6D vs 70D High ISO performance comparison

TeamSpeed wrote in post #16781630
What software to process the raw?

Photoshop CC/Adobe Camera RAW

palad1n wrote in post #16781544
you mean jpeg settings in camera or RAW process without NR to jpeg?

i used jpeg camera settings standard NR.
RAW conversion without any NR, just cropping in PS.


I don´t see much difference on RAW images even without NR except shadow areas... but this is subjective... true fact is : post-processed RAW picture from 70D beats SOOC jpeg 6D very easily, so if anyone is shooting FF SOOC jpeg only, he/she should reconsider shooting RAW, otherwise FF image quality suffers a lot loosing potential.


and after some PP work on both images (using several layers with different NR settings and masks, there are literally no differences at all.. All details can be retained on both cameras...


MakisM1: true, 70D simply can´t beat 6D in ISO 25600.


gabebalazs: Gabe, it seems logical, but this maybe applies on SOOC jpeg quality, RAW with PP looks almost equally the same. You have to look very carefully on 100% level to see any minor difference.

Mutually exclusive statements. Both can't be true.

SqueekyBoy wrote in post #16781740
Mutually exclusive statements. Both can't be true.

Whole test was about ISO 6400... My statement was logically related to ISO 6400...

70D can´t beat 6D in ISO 25600 because hardware limitation... It is the same thing as 6D ISO 102400 is still boosted version of ISO 25600, same as 70D ISO 25600 is still actually ISO12800, just boosted by camera software...

palad1n wrote in post #16781729
Photoshop CC/Adobe Camera RAW

I have to ask because I see no color noise at all on both, and if there were no NR at all, and you just looked at the raw data, there would indeed be chroma noise.

We know that based on raw content alone, that:
- if we use the 7D as a baseline
- the 70D is about 1/3 stop better or so
- the 5D3 is about 1 2/3 stops better than the 7D
- the 6D is about another 1/3 stop better than the 5D3

Therefore it is logical that the 6D is about 1 2/3 stop better than the 70D. It also has been shown a few times before by various members, as well as downloadable from DPReview.

I am not sure what you are exactly doing, but what I am seeing cannot be the straight raw without NR results in your crops, both cameras will show some sort of chroma noise if there was no NR. There isn't any noise at all really, other than a tad bit of luminescent noise, which means those images have been processed by something somewhere. I am a big advocate of crop performance, but these results go against everything I have seen, processed myself, and read to this point.

EDIT: also when I take the two crops and equalize them in photoshop (which will expose everything, noise, dirt on the sensor, etc), there is nothing, the crops are too clean to be just straight raw output without any processing. I would post those results, but your imaging edit profile isn't set to OK. This is the function I use to look for issues in images or from the sensor itself or dirt on the sensor, and any noise or issues become accentuated. For example, the 7D at 6400 with ZERO NR looks like the below, and when you equalize it, it shows the noise characteristics in an exaggerated fashion, but is very useful for comparison. There is entirely no chance that the 70D is 2 stops better than the 7D.

Could you please provide the raw files from the 6D and 70D so that those of us that have done these back to back comparisons over the years could take a look?

palad1n wrote in post #16780222
70D crop ISO6400 NO_NR (RAW)

Hosted photo: posted by palad1n in
./showthread.php?p=167​80222&i=i177863929
forum: Canon Digital Cameras

6D crop ISO6400 NO_NR (RAW)

Hosted photo: posted by palad1n in
./showthread.php?p=167​80222&i=i4963188
forum: Canon Digital Cameras

differences are visible, especially in shadow areas, but this is 100% crop.

Lukas, your full frame shots are either already cropped, or your 100% crops aren't 100%, not even close; I suspect the latter.

When I first looked at the the claimed "100% crops", I already saw that they weren't true 100% crops just based on the size of the noise grain specks and actually how clean they looked for ISO6400.

Then I saved both images, blew up the whole frame to 5472 pixels, then superimposed the claimed 100% crop. If the 100% crop was really a true 100% crop, then the Canon logo would be exactly the same size on both. But the whole frame logo after uprezzing to 5472 is about twice as big as the claimed 100% crop. So the crop is either not a 100% crop, or the whole frame image had been cropped too.

TeamSpeed wrote in post #16781801
I have to ask because I see no color noise at all on both, and if there were no NR at all, and you just looked at the raw data, there would indeed be chroma noise.

