high iso comparisonHigh ISO testing (Canon 6D, Fuji X-T1, Sony A7R)

For years now I've been impressed that I can even create an image at all at ISOs of 6400 and higher. It reminds me of the old talking dog story – it's not what he says so much as it is that he can talk at all. Lately, though, I find that I can make truly usable images at such stratospheric ISO settings.

Some of the work I do is event coverage and it often requires shooting under very low light conditions without strobes. Most often the images are for electronic use only – web viewing, internal company recaps and presentations, projected slideshows and PR related web placement. So for quite a while now I've realized that (in many cases) I can set my cameras to auto ISO (6400) and be comfortable that the images will hold up just fine for their intended use. For these such events I shoot RAW+JPG and only deal with the RAW files if the camera's JPG engine didn't do a good enough job – and I'm finding that the Fujis that I usually use for these events usually do a good enough job in that regard.

This Sunday morning I thought that I would, before starting my print work, play a little bit and compare the high ISO out of camera JPGs of my three newest cameras:

• Canon 6D
• Fuji X-T1
• Sony A7R

This is not so much of a test as it is a simple, very unscientific and non-rigorous comparison of the OOC JPG output of these three rather different photographic tools.

As for lenses, I used the Canon EF 50/1.4 on both the 6D and A7R (via the Metabones III adapter). On the Fuji I used the Fujinon 35/1.4. In both cases the aperture was set to f/8.

First off I used an external, incident light meter to determine exposure since each camera's built in meter came up with exposure. On one hand there is some value in comparing the cameras' exposure choices but what I wanted to compare was the amount and type of noise they produced – and in some cases the exposure difference was an entire stop. That might have made a difference as to how one camera better retained highlight or shadow detail however that's not the route I took.

Second, I turned off high ISO noise reduction for each camera (or set it to the lowest available setting). That's a personal choice, I guess, as I don't mind "grain like" noise – at least I prefer it infinitely over detail smearing noise reduction. White balance was set to auto.

Third, I re-sized all of the images to the pixel dimensions of the smallest camera's file size – the 16MP Fuji X-T1. I'm sure there are better ways to down sample the larger files that would be advantageous to the 20MP Canon or the 36MP Sony but again – this is not a robust scientific comparison.

Could the results from each camera have been improved by dealing with the RAW files instead of relying on the OOC JPGs? Almost certainly. 

I'm offering this comparison because  A) someone out there might find it useful or satisfy some curiosity they may have and  B) I actually find this sort of thing interesting and, on some level, mildly entertaining. Sure I prefer to shoot "real" things whenever I can but I also find value in gaining a better understanding of the capabilities of my equipment in my down time.

Feel free to look at these images and draw whatever conclusions you may from them.

Here are a few things that I think I see:

• The Sony's JPG dynamic range and color good but I don't like the "texture" of it's JPG rendering – especially at higher ISOs.
• The Canon high ISO noise pattern is the most pleasing.
• Both the Sony and Canon show much more chroma noise than the Fuji.
• The Fuji's color appears to me to be the most accurate of the three, under these conditions. The Canon is probably the least accurate (which surprised me. I can probably tweak the JPG color in-camera fairly easily. I think I'll have a look at that.)
• For an APSC sized sensor I think the Fuji is exceptional. Again, we're looking at ISO 6400, 12,800 and 25,600 shots!

My conclusions:

• Usable images at ISO 25,600 from them all (certainly for web sized viewing). To me that is amazing.
• At ISO 12,800+ the most pleasing overall is probably the Canon, but the Fuji is a very close second. I find the Canon has better noise structure and significantly better detail but the Fuji has better color and much cleaner chroma noise.
• In my opinion it's pretty easy to improve on the Sony's JPG rendering by using the RAW file. Even still, as bad as I think the high ISO Sony JPGs are their detail is still as good as that of the Canon in this comparison.

Again, I'm dumbing down the Canon and Sony files for the purpose of this test. Unfair, sure.

I'd appreciate hearing anyone else's opinions and conclusions as well.