Canon 5D3 vs Nikon 800E noise tests

Started 6 months ago | Discussions
kymarto
Contributing MemberPosts: 632
Like?
Canon 5D3 vs Nikon 800E noise tests
6 months ago

Hi folks,
Vis-a-vis the endless debate about these two cameras: I am privileged to use both regularly, one for myself (Nikon) and the other for work (Canon). I was reading again as the opposing fanboys went at each other. I actually like both. The Canon feels better in my hand, but I like the Nikon layout much more. On the other hand the Canon has (to my mind) a more sensible menu layout, and the bracketing controls are much better, allowing more than 1 EV between shots. Advantages and disadvantages both ways...

But for me the main point is image quality. I was very interested to see how the two cameras compare in terms of noise. I had read that the Canon was noisy in the shadows at base ISO, but performed better than the Nikon at high ISOs. So yesterday I took the Canon home and performed two head-to-head tests.

First I wanted to compare shadow noise. I shot two frames (raw) each, at 100 and 800 ISO, underexposed by 3 EV. I then imported them into ACR and brought them up 3 stops, and performed some minor tweaks to get the contrast presentable and the the color balance similar between the two cameras.
The shots are all at 24mm @ f5.6, the Nikon with the 14-24 and the Canon with the 16-35 II. Nikon shooters, be glad, the Canon lens can't hold a candle to the Nikon in the corners
Anyway I then cropped out similar areas in the four shots and lined them up. If you look at the file I uploaded at original size, the image magnification is at 150%
The second image is done the same way, but shooting both cameras at 6400 and 12800 ISO (no exposure compensation).
My impressions are these: as to the underexposure noise test, I was shocked. The Nikon at 100 is very clean, but the Canon is pretty miserable, especially in terms of shadow noise (both luma and chroma). The Nikon at 800 is almost comparable to the Canon at 100, actually noisier in the mids and highs, but cleaner in the shadows.
The high ISO test also surprised me (second image). I kept hearing that the Canon was clearly superior to the Nikon in terms of high ISO noise, but that is not my impression from my test. Perhaps Canon has better noise reduction algorithms in the processor for jpgs, but this was a head-to-head test with CR2 and NEF. To my eye the Nikon is actually slightly better than the Canon at both ISOs, especially in the shadows. Have a look and see what you think.

Tests at -3 EV underexposure

High ISO tests (no exposure compensation)

Canon EOS 5D Mark III Nikon D800E
If you believe there are incorrect tags, please send us this post using our feedback form.
ImageAmateur
Senior MemberPosts: 1,488
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to kymarto, 6 months ago

Jujst wondering, would be nice to see the original image also (at least to see size and crop area).

Also, confirming that you shot RAW?

That said,

- the Canon 12800 shots actually, despite on first glance looking to have less detail, actually are cleaner (marginally). Looking at the central window mesh, the Nikon image loses a portion of mesh to the right of the white oval in the window, whereas the mesh is barely visible in the Canon image, so noise is stronger in the Nikon image.

Also in the 12800 image, under the window chroma noise seems stronger in the  Nikon image, Canon image looks to have retained a general level of less chroma noise.

Query, both 12800 images look more saturated than the 6400 images, has this come straight from the RAW conversion?

I think both have done well, considering one is talking up to ISO 12800 and pulling up shadows 3 stops.

Not much between them.

Cheers.

-- hide signature --

Wishing You Good Light.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,943Gear list
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to kymarto, 6 months ago

kymarto wrote:

My impressions are these: as to the underexposure noise test, I was shocked. The Nikon at 100 is very clean, but the Canon is pretty miserable, especially in terms of shadow noise (both luma and chroma). The Nikon at 800 is almost comparable to the Canon at 100, actually noisier in the mids and highs, but cleaner in the shadows.