We know that based on raw content alone, that:
- if we use the 7D as a baseline
- the 70D is about 1/3 stop better or so
- the 5D3 is about 1 2/3 stops better than the 7D
- the 6D is about another 1/3 stop better than the 5D3

Therefore it is logical that the 6D is about 1 2/3 stop better than the 70D. It also has been shown a few times before by various members, as well as downloadable from DPReview.

I am not sure what you are exactly doing, but what I am seeing cannot be the straight raw without NR results in your crops, both cameras will show some sort of chroma noise if there was no NR. There isn't any noise at all really, other than a tad bit of luminescent noise, which means those images have been processed by something somewhere. I am a big advocate of crop performance, but these results go against everything I have seen, processed myself, and read to this point.

EDIT: also when I take the two crops and equalize them in photoshop (which will expose everything, noise, dirt on the sensor, etc), there is nothing, the crops are too clean to be just straight raw output without any processing. I would post those results, but your imaging edit profile isn't set to OK. This is the function I use to look for issues in images, and any noise at all gets accentuated.

Could you please provide the raw files from the 6D and 70D so that those of us that have done these back to back comparisons over the years could take a look?

i tested it once again and you are right, even if the NR slider in Camera RAW is 0, ACR removes all chroma noise. But still... No major details or color information are lost, ACR is very effective in this and still it looks much much better than any SOOC jpeg.

I think Photoshop/ACR/Lightroo​m is great investment for improving IQ from any camera.

Here are my samples, ISO 6400.
RAW, from ACR, no NR (not even chroma), sharpness at 0

Whole image, then 100% crop:
70D, then 6D

gabebalazs wrote in post #16781815
Lukas, your full frame shots are either already cropped, or your 100% crops aren't 100%, not even close; I suspect the latter.

When I first looked at the the claimed "100% crops", I already saw that they weren't true 100% crops just based on the size of the noise grain specks and actually how clean they looked for ISO6400.

Then I saved both images, blew up the whole frame to 5472 pixels, then superimposed the claimed 100% crop. If the 100% crop was really a true 100% crop, then the Canon logo would be exactly the same size on both. But the whole frame logo after uprezzing to 5472 is about twice as big as the claimed 100% crop. So the crop is either not a 100% crop, or the whole frame image had been cropped too.

Both images from both camera have same crop area.

palad1n wrote in post #16781832
i tested it once again and you are right, even if the NR slider in Camera RAW is 0, ACR removes all chroma noise. But still... No major details or color information are lost, ACR is very effective in this and still it looks much much better than any SOOC jpeg.

I think Photoshop/ACR/Lightroo​m is great investment for improving IQ from any camera.

The tools make a huge difference in results, and even though the canon JPG engine is much improved, it still removes a ton of detail and they run very aggressive NR values with the latest bodies.

If you are willing to share the 70D and 6D raw images, I can do my usual comparisons between them and post the results here in this thread. I will also run my custom noise actions to see how well they clean up while keeping details.

gabebalazs wrote in post #16781834
Here are my samples, ISO 6400.
RAW, from ACR, no NR (not even chroma), sharpness at 0

Whole image, then 100% crop:

HTTP response: NOT FOUND | MIME changed to 'text/html'

Yes, that was my fault, the difference was i didn´t pulled the chroma levels in ACR, so it left the noise itself, but removed all chroma noise... Still, ACR does terrific job...

palad1n wrote in post #16781775
Whole test was about ISO 6400... My statement was logically related to ISO 6400...

70D can´t beat 6D in ISO 25600 because hardware limitation... It is the same thing as 6D ISO 102400 is still boosted version of ISO 25600, same as 70D ISO 25600 is still actually ISO12800, just boosted by camera software...

Ok, thanks for the clarification. I will compare my best processed 60D to SOOC from my 6D at 6400. Interesting proposition. Although I must admit, if I know I'm going to that extreme, I'm taking the 6D. I always shoot raw.

SOOC jpegs, 100%

Gabe, would you mind if I post 2 small sections from your example above, where I show what I mean about both the chroma noise and equalizing to show the differences?

palad1n wrote in post #16781835
Both images from both camera have same crop area.

That's perfectly fine and fair, but you still shouldn't state that they are 100% crops.
Just imagine people looking at them with amazement that WOW, both the 6D and 70D are awesome at 100% crop at ISO 6400. It may confuse people.

TeamSpeed wrote in post #16781847
Gabe, would you mind if I post 2 small sections from your example above, where I show what I mean about both the chroma noise and equalizing to show the differences?

Sure, go ahead, thanks for asking. I can even send you the RAW files if you'd like me to.

My NR'd samples will be next I post