If you read my write-up on the camera in the Nokia 1020 from back in November, you’ll know that I found a number of shortcomings with it. At that time, I noted that there would be a firmware update coming that would give RAW capture to the phone in the form of DNG. It was called the Lumia Black update.
I got the firmware update for my phone recently so wanted to do a write-up on how it performs and do some comparisons. I’ll be comparing to the in-camera JPEG rendering as well as comparing the DNG quality from the 1020 to RAW capability from a Nikon D800 and a Fuji X100s. In all cases White Balance was left on Auto. For the daylight shots I selected an ISO of 100 on the Nokia and Nikon and 200 (the lowest setting) on the Fuji. For the night shots, the ISO was set to 3200 on the Nikon and Fuji and I left it on Auto for the Nokia. On the D800 f8 was used for all images and I used f5.6 on the Fuji.
Now, is it really fair to compare a $600 smartphone to a $3,000 DSLR with a $1,200 lens atrached or to an $1,100 high quality fixed-lens APS-C camera? No, not really. Which is why I’ve tried to remove lens issues as much as possible. I’m not looking at things like barrel/pincushion distortion. I’m not going to look at centre vs. edge clarity. Nor is doing something like measuring resolution charts going to be overly useful. What I have tried to do is look at issues that relate primarily to the sensor and the more subjective issues around overall image quality.
When it comes to converting the RAW images from these cameras, unless otherwise specified, all have been converted to JPEG with only the default settings in Lightroom/Adobe Camera RAW. No other adjustments – again, unless otherwise specified – have been made to the RAW images. No sharpening was applied when resizing the images and exporting in Lightroom.
So, is Black the new RAW? Let’s find out.
I did both daytime and night time shots with all cameras. There will be some slight differences in framing due to the differing focal lengths. On the D800 I used the 17-35 f2.8D AF-S ED. I tried to keep the focal length around 27/28mm but the zoom ring did get bumped around a little on a couple shots. The slight differences in framing won’t have an impact on the image results.
RAW Compared to JPEG
First we’ll look at a comparison of the 1020 DNG to in-camera JPEG. The DNG images from the Nokia are 10 bit as compared to the JPEGs which are 8 bit. This should allow for better dynamic range and smoother colour gradation. In terms of file size, the DNG images are around 48.5MB. A bit smaller than I initially expected but still large and will still fill up the internal storage of the phone fairly quickly.
In each case, the converted RAW image is first followed by the in-camera JPEG. All images are clickable to a larger version.
What do we see? Right off, several of the shortcomings of the in-camera JPEGs disappear in the RAW images. The RAW images have better saturation out of camera, better contrast, greater colour accuracy and markedly better dynamic range. In the second image, in particular, the dynamic range advantage is noticeable. The snow on the ground is overexposed in several areas where in the RAW image there is no overexposure in the snow. What is also evident is the difference in colour. All of the in-camera JPEG images exhibit a marked cyan colour cast that is not present in the RAW images. In the last set of images, we can see how the Nokia handled the mixed lighting and we can see that the RAW image is much superior to the JPEG. It appears that there may be some significant issues with the way that Nokia has programmed the in-camera JPEG rendering engine.
Now let’s take a closer look at the RAW images specifically.
First we’ll look at the histogram for one of the images. What we see is that the histogram is pulled in slightly at both ends. If we were to look at the graph for the JPEG of the same image, we would see that it is hitting at both ends. This will, in some respect, be due to the much higher contrast that is in the JPEG image but it also points to the extra dynamic range that exists in the RAW images. This histogram for the RAW images is consistent across all the images I captured with the exception of things like streetlights and neon lights in the night images which can be considered similar to specular highlights. The RAW images, exhibit noticeably more dynamic range than the JPEGs. This dynamic range advantage could even allow the user to take advantage of some Expose to the Right shooting if so desired.
We’ll take a closer look at a cropped area to see what else we find. Cropping in we see that, even at ISO 100 and with the default noise reduction, there is still some luminance noise remaining in the shot. We also see that there is a bit of chromatic aberration. In the second version of this crop, a small additional amount of luminance noise reduction has been applied along with some capture sharpening and the Auto-CA removal has been turned on. The noise and CA have been cleaned up quite effectively and there has been no loss of resolution from using the noise reduction. A very good result.
What about editing? Can we get a RAW version to look more like the JPEG, or better than the JPEG with some optimisation? The image on the left is the RAW image after editing. The comparison is, of course, subjective but overall I would say that the image on the left is much better than the one on the right. There is none of the unpleasant cyan colour cast. The blue in the clouds is a more accurate representation based on white balance. The band of colour along the horizon was, in fact, more orange as it appears in the left image than yellow as it appears in the JPEG image on the right. The reds in the grass and the yellow/green of the grass showing through the ice are also much more accurate than with the in-camera JPEG.
Now we’ll take a look at how the RAW images from the 1020 compare to those from a D800 and an X100s. In each instance, the Nokia image is first followed by the Nikon and lastly the Fuji. Differences in the image sizes are due to differences in the aspect ratios used by the cameras.
There are certainly differences in the white balance rendering between the cameras and this is to be expected. There are some exposure differences that can be attributed to metering. But what we don’t see are significant differences in the overall colour quality of the images and certainly, at this level of examination, the Nokia holds up quite well.
Now let’s take a closer look. The Nikon is clearly the superior image in all respects. The enhanced noise levels of the Nokia are evident as compared to the other two. But, the Nokia image, despite the noise, has quite good resolution and overall image quality is strong. There is a hint of CA in both the Nikon and Nokia images but none in the Fuji.
There is another positive here as well insofar as the lens of the Nokia is concerned. This sign is closer to the egde of the frame. If we were going to see barrel distortion, it should start to show here. And it doesn’t. We should also, if it were going to happen, see a drop-off in the resolution in this area of the image. And we don’t, really. I know I said I wasn’t going to address those things, but I felt it was worth mentioning in passing.
At the end of it all, there is a pretty startling difference between the RAW capture capability of the Nokia Lumia 1020 and the JPEG quality. While I’m not going to run out and start using the 1020 for commercial shoots, the quality is certainly good enough to be used as a travel camera or for street shooting. Noise is, as is to be expected, an issue even at low ISO but it does clean up very nicely without a loss of resolution, as does the hint of CA that may creep into some shots. It’s not up to top level DSLR quality but no one should expect it would be. It’s not up to very good APS-C camera quality but, again, no one should expect it to be. It is, when using RAW capture, perhaps the best smartphone camera around. And RAW is certainly the way I’ll be using mine going forward.
Nokia has definitely upped the ante in the smartphone photography game with the RAW capability for some of its devices. Now, Apple and Android…. what have you got?