iso 12.800 test sony a7 vs nikon df comparison at Steve Huff

[S3ZAi]

I was taking a look at this topic here at Steve Huffs page:

http://www.stevehuffphoto.com/2013/...noise-reduction-and-low-light/#comment-268421

and I was a bit surprised with a few things:

- the nikon photo is quite a bit brighter

- the sony looks really bad at 100% due to color noise

- sony noise might be due to a darker photo or to 24mp vs 16mp, but there is definitely more details with the sony photo.

I downloaded both photos and checked the exif data, for some reason sony photo shows 0 evf, whereas the nikon photo shows -1.3 evf, this got me thinking that there was more light when the nikon shot was taken, but Steve reassured me this was definitely not the case. Of course there are many other variables and I'm not the biggest pro myself, so I went experimenting with both photos. What I did was, I opened both in photoshop and downsampled both photos to 2000mp wide. As most peoples monitors are 1920x1080, I thought this was a better option than to go pixelpeeping, especially because the photos are jpg's and not raw files. I think the pixelpeeping should be done with the raw files. Anyway, after that I applied topaz noise reduction software, nothing too fancy, I didn't change any variables, but used the strong noise reduction preset with both files. The result was stunning, while everyone praised nikon photo for much more pleasing, I think it sucks after these steps. Whereas the sony file seems very very good, perfectly usable for anything but large prints. But if one would want to make large prints, I'd suggest they use a tripod and raw files.

Anyways, these are the results:

nikon df:


Nikon df, downsampled to 2000mp at iso 12.800 external nr applied

sony A7:


Sony A7, downsampled to 2000mp at iso 12.800 external nr applied

I know this all doesn't say anything, the way I did this is very questionable, but so is the comparison Steve Huff made. I think when pixelpeeping sony photo seems worse, but with some work it is the other way around. And this is just the jpg files, I think the raw files will really show the difference. Also, the nikon set up is 1700$ more expensive than the sony setup.

Also, take into account that dpreview compression makes both photos a bit worse I believe.

Copyright both photos belongs to Steve Huff.

 

[deep7]

I downloaded Steve's original files too. It is patently clear that the Sony's jpeg conversion (in-camera) is quite messy, producing an ugly, artificial pattern as sharpening and noise control fight each other, despite both being set to zero. The Nikon's file was also quite ugly and lacked detail, though the grainy look seems less processed. I am convinced, though, that a well-converted RAW file from the Sony would have trounced the Nikon, which lacked so much detail and had cramped dynamic range.

I don't have a foot in either camp. I shoot Canon but am contemplating either moving back to Olympus or trying Fuji. I'm still interested in seeing these comparisons and like your practical take on this one. Really amazing how usable your worked images are!

 

[Donny out of Element here]

you don't have to do what you have done. Just look at originals at Steve Huff's site and you will see that there is just no comparison between the 2 cameras: A7 blows Df in details. Df just can't even come close and that is one of the best low-light sensors out there. That is why Sony A7/R cameras are actually astonishing in many ways. In fact that ISO12800 shot on Df looks like a P&S cheap camera, while Sony A7 remained so much detail. It's like night and day. In my opinion after that comparison Nikon should be ashamed - yes ashamed, since their main strength was not in Mp (Sony wins here too) but in high ISO performance. And Nikon screwed big time, compared to Sony A7. That is why in real life comparisons some cameras don't look that good like they should on paper (Df), while other cameras (A7) outperform and shine even in something that they are not considered to be best of the best.

 

[Clayton1985]

I downloaded Steve's original files too. It is patently clear that the Sony's jpeg conversion (in-camera) is quite messy, producing an ugly, artificial pattern as sharpening and noise control fight each other, despite both being set to zero. The Nikon's file was also quite ugly and lacked detail, though the grainy look seems less processed. I am convinced, though, that a well-converted RAW file from the Sony would have trounced the Nikon, which lacked so much detail and had cramped dynamic range.

I don't have a foot in either camp. I shoot Canon but am contemplating either moving back to Olympus or trying Fuji. I'm still interested in seeing these comparisons and like your practical take on this one. Really amazing how usable your worked images are!

 

[S3ZAi]

I downloaded Steve's original files too. It is patently clear that the Sony's jpeg conversion (in-camera) is quite messy, producing an ugly, artificial pattern as sharpening and noise control fight each other, despite both being set to zero. The Nikon's file was also quite ugly and lacked detail, though the grainy look seems less processed. I am convinced, though, that a well-converted RAW file from the Sony would have trounced the Nikon, which lacked so much detail and had cramped dynamic range.

I don't have a foot in either camp. I shoot Canon but am contemplating either moving back to Olympus or trying Fuji. I'm still interested in seeing these comparisons and like your practical take on this one. Really amazing how usable your worked images are!

--
Don.
A Land Rover, a camera ... I'm happy!

I agree.... why not just do this comparison with RAW or at least show it both ways?

The thing is, there is still no software to process sony raw files properly, perhaps the same applies to the nikon raw files, so any comparison for now has to made with jpg's. At least I'm sure the LR 5.3 RC doesn't do a proper job yet, dunno about other software for sure though.

I was just very very surprised about the fact that the nikon exif shows -1.3 evf. I added 1.3 exposure to the nikon photo and it got totally way more bright than the sony photo. I am almost sure something didn't add up and while shooting the nikon there was more light. The sony photo seems more noisy because it was darker. I don't know what else to make of it.

But both being jpg's it doesn't really matter anyway, downsampled I think sony looks better to my taste, but the nikon is just as astonishingly good at iso 12.800. I will be very very happy when I finally receive my A7r to shoot some 'christmassy' scenes. I just wish I had the 55 1.8 already though...

 

[deep7]

I downloaded Steve's original files too. It is patently clear that the Sony's jpeg conversion (in-camera) is quite messy, producing an ugly, artificial pattern as sharpening and noise control fight each other, despite both being set to zero. The Nikon's file was also quite ugly and lacked detail, though the grainy look seems less processed. I am convinced, though, that a well-converted RAW file from the Sony would have trounced the Nikon, which lacked so much detail and had cramped dynamic range.

I don't have a foot in either camp. I shoot Canon but am contemplating either moving back to Olympus or trying Fuji. I'm still interested in seeing these comparisons and like your practical take on this one. Really amazing how usable your worked images are!

 

[Lightshow]

I downloaded Steve's original files too. It is patently clear that the Sony's jpeg conversion (in-camera) is quite messy, producing an ugly, artificial pattern as sharpening and noise control fight each other, despite both being set to zero. The Nikon's file was also quite ugly and lacked detail, though the grainy look seems less processed. I am convinced, though, that a well-converted RAW file from the Sony would have trounced the Nikon, which lacked so much detail and had cramped dynamic range.

I don't have a foot in either camp. I shoot Canon but am contemplating either moving back to Olympus or trying Fuji. I'm still interested in seeing these comparisons and like your practical take on this one. Really amazing how usable your worked images are!

 

[deep7]

Worst test I've seen in a while, ISO the same? Check. Aperture the same? Check. Shutter speed the same? Check, but not that important. EC the same? Nope, EC changes the ISO/gain so the noise levels don't look right for the ISO. Good target subject? Nope, sigh..... Waste of effort and bandwidth.

--
http://www.flickr.com/photos/lightshow-photography/
My lenses:
http://www.fredmiranda.com/forum/viewprofile.php?Action=viewprofile&username=LightShow

That's not quite right, unless you are using a manual setting with "auto ISO", which is definitely not the case here. Exposure compensation only changes the shutter speed or aperture, depending on how you've set your camera.

However, in this situation, the main fixed variable being tested is image quality at a particular "ISO". Steve set both cameras with the same sensitivity, aperture and shutter speed, which makes the test very valid indeed. We see a difference in detail rendition, noise control, exposure and vignetting, so it's not only the qualities of the sensors themselves affecting the image but also the lenses and the manufacturers decisions regarding correct exposure and sensitivity ratings. This is much more relevant for potential purchasers than you seem to realise.

If you read what I wrote above, you'll note that lots of factors can be involved in the reported exposure compensation difference but the one I neglected, which seems to have some bearing in this case, is simply how the cameras meter.

Why do you think the target subject is not good? I think it's excellent as it has the sort of dark tones which are excruciatingly difficult for many sensors to cope with in bad light, plus enough highlight to show which sensor has most headroom at that setting.

--
Don.
A Land Rover, a camera ... I'm happy!

 

[commiebiker]

I have a D4 and now an A7 (not R) Much of my photography is done under very poor lighting conditions, and as soon as this stupid ice storm passes, I'll take both out together and see how it goes

 

[Dr hatem Ayman]

well guys, relax, who cares if nikon has a slightly better edge re: noise. the A7 still looks awesome at 12800 for a camera that is at least 1/3 rd the size of the nikon dinosaur. I already have the d800 but I am sick and tired of the weight and the sheer size of that thing, the d800 is so 1890s to me not 21st century at all. Miniaturization comes at a price guys but luckily sony made us pay a very little price for the size (noise and IQ wise)

 

[Dr hatem Ayman]

but if you DO want to compare there is significantly MORE detail on the Sony which I cant understand, I thought the nikon would 'theoretically' be leaps ahead but if you look at the logo you will be shocked