5D3 vs D800: Comparing High ISO RAW

Started Mar 14, 2012 | Discussions
cs hauser
Contributing MemberPosts: 713
Like?
5D3 vs D800: Comparing High ISO RAW
Mar 14, 2012

The Test:

  • These came from ISO 25600 RAW files

  • Used dcraw to convert to 16-bit TIFF before JPEG

  • No sharpening, No NR

  • Downsized both to 8 megapixels (PS Bicubic)

  • 5D3 exposure is much longer than D800. We will ignore because testers admitted that light source changes between tests.

The 100% Crops of 8 megapixels

So to summarize:

  • D800 has 63% more pixels

  • D800 has at least 2 stops more dynamic range at ISO 100

  • D800 looks equal or better than the 5D3 at ISO 25600

Sooo... exactly what did Canon gain by staying at 22 megapixels?

Canon EOS 5D Mark III Nikon D800
If you believe there are incorrect tags, please send us this post using our feedback form.
Randplaty
Contributing MemberPosts: 504
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 14, 2012

FPS probably. Well at least Nikon fans are saying the files are equal now. Before they were saying the D800 beats the 5D3.

Reply   Reply with quote   Complain
phiri
Contributing MemberPosts: 967Gear list
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 14, 2012

So to summarize:

  • D800 has 63% more pixels

  • D800 has at least 2 stops more dynamic range at ISO 100

  • D800 looks equal or better than the 5D3 at ISO 25600

Sooo... exactly what did Canon gain by staying at 22 megapixels?

Yes, I wonder why no one is participating in this raw thread. Just borrowed from someone's comparisons at Nikon. Let everybody be the judge. These normalised to the D4 size

 phiri's gear list:phiri's gear list
Nikon D7000 Nikon 1 J1 Nikon D600 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 50mm f/1.8G +6 more
Reply   Reply with quote   Complain
etto72
Regular MemberPosts: 316Gear list
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 14, 2012

If you post the same files with the ACR setting to default than Canon looks better!!

cs hauser wrote:

The Test:

  • These came from ISO 25600 RAW files

  • Used dcraw to convert to 16-bit TIFF before JPEG

  • No sharpening, No NR

  • Downsized both to 8 megapixels (PS Bicubic)

  • 5D3 exposure is much longer than D800. We will ignore because testers admitted that light source changes between tests.

The 100% Crops of 8 megapixels

So to summarize:

  • D800 has 63% more pixels

  • D800 has at least 2 stops more dynamic range at ISO 100

  • D800 looks equal or better than the 5D3 at ISO 25600

Sooo... exactly what did Canon gain by staying at 22 megapixels?

 etto72's gear list:etto72's gear list
Canon EOS-1D X Canon EOS 5D Mark III Canon EOS M Sony Alpha 7R Fujifilm X-T1
Reply   Reply with quote   Complain
cs hauser
Contributing MemberPosts: 713
Like?
You sample doesn't look valid
In reply to phiri, Mar 14, 2012

phiri wrote:

Yes, I wonder why no one is participating in this raw thread. Just borrowed from someone's comparisons at Nikon. Let everybody be the judge. These normalised to the D4 size

Hmmm... I don't know what RAW conversion was used in the samples you borrowed from the Nikon forum. But it looks very much like there has been NR applied to all three images. There's hardly any chroma noise on all three images, which is suspicious for ISO 6400.

I would guess that those were processed in ACR with default NR settings.

The ones I posted did not have any NR applied at all.

Reply   Reply with quote   Complain
studio311
Contributing MemberPosts: 553
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 14, 2012

D800 is damn impressive. Hard for me to believe the 5d3 won't be a bit better than d800 at ISO 6400 and up...but for me ... the d800 resolution and DR(!!!) at low iso is jaw dropping. Stunning actually.

Reply   Reply with quote   Complain
sandy b
Veteran MemberPosts: 7,389Gear list
Like?
yes, you are right
In reply to cs hauser, Mar 14, 2012
 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
phiri
Contributing MemberPosts: 967Gear list
Like?
Re: You sample doesn't look valid
In reply to cs hauser, Mar 14, 2012

cs hauser wrote:

phiri wrote:

Yes, I wonder why no one is participating in this raw thread. Just borrowed from someone's comparisons at Nikon. Let everybody be the judge. These normalised to the D4 size

Hmmm... I don't know what RAW conversion was used in the samples you borrowed from the Nikon forum. But it looks very much like there has been NR applied to all three images. There's hardly any chroma noise on all three images, which is suspicious for ISO 6400.

I would guess that those were processed in ACR with default NR settings.

The ones I posted did not have any NR applied at all.

All information on how the pictures have been arrived at. They are downsampled to 16 MP to match the D4 resolution
--
http://www.flickr.com/photos/39182144@N03/

 phiri's gear list:phiri's gear list
Nikon D7000 Nikon 1 J1 Nikon D600 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR Nikon AF-S Nikkor 50mm f/1.8G +6 more
Reply   Reply with quote   Complain
Mike K
Veteran MemberPosts: 5,495Gear list
Like?
Re: 5D3 vs D800: Exposure and resolution difference due to...
In reply to cs hauser, Mar 14, 2012

These crops from the Imaging Resource comparometer?

The top of the last long thread identifies another source of difference in the shots. The Canon was shot with the f 2.5 compact macro with Life Size Converter. This is essentially an extension tube on top of a 1.4x teleconverter behind the macro lens. There is loss of light and extra lens elements involved, so clearly not an exact apples to apples comparison. Pretty hard to do that anyway unless a fully manual Zeiss or Nikon lens is used on both cameras with a N to C adapter.

Camera comparisons like these are never perfect, there are simply too many variables to control. After equalizing lenses, exposures, then there are the differing RAW conversions, resizing and display choices, etc. It will never truly be changing just one variable at a time.

 Mike K's gear list:Mike K's gear list
Canon EOS-1D Mark II Canon EOS 5D Mark II Canon EOS D60 Canon EOS 6D Canon EF 135mm f/2.0L USM +9 more
Reply   Reply with quote   Complain
MASTERPPA
Contributing MemberPosts: 859
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to studio311, Mar 14, 2012

I did the exact same test, and the camera are nearly identical. The D800 has more noise, but that is negated when down sampled. I don't know about the DR.

If only Nikon had sRAW and a 70-200 F4 travel lens... sniff sniff...

studio311 wrote:

D800 is damn impressive. Hard for me to believe the 5d3 won't be a bit better than d800 at ISO 6400 and up...but for me ... the d800 resolution and DR(!!!) at low iso is jaw dropping. Stunning actually.

Reply   Reply with quote   Complain
xerojay
Regular MemberPosts: 256
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 14, 2012

This is in-line with my own observations after comparing the RAWs.

I actually found the D800 to have a slight advantage over the 5D3, but my method was a bit different:

-Load the files into ACR
-Zero-out all sharpening & NR sliders
-set the file size to 11mp (in ACR)
-Open the images in Photoshop and pixel-peep away.

I can't defend Canon on this one; they're falling behind the times on sensor tech. The D800 sensor is an all-round better performer.

I also can't defend the Nikon D4 either, which is within about 1/3-stop of the D800, but without the major DR and MP advantage...Why would anyone buy the D4?!

Reply   Reply with quote   Complain
cs hauser
Contributing MemberPosts: 713
Like?
Re: 5D3 vs D800: Exposure and resolution difference due to...
In reply to Mike K, Mar 14, 2012

Mike K wrote:

These crops from the Imaging Resource comparometer?

Crops were from Imaging Resource RAW files that I converted myself. It's not from the comparometer.

Mike K wrote:

The top of the last long thread identifies another source of difference in the shots. The Canon was shot with the f 2.5 compact macro with Life Size Converter.

This issue has been brought up before. I believe Imaging Resources use Sigma 70mm Macro lenses on all their recent tests. However, some Canon cameras erroneously report Sigma lenses as something else. They don't support third party lenses, after all. In this instance, they reported it as the old 50m macro.

Reply   Reply with quote   Complain
Dominique Dierick
Veteran MemberPosts: 4,467Gear list
Like?
Not again, pleazze...
In reply to Mike K, Mar 14, 2012

Please check DXOmark and compare Canon sensor (eg 5DmkII) versus Sony (eg Nikon D7000). You will see the exact result that Canon is lower than Sony's versus nominal ISO. About 2/3rd stop. Meaning, this Imaging Resource only confirms what has been the case before.

Mike K wrote:

These crops from the Imaging Resource comparometer?

The top of the last long thread identifies another source of difference in the shots. The Canon was shot with the f 2.5 compact macro with Life Size Converter. This is essentially an extension tube on top of a 1.4x teleconverter behind the macro lens. There is loss of light and extra lens elements involved, so clearly not an exact apples to apples comparison. Pretty hard to do that anyway unless a fully manual Zeiss or Nikon lens is used on both cameras with a N to C adapter.

Camera comparisons like these are never perfect, there are simply too many variables to control. After equalizing lenses, exposures, then there are the differing RAW conversions, resizing and display choices, etc. It will never truly be changing just one variable at a time.

-- hide signature --

My equipment: Nikon D3s, D700, D7000, EOS 450D
The ladies equipment: Panasonic LX5 and GH2 for video

 Dominique Dierick's gear list:Dominique Dierick's gear list
Fujifilm X100S
Reply   Reply with quote   Complain
studio311
Contributing MemberPosts: 553
Like?
Re: yes, you are right
In reply to sandy b, Mar 15, 2012

Thanks for the link. Wow,wow,wow.

sandy b wrote:

http://forums.dpreview.com/forums/read.asp?forum=1021&message=40909199

Reply   Reply with quote   Complain
Dominique Dierick
Veteran MemberPosts: 4,467Gear list
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to xerojay, Mar 15, 2012

Well, fps, robustness, lean and mean file size for PJ to name a few reasons. I have a D800 on order, but 70% of my work will remain done with D3s.

-- hide signature --

My equipment: Nikon D3s, D700, D7000, EOS 450D
The ladies equipment: Panasonic LX5 and GH2 for video

 Dominique Dierick's gear list:Dominique Dierick's gear list
Fujifilm X100S
Reply   Reply with quote   Complain
cs hauser
Contributing MemberPosts: 713
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to xerojay, Mar 15, 2012

xerojay wrote:

I can't defend Canon on this one; they're falling behind the times on sensor tech. The D800 sensor is an all-round better performer.

I don't think Canon is falling behind the times in sensor tech. They're behind Sony, and really that's about all you can say. Other sensor makers (Nikon, Panasonic, Samsung, Fuji) have comparable DR, comparable deep shadow banding problems, and comparable high ISO performance as the latest Canon sensors.

The 1DX with its 16-channel readout would probably perform closer to D4 than the 5D3 does.

xerojay wrote:

I also can't defend the Nikon D4 either, which is within about 1/3-stop of the D800, but without the major DR and MP advantage...Why would anyone buy the D4?!

Nikon was thinking the exact same thing. That's why they limited the D800 to only 4.0 fps. As for why Nikon didn't just use the D800 sensor in the D4... I'm guessing the the sensor and/or image processors can't handle 36 MP at 10 fps.

Reply   Reply with quote   Complain
KatManDEW
Senior MemberPosts: 1,579Gear list
Like?
Re: 5D3 vs D800: Comparing High ISO RAW
In reply to cs hauser, Mar 15, 2012

cs hauser wrote:

The Test:

  • These came from ISO 25600 RAW files

  • Used dcraw to convert to 16-bit TIFF before JPEG

  • No sharpening, No NR

  • Downsized both to 8 megapixels (PS Bicubic)

(images snipped)

  • 5D3 exposure is much longer than D800. We will ignore because testers admitted that light source changes between tests.

So to summarize:

  • D800 has 63% more pixels

  • D800 has at least 2 stops more dynamic range at ISO 100

  • D800 looks equal or better than the 5D3 at ISO 25600

Sooo... exactly what did Canon gain by staying at 22 megapixels?

I'm with you on this... You didn't mention one other thing in your summary - The D800 is $500 cheaper...

I just looked at the images in this thread and the other comparison thread, on my calibrated NEC PA301WSV, and they confirm what I've thought since seeing the first images from both cameras - neither camera has hands down, undeniably better image quality . It's close enough that folks will be debating it when the replacements for each camera come out years from now.

The disturbing thing about this is that the Nikon has IQ equal to the 5D3, AND it has 63% more pixels, AND IT COSTS $500 LESS.

The 5D3's "saving grace" was supposed to be that would be the low light king, and that just does not appear to be the case.

Sooo... exactly why does the Canon cost $500 more?

More Canon kool aid, anyone?

 KatManDEW's gear list:KatManDEW's gear list
Canon EOS 300D Canon EOS 7D Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF-S 18-55mm f/3.5-5.6 +14 more
Reply   Reply with quote   Complain
Guidenet
Forum ProPosts: 12,588Gear list
Like?
Who cares?
In reply to cs hauser, Mar 15, 2012

What I want to know is who cares? I've been a Nikon shooter since 1968 and would never move to Canon unless there were a compelling reason to switch. I don't see a compelling reason with these images or any others. On the other hand, if I were a Canon shooter and made my living with those, I see no compelling reason to switch to Nikon based on these images or any others. I wouldn't shoot at that ISO anyway and expect to sell the image.

When you get into the better cameras, there's not much wrong with them. My Nikon D700 will eat any Canon alive in dim light, but I'm in trouble with a deep crop on that distant heron over there. Can anyone here really say Canon or Nikon would make any difference in their professional abilities. Could any amateur here actually claim that one or the other would make a compelling difference in their photography. Remember, we're talking close to Pro-level gear from a 7D up and a D300S up and I'm only including those because of build quality.

-- hide signature --

Cheers, Craig

Follow me on Twitter @craighardingsr : Equipment in Profile

 Guidenet's gear list:Guidenet's gear list
Nikon D300 Nikon D700 Nikon D3S Nikon D800 Nikon AF-S DX Nikkor 18-135mm f/3.5-5.6G ED-IF +24 more
Reply   Reply with quote   Complain
Guidenet
Forum ProPosts: 12,588Gear list
Like?
The D4 was designed for a very specific crowd.
In reply to cs hauser, Mar 15, 2012

cs hauser wrote:

Nikon was thinking the exact same thing. That's why they limited the D800 to only 4.0 fps. As for why Nikon didn't just use the D800 sensor in the D4... I'm guessing the the sensor and/or image processors can't handle 36 MP at 10 fps.

No doubt that's the ticket. The D4 and IDX were designed for photojournalists and sports photographers and just in time for the Olympics. I doubt a 36 MP camera could store images fast enough to be a good choice. Moreover that much is not needed.

The D4 and IDX are big robust cameras that will take abuse, climate issues and most anything else you can throw at them. In the trenches of war or the photo-box at track and field these cameras are at home and will give years of trouble free service most of the time. That's who they were designed for. Getty Images and a local Newspaper will buy these for their staff photographers.

Imagine standing at the final turn in a track event. You've got your IDX or D4 and telephoto locked into that turn. As the runners appear, you start machine gunning the leader until they pass. Hopefully, your camera tracked dead on and all those will be in focus. That's what these were designed for. Nobody gives a rat's behind about much else like megapixels. Can it track the runner until he passes. That's the key and there's your living. Huge buffer and high framerates with superb tracking in the AF.

For me, I find my D700 or even my D300 to be robust enough for my uses, especially in the studio.

-- hide signature --

Cheers, Craig

Follow me on Twitter @craighardingsr : Equipment in Profile

 Guidenet's gear list:Guidenet's gear list
Nikon D300 Nikon D700 Nikon D3S Nikon D800 Nikon AF-S DX Nikkor 18-135mm f/3.5-5.6G ED-IF +24 more
Reply   Reply with quote   Complain
KatManDEW
Senior MemberPosts: 1,579Gear list
Like?
Re: Who cares?
In reply to Guidenet, Mar 15, 2012

Guidenet wrote:
My Nikon D700 will eat any Canon alive in dim light,

I think many were expecting Canon to makeup that difference with the 5D3, and were willing to give up an increase in resolution to gain that low light capability. It appears that we got neither (low light or high res), and we got a $500 price premium to add insult to injury.

 KatManDEW's gear list:KatManDEW's gear list
Canon EOS 300D Canon EOS 7D Canon EOS 5D Mark III Canon EF 50mm f/1.4 USM Canon EF-S 18-55mm f/3.5-5.6 +14 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads