panel is longer than my screen is high.  My screen is 1050 pixels.  I’m guessing the tonemapping panel when fully open is about 1200 pixels or a bit less.  Activating a tonemapping adjustment doesn’t decouple the scroll wheel from the tonemapping panel so as you’re adjusting a particular tonemapping operator you’re also scrolling in the panel and can’t see what you’re adjusting because the operators keep moving up and down.  I’m hoping this is merely a bug and gets fixed before the final version.

I’ve not done a comparison of the change to the alignment based on matching features between the current version and the new version.  I’ve not found the current alignment process to be lacking.  In terms of speed, the converting of RAW files is much quicker than in the current version.  Excluding the semi-manual deghosting, other aspects of the application – loading a 32 bit image, alignment, saving images, tonemapping adjustments, are also quicker.  The tonemapping adjustment sliders are live in the new version where they aren’t in the current version.  Increasing speed is a positive.

I also took a look at the tonemapping to see if there were any differences between the current version and the new version.  There’s one difference in the Default tonemapping settings in Details Enhancer.  The White Point is at .175% in the current version and .25% in the new version.  Adjusting the WP back to .175% shows no difference in the two images after processing the tonemapping settings.

That’s a quick overview of some of the new features and comparison with the current version.  Will it be worth the cost to upgrade?  That’s a decision each user will have to make on their own.  For me, as it sits right now, probably not.  Many of the enhancements are cosmetic.  The functional changes that matter to me are an uptick in speed but that speed increase isn’t significant enough in itself to justify the upgrade.  The new semi-manual deghosting isn’t something I could find myself using given the slowness and that there’s no benefit over the automated approach – at least in the one image set I used.  Even if there were a better result in other images, the slowness is still a factor.

Based on this beta, I’d have to say that HDRSoft hasn’t definitively regained it’s spot as the ‘gold standard’ in HDR software.

EDIT:  In the 3rd beta of v4, the option was included to close the tonemapping preview thumbnails panel.  Definitely a welcome change.

UPDATE:  August 9, 2010 – SNS-HDR continues to be improved.  The latest (v1.2) has improved speed over previous versions.  While still not quite as fast as some others, it’s now fast enough that speed isn’t a concern.  Also, something I had overlooked previously is that at the bottom of the tonemapping panel, to the right of the colour profile dropdown is a small monitor icon.  Clicking on this brings up a dialogue box to select the appropriate display profile.  This is only necessary in WinXP due to limitations on colour management support in that OS.  Later versions of Windows don’t need to do this.

SNS-HDR is a relative newcomer to the game. Since writing the introductory piece for this review series, checking the search stats for my site, SNS has been the single most searched for term that brought people to my blog. It’s obviously getting a fair bit of attention. Based on what I can see tracking a few IP addresses, most of the attention is coming from Europe but it’s likely that the buzz will move across the pond in due course.  The website is in Polish but there’s a Google Translate dropdown in the upper right that will convert to English (or other languages).

I should state upfront that, after writing an email to Sebastian Nibisz with a question, he looked at my blog, saw the review series and sent me a license key to unlock a few features that aren’t availabe in the demo version. Those features are full access to the batch processing mode and the ability to save files at full size and bit depth.

There are 3 versions of the software:  Pro; Home; and Lite.  Pro is the full featured version.  Home has the GUI of Pro but without the batch function.  Lite has no GUI or other editing functions and you drag and drop files onto the icon to activate the merging/tonemapping functions which are fully automated.

I became aware of SNS via a board posting on The Luminous Landscape forums.  Downloaded it, checked it out and decided to add it to the review series.

SNS works differently from other applications.  It’s more automated for one.  It’s not true HDR in that it doesn’t create a 32 bit file; but rather is an exposure fusion tool.  It loads a set of images, merges then goes through an automated step called tonemapping.  It’s a process to blend and distribute brightness for an optimal result.  There’s no menu bar at the top of the screen either, which takes a bit of getting used to.  The GUI is simple and clean, as the screen shot below shows.

While it doesn’t create 32 bit files in it’s exposure merging process, it can read and tonemap 32 bit HDR and EXR files.  Once the program is open, all your options are contained in the panel on the right of the screen.  This is fixed in place and the program doesn’t support multiple monitors.  The three icons at the left of the Image panel allow you to open a single or multiple files, save an image or open batch processing.  Below that you have a histogram.  Next you get into your tonemapping adjustment operators.  There are 4 presets – Dramatic, Natural, Night and Portrait.  We’ll get into these a little later.  Below that you have some of the usual suspects for tonemapping operators.  In addition to the Brightness slider which is a global adjustment, you can adjust the brightness of any individual colour by clicking on the line in the color swatch and dragging upward or downward.  Each click creates a control point which can be dragged.  To delete a control point, simply right-click on it.  Next are Contrast, Highlights, Highlight Details, Shadows, Midtone Contrast and Microcontrast.  Below this is the Color panel with Saturation and Hue controls.  Like with Brightness, there are global sliders as well as colour boxes that allow you to more finely adjust individual colours.  Click and drag to adjust.  Right-click to delete a control point.  Next is Warmth, which is a white balance adjustment.  At the very bottom is a dropdown menu for Profiles.  This allows you to tag your file with one of 4 colour profiles – sRGB, AdobeRGB, ProPhotoRGB or Wide Gamut RGB.  SNS is colour managed.  Big thumbs up!  Personally, I’d like to see this higher up in the hierarchy of options because it may get missed at the bottom.  But having it anywhere is much better than not having it at all.

On most screens, this adjustment panel is longer than the screen is high.  You can scroll up and down in the panel with your mouse.  Be careful though because positioning the mouse on or near any of the adjustment sliders will cause that slider to be affected by the scrolling of the mouse wheel.  If you want to scroll the adjustment bar up and down, make sure your mouse is positioned on the scroll bar at the right.  In the very bottom, lower right of the screen is a small button with a menu triangle.  This allows you to switch the language between English and Polish.  Just to the left of that is a small envelope icon.  This is an email link to the program’s creator, Sebastian Nibisz, to enable users to easily send an email with questions.  Nice touch.

To make and adjustment, you can click and move the slider, position your mouse over the slider and use the scroll wheel or input a number into the box.  There is a reset button at the right of each adjustment.  Right-clicking on the slider will also reset it to its default position.  There is no ‘Undo’ to take you back one step at a time. The sliders are live.  Moving the sliders with the mouse is a bit choppy; however, and getting it to move in fine, 1 point increments can be difficult.  Improving the slider function would be beneficial.

To load a set of images for merging, click on the Open Image icon and select the multiple images.  You can only have one image open at a time.  If you try to open a second or new set, you’ll be asked if you want to save the current image.  Loading a set of images brings up a dialogue for Size Reduction for onscreen viewing (leaving it at 1X is fine, you can always zoom in and out later if needed) and a checkbox for Auto Alignment.  The program does support RAW files using dcraw which is a well known RAW conversion utility used by other software applications.  Loading/converting of the RAW files and even alignment if checked doesn’t take an inordinate amount of time.  It’s a bit slower than other applications but not by a significant amount.  Where SNS-HDR has issues is with the initial tonemapping.  This process is quite slow.  This is an automated process that can’t be bypassed.  The change history on the SNS website indicates that speed has been improved a number of times over various version iterations.  I’d suggest that this may be the single most important thing to be addressed in future versions.  Even with the batch processing option, the tonemapping stage is slow enough that it renders the software basically not usable in a setting where time and turnaround are important.  Which is really a shame because the results are pretty darned good.

The image below is straight out of the merge/tonemap process with no additional adjustments.  Excellent starting point.

The key things to note are the windows.  There is extremely good detail in the windows, particularly on the left.  The histogram shows that there are zero blown pixels.  Colour is excellent.  The blue in the windows on the right is exactly what it should be.  This is the same set of images that have been used for all the other apps tested to see how well they handle a very wide brightness range.  I decided to throw another set of somewhat difficult images at SNS to see how it handled those.  This set doesn’t have a hugely wide d-range but it does have some strong, specular highlights on metal which can be difficult for some software to deal with.

According to the histogram there are no blown pixels.  Placing the picker over the specular highlights shows they’re well below blown at about a 228 value.  Again this is the default image with no additional adjustments.  Once again, a very good starting point.

Let’s get into the guts and see how SNS works when we use the available editing tools.  As usual, we’ll start with an exisiting HDR image.  Opening the image and the program working through its automated tonemapping process is also a slow undertaking.  Below is the default upon opening this image in SNS.

It’s quite washed out, lacking in colour saturation but not lacking in contrast.  Overall it’s on the dark side.  It’s not quite as good a starting point as what you get from merging a bracketed set of files right in SNS.  But we can work with it.  Once the file is open in SNS, positioning the mouse cursor inside the image and using the scroll wheel will zoom the image in and out.  Right-clicking inside the image will show the ‘untonemapped’ version so you can get a sense of the difference if you want.  One thing to note about this existing image is the colour balance.  It’s very good and very neutral.  Photomatix and other HDR applications tend to exaggerate some colours and can introduce colour casts.  In the case of this particular image, the water can be very cool/blue.  Not here.  No colour cast, no cool hue.  Very neutral.  This makes the editing easier.

Working with the adjustment tools available in SNS from the Default tonemapping we can generate a pretty good, realistic result as shown below.

It’s got good contrast, much better colour.  There are a few pixels on the verge of full saturation (blown out) in the waterfall pools, but that’s OK.  The reds/oranges are still a little weak but that can be adjusted with further editing in PS.  This is a very usable result.

How does SNS do at moving into the more surreal end of the scale?  For this approach, I started with the Dramatic preset and worked from there.  This result almost has a bit of a post-apocalyptic look.  It’s kind of cool and different from the less realistic images other applications generate.  Strong colour saturation in some colours, less in others.  Strong contrast.  And we’ve introduced that cool, blue hue to the water.  Overall a bit dark but that can be adjusted if desired.

Anytime you make a change from a preset the preset dropdown menu changes to +Custom.  Clicking on the + sign allows you to save those settings as a custom preset.

Let’s work on the image from the bracketed set merged in SNS.  The result below is the ‘realistic’ image.

The colours are very good and very accurate.  Other HDR apps tend to either make the scene too cool or too warm.  SNS keeps it very neutral.  As with the earlier images it may be overall a bit dark but that’s easily fixed.  Very good detail and colour is retained in the windows even after adjustment.  Again, a very usable image.  Getting to this point didn’t take a lot of work.  It would take a bit less work if the action of the sliders and colour bars were a bit smoother in the fine adjustment.

At the other end of the spectrum, the image below shows the more surreal result out of SNS.

It’s a bit more extreme than the realistic version but definitely not the wild result you can get with other HDR applications.  That kind of wild result may not be achievable with SNS.  And that’s OK.  There’s nothing wrong with an application that keeps things in the more conservative end of the pool since there are plenty of folks who don’t want to go near the wild end.

Loading TIFF files rather than RAW does speed the process up somewhat.  There’s no RAW conversion so that step is bypassed.  The alignment step is no quicker.  The merging step appears to be faster.  The tonemapping stage is no quicker.  So a bit of time can be saved by loading TIFF files but not a significant amount.  Loading TIFFs vs RAW images seems to make no difference to the merged result.

As noted above, there are 4 tonemapping adjustment presets – Natural, Dramatic, Night and Portrait.  The set of images below show each of the presets and the default setting for comparison.

The Default setting is going to be a good, general purpose starting point.  Natural increases colour saturation and contrast slightly.  Dramatic increases overal brightness, contrast and detail and would be a good starting point for moving into the more surreal look.  Night increases brightness, contrast and detail a bit more than Dramatic and would also be a good starting point for a more surreal approach.  Portait is the softest of the presets.  As the name suggests it’s best use is likely for images where people play a strong role in the composition and in images where a softer look is desired.  I can’t see using this one much.

Invoking the Batch mode in SNS brings up a simple dialogue box that allows you to point to the source folder, choose how many images in each set to merge (from 1 to 12 or All), whether you want to apply alignment.  Below that in the Destination you choose your saving location and file type.  The options are JPEG with quality settings, 8 bit TIFF, 16 bit TIFF and 16 bit TIFF+SNS.  This last one is interesting.  It allows you to save the tonemapping parameters into the image file so that you can re-open the file in SNS without having to make any recalcuations.  This would speed the process up.  Click on Run to start the batch.  The Batch icon will turn green and a Batch progress bar appears at the bottom of the screen.  Click the Batch icon again and you have the option to Stop the batch process.  I didn’t actually test the batch process but expect the same speed issues would be present.  If you want to use it and are processing a large number of bracketed sets, running it overnight is probably advisable.

There’s no deghosting function in SNS so that can’t be tested.

At this point there’s no Help menu within SNS and no user manual or other documentation on the website.  Sebastian told me that is in the works.  Hopefully it’ll be ready soon.  Sebastian did respond very quickly to the couple of emails I sent him with questions.  And very quickly particularly considering the time difference.  His answers were complete and helpful.  As more people learn about SNS and start using it, I think it’ll be important for that documentation to be available otherwise he may end up being flooded with questions via email  Perhaps a user forum or wiki would help in that regard as well.

Overall, SNS-HDR Pro is a very good piece of software.  As I’ve noted, the only real downfall I’ve found thus far is the speed.  At $85 Euros for the full version (Home is $30 Euros, Lite is free) which is about $105 USD at the time of writing it’s in the ballpark with the other applications available.  Given how well it works at exposure fusion, if the speed could be addressed, I’d definitely consider buying it and using it in my work.  The quality of the exposure fusion is excellent.  It’s better than the Exposure Fusion function in Photomatix and better than Enfuse which are two others I’ve used.  From a personal standpoint, I’d love to see this exposure fusion technology implemented into Kolor’s AutoPano software.  HDR or blended panoramics has become pretty popular and having this kind of exposure fusion technology built into a pretty decent pano app like AutoPano would be great combination.  While not truly HDR since there’s no 32 bit file created, SNS would fit into what JP Caponigro calls XDR (Xtended Dynamic Range).

As usual, if you note any glaring (or non-glaring) errors or I’ve missed something, let me know and I’ll make the necessary correction.

After three weeks without a proper monitor courtesy of Dell and their poor customer/warranty service, I’m finally back up and running, caught up on the work that got behind and ready to dive back into the HDR app. review series. In this instalment Artizen HDR from Supporting Computers.  One note before continuing:  In the introductory post to this series, I commented on my dislike of the practice some of these app developers have of watermarking finished images.  Well, Artizen does that but goes one step further.  If you’re working with a trial, you’ll get screen that pops up on a regular basis asking if you’re ready to buy the software now.  This is incredibly annoying and were I in the market for an HDR app would completely turn me off buying it.

I first became aware of Artizen a few years ago.  At the time, the concept was interesting.  An HDR application but also a fairly fully featured photo editing suite.  Checked it out back then and while it was interesting, it really wasn’t ready for prime time.  The HDR part of the software wasn’t as good as other options available and in terms of an editing package, Photoshop Elements or Paint Shop Pro were superior.

Fast forward a few years and not a lot has changed, unfortunately.  Downloaded the latest trial of the software and the GUI looks basically the same as it did back then.  The GUI looks deceptively appealing – nice dark grey backgrounds, colour swatch on the right, various editing icons on the left but once you delve into it, it’s not as attractive as it first appears.

There aren’t many Preferences available to be set in Artizen.  You’ll find the two there are (scratch disk location and memory usage) under the File menu.  On the upside, Artizen does have a batch function.  You can have it tonemap or not and can save the 32 bit HDR files or not.  In this regard, the feature set is much the same as what you’ll find in the batch function of Photomatix.  The software doesn’t support multiple monitors.  It is also a Windows-only application.  The recommended system configuration indicates a dual core processor but I could find nothing to indicate that it’s a multi-threaded app.

We’ll come back to talk more about the GUI as we work through the software’s features.  Let’s start to have a look at how well it works.  Opening an existing HDR file is reasonably quick.  It’d be quicker if you didn’t have to click out of the ‘Buy Now?’ dialogue.  Obviously that will go away if you’re working with a fully licensed version rather than a trial.  Like some other applications, Artizen works with a series of nested screens.  Can you drag the main screen to a second monitor and the nested screen back to your main monitor?  Nope.  Below is a screen capture of the nested screens with an open HDR file and as you can see, it’s already got a lovely, large watermark across it.  On the right hand side of the screen, you’ll panels for a histogram, colour swatches, metadata and layers.  Each of these can be opened and closed with the small triangle on the left of the panel bar.  If all are open, they won’t all fit on screen.  There’s also no way to scroll down with the mouse to view all of the information.  You have to open and close panels as you might need them.  This can be a bit tedious and inconvenient.  At the bottom left you’ll see an explorer window.  The reason for this, I guess, is to make the workflow faster by giving you a thumbnail of images in a particular folder.  It takes a very long time to load the thumbnails which actually slows down the workflow.  This explorer bar also takes up valuable screen real estate.  Personally, I’d click the triangle in the centre of the screen and close this bar or deselect the Image Browser Bar in the Panels menu.

You can get into the tonemapping functions either from the HDR’s>Tonemapping menu or by clicking the HDR TM button at the top of the screen.  The tonemapping function opens in yet another screen.

The usual suspects in a set of tonemapping operators are here.  You have your global and local controls.  In a nice touch, the local controls have actually been labeled as such so there’s no confusion for someone newer to HDR.  Oddly, some of the sliders offer a live preview but most do not.  The individual colour sliders do, as does the Temp slider.  The global saturation slider doesn’t.  Nor do any of the others.  With those you have to release the mouse button to apply the adjustment.  There is an Undo button in the top left of the screen and you can set the number of ‘Undos’ in the Preferences up to 1000 (who’d ever need THAT many?).

There are a few presets available – Display, Dramatic and Natural.  Dramatic is where you’d start if you’re going for the more surreal or grunge look.  Display is a decent general starting option.  Natural isn’t as natural as you might think, it actually tends a bit toward the surreal side of the tonemapping spectrum.  There are all only starting points; however, and you can adjust from there.  These are decent starting points and can provide a nice basis for going down a particular tonemapping path.  The default tonemapping preset is Display.  On the downside, the settings are not ‘sticky’.  That’s to say that if you switch from one preset after making adjustments then go back to it, you have to start over again.  There are two ways to select a preset.  On the upper left of the tonemapping operators you can choose from one of the three in the dropdown menu.  Also, in the main panel of the tonemapping screen in the upper section there is a dropdown menu with the three presets.  Duplicating this way seems a bit unnecessary.  It would be nice if the presets were available in the upper left of the tonemapping operators and stayed at the default settings but in the upper part of the main screen if you switch back and forth, any adjustments away from the default were retained with the most recent changes.  The tonemapping adjustments aren’t ‘sticky’ from image to image either.  If you want to apply the same settings to a new image, you’re forced to save the settings as a preset.  It would be preferable if, as with Photomatix, the last used settings were retained from image to image.  It’s simple enough to override by selecting preset defaults from the upper dropdown menu.

What do the default presets look like?  Check them out below.  Artizen is not a colour managed application.  Tonemapped files saved out of Artizen will not be tagged with a colour space.  Artizen also appears to strip any profile from an image opened and saved in the program.  I tried opening and saving a TIFF that had been tagged with ProPhotoRGB, making no adjustments, and when I re-opened the saved file in PS, the missing profile dialogue box appeared.  This essentially makes Artizen unuseable for any serious work.

None of these look overly appealing to my eye.  That said, any may prove a reasonable starting point for further adjustment.  As noted above, opening an existing HDR file doesn’t take much time.  Clicking on the HDR TM button brings up the tonemapping screen fairly quickly as well on my system.  Once you’re in the tonemapping screen, the actual image doesn’t take up a lot of the screen real estate.  The default Preview setting is Balanced.  This is supposed to be a balance between fast updates and accurate updates.  The downside is the on screen image is small.  If you want the on screen image to be larger, you need to up the zoom or select Slow & Detailed from the Preview dropdown.  I’d advise against this.  The ‘slow’ part is bang on.  Balanced or Fast is the way to go.  Selecting ‘Fast’ zooms the image to over 200% in my case which causes a loss of detail.  You can click on the 100 icon which brings it back the same size image as the Balanced default. If you switch from one tonemapping preset to another, the preview quality defaults back to Balanced as well and you have to change it again manually.  These steps to ‘correct’ for the software do take time and shouldn’t be needed.  Adjustment updates in this setting are much quicker.  Selecting ‘Very Fast & Low Detail’ causes the preview image to shrink to the point of not being overly useful.  The supposed 100% preview is oddly quite small and, for me, too small to be overly useful.  More on why this is in a bit.  One really good thing about the tonemapping screen is the histogram at the bottom.  This isn’t just a histogram but also a Levels control.  Like the Levels control you’d see in PS you can move the white point, black point and midpoint sliders to adjust the image contrast.  There are also white and black eyedroppers that you can use in the image to select your white and black points.  This is a REALLY nice tough and something that other HDR apps should consider implementing.  I’d even go so far as to say that it’s more useful here than in PS.  Here’s where we come back to the size of the preview image.  Because even at 100% it’s so small on screen (this will be affected to some extent by the resolution of the monitor) it may be difficult to select just the right white or black point with the eyedroppers.  Yes you can increase the zoom level but that results in a loss of detail.  Having the preview image at 100% fill more of the screen would be very helpful.  Stuck down at the very bottom of the tonemapping screen in Additional Options are two radio buttons for Output – one for 8 bit and one for 16 bit.  The default is 8 bit.  Be sure to check these and get the output bit depth you want, particularly if you want 16 bit.  Really, these should be in a set of global preferences that you can set up once and not have to worry about again rather than having to set it for every image processed.  In discussions with Supporting Computers, they acknowledge that the GUI needs some work but it’s needed work for about 3 years now so no telling when they may get to it.  Make all of your tonemapping adjustments then click on Apply at the top of the screen… and wait.  Applying the tonemapping settings and rendering a finished output file does take a bit of time.  More than one might like.  There appears to be no difference between the preview and the final image even using the Fast preview setting.  This is someting that the folks at HDRSoft would do well to correct in their Details Enhancer tonemapper.

What’s the final result look like?  See below.  This image was resized in Artizen and saved at 70% JPEG quality.  The tonemapping is alright.  Not much in the way of quibble there.  This was effected working from the Display default and adjusting from there.  If you click through to the 100% image, there is some very evident JPEG artifacting that isn’t at all appealing.  By comparison, the second image in this set was saved at 100% size and quality in Artizen as a JPEG then resized in PS and saved at quality level of 8 (roughly 75%).  There is none of the artifacting present.  There are several compression methods available in Artizen.  I chose a different method from the default which is supposed to give smoother results and the artifacting was really not much better.  I’d surmise then that the resizing algorithms in Artizen aren’t overly good.  The software can create a good ‘realistic’ image from an existing HDR file.  I’d have to do some more work on this either in LR or PS to get it to where I wanted it as a final version but this is a pretty good starting point for final tweaking.

Moving to the surreal side of things, selecting the Dramatic preset as a starting point starts taking the result in that direction.  It also changes some of the tonemapping operators available.  The folks at Supporting Computers have tried to customise the tonemapping operators for each preset which is a nice touch.  Using ‘Fast’ as the preview option will result in the final output being significantly different from the preview.  Using ‘Balanced’ as the preview option corrects this.  The result is a bit slower but more accurate preview.  In this regard, it seems Artizen is in muc the same place as PM and would do well to try to improve the preview accuracy with the preview speed.

The result of pushing the tonemapping more into the surreal end of the pool is below.  Clearly, with this image, Artizen can do both pretty well.  Not my style and not a result I’d work toward with tonemapping but a lot of people like it so Artizen can do that for them too.

We’ll move on now to see how Artizen does merging and tonemapping a set of image files from scratch.  The file loading dialogue provides a number of options.  There is a manual alignment feature which can be bypassed if desired.  The automatic alignment offers three possibilities.  The one labeled ‘Use Slow Method’ makes one wonder whether the other options are faster but not as good.  The images I’ve used were shot on a tripod with a cable release so there should be no alignment issues and as such I’ve selected ‘Use Region for Auto Alignment’.  This, to me, would be similar to the ‘matching features’ option in Photomatix.  In the Merging Options the one to choose will be, pretty obviously, Create HDR File.  The second option ‘Luminosity Precision for 3D IBL’ is for 3D graphics.  When choosing ‘Use Region for Auto Alignment’, you have the option to select an area of the image in the small preview window in the lower right which the software will use for the alignment process.  You can also simply click Apply and let the software do what it will with the alignment.  Reduce Micro Noise and Attempt to Reduce Ghosting are the two final selections in the Merge Options.  We’ll take a look at how the deghosting works later but I’ve left both of these unchecked for this set of images.

The file merging and alignment process is not fast.  I didn’t try the Use Slow Method and given how long the ‘faster’ method took, I wouldn’t want to.  It’s quite slow to merge and align.  Easily the slowest of all the applications reviewed thus far.  Selecting a smaller region within the full image to use for alignment doesn’t make the process go any faster.  While DPHDR had problems aligning the same source images, Artizen didn’t.  And it shouldn’t given the way they were shot.  As odd as it may seem, selecting ‘Create HDR File’ in the Merge Options is NOT what you want to do.  Even with this box unchecked, the program will still merge the files into a 32 bit HDR file.  With it checked; however, after the alignment stage but before the merging stage, you’ll be presented with a dialogue to name and save the 32 bit file.  You CANNOT cancel out of this.  If you click Cancel, the dialogue will simply re-open.  You’re forced to save the file at this stage whether you want to or not.  Leaving this box unchecked will allow you to create the 32 bit file and save it later if you want to.  Seems counterintuitive not to check a box to create a 32 bit file when that’s actually what you want to do, but leaving it unchecked is perhaps the better way to go.

I chose the images for this portion of the test on purpose.  I selected this set of images because there was a very broad dynamic range and I wanted to see how well each application could crunch the brightness range to give a useful result.  As can be seen below, Artizen didn’t do so well in either a realistic interpretation or a more surreal interpretation.

In the realistic version, I tried all three presets as a starting point.  In the end, I settled on Natural and worked from there.  This is the best I could get and it’s not that good.  The windows are still very blown out and the colour is far off.  For the surrealistic image I started with the Dramatic preset.  The colour of the floor and brick of the walls is a bit better but the windows are still blown out.  Neither of these is a result that could be used for anything.

There should be no reason to believe that the result would be any different if the source images had been merged in another app. then tonemapped in Artizen.  Artizen doesn’t appear to be able to handle a very wide dynamic range as well as other applications.  To test this theory, I created an HDR file in Photomatix of the same 9 source images to tonemap it in Artizen.  The results are below, a realistic version first and a more surreal version second.  For the realistic version I used the Display preset as a starting point and for the surreal version the Dramatic preset was the jumping off point.

Surprised?  I know I was.  My initial thought was that it was the tonemapping operators of Artizen that had difficulty with the wide dynamic range.  Given these results, both of which could be used with a little further tweaking, it appears that it’s actually the merging of the source images in Artizen that appears to have some problems.  Colour still isn’t quite as good as with the other applications tested thus far but the detail in the windows is very good in both cases.  The colour could be addressed in post-tonemap editing.

To try to see if this is a general issue with Artizen or a particular issue with this set of images, I decided to try another set.  These also have a fairly wide dynamic range.  The first is the result from merging and tonemapping in Artizen.  The second was merged in Photomatix and tonemapped in Artizen (ignore the reflection of the idiot in the front fender).

There is definitely more detail retained in the highlight areas of the version that was merged in Photomatix.  The colour is also better.  The merging of source images in Artizen does appear to be less effective than other applications.

Artizen does have a deghosting option that can be turned on during the HDR merge.  There is no user input, it’s fully automated.  How well does it work?  Not too badly at all, in fact.  The original and deghosted images are below for comparison.

Once again, you can see comparing the original image to the merged image from Artizen that colours are off.  Pretty far off in this case.  But the deghosting process has worked pretty well.

In terms of support, I sent a couple questions to the general Artizen support group about the software; and I have to admit to sandbagging a little when asking the questions, to see how quickly they would respond and how good the responses were.  They responded very quickly in both cases.  The quality of the response was less positive.  Not all of my questions were answered and the ones that were answered weren’t necessarily answered as well as I might have liked.  They also seemed more eager to offer me a discount code that readers of the blog could use to buy the software than in answering my questions accurately.  There are a number of tutorials on their website which is good but the screen shots of some of the tutorials appear to be out of sync with the latest version of the software.

At $46, Artizen is one of the less costly applications out there.  Even at that price, despite some of its positives, there are too many negatives to be able to recommend it for use.  The speed is a big issue.  The apparent problems with merging source images is another.  As noted in the opening, it doesn’t really appear that a lot has changed with the program in the few years since I first had a look at it.  Good idea at the time but not executed as well as it could be and there are now other, better options that do as much and more.  Artizen is a Windows-only application and is not built to take advantage of a 64 bit system.  Next up, Ariea HDRMax.

As with the previous pieces in the series, if you see any glaring errors, please let me know and I’ll re-evaluate and correct as necessary.  Happy to read your feedback.

Dynamic Photo HDR is from a company called Mediachance.  It has a number of photo/video/graphics related applications available but DPHDR is probably the most well known.  DPHDR was the first HDR application I purchased.  When I bought it, I was trying out a few different alternatives and found that DPHDR had a really interesting manual alignment function.  In my case, I used it for manual ‘mis’alignment and to blend multiple image layers into a single impressionistic photo result.  The first eight images in my Impressionism II Gallery were made using this method.

For this review, I’m using a trial version of their latest iteration of the software (v4.7).  Since I’ve primarily used Photomatix up to now for HDR work, my last paid version of DPHDR is v3.x. Dynamic Photo HDR is available for both Windows and Mac but the editing add-on Photo-Bee is Windows only (more below on this).

I’ll say upfront that speed has never been a strong suit of DPHDR.  Loading an existing .hdr file takes much longer than with either of the previous two applications in this review series.  Loading a set of RAW files and providing a preview is a bit quicker.  But the speed to merge the files from the preview is slow.  More on speed later.

DPHDR does not employ colour management at any stage in the process.  You can’t have it tag your input RAW files with a colour space during conversion and you can’t tag the saved output with a colour space.  If you load tagged TIFFs as your input images, the saved output will not have a colour profile embedded.  DPHDR doesn’t even honour embedded profiles.  The net result is that what you see when you’re tonemapping the image may be quite different from what you see when you open it for further editing and have to assign a profile.  Not allowing the ability to tag input files is one thing but not providing the option to tag output and stripping profiles from tagged input files is something entirely different and not different in a good way.  This is a significant negative, in my opinion, in comparison to the previous two applications.

The DPHDR GUI is a bit hackneyed.  There are panels within the main screen and while these panels can be moved to a second monitor, the setting is not ‘sticky’.  That is, when you close and re-open the program, the panels are back on the main monitor and have to be moved again.  So there is some multi-monitor support but it’s not complete.  Intuitvely you might think the small pushpin icon at the top of each panel would allow you to ‘pin’ the panel to a particular monitor.  Not so.  What this does is determine whether the panel is expanded all the time or contracted and expands when you move your  mouse over it.  Each step you go through in the process in DPHDR brings up a new sub-screen.  No work is actually done in the main screen and all these sub-screens are much smaller which means getting a good look at what you’re doing is more difficult.  The three images below show some of the screens at different stages.

There are no user-adjustable preferences for image processing.  You take what Mediachance has decided to give you.

So now let’s take a look at the program in action.  Opening our existing HDR file is, as noted above, a fairly slow process.  Once it’s open on screen, you have a 100% view in the middle of the screen and a small preview in left panel.  There’s no way to adjust the zoom level of the preview in the main screen to fit the whole thing in view without actually reszing the file.  If you go to Image>Resize HDR, the options there will actually rez down the image file.  If you do this, there’s no ‘Undo’.  You’ve rezzed down the file and you’re cooked.  You can rez it back up but we know what problems that causes.  The only way to get back to your original is to go to File>Reload Current File.

Once we go into the tonemapping screen, there are several options.  Tonemapping operators are grouped into Local and Global.  Local is going to get you more easily to the surreal/grunge look whereas Global is intended to provide a more realistic output.  There are labels for the tonemapping operators such as Eye Catching, Halo Matix, Photographic, Human Eye, etc.  The idea is to give you a sense of what the particular tonemapper is going to generate in terms of a result.  In general, while you can get a ‘realistic’ result using the Local tonemappers, you’re going to want to try the Global options first.  These should give a more realistic output.  If you’re not getting a result you’re pleased with, try the Local operators.  Getting a photorealistic result with DPHDR isn’t a simple process.

New in v4 of the program is that the tonemapping adjustment sliders now offer a live view of the adjustment.  This is a positive.  Less hit and miss.  There’s still a small delay; however, after the slider is released for the adjustment to be applied.  On the right side of the tonemapping screen are some other controls.  There’s a Gamma slider and a Curve adjustment.  Both can be useful in getting a good result.  Neither of these are live; however.  You make the adjustment and only after releasing it do you get the result on screen.  Above these is a Filter Color button.  Clicking this opens (yet) another screen where you can apply a variety of effects to the image from B&W to Orton to Vignette.  Generally I’d ignore this and if I wanted to do any of these types of things, I’d do them later in a more fulsome editing application.  Below the Curve adjustment are Color Equalizer and Hue Shift adjustments.  Color Equalizer is basically a saturation adjustment.  You click on the line and drag down to reduce saturation of a particular colour and up to increase it.  Placing a single control point will cause a curved saturation line to be created.  If you want to isolate a particular colour more, you need to place additional control points on the line and drag those back in the other direction.  You can toggle the result off and on with the checkbox.  Clicking the x1 icon will toggle between x1 and x2 and x2 is a more intense application of the adjustment.  The Hue Shift adjustment, as you’d expect, adjusts the hue of a particular colour.  It works the same way as the Color Equalizer.  Click and drag on the line to change the hue of a colour or colour group.  Place more control points and drag to affect a particular colour more or less.  The .. icon allows you to invoke a number of presets but there are no indications of what the presets are till you keep clicking the button.  This really seems a useless option to me.  Along the bottom of the tonemap preview are clickable icons for further adjustments.  Kelvin is a colour temperature (White Balance) adjustment.  Opening it will show you a variety of options to select from.  To cancel a change, click on Kelvin and select Default.  The eyedropper allows you to click inside the image to adjust the white balance.  Match Color opens a screen that lets you match the colour balance in your current image with another image.  There are a number of presets to choose from (useless) but you can also load another image of your own to match to.  This could be useful if you have a number of images from the same location and easily want to match the colour balance of all of them.  It may be a decent starting point from which you can tweak further.  In the lower right you have several checkboxes, the only two of which that are useful are the Add Clarity and Dehaze options.  Both add a touch of contrast/sharpness (not a lot) and can give the image a bit more punch and remove some of the flatness that can be imparted by crunching a ton of brightness into a small space.  NR is a noise reduction feature.  Don’t use it.  It works by adding blur and you’ll end up with an overly soft image.  In addition, when you invoke it, the ability to move around the image is done via X and Y sliders which don’t give you a great deal of visual control because of the zoom level.  There are better noise reduction options elsewhere.

As you can see, there are a lot of options and possibilities built into DPHDR.  How do all those options work in practice?  The image below is the ‘realistic’ version of the existing HDR file.

It’s not bad.  The contrast is pretty flat which can be adjusted later.  The blues in the water are a little strong which can also be adjusted.  I didn’t use the Color Equalizer tool during tonemapping to try to correct the blue in the water but I could have.  Because the working image is so small, it’s a bit difficult to see what you’re getting in smaller areas so I left any of that kind of editing for later.  Is this result as good as what was obtained in either Photomatix or CS5 HDR Pro?  To my eye not quite but after all of the applications have been reviewed, I’ll provide a summary with all of the images side by side.  This is certainly a result that could be worked with to optimise further.  Getting to this result took more time than in either of the other applications.  I had to work with different tonemappers to find a result that gave me something I was reasonably happy with.  Getting a photorealistic result in DPHDR is a bit more work than in the other two apps. reviewed so far.

What about a more extreme result?  Yep, DPHDR can definitely do that and this is probably the strength of this piece of software.  Below is a fairly grunge-look result on the existing HDR file.  This took all of about 10 seconds to achieve.  But it took over a minute to process and save out the 8 bit JPEG file.  So again, we see that speed isn’t a strength of this application.

When it comes to saving the tonemapped file, there are two options.  Clicking on the small triangle beside Process & Save will allow you to select a second option which is Process & Edit.  Selecting Process & Edit will open the tonemapped image in another application called Photo-Bee.  This is a new feature that’s been added in v4 and my guess is it’s an attempt to create an ‘all-in-one’ application where you can create and tonemap your HDR file then continue editing it all in the same family of software applications.  Don’t bother.  While Photo-Bee has some of the usual editing tools, implementation of those tools isn’t good.  Hue/Saturation, for example, is a global function.  You can’t select individual colours to adjust.  The Curves function is a global up or down (i.e., no S-curve).  When saving a file edited in Photo-Bee, there are still no colour management options.  Photo-Bee is basically a waste of time.  Stick with Process & Save then do any additional editing in a proper editing application.

DPHDR can handle existing HDR files reasonably well.  It’s slow and getting a realistic result is more difficult and takes more time/trial and error.  Generating a more extreme result is very easy and, as noted above, this is probably the strength of the software.

Let’s move on to look at how it works with a merge from scratch.  There are several options available when loading the files to be merged.  If  you load RAW files, in the bottom left of the preview screen you can select from Auto Developer, Zero Processing or Camera Curves.  This will determine how the RAW files are interpreted before merging together.  You can select each of the different options and see how it impacts the preview.  If you want to change the EV setting you can do this by selecting the individual file and adjusting the EV.  For auto-bracketed sequences this shouldn’t be necessary.  Sometimes for manually bracketed sequences it may be.  The HDRI Simulator on the right gives you a preview of how the tonemapped image will look based on the default settings of either the Eye Catching or Auto Adaptive tonemappers.  I wouldn’t pay a lot of attention to this.  In the very lower left of this screen is a checkbox for Align Files in Next Step.  You do want to have this checked.  The auto-alignment functionality of DPHDR isn’t overly good.  Even images taken on a tripod with a cable release will often not be aligned properly in the automatic setting.  This is a pretty major downfall of the program and development time would be better spent in this area than in something less worthwhile like Photo-Bee.  Click OK and the merge begins.  Be prepared to wait.  And wait.  Loading TIFFs as the input files makes the preview quicker because DP isn’t processing the RAW files but none of the rest of the workflow is any faster.  When the alignment screen comes up you can choose to align for Movement, Rotation or both.  Movement should generally be sufficient.  You can move the image around to get an area where determining alignment is easier (i.e., high contrast areas).  On the right side of the screen you see a set of image pairs.  The application chooses a base image and compares alignment of that base image to all others in sets of two.  There is a very good chance that you will have to do some manual alignment.  Click on an image pair to select it then click Align View which is supposed to automatically align the pair of images.  It works most of the time but if it doesn’t, you’ll have to align manually.  You can use the arrow keys on your keyboard to move the second image of the pair around.  You can also use the sliders on screen.  If there’s rotational movement, you’ll have to use the sliders.  If you end up having to do manual alignment, you can click on the Diff button.  This will change the contrast of the two images and show a number in the upper left corner of the screen.  The smaller this number is, the better the alignment.  It won’t be zero.  Adjust the image pair till the number hits the smallest value in all directions.  It can be tedious.  Believe it or not, sometimes the number will keep getting smaller and visually you can see the alignment getting worse.  Suffice it to say that the alignment functionality of DPHDR really is a major weakness.  Clicking OK will apply the movement corrections to each image pair then finalise processing the merged result for tonemapping.  Be prepared to wait.  And wait.  If the image opens and it still doesn’t appear to be aligned properly, you can go back to the alignment step by clicking the back arrow icon in the upper left of the screen or by choosing Image>Back to Align from the top menus.  The ‘unalignment’ will take time.  In the alignment screen, there is a button for Advanced Options.  Clicking this will open another screen where you can refine how the images will blend together.  The left is an exposure weighting function which determines how the various exposures are merged together.  Given the alignment issues that can crop up, the default of Smoother probably makes sense.  Choosing Custom and adjusting the shape of the distribution curve probably isn’t a good idea since the preview you get is small and you can’t tell how the change will impact the entire image.  On the right are some selections for tonal response.  If you’re loading RAW files, selecting the second option, Ideal Camera, makes some sense since RAW files are supposed to have linear response curves.  Normal Camera is the default and probably works for using JPEGs as input files.  It will assume a response curve based on whatever has been programmed into the software.  There’s a Custom option here as well.  If you loaded TIFFs as your input images, if those TIFFs were processed in ACR or Lightroom and you knew the tone curve that ACR applies you could input that tone curve in the Custom option.  While these may be interesting options to include, they’re likely far more than the average user is going to want or need.  An argument could be made that the user shouldn’t be adjusting tone curves of the input images either because the merging and tonemapping functions are supposed to make allowance for the less than perfect tone response of cameras and that making adjustments in a 32 bit environment is preferable to doing it on lower bit depth input images.

The image below is the realistic version with the source files merged in DPHDR.  In the essense of full disclosure, I had to work through this three separate times to come up with something reasonable and this is the third of the three.

Even though I’ve given DP more attempts than the previous two applications, the result isn’t as good.  It cruched the brightness range well, no question.  The overall contrast is very flat and the colours are extremely muted.  The blue in the large window on the right that I was able to retain in both PM and HDR Pro is gone here.  In order to keep that blue in the windows, too many other sacrifices would have been necessary.  This image would require a fair bit of additional work to be useable.  And due to misalignment problems it may not be useable at all.

The less than realistic result is a different story and again we see that this may be where the strengths of this program are.  On a full size version, zooming in reveals some alignment issues.

There is a deghosting option in DP.  Unlike PM and HDR Pro, it’s not automatic.  The user needs to manually mask out areas where ghosting appears.  You do this on one image pair and can apply the mask to all other image pairs but depending on the type of movement, this may work or it may not.  Suffice it to say this can be an incredibly tedious task.  Building in an automated anti-ghosting algorithm wouldn’t be easy by any stretch.  But the manual option in DP doesn’t work overly well and because it’s there people are going to be tempted to use it and be very frustrated with the tedium and the less than satisfactory result.  The image below shows the artifacting that remains from attempting to manually mask out some of the ghosting in the windblown branches.  Having no deghosting option would be better than having one that doesn’t work very well, as the one here doesn’t.

The last item in the review is support/documentation.  The user manual that comes with the program and that can be called up via Help>Tutorial is quite extensive and walks the user through the process of creating and tonemapping HDR files with Dynamic Photo HDR and does a good job of explaining all the various features and options.  The website has some video tutorials and other good information on using the program.  If you want tech support via email or telephone, forget it.  It doesn’t exist.  Any questions you can’t find answers to on the website or from other users are going to go unanswered.  There is a user forum and the link to it is on the Mediachance website.  The forum isn’t hugely active but you may be able to get some answers there.

Dynamic Photo HDR is a complex piece of software.  If offers the user a plethora of options and functions to choose from.  Probably more than most users need.  The look and feel of the software is of a piece of software that was written by a programmer with programmers in mind as the end user.  Development time has been spent on things like Photo-Bee and allowing the user to adjust the exposure weighting and tone response curves that could likely have been better spent improving things like image file alignment and an automated deghosting algorithm.  At $55 it isn’t overly expensive ($25 to upgrade from a previous version) but I’m not sure the performance is up to the price even at that level.  It’s very slow relative to others tested so far.  Getting a photorealstic result is more difficult with it than the others tested so far.  And there are basic functionality problems (e.g., alignment).  Several years ago when HDR software was first becoming widely available to the masses and the applications were in the infancy stage, DPHDR was a very good choice.  Its primary competition from the early days, Photomatix, has been improved more than DP.  Adobe has upped its game in HDR with CS5 HDR Pro.  And there are other, newer, competitors on the market that, arguably, work better in many respects than DP.  Unlike Photomatix, there is no Lightroom plugin to export files from LR to DP.  The website indicates that speed has been improved in v4 and that the GUI has been enhanced.  I’m not really seeing a pick up in speed and the GUI doesn’t look much different from previous versions.  There’s a tonemapping paintbrush that’s been included that lets the user paint away tonemapping in specific areas to varying degrees.  This is really a gimick and not of much practical use.  There’s also a feature to help remove halos that can result from overly aggressive tonemapping.  It’s primarily designed for skies.  Since neither of the test images had skies I couldn’t give it much of a go.  There’s also a setting for skin but HDR images of people generally aren’t all that appealing anyway (unless you’re Luke Kaven) so I’m not sure why this was included.

Thanks for reading.  As with the others, if you see any glaring (or non-glaring) errors, let me know.

Instalment two in this series is to look at Photomatix Pro from HDRSoft.  Photomatix is probably the most used HDR application out there.  It’s been continually improved over time and integration with other applications has been added (more later).

Photomatix has always been reasonably fast in terms of loading and processing a bracketed sequence, applying tonemapping adjustments and rendering out the tonemapped LDR file.  Until HDR Pro in CS5, it was certainly the fastest I’ve used but now it seems that CS5 has moved to the top step in terms of speed.

When creating a merged HDR file, Photomatix works best with TIFF files.  You can load RAW files into it and it will interpret the RAW files but it works faster if you load it with TIFFs even though the TIFF files are much larger.  The Lightroom plugin for Photomatix uses LR to convert the RAW files to TIFF before loading to PM.  Using this route, your speed will be limited by how fast Lightroom converts the files.  For users of LR this is a convenient workflow though.

PM is a colour managed application.  You can tag your HDR files with a colour space so that everything stays in a colour managed loop from start to finish.  If you load RAW files into PM directly, you can choose what colour space to assign to the primary input files.  Your choices are sRGB, AdobeRGB and ProPhotoRGB.  If you’re using the LR plugin to load files into PM, it’s a bit wonky to get the files tagged with a profile.  There are two routes in LR to export the files.  You can go to File>Plugin Extras>Export to Photomatix Pro but using this, what might seem the logical route, doesn’t give you the option to tag the input files.  If you then open the tonemapped file for further editing you’ll be presented with the Missing Profile warning.  In order to get the files properly tagged, you need to go to File>Export, choose Photomatix Pro on the left and you’ll be presented with the option to select file type (using the File>Export>Plugin Extras>Export to Photomatix Pro, the file type is automatically TIFF and 16 bit, no choices), bit depth and colour space.  It would be nice if the folks at HDRSoft would incorporate these choices into their own export dialogue.

The Photomatix GUI is pretty utilitarian.  But it works.  It does not fully support multiple monitors which means that you can’t get a nice big, full screen version of the image to tonemap because some of the screen real estate is taken up with the tonemapping panel.  Interestingly, when you’re in tonemapping mode, you can move the histogram to a second screen.  Multiple monitor support would be a postive since a growing number of photographers do use multiple monitors.  Below are two screen captures showing the interface when the program is launced and when a file is open for tonemapping.

In the program preferences (View>Preferences) you can set up some default working conditions.  In the General tab, you can set up, among other things, how many cores the program uses to process.  In my situation, with a dual core processor it defaults to 2. Multi-core support only comes into play with tonemapping.  It doesn’t impact the merging of the source files (including alignment) or the Exposure Fusion option (more on that later).  I believe HDRSoft have plans to enhance multi-core support for the merging/alignment functions in the future.  In the RAW tab you can set up defaults for white balance and colour space when importing RAW files into PM directly.

Let’s get into how the program works with an existing 32 bit file.  In the previous instalment on CS5 HDR Pro, I used an HDR file that had been merged in PM.  It would kind of defeat the purpose to use that same file for testing PM so I’m going to use a file that was merged in HDR Pro.  It’s the same image and the same input files were used for consistency.

There are two options for tonemapping in PM.  The main one, with the most controls and options is the Details Enhancer.  This is the one most will be familiar with (and the one I’ve used here) and it has all of the options you’d expect to see.  You can adjust the light smoothing, luminance, strength of the operators, saturation, white point, black point, microcontrast and more.  This is the PM equivalent of the Local Adaptation tonemapper in CS5 HDR Pro.  Light smoothing can be approached in two ways.  By choosing Light Mode (clicking the checkbox) you can then select from 5 preset smoothing options.  Unchecking Light Mode takes you to a smoothing slider which gives you more control.  Using the presets in Light Mode is generally the route to go if you want to work to a less than completely photorealistic result or a completely unrealistic result.  The finer control offered by the smoothing slider generally allows for a more realistic result; although a realistic rsult can be achieved in Light Mode as well.

Preview of any of the adjustments in tonemapping is not instantaneous.  You have to make the adjustment then there’s a slight delay (not long) before seeing the image update with the effect of the change.  This can make tonemapping a bit of a hit and miss affair but the delay isn’t very long so overall the tonemapping stage can still be accomplished reasonably quickly.  Once you’ve applied your tonemapping adjustments and click Process at the bottom of the panel, there is a delay while the tonemapping adjustments are fully applied to the image.  Probably the single biggest issue with PM is that there can be a significant difference between the preview and the processed output.  You can think you have a nice looking result in preview only to find it doesn’t look like what you thought once the file is processed.  This only happens using the Details Enhancer option and the difference between preview and final image tends to be greater when you’re on the edge of (or beyond) the completely photorealistic result.  The folks at HDRSoft are aware of the problem but to date haven’t implemented a solution.  The net result is that you can, sometimes, have to go back and re-tonemap the image again which is time spent on duplication of work.  It would be very beneficial if HDRSoft could find a solution to this so that the preview and processed image were the same.  They indicate that it’s an issue with the preview image being a smaller version and that the preview, as a result, isn’t entirely accurate.

The image below is the ‘realistic’ result from PM using the Details Enhancer tonemapper.  This is the ‘exisiting’ HDR file which, as noted above, was merged in CS5 HDR Pro and saved as a .hdr file then opened and tonemapped in PM.

Contrast is quite flat (not a bad thing).  From here I’d do some further editing to adjust colour levels, contrast and sharpen to create a final output.  Getting this result wasn’t difficult.  Knowing that the final image can be a bit more extreme than the preview, I purposely kept the tonemapping quite tame so that the final processed image would still be useable without having to go back and do the work again.

Achieving a more ‘surreal’ result with PM is also quite easy.  You can apply the tonemapping operators multiple times to enhance the effects as you want.  The image below is the less than realistic result.  It’s pretty ugly (to me).  It’s got virtually no shadow contrast, blown highlights and the colour is whacked but a lot of people like this effect and it can be done with PM, no question.  This was tonemapped a single time.

Now let’s turn to the file that’s both merged and tonemapped in PM.  Same input files as were used in the CS5 HDR Pro review previously.  The image below is the realistic version.  As in HDR Pro, the window on the left is pretty muc devoid of detail (no surprise).  Comparing the two images, there is slightly more colour from the dirt on the window visible in the PM version than the PS version.  The difference between the two isn’t significant enough to say PM is superior to HDR Pro or that PM can crunch more dynamic range than HDR Pro.

The colour balance between this and the HDR Pro version isn’t markedly different.  The HDR Pro version has slightly more saturated colour but that can be addressed in final editing with the PM version.  The HDR Pro version shows more fine detail (i.e., sharper) but, again, that can be addressed with sharpening of the PM image in final editing.  The biggest difference between the two is in the granite of the floor.  The detail and colour in the HDR Pro version is a bit better and more accurate.  It would take some extra work in final editing with some masked adjustment layers to get the PM version to the same point but I’m confident it could be done.

Getting to the ‘surreal’ result both merging and tonemapping in PM isn’t difficult.  The image below is the less than realistic result.

This one isn’t that extreme.  It’s got more of an illustration/graphic look and in some situations could be useable depending on the desired look.  Below is a more extreme version that was tonemapped twice.

As you can see, going from mild to wild is certainly possible with Photomatix.

We’ll take a quick look at the Tone Compressor tonemap option of Photomatix as well.  There aren’t as many options here.  The goal of this tonemapper, really, is just to crunch as much dynamic range as possible into a visible space.  It works reasonably well.  It will generate more realistic results and it seems to tend to sacrifice highlights in some cases.  There aren’t the same preview/processed differences with the Tone Compressor as there are with the Details Enhancer.  If you’re looking for a realistic result and don’t have extreme highlights to deail with (or are willing to fix those later) then the Tone Compressor may work for you.  The image below was processed using the Tone Compressor option.

Once again, the contrast is a bit flat and colours are muted.  Both of these can be addressed in final editing.  The waterfall pool didn’t blow out and the result is photorealistic.  The Tone Compressor option works quickly and the time to process the preview image is quite short.

Photomatix also has an image blending option.  Not true HDR since it doesn’t create a 32 bit file, it’s an automated exposure blending tool.  Exposure Fusion, as it’s called, can be launched via the Lightroom plugin just as the merge to HDR option.  There are several possibilities with Exposure Fusion in terms of how it blends the images from fully automated to blending based on user input.  It can use just two files or multiple files.  You can load multiple files in and then select which two to use if you wish.  It’s not overly fast, either to preview adjustments in the more interactive modes nor to process the blended image.  It doesn’t crunch as much brightness range as the actual HDR process, nor as much as other automated blending applications like, for example, Enfuse.  Generally the Exposure Fusion functionality is going to be of more limited use but the end result is definitely a realistic (as opposed to surrealistic) look.

Photomatix has a very good batch function.  If you’ve got a large number of images to merge into several HDR files, using the batch function can make pretty quick work of the process.  You point the program to the folder that holds the input images, tell it how many to use in each merge and it goes to work.  You can have it batch tonemap as well.  To do this you set up a tonemapping preset and tell the batch routine what preset to use.  Alternatively, when you select the tonemapping option in batch processing, you’re presented with the tonemap panel and can set the various parameters to be used.  There’s no preview though so you have no idea what the final result will look like.  The better choice, if you want tonemapping applied automatically, is to merge one set of files, tonemap those, save the settings as a preset and invoke that preset.  You can set the batch function to work overnight and come back to the computer in the morning with a folder of fully mergd (and tonemapped if desired) files ready to work on.  It’s a nice feature and it works well.

Photomatix does have an anti-ghosting function which can be used to reduce ghosting due to movement in source images.  There are two options – Background Movements and Moving Objects/People.  The Moving Objects/People option is the more robust and can be set to two levels of detection – Normal and High.  For the deghosting of the tree branches in the test image, I selected Moving Objects/People and set the detection level to High because I wanted to compare the most robust deghosting routine in PM to CS5 HDR Pro.  It works reptty well.  Not as well as HDR Pro, but still pretty well.  There is a bit more remaining, visible movement and a few more ghosting artifacts than in HDR Pro but overall it’s still good.  The detail crop below shows the result.

One problem that does crop up with Photomatix is memory use.  Even now, with a 64 bit OS installed and the 64 bit version of PM installed, I have had the program crash due to insufficient memory.  It’s only happened once, but it has happened.  It happened more often in a 32 bit operating environment.  HDRSoft are aware that users do have issues with memory; particularly in a 32 bit environment where more limited access to RAM exists.  The standard workaround is to use the batch function, which does work.  In the 32 bit version, the programmers at HDRSoft did make a change to the program to enable it to take advantage of the /3 gig switch which can be turned on in the boot.ini menu (more on this in a bit).  This did most definitely help resolve some of the insufficient memory issues but didn’t eliminate them entirely.

Overall, Photomatix is a terrific program.  Up to now it’s been my main HDR processing application.  Whether HDR Pro in CS5 takes over that slot is still to be determined.  It can produce excellent results, up to HDR Pro it was the fastest HDR application I’d used and use support is terrific.

The folks at HDRSoft truly do understand the concept of user support.  They respond very quickly to email inquiries (often the same day or at most the next day).  There is a very good FAQ section on their website as well as an extensive set of HDR-related links.  They may even change the program code if it will help a user.  This is where I’ll come back to the /3 gig switch item noted above.  I had been having problems with PM and insufficient memory.  In trading emails with their support group I mentioned I had this option turned on in the boot.ini menu.  At the time, PM was not set up to use this option.  Based on their discussion with me, they did some research into it, determined how to implement it in Photomatix and included it in their next release version of the program.  That’s damn good support!  There’s no manual alignment option but the alignment function in PM is good enough that it shouldn’t be needed.

PM Pro is available for both Windows and Mac systems, in both 32 and 64 bit versions.  It includes the Lightroom plugin.  Cost is $99 and upgrades are free for ‘at least’ a year according to their website.  I’ve been using the program for a couple of years now at least and have yet to have to pay for an upgrade which includes a migration from v2.x to v3.x.  A version which includes a Photoshop (or Aperture) tonemapping plugin is available for $119.  In my opinion, it’s well worth the price.

As with the previous instalment on HDR Pro, if you see any glaring (or non-glaring) errors, let me know.  Happy to read your comments.