The high ISO test also surprised me (second image). I kept hearing that the Canon was clearly superior to the Nikon in terms of high ISO noise, but that is not my impression from my test. Perhaps Canon has better noise reduction algorithms in the processor for jpgs, but this was a head-to-head test with CR2 and NEF. To my eye the Nikon is actually slightly better than the Canon at both ISOs, especially in the shadows. Have a look and see what you think.

I assume you downsampled the D800e to the same size as the 5DIII's images and applied some sharpening - the noise shows the characteristic artifacts introduced by deconvolution and may actually make the images look noisier than they are, so perhaps it would be better to either not use any sharpening at all for such a test - or just only minimally sharpen both images to make them look approximately equally 'sharp'.

At the lower ISOs underexposure exposes the more limited DR of the 5DIII, amplifying its weakness there. Things would not look so bad around middle gray without underexposure, as shown in your higher ISO shots which confirm results from DxO's more controlled tests: the 'e' has (slightly) better 18%SNR throughout the range...

Perhaps the 5DIII's reputation in the shadows has more to do with its excellent focusing system than with sensor related IQ?

Jack

Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to ImageAmateur, 6 months ago

ImageAmateur wrote:

Jujst wondering, would be nice to see the original image also (at least to see size and crop area).

Also, confirming that you shot RAW?

That said,

- the Canon 12800 shots actually, despite on first glance looking to have less detail, actually are cleaner (marginally). Looking at the central window mesh, the Nikon image loses a portion of mesh to the right of the white oval in the window, whereas the mesh is barely visible in the Canon image, so noise is stronger in the Nikon image.

Also in the 12800 image, under the window chroma noise seems stronger in the Nikon image, Canon image looks to have retained a general level of less chroma noise.

Query, both 12800 images look more saturated than the 6400 images, has this come straight from the RAW conversion?

I think both have done well, considering one is talking up to ISO 12800 and pulling up shadows 3 stops.

Not much between them.

Cheers.

-- hide signature --

Wishing You Good Light.

Yes, both shot RAW (CR2 and NEF).
There were differences in the characteristics in ACR. The Canon shots were more saturated and darker by about 1/2 stop (at rated ISO, f stop and speed). I adjusted so that they were more or less similar-looking, meaning boosting brightness a bit in the Canon images and color balance in the Nikon (+5 red and +5 yellow, midtones).
This was not a formal test, but a perceptive one. I just wanted to see what I could expect from images in each camera at those ISO when I went in to post process and "optimize" them.
Yes, I agree that they are quite similar, which surprised me since I keep hearing how much better the 5D3 is at high ISOs.
The underexposure test is another matter. The Canon really isn't in the league of the Nikon at base ISO, although the Nikon images deteriorate more rapidly as ISOs are boosted. Still, at 800 the Nikon is clearly better than the Canon.

Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to Jack Hogan, 6 months ago

I believe that ACR has a default sharpening of 25% @ 1px, which I unfortunately did not remember to turn off. No other sharpening. The Nikon was downrezzed to 5760 to equal the Canon, but then in this example both were uprezzed to 150% at 100% viewing size of the composite image, so it's all a bit dodgy, but indicative, I think of what one can expect.

Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: Canon 5D3 vs Nikon 800E noise tests--full image
In reply to kymarto, 6 months ago

Here is the original Nikon 6400 ISO image, downrezzed to 3000x2002 px so it would upload, straight from NEF to ACR to medium quality jpg in PS.

Reply   Reply with quote   Complain
cpkuntz
Contributing MemberPosts: 671Gear list
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to kymarto, 6 months ago

The general consensus I've always heard is that noise may be a bit more noticeable at 100% in the D800, but its greater resolution effectively cancels any slight advantage at 100% a 5D III file may have, so it turns out either to be a wash or, more likely, cleaner overall at high ISO in the D800.  These tests confirm that consensus.  Who ways that the 5D3 is cleaner?    It is a minority view in conversations I've had, and wrong, too.  I suppose it does have more aggressive jpeg noise reduction, but this is a bad thing, and the 5D3 is known for its mushy, undetailed jpegs.  DXO Optics is good for getting every last drop of detail from a 5D3 raw file, and sometimes they give results that can be printed to about 19 x 11" and look quite detailed.  You can get away with 16* by 24* and have the same detail in a D800 file, though.  Of course, the significant advantage in read noise means that it's much easier to get clean shadows at base ISO after positive exposure compensation from the D800.  Canon needs to work on read noise if it hopes to ever catch up with Nikon in image quality.  That said, both are nice cameras.  I have the 5D3 and enjoy it but a good friend has the D800 - I shoot with it once in a while and I love those D800 files.

 cpkuntz's gear list:cpkuntz's gear list
Canon EOS 7D Canon EOS 5D Mark III Nikon D810 Canon EF 35mm f/1.4L USM Canon EF 50mm f/1.2L USM +8 more
Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to cpkuntz, 6 months ago

I use one for work and the other personally. Yes, both are really nice cameras. I find also that there was a big improvement in build quality and feel, to the point where I prefer the feel of the Canon in my hand, although I like the button layout and dual wheels of the Nikon much more.

Reply   Reply with quote   Complain
sandy b
Veteran MemberPosts: 7,486Gear list
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to kymarto, 6 months ago

Feel is subjective, I have Canon buds who love the feel of the Canon, I kind of agree, although I'm used to the D600. As far as build quality, the 5D3 is more akin to the D600 than D800, less steel, more plastic, although thats not always bad, I like the quality of My D600, but the D800 is clearly better.

Thank you for your interesting project.

 sandy b's gear list:sandy b's gear list
Nikon Coolpix P330 Nikon D7000 Nikon D600 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Teleconverter TC-14E II +3 more
Reply   Reply with quote   Complain
aut0maticdan
Contributing MemberPosts: 672Gear list
Like?
for some reason I came here expecting shutter sound decibels :) /nt
In reply to kymarto, 6 months ago
No text.
 aut0maticdan's gear list:aut0maticdan's gear list
Fujifilm X100S Nikon D800 Sony Alpha 7S Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED +8 more
Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: for some reason I came here expecting shutter sound decibels :) /nt
In reply to aut0maticdan, 6 months ago

Oops, sorry to be imprecise. I could do that too, but I don't really care about the decibel levels. I think the Canon is slightly quieter but the Nikon shutter is crisper--is my subjective feeling...

Reply   Reply with quote   Complain
Kabe Luna
Veteran MemberPosts: 9,351Gear list
Like?
No surprise here
In reply to kymarto, 6 months ago

The D800 beats the snot out of the 5DIII in any measure of or aspect of IQ dependent upon dynamic range and low read noise through ISO 1600. At this point, for anyone with eyes to see realizes that. However, high ISO in ample daylight is not the test of a camera's high ISO capabilities, I feel. Test these two in dim artificial (tungsten) light beyond ISO 1600, and you'll see the 5DIII fairly easily run away from the D800. The D800's high ISO weakness is the reason I've supplemented it with a Dƒ.

kymarto wrote:

Hi folks,
Vis-a-vis the endless debate about these two cameras: I am privileged to use both regularly, one for myself (Nikon) and the other for work (Canon). I was reading again as the opposing fanboys went at each other. I actually like both. The Canon feels better in my hand, but I like the Nikon layout much more. On the other hand the Canon has (to my mind) a more sensible menu layout, and the bracketing controls are much better, allowing more than 1 EV between shots. Advantages and disadvantages both ways...

But for me the main point is image quality. I was very interested to see how the two cameras compare in terms of noise. I had read that the Canon was noisy in the shadows at base ISO, but performed better than the Nikon at high ISOs. So yesterday I took the Canon home and performed two head-to-head tests.

First I wanted to compare shadow noise. I shot two frames (raw) each, at 100 and 800 ISO, underexposed by 3 EV. I then imported them into ACR and brought them up 3 stops, and performed some minor tweaks to get the contrast presentable and the the color balance similar between the two cameras.
The shots are all at 24mm @ f5.6, the Nikon with the 14-24 and the Canon with the 16-35 II. Nikon shooters, be glad, the Canon lens can't hold a candle to the Nikon in the corners
Anyway I then cropped out similar areas in the four shots and lined them up. If you look at the file I uploaded at original size, the image magnification is at 150%
The second image is done the same way, but shooting both cameras at 6400 and 12800 ISO (no exposure compensation).
My impressions are these: as to the underexposure noise test, I was shocked. The Nikon at 100 is very clean, but the Canon is pretty miserable, especially in terms of shadow noise (both luma and chroma). The Nikon at 800 is almost comparable to the Canon at 100, actually noisier in the mids and highs, but cleaner in the shadows.
The high ISO test also surprised me (second image). I kept hearing that the Canon was clearly superior to the Nikon in terms of high ISO noise, but that is not my impression from my test. Perhaps Canon has better noise reduction algorithms in the processor for jpgs, but this was a head-to-head test with CR2 and NEF. To my eye the Nikon is actually slightly better than the Canon at both ISOs, especially in the shadows. Have a look and see what you think.

Tests at -3 EV underexposure

High ISO tests (no exposure compensation)

-- hide signature --
 Kabe Luna's gear list:Kabe Luna's gear list
Nikon D800 Nikon Df Nikon AF-S Nikkor 16-35mm f/4G ED VR Sigma 50mm F1.4 EX DG HSM Nikon AF-S Nikkor 85mm f/1.8G +8 more
Reply   Reply with quote   Complain
Horshack
Veteran MemberPosts: 4,226
Like?
Re: No surprise here
In reply to Kabe Luna, 6 months ago

Kabe Luna wrote:

The D800 beats the snot out of the 5DIII in any measure of or aspect of IQ dependent upon dynamic range and low read noise through ISO 1600. At this point, for anyone with eyes to see realizes that. However, high ISO in ample daylight is not the test of a camera's high ISO capabilities, I feel. Test these two in dim artificial (tungsten) light beyond ISO 1600, and you'll see the 5DIII fairly easily run away from the D800. The D800's high ISO weakness is the reason I've supplemented it with a Dƒ.

Can you show a demonstration of the 5DIII "running away" vs the D800 at High ISO?

Reply   Reply with quote   Complain
Kabe Luna
Veteran MemberPosts: 9,351Gear list
Like?
If I still owned a 5DIII, I could...
In reply to Horshack, 6 months ago

Horshack wrote:

Kabe Luna wrote:

The D800 beats the snot out of the 5DIII in any measure of or aspect of IQ dependent upon dynamic range and low read noise through ISO 1600. At this point, for anyone with eyes to see realizes that. However, high ISO in ample daylight is not the test of a camera's high ISO capabilities, I feel. Test these two in dim artificial (tungsten) light beyond ISO 1600, and you'll see the 5DIII fairly easily run away from the D800. The D800's high ISO weakness is the reason I've supplemented it with a Dƒ.

Can you show a demonstration of the 5DIII "running away" vs the D800 at High ISO?

...the OP has both, so maybe he'll supplement his initial test with a series in low tungsten light. Amp noise on the D800 alone makes dynamic range and color fidelity at ISOs higher than 3200 progressively worse than one might anticipate.

Have you used both at high ISO?

-- hide signature --
 Kabe Luna's gear list:Kabe Luna's gear list
Nikon D800 Nikon Df Nikon AF-S Nikkor 16-35mm f/4G ED VR Sigma 50mm F1.4 EX DG HSM Nikon AF-S Nikkor 85mm f/1.8G +8 more
Reply   Reply with quote   Complain
Horshack
Veteran MemberPosts: 4,226
Like?
Re: If I still owned a 5DIII, I could...
In reply to Kabe Luna, 6 months ago

Kabe Luna wrote:

Horshack wrote:

Kabe Luna wrote:

The D800 beats the snot out of the 5DIII in any measure of or aspect of IQ dependent upon dynamic range and low read noise through ISO 1600. At this point, for anyone with eyes to see realizes that. However, high ISO in ample daylight is not the test of a camera's high ISO capabilities, I feel. Test these two in dim artificial (tungsten) light beyond ISO 1600, and you'll see the 5DIII fairly easily run away from the D800. The D800's high ISO weakness is the reason I've supplemented it with a Dƒ.

Can you show a demonstration of the 5DIII "running away" vs the D800 at High ISO?

...the OP has both, so maybe he'll supplement his initial test with a series in low tungsten light. Amp noise on the D800 alone makes dynamic range and color fidelity at ISOs higher than 3200 progressively worse than one might anticipate.

Have you used both at high ISO?

Yes. I did a comparison here: http://www.dpreview.com/forums/post/41811059

Reply   Reply with quote   Complain
benjaminblack
Contributing MemberPosts: 543Gear list
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to cpkuntz, 6 months ago

cpkuntz wrote:

The general consensus I've always heard is that noise may be a bit more noticeable at 100% in the D800, but its greater resolution effectively cancels any slight advantage at 100% a 5D III file may have, so it turns out either to be a wash or, more likely, cleaner overall at high ISO in the D800. These tests confirm that consensus. Who ways that the 5D3 is cleaner? It is a minority view in conversations I've had, and wrong, too. I suppose it does have more aggressive jpeg noise reduction, but this is a bad thing, and the 5D3 is known for its mushy, undetailed jpegs. DXO Optics is good for getting every last drop of detail from a 5D3 raw file, and sometimes they give results that can be printed to about 19 x 11" and look quite detailed. You can get away with 16* by 24* and have the same detail in a D800 file, though. Of course, the significant advantage in read noise means that it's much easier to get clean shadows at base ISO after positive exposure compensation from the D800. Canon needs to work on read noise if it hopes to ever catch up with Nikon in image quality. That said, both are nice cameras. I have the 5D3 and enjoy it but a good friend has the D800 - I shoot with it once in a while and I love those D800 files.

This is pretty spot on. When shooting RAW they're pretty close, maybe with the D800 gaining the upper hand. This is an achievement that most of the biggest names in our industry said was 100% impossible. They said "you can't beat physics, and when you cram that many pixels into a sensor you're going to get a lot of noise." The big names were wrong. The D800 is an engineering marvel in that sense. Even if you conclude the 5D3 and D800 are even in high ISO, well then you have to look at DR and resolution. And we all know who wins that battle

However, when shooting jpg, I understand that the 5D3 has less noise at high ISO's. According to Chuck Westfall this is because "the 5D3 was designed for journalists and entry level sports photographers." And we all know many of those photographers shoot in jpg and hand their files off to their editors.

 benjaminblack's gear list:benjaminblack's gear list
Nikon D700 Nikon D800 Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Micro-Nikkor 60mm f/2.8G ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II +6 more
Reply   Reply with quote   Complain
Thoughts R Us
Contributing MemberPosts: 544
Like?
Re: Canon 5D3 vs Nikon 800E noise tests
In reply to kymarto, 6 months ago

These test results confirm what everybody else found out when these two cameras were introduced and a spate of comparison tests hit the web.

As others have mentioned, the sensor in the D800 is an absolute engineering breakthrough.  Kudos to Sony.

I have shot with both cameras, and i would contend that the major advantage of the 5d3 is it's superior AF.  The 5d3 AF is really superlative, and trumps the D800.

If someone were having to choose between the 2 cameras for a purchase, that would be the main selling point for the 5d3 over the D800.

It's funny because before Canon trailed Nikon in AF, whereas Nikon trailed Canon in megapixels.  It was debatable as to overall IQ.  But now they've kind of switched places...it seems they listened to their critics and reacted.

I would expect the next Canon 5d iteration to have a significantly improved sensor that remedies most of not all of their shortcomings.  If Canon can make a sensor leap equal to what they did in going from the  9 point AF of the 5d2 to the 61 point AF of the 5d3, that will be a significant leap indeed.  Then they may be able to challenge Nikon in the sensor department.

Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: No surprise here
In reply to Kabe Luna, 6 months ago

The D800 beats the snot out of the 5DIII in any measure of or aspect of IQ dependent upon dynamic range and low read noise through ISO 1600. At this point, for anyone with eyes to see realizes that. However, high ISO in ample daylight is not the test of a camera's high ISO capabilities, I feel. Test these two in dim artificial (tungsten) light beyond ISO 1600, and you'll see the 5DIII fairly easily run away from the D800. The D800's high ISO weakness is the reason I've supplemented it with a Dƒ.

kymarto wrote:

Hi folks,
Vis-a-vis the endless debate about these two cameras: I am privileged to use both regularly, one for myself (Nikon) and the other for work (Canon). I was reading again as the opposing fanboys went at each other. I actually like both. The Canon feels better in my hand, but I like the Nikon layout much more. On the other hand the Canon has (to my mind) a more sensible menu layout, and the bracketing controls are much better, allowing more than 1 EV between shots. Advantages and disadvantages both ways...

But for me the main point is image quality. I was very interested to see how the two cameras compare in terms of noise. I had read that the Canon was noisy in the shadows at base ISO, but performed better than the Nikon at high ISOs. So yesterday I took the Canon home and performed two head-to-head tests.

First I wanted to compare shadow noise. I shot two frames (raw) each, at 100 and 800 ISO, underexposed by 3 EV. I then imported them into ACR and brought them up 3 stops, and performed some minor tweaks to get the contrast presentable and the the color balance similar between the two cameras.
The shots are all at 24mm @ f5.6, the Nikon with the 14-24 and the Canon with the 16-35 II. Nikon shooters, be glad, the Canon lens can't hold a candle to the Nikon in the corners
Anyway I then cropped out similar areas in the four shots and lined them up. If you look at the file I uploaded at original size, the image magnification is at 150%
The second image is done the same way, but shooting both cameras at 6400 and 12800 ISO (no exposure compensation).
My impressions are these: as to the underexposure noise test, I was shocked. The Nikon at 100 is very clean, but the Canon is pretty miserable, especially in terms of shadow noise (both luma and chroma). The Nikon at 800 is almost comparable to the Canon at 100, actually noisier in the mids and highs, but cleaner in the shadows.
The high ISO test also surprised me (second image). I kept hearing that the Canon was clearly superior to the Nikon in terms of high ISO noise, but that is not my impression from my test. Perhaps Canon has better noise reduction algorithms in the processor for jpgs, but this was a head-to-head test with CR2 and NEF. To my eye the Nikon is actually slightly better than the Canon at both ISOs, especially in the shadows. Have a look and see what you think.

Tests at -3 EV underexposure

High ISO tests (no exposure compensation)

-- hide signature --

I actually did this in a night scene, and my feeling was that the 5D3 was very slightly superior because the noise was a bit less visible in the midtones and light areas. I will try it again at some point.

Reply   Reply with quote   Complain
kymarto
Contributing MemberPosts: 632
Like?
Re: If I still owned a 5DIII, I could...
In reply to Horshack, 6 months ago

Very nice work. Thanks for posting the links!

Reply   Reply with quote   Complain
KnightPhoto2
Senior MemberPosts: 1,369Gear list
Like?
Re: If I still owned a 5DIII, I could...
In reply to kymarto, 6 months ago

kymarto wrote:

Very nice work. Thanks for posting the links!

+1!

-- hide signature --

Best Regards,
SteveK
'A camera is an instrument that teaches people how to see without a camera.' -- Dorothea Lange
http://images.nikonians.org/galleries/showgallery.php/cat/500/ppuser/119002

 KnightPhoto2's gear list:KnightPhoto2's gear list
Nikon D4 Nikon D800E Nikon 1 V3 Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II +20 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads