5D-II vs D800E test - with SAME lens.

Started May 14, 2012 | Discussions
davexl
Senior MemberPosts: 1,050
Like?
5D-II vs D800E test - with SAME lens.
May 14, 2012

Bookshelf Test.

Having just upgraded from the 5D-II to the D800E, I have been running some tests I thought I would share. Lots of tests are floating about, but I have not seen any like this - done with the exact same lens . Apples to Apples.

This test is with a older 50mm lens that is still outresolving the sensor, but unlike recent Nikon lenses has a manual aperture ring - so f8 is as consistent on both cameras as we can get.

I have tried to make things as equal as possible. Same lens, shot a few minutes apart using an adapter to fit it on the Canon, with aperture and focus manually set, and checked with live view, and the best one used. As it happens I did manage to swap lenses without bumping anything.

In Lightroom 4 RC2 - Colour Calibration Profiles are custom made using the Xrite Colorchecker passport - works very well, I have never had colour as consistent between brands before.

Slight lighting change outside due to clouds.

Sharpening is 150, radius 0.9, detail 0. Not my personal preference, but I have found the detail slider can give very different effects on these cameras, (5DII is way more sensitive) so I wanted to eliminate sharpening differences.

100% Crops - but both uprezzed to 10,000 pixels, so that the 5D-II is not disadvantaged by being the only one resampled. (click to see all of image)

The black Cinefex magazines show the most difference - the spine text is mostly illegible on the 5D-II, and legible on the D800E.

Crops at 100%:

Thoughts:

Resolution

An easy win to the D800E, as you would expect, but as the uprez test shows, it is not as large as 22 vs 36 sounds (of course - area vs linear). Nothing the Canon has to be too ashamed of I think.

My Mum could probably not tell the difference, but then, she is not the one paying me to do great work and print it big either.

The D800e seems to me to be somewhere in between cameras lacking an AA filter and ones with. You do not seem to get shocking moire surprises, and the extra detail is worthwhile. (But the D800 can be sharpened up into something very, very close)

I think AA filters are becoming less relevant. On a 6MP sensor, removing it makes a big deal. On 36MP, not so much.

The moire in these images is surprisingly comparable - not a big deal, and it cleans up perfectly in Lightroom with the brush.

Worth switching brands over?

For me, yes - the MP plus the DR plus the world class high ISO plus incredibly robust files for heavy PP - but that hardly counts, as it is a switch back to Nikon - my 5D-II was a stopgap to give me high MP for a reasonable price while Nikon was asleep. I had decided to hold off on selling the rest of my Nikon glass (except the 14-24) until I saw the 5D-III and the D800, and thankfully Nikon decided to really swing for the fences. I had resigned myself to Nikon ignoring me again, and was mucho surprised.

If you have a bunch of Canon glass then the decision is much harder, and I would think for the vast majority the sensible answer would be no. But hey, this is the DPR forums... sensible discussion left the building long ago

I am working on the really interesting test - the Nikon 14-24 to follow, when I can work out how to keep the adapted aperture consistent.

-- hide signature --

Things I have clicked with: Agfa 110, Minolta X-300, X700, Nikon F3, F4, Hassleblad 500C/M, 4x5", D70, D200, Canon 1Ds, II, D2x, D3, Canon 5D-II, D800E

yabokkie
Regular MemberPosts: 483
Like?
Re: 5D-II vs D800E test - with SAME lens.
In reply to davexl, May 14, 2012

I think 14-24/2.8 is a must-have lens for anyone who shoots wide, including those born as a Canon photographer with no Nikon gear.

I haven't decided yet but I think D800 should be a better choice. by not applying AA filtering, what D800E does is really adding noise which cannot be removed.

Reply   Reply with quote   Complain
JackM
Veteran MemberPosts: 6,276Gear list
Like?
sigh. yet another invalid test.
In reply to davexl, May 14, 2012

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

 JackM's gear list:JackM's gear list
Fujifilm X100S Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF 100mm f/2.0 USM Canon EF 300mm f/2.8L IS USM +5 more
Reply   Reply with quote   Complain
WalterSrChat
Senior MemberPosts: 1,056Gear list
Like?
Another Crappy D800 Comparison In This Forum
In reply to davexl, May 14, 2012

Before you buy, check out these problems!
Problems;

Nikon lockup issue
http://www.engadget.com/2012/05/04/nikon-confirms-woes-with-d4-and-d800/

Blurry View finder
http://fstoppers.com/...news-nikon-d800-has-confirmed-focusview-finder-issues

D800 and wireless flash photography problems
http://lumenatic.com/2012/03/25/d800-and-wireless-flash-photography-problems/

Nikon D800 long exposure issue:
http://emmanuelcoupe.wordpress.com/2012/04/04/nikon-d800-long-exposures-issues/

AF issues
http://nikonrumors.com/forum/topic.php?id=5888
http://forums.dpreview.com/forums/read.asp?forum=1021&message=41325311

Another good list of other issues
http://nikonrumors.com/forum/topic.php?id=5517

Oh! Forgot to mention the exploding battery

Walter Sr

An individual may need to be on the receiving end, of what they do, to understand why they should not do it. (Walter Sr 2012)
--
I am out to take the perfect picture, if it exits!

 WalterSrChat's gear list:WalterSrChat's gear list
Canon EOS-1Ds Mark III Canon EOS 20D Canon EOS-1D X Canon EF 50mm f/1.8 II Canon EF 16-35mm f/2.8L II USM +3 more
Reply   Reply with quote   Complain
kevindar
Veteran MemberPosts: 3,482Gear list
Like?
thanks for the comparison
In reply to davexl, May 14, 2012

Nikon, without a doubt has a great sensor in d800/e, with resolution, noise and dynamic range, and over all a decent body to go with it, although there appear to be some bugs which will hopefully get sorted out. I do wish canon was a 28MP version of d800, shooting at 5fps at full resolution, and 6-8 in dx mode. that would have been the perfect camera for me.

The one true disappointment for me with 5d3 remains dynamic range, but over all, I love the camera
--
http://razzi.me/kevindar/photos

 kevindar's gear list:kevindar's gear list
Canon EOS 5D Mark III Canon EF 85mm f/1.2L II USM Canon EF 16-35mm f/2.8L II USM Nikon AF-S Nikkor 14-24mm f/2.8G ED Canon EF 70-200mm f/2.8L IS II USM +21 more
Reply   Reply with quote   Complain
qianp2k
Senior MemberPosts: 8,277Gear list
Like?
Re: sigh. yet another invalid test.
In reply to JackM, May 14, 2012

+1. How about you downsampling both to 22mp or 18mp size, and I bet the gap will be much narrower or even indistinguishable. Not everyone needs huge print nor want to deal with huge files.

In addition using the same lens sounds fair but doesn't reflect reality. I don't know if any IQ lost in Canon camera using Nikon lens thru an adapter. You should use the Canon native corresponding lens on 5D3 for this test. Why DPR doesn't use the same lens but native lens for each respective brand in its lab tests?

JackM wrote:

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

-- hide signature --
 qianp2k's gear list:qianp2k's gear list
Canon EOS-1D Mark III Canon EOS 5D Mark II Canon EOS 5D Mark III Sony Alpha 7R Canon EF 15mm f/2.8 Fisheye +19 more
Reply   Reply with quote   Complain
gipper51
Veteran MemberPosts: 3,922Gear list
Like?
Re: sigh. yet another invalid test.
In reply to JackM, May 14, 2012

JackM wrote:

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

No kidding. They are scaled to the same output dimensions, which is the best way to do it. Of course the 5D3 will be enlarged more, it's a smaller file.
--

On the 5D3 and D800: "The fact there is still so much to criticize with cameras this good only proves the human race can never be satisfied with anything." (Me)

My sites:
http://www.gipperich-photography.com
http://www.pbase.com/gipper51/portraits

 gipper51's gear list:gipper51's gear list
Canon EOS 5D Mark II Canon EOS 6D Canon EF 17-40mm f/4.0L USM Canon EF 28-135mm f/3.5-5.6 IS USM Tamron AF 28-75mm F/2.8 XR Di LD Aspherical (IF) +3 more
Reply   Reply with quote   Complain
gipper51
Veteran MemberPosts: 3,922Gear list
Like?
Re: sigh. yet another invalid test.
In reply to qianp2k, May 14, 2012

qianp2k wrote:

+1. How about you downsampling both to 22mp or 18mp size, and I bet the gap will be much narrower or even indistinguishable. Not everyone needs huge print nor want to deal with huge files.

JackM wrote:

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

-- hide signature --

Why? You would intentionally throw away any pixel count advantage the D800 may have and compare it to a lower res body? That proves nothing. I could reduce a 90mp stitched panoramic I have down to 22mp also, and the 5D3 would look "very close". Any large file will look close to a smaller one if you downsize it. I'm as tall as Shaquille O'neall if you cut him off at the knees!

Reducing larger files down to a smaller one is NEVER the proper way to test for resolution, at least in the real world (meaning not DXO tests).

 gipper51's gear list:gipper51's gear list
Canon EOS 5D Mark II Canon EOS 6D Canon EF 17-40mm f/4.0L USM Canon EF 28-135mm f/3.5-5.6 IS USM Tamron AF 28-75mm F/2.8 XR Di LD Aspherical (IF) +3 more
Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 20,696
Like?
Nicely done!
In reply to davexl, May 14, 2012

Of course, a few responses below show that some still can't quite figure out that a proper comparison is done at the same display size, as you have done.

However, at f/8, both systems suffer from diffraction softening narrowing the differences between the systems. A better aperture would be f/5.6, although, I will admit that the resolution lost to diffraction from f/5.6 to f/8 is trivial.

Of more importance is the sharpening applied. I would process both photos optimally, which most likely means different levels of sharpening applied to each.

In any case, nicely done!

davexl wrote:

Bookshelf Test.

Having just upgraded from the 5D-II to the D800E, I have been running some tests I thought I would share. Lots of tests are floating about, but I have not seen any like this - done with the exact same lens . Apples to Apples.

This test is with a older 50mm lens that is still outresolving the sensor, but unlike recent Nikon lenses has a manual aperture ring - so f8 is as consistent on both cameras as we can get.

I have tried to make things as equal as possible. Same lens, shot a few minutes apart using an adapter to fit it on the Canon, with aperture and focus manually set, and checked with live view, and the best one used. As it happens I did manage to swap lenses without bumping anything.

In Lightroom 4 RC2 - Colour Calibration Profiles are custom made using the Xrite Colorchecker passport - works very well, I have never had colour as consistent between brands before.

Slight lighting change outside due to clouds.

Sharpening is 150, radius 0.9, detail 0. Not my personal preference, but I have found the detail slider can give very different effects on these cameras, (5DII is way more sensitive) so I wanted to eliminate sharpening differences.

100% Crops - but both uprezzed to 10,000 pixels, so that the 5D-II is not disadvantaged by being the only one resampled. (click to see all of image)

The black Cinefex magazines show the most difference - the spine text is mostly illegible on the 5D-II, and legible on the D800E.

Crops at 100%:

Thoughts:

Resolution

An easy win to the D800E, as you would expect, but as the uprez test shows, it is not as large as 22 vs 36 sounds (of course - area vs linear). Nothing the Canon has to be too ashamed of I think.

My Mum could probably not tell the difference, but then, she is not the one paying me to do great work and print it big either.

The D800e seems to me to be somewhere in between cameras lacking an AA filter and ones with. You do not seem to get shocking moire surprises, and the extra detail is worthwhile. (But the D800 can be sharpened up into something very, very close)

I think AA filters are becoming less relevant. On a 6MP sensor, removing it makes a big deal. On 36MP, not so much.

The moire in these images is surprisingly comparable - not a big deal, and it cleans up perfectly in Lightroom with the brush.

Worth switching brands over?

For me, yes - the MP plus the DR plus the world class high ISO plus incredibly robust files for heavy PP - but that hardly counts, as it is a switch back to Nikon - my 5D-II was a stopgap to give me high MP for a reasonable price while Nikon was asleep. I had decided to hold off on selling the rest of my Nikon glass (except the 14-24) until I saw the 5D-III and the D800, and thankfully Nikon decided to really swing for the fences. I had resigned myself to Nikon ignoring me again, and was mucho surprised.

If you have a bunch of Canon glass then the decision is much harder, and I would think for the vast majority the sensible answer would be no. But hey, this is the DPR forums... sensible discussion left the building long ago

I am working on the really interesting test - the Nikon 14-24 to follow, when I can work out how to keep the adapted aperture consistent.

-- hide signature --

Things I have clicked with: Agfa 110, Minolta X-300, X700, Nikon F3, F4, Hassleblad 500C/M, 4x5", D70, D200, Canon 1Ds, II, D2x, D3, Canon 5D-II, D800E

Reply   Reply with quote   Complain
JackM
Veteran MemberPosts: 6,276Gear list
Like?
Re: sigh. yet another invalid test.
In reply to gipper51, May 14, 2012

gipper51 wrote:

qianp2k wrote:

+1. How about you downsampling both to 22mp or 18mp size, and I bet the gap will be much narrower or even indistinguishable. Not everyone needs huge print nor want to deal with huge files.

JackM wrote:

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

-- hide signature --

Why? You would intentionally throw away any pixel count advantage the D800 may have and compare it to a lower res body? That proves nothing. I could reduce a 90mp stitched panoramic I have down to 22mp also, and the 5D3 would look "very close". Any large file will look close to a smaller one if you downsize it. I'm as tall as Shaquille O'neall if you cut him off at the knees!

Reducing larger files down to a smaller one is NEVER the proper way to test for resolution, at least in the real world (meaning not DXO tests).

Agreed and well put. Leave the resolutions alone and let us see and judge for ourselves.

However, looking at 100% crops only tells you how each would look when printed to about 100dpi. Or 58" wide for the 5D3 and 74" wide for the D800. I print no larger than 30" wide.

 JackM's gear list:JackM's gear list
Fujifilm X100S Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF 100mm f/2.0 USM Canon EF 300mm f/2.8L IS USM +5 more
Reply   Reply with quote   Complain
JackM
Veteran MemberPosts: 6,276Gear list
Like?
Re: Nicely done!
In reply to Great Bustard, May 14, 2012

Great Bustard wrote:

Of course, a few responses below show that some still can't quite figure out that a proper comparison is done at the same display size, as you have done.

Umm, maybe, but 10,000 pixels wide?? That's only relevant to people printing 8 feet wide from a single exposure.

 JackM's gear list:JackM's gear list
Fujifilm X100S Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF 100mm f/2.0 USM Canon EF 300mm f/2.8L IS USM +5 more
Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 20,696
Like?
Re: Nicely done!
In reply to JackM, May 14, 2012

JackM wrote:

Great Bustard wrote:

Of course, a few responses below show that some still can't quite figure out that a proper comparison is done at the same display size, as you have done.

Umm, maybe, but 10,000 pixels wide?? That's only relevant to people printing 8 feet wide from a single exposure.

The purpose, as the OP explained, was so people would not cry foul that only one of the photos had been upsampled. A better way about it is to upsample the 5D2 photo to the dimensions of the D800E photo, without using interpolation , which makes for the most fair detail comparison.

Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 20,696
Like?
Ideally...
In reply to JackM, May 14, 2012

JackM wrote:

However, looking at 100% crops only tells you how each would look when printed to about 100dpi. Or 58" wide for the 5D3 and 74" wide for the D800. I print no larger than 30" wide.

...one would print both photos at a variety of sizes, scan in the prints (or portions of the prints), and post 100% crops of the scans.

This not only compares the photos at the same display size for a variety of print sizes, it includes what the process of printing does with (to) the photo.

Reply   Reply with quote   Complain
tomboy
Contributing MemberPosts: 924Gear list
Like?
Re: Another Crappy D800 Comparison In This Forum
In reply to WalterSrChat, May 14, 2012

I got my d800 ..no problem with any of the issue u posted and the image from d800 at iso 25,600 is very good ..

here a test photo..and no noise reduction even used
F

Reply   Reply with quote   Complain
JackM
Veteran MemberPosts: 6,276Gear list
Like?
Re: Ideally...
In reply to Great Bustard, May 14, 2012

Great Bustard wrote:

JackM wrote:

However, looking at 100% crops only tells you how each would look when printed to about 100dpi. Or 58" wide for the 5D3 and 74" wide for the D800. I print no larger than 30" wide.

...one would print both photos at a variety of sizes, scan in the prints (or portions of the prints), and post 100% crops of the scans.

This not only compares the photos at the same display size for a variety of print sizes, it includes what the process of printing does with (to) the photo.

Agreed. And on that note, I just received two 20x30" prints, one from my (now sold) 7D and one from my 5D3. They are both low ISO images. You have to put your nose to them to see the advantage to the 5D3, and even then an untrained eye probably couldn't.

 JackM's gear list:JackM's gear list
Fujifilm X100S Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF 100mm f/2.0 USM Canon EF 300mm f/2.8L IS USM +5 more
Reply   Reply with quote   Complain
qianp2k
Senior MemberPosts: 8,277Gear list
Like?
Re: sigh. yet another invalid test.
In reply to JackM, May 14, 2012

Evaluate at the same size said by many sounds fair. However I never print more than 30x20" so why not compare at the same size at 30x20" or less? Then the advantage of 36mp over 22mp is not that much or even distinguishable anyway. I have already read several actually did that said so. I agree if you print 40x30" or larger, then D800 has a clear advantage but this is not the case by many including me. Another argument is 36mp cropping ability. Sure more pixels better cropping to the same size but this will not replace the crop camera and long lens in scenarios of wildlife and birding. Pixel density is still the key factor in cropping.

JackM wrote:

gipper51 wrote:

qianp2k wrote:

+1. How about you downsampling both to 22mp or 18mp size, and I bet the gap will be much narrower or even indistinguishable. Not everyone needs huge print nor want to deal with huge files.

JackM wrote:

You upscaled the 5D3 image much more than the D800 image. 74% vs 36%, respectively.

Next.

-- hide signature --

Why? You would intentionally throw away any pixel count advantage the D800 may have and compare it to a lower res body? That proves nothing. I could reduce a 90mp stitched panoramic I have down to 22mp also, and the 5D3 would look "very close". Any large file will look close to a smaller one if you downsize it. I'm as tall as Shaquille O'neall if you cut him off at the knees!

Reducing larger files down to a smaller one is NEVER the proper way to test for resolution, at least in the real world (meaning not DXO tests).

Agreed and well put. Leave the resolutions alone and let us see and judge for ourselves.

However, looking at 100% crops only tells you how each would look when printed to about 100dpi. Or 58" wide for the 5D3 and 74" wide for the D800. I print no larger than 30" wide.

-- hide signature --
 qianp2k's gear list:qianp2k's gear list
Canon EOS-1D Mark III Canon EOS 5D Mark II Canon EOS 5D Mark III Sony Alpha 7R Canon EF 15mm f/2.8 Fisheye +19 more
Reply   Reply with quote   Complain
JackM
Veteran MemberPosts: 6,276Gear list
Like?
Re: Nicely done!
In reply to Great Bustard, May 14, 2012

Great Bustard wrote:

The purpose, as the OP explained, was so people would not cry foul that only one of the photos had been upsampled.

No I got that, but it's still foul because the 5D2 image was so severely upsampled, while the D800 was lightly upsampled.

A better way about it is to upsample the 5D2 photo to the dimensions of the D800E photo, without using interpolation , which makes for the most fair detail comparison.

I think upsampling the smaller image puts it at a bigger disadvantage than if the larger image is downsampled.

 JackM's gear list:JackM's gear list
Fujifilm X100S Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF 100mm f/2.0 USM Canon EF 300mm f/2.8L IS USM +5 more
Reply   Reply with quote   Complain
qianp2k
Senior MemberPosts: 8,277Gear list
Like?
Re: Nicely done!
In reply to Great Bustard, May 14, 2012

Nobody denies D800 has more resolution over 5D3. The issue is that if you really need 36mp by individuals? I don't need it as I never print more than 30x20" and I believe crop camera and long lens in wildlife and birding than heavy cropping. So in my case (and I am not alone) 36mp is not a huge deal, no mention it has cons - larger files and slower burst rate. Between D800 and 5D3, I'd pickup latter as I use the camera for multiple purposes not just landscape and studio, and I prefer Canon's lens choice.

Great Bustard wrote:

JackM wrote:

Great Bustard wrote:

Of course, a few responses below show that some still can't quite figure out that a proper comparison is done at the same display size, as you have done.

Umm, maybe, but 10,000 pixels wide?? That's only relevant to people printing 8 feet wide from a single exposure.

The purpose, as the OP explained, was so people would not cry foul that only one of the photos had been upsampled. A better way about it is to upsample the 5D2 photo to the dimensions of the D800E photo, without using interpolation , which makes for the most fair detail comparison.

-- hide signature --
 qianp2k's gear list:qianp2k's gear list
Canon EOS-1D Mark III Canon EOS 5D Mark II Canon EOS 5D Mark III Sony Alpha 7R Canon EF 15mm f/2.8 Fisheye +19 more
Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 20,696
Like?
Re: Nicely done!
In reply to JackM, May 14, 2012

JackM wrote:

Great Bustard wrote:

The purpose, as the OP explained, was so people would not cry foul that only one of the photos had been upsampled.

No I got that, but it's still foul because the 5D2 image was so severely upsampled, while the D800 was lightly upsampled.

The 5D3 photo is only upsampled 28% more than the D800E photo, which, of course, will be the same no matter what size they are resampled to if displayed at the same size.

A better way about it is to upsample the 5D2 photo to the dimensions of the D800E photo, without using interpolation , which makes for the most fair detail comparison.

I think upsampling the smaller image puts it at a bigger disadvantage than if the larger image is downsampled.

Downsampling removes detail from the photo made with more pixels. The only purpose to downsampling is if the final photo is to be downsampled for web display, or the print size is so small that the printer cannot handle the PPI of the original file.

Reply   Reply with quote   Complain
Great Bustard
Forum ProPosts: 20,696
Like?
Re: Nicely done!
In reply to qianp2k, May 14, 2012

qianp2k wrote:

Nobody denies D800 has more resolution over 5D3. The issue is that if you really need 36mp by individuals? I don't need it as I never print more than 30x20" and I believe crop camera and long lens in wildlife and birding than heavy cropping. So in my case (and I am not alone) 36mp is not a huge deal, no mention it has cons - larger files and slower burst rate. Between D800 and 5D3, I'd pickup latter as I use the camera for multiple purposes not just landscape and studio, and I prefer Canon's lens choice.

The purpose of the OP was not to tell you which camera to get, but to display the differences in resolution between the two systems. If you do not need the resolution of the D800, or find the resolution advantage of the D800 is not "worth it" for whatever disadvantages the D800 has compared to another system, then the D800 is not the best choice for you.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads