canon rumors FORUM

Gear Talk => EOS Bodies - For Stills => Topic started by: Canon Rumors on September 22, 2012, 09:02:35 AM

Title: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Canon Rumors on September 22, 2012, 09:02:35 AM
A big difference
A reader sent in a quick comparison between the Canon EOS-1D X and Canon EOS 5D Mark III showing a very different color rendition as well as a significant difference in noise performance at ISO 25,600.

The Quick Test
The following test was conducted at sunset. This is the reason for the white balance setting, which on my 5d3 most closely approximated the warm tone of the light. Both files were dialed in at 8000k, +15 magenta. Both use the exact same lens, and were shot within 1 minute of each other. Both files are DNG 7.1, process version PV2012, and use the camera calibration setting “Camera Neutral.” All other settings including sharpening, noise reduction, tone curve, etc are all zero’d out for consistency. I’m using the latest version of Adobe Photoshop Lightroom available as of this writing, which is 4.1.

The Test with 100% Crops | Click for Larger

Canon EOS-1D X at B&H | Amazon | Adorama | Canon EOS 5D Mark III at B&H | Amazon | Adorama

Source: [PPT]

cr

Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: kevl on September 22, 2012, 09:19:46 AM
The 1Dx is a high ISO monster! Perhaps I need to remortgage the house so I can upgrade from the 5D3 to the 1Dx.

Frankly, the 5D3 did an amazing job in that photo too. Color shift doesn't bother me too much because I always play with that in post even if it looks good as shot. So no color shift wouldn't even save me on workflow.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: quartzie on September 22, 2012, 09:37:44 AM
IMHO, there is a noticeable difference in lighting color - as evidenced by the reflection in the windshield.
One minute can make a huge difference during sunset, and given that the shots aren't perfectly aligned, I'd wager that each camera measured the light somewhere else.

The difference between their high ISO performance is vast, but I'd prefer to see a slightly more reasonable exposure. Pulling 1/6400s just to be able to use ISO 25600 during the day seems a bit outlandish and quite unlikely in most settings.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: pharding on September 22, 2012, 09:39:07 AM
I was at the WFX 2012 Trade Show and our booth was immediately next to Canon's large booth. I spent a fair amount of time chatting with their camera representatives. They stated for normal situations the 1DX and 5D Mark III perform equally well. For low light/high ISO situations the 1DX performs better.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: tron on September 22, 2012, 09:47:26 AM
The difference is obvious but ... the difference in cost is even higher!

Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: pedro on September 22, 2012, 09:49:23 AM
I was at the WFX 2012 Trade Show and our booth was immediately next to Canon's large booth. I spent a fair amount of time chatting with their camera representatives. They stated for normal situations the 1DX and 5D Mark III perform equally well. For low light/high ISO situations the 1DX performs better.

Due to it specs it must not perform differently than this! So I am happy for the PJs and everyone who can afford this cam. The 5D3 for me is plenty of camera, though. At almost half of the price. Coming from a 30D, ISO 25k is awesome...I even dare to go up as high as ISO 51kk and 102 k for b/w.
It's taking me back to my Contax 139 Quartz days some 30 years ago  8) 102k look similar to Tri-X pan pushed to ISO 3200  ;D Cheers, Pedro
102k
http://www.flickr.com/photos/guatitamasluz/7925050234/#in/photostream (http://www.flickr.com/photos/guatitamasluz/7925050234/#in/photostream)
51k
http://www.flickr.com/photos/guatitamasluz/7925036148/#in/photostream (http://www.flickr.com/photos/guatitamasluz/7925036148/#in/photostream)

Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: bdunbar79 on September 22, 2012, 10:07:24 AM
In using both I can tell you the 1DX will always auto ISO at 1/3 lower than the 5D Mark III.  I've done indoor controlled tests and auto ISO on the 1DX always gave 5000 with 5D3 at 6400, etc.  Every ISO setting.  So when you say the metering is different, that is certainly correct. 
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Woody on September 22, 2012, 10:27:38 AM
Regardless of how good the 1DX sensor is, it is still slightly behind the Nikon D4:
http://www.juzaphoto.com/article.php?l=en&article=113 (http://www.juzaphoto.com/article.php?l=en&article=113)

"The performance of the two cameras is similar, but from ISO 6400 upwards the Nikon D4 is the winner by a little margin: I'd say that it has about 0.5 stop better high ISO performance than the 1DX; the D4 shows a little more noise and it has a more easy to remove, fine granulated noise."
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: DB on September 22, 2012, 10:42:04 AM
Regardless of how good the 1DX sensor is, it is still slightly behind the Nikon D4:
http://www.juzaphoto.com/article.php?l=en&article=113 (http://www.juzaphoto.com/article.php?l=en&article=113)

"The performance of the two cameras is similar, but from ISO 6400 upwards the Nikon D4 is the winner by a little margin: I'd say that it has about 0.5 stop better high ISO performance than the 1DX; the D4 shows a little more noise and it has a more easy to remove, fine granulated noise."

The 1DX exposure is darker in every shot vs the D4 (look at light flare and sunlight reflections as well as more color saturation in the 1DX pictures).

Perhaps, as Brett has said, there is also a metering differential between the 1DX and the Nikon D4 or the gap between the actual observed ISO and nominally quoted ISO vary between the 2 models.

I prefer the 1DX photos and fail to see hardly any difference in ISO performance (especially at say 12,800)
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: bdunbar79 on September 22, 2012, 10:43:58 AM
Regardless of how good the 1DX sensor is, it is still slightly behind the Nikon D4:
http://www.juzaphoto.com/article.php?l=en&article=113 (http://www.juzaphoto.com/article.php?l=en&article=113)

"The performance of the two cameras is similar, but from ISO 6400 upwards the Nikon D4 is the winner by a little margin: I'd say that it has about 0.5 stop better high ISO performance than the 1DX; the D4 shows a little more noise and it has a more easy to remove, fine granulated noise."

I'd love to see keepers from a football game at ISO 25,600 from a D4, especially printed out 8 x 10.  The overall performance of the two cameras aren't the same either.  This of course coming from actual real-life users who shoot every weekend at night games and not an article on the internet.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: westr70 on September 22, 2012, 10:54:49 AM
The difference is obvious but ... the difference in cost is even higher!

+1 Right on.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: emag on September 22, 2012, 10:59:00 AM
WOW!!!  Someone doesn't know how to park! 
Oh......yeah........both cameras are wonderful, choose one and enjoy.  You won't be disappointed.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: tron on September 22, 2012, 11:14:30 AM
WOW!!!  Someone doesn't know how to park! 
Oh......yeah........both cameras are wonderful, choose one and enjoy.  You won't be disappointed.
+1   (on both comments) ;D
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: tron on September 22, 2012, 11:17:12 AM
A big difference
A bid difference maybe but at ISO 25K. How about 12800 and 6400?
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: bdunbar79 on September 22, 2012, 11:18:40 AM
A big difference
A bid difference maybe but at ISO 25K. How about 12800 and 6400?

I did a test there and the difference from ISO 6400 to 12800 seems to flatten out and the difference doesn't seem to change at all.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: jebrady03 on September 22, 2012, 11:21:03 AM
IMHO, there is a noticeable difference in lighting color - as evidenced by the reflection in the windshield.
One minute can make a huge difference during sunset

Completely agree.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: mrsfotografie on September 22, 2012, 11:40:47 AM
IMHO, there is a noticeable difference in lighting color - as evidenced by the reflection in the windshield.
One minute can make a huge difference during sunset

Completely agree.

+1. Not a scientific test ;)
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Sony on September 22, 2012, 11:42:13 AM
Regardless of how good the 1DX sensor is, it is still slightly behind the Nikon D4:
http://www.juzaphoto.com/article.php?l=en&article=113 (http://www.juzaphoto.com/article.php?l=en&article=113)

"The performance of the two cameras is similar, but from ISO 6400 upwards the Nikon D4 is the winner by a little margin: I'd say that it has about 0.5 stop better high ISO performance than the 1DX; the D4 shows a little more noise and it has a more easy to remove, fine granulated noise."
Take another look at http://www.imaging-resource.com/IMCOMP/COMPS01.HTM (http://www.imaging-resource.com/IMCOMP/COMPS01.HTM)
We'll see that at high ISO the black in Nikon's images shift to red while Canon's maintain.
But it's not a big deal as rarely high ISOs are used. It's just a technical fact.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: clicstudio on September 22, 2012, 11:48:21 AM
Glad I own a 1D X.   :o
But then again the noise and color accuracy are fixable.
You can't fix the price difference.  ::)
Still, I must say the settings are kind of ridiculous.
1/6400 at F2.2?
My best friend has a 5D III. I will try to do a comparison too to see how they perform with the new 24-70 II
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: cervantes on September 22, 2012, 11:59:38 AM
Since it doesn't seem to be possible to post on the website of the tester I'd like to post here:

Due to the fact that the 5D3 pictures are brighter than the 1DX it appears to me that Canon used a little trick to "improve" the 1DX's noise performance, which is that the Camera displays ISO 25,600 but in fact it uses just ISO 20,000 or so. If both cameras were using the same ISO-speed the brightness should match in my opinion. Of course the 1DX is supposed to have better noise performance at ISO 20,000 than the 5D3 on 25,600.

So i think that a second comparison should be done using exactly the same brightness (which means the same ACTUALLY ISO-speed) while disregarding displayed ISO numbers. This should show the real difference in noise between the two cameras.

I'd love to see that comparison!
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Viggo on September 22, 2012, 12:01:07 PM
The metering difference I don't get why people are using as an argument when it comes to iso evaluation. First off, which metering is more accurate? Absolutely no contest ... and in the X you can offset your 0 ev, mine is at +5/8. I've shot 20k shots on the 5d3 and 11k on the 1d X and the difference is huge in metering precision and noiseperformance. I always need to to adjust the metering back and forth on the 5d to get what I want which really slows down shooting, is very annoying and makes me miss moments. The 1d is consistent no matter what light it's crazy, i hardly ever touch the ev scale.  I guess all the 5d owners don't know or don't want to know and needs to defend their purchase. I too defend the 1d X over the 5d, and i have used them both to the limits. And it just isn't right the difference is minor and not worth the price. It might not be worth it to some people, but the 1d is twice the camera.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: dafrank on September 22, 2012, 12:11:20 PM
As a professional who has shot cars and trucks for car makers and their agencies for over 25 years, let me say that it is highly unlikely that there would be much if any difference in the color rendition of the two cameras had the color of the lighting been identical. As others have already pointed out, the color of the cloud bank reflected in the windshield of the 5D3 is obviously hugely warmer than the clouds reflected in the windshield of the 1DX. There is no question in my mind that this accounts for the vast majority, if not 100%, of the color difference.

At sunset, when clouds across the horizon are moving quickly, dissapating due to weather front movement or, more likely, atmospheric cooling that reduces the amount of evaporated water vapor in the air, the sky color often changes very rapidly as the very warm (low color temperatrure expressed in Kelvin degrees) source of light from the low-sky direct sunlight may be revealed as it peeks out between low hanging cumulus clouds, or is revealed as the clouds covering it are thinned out or disappear entirely. If you look at the reflection in the 5D3 windshield, one can see this effect in the reflected clouds which have obviously been lit by much warmer light from the setting sun. This is all very clear from the examples.

As to the color noise, it was also very much affected by the color temperature. The "grain" size and frequency of the noise was obviously affected by the warmer color light; because the subject was much warmer (redder) in the 5D3 shot, it was recorded much more heavily in the red channel, the channel which always reproduces much more noise than either the blue or green channels. This is the same reason why images lit by low light tungsten lighting are always noisier than those lit by low light daylight-colored light sources. You can easily see this by the vastly greater number of red noise clumps in the 5D3 shot, making the image look more "noisy" than the image from the cooler (less red, more blue and green) lit 1DX shot.

To summarize, the color rendition issue is probably due entirely to ambient color temperatures at the time of each shot, and the noise differences are probably almost as much due to the same influence; without the color change, the 1DX probably would still have had less noise, but only a tiny fraction less, and the noise evident on the 5D3 would have been of the same general color range as visible in the shot taken with the 1DX.

Regards,
David
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: tron on September 22, 2012, 01:37:56 PM
As a professional who has shot cars and trucks for car makers and their agencies for over 25 years, let me say that it is highly unlikely that there would be much if any difference in the color rendition of the two cameras had the color of the lighting been identical. As others have already pointed out, the color of the cloud bank reflected in the windshield of the 5D3 is obviously hugely warmer than the clouds reflected in the windshield of the 1DX. There is no question in my mind that this accounts for the vast majority, if not 100%, of the color difference.

At sunset, when clouds across the horizon are moving quickly, dissapating due to weather front movement or, more likely, atmospheric cooling that reduces the amount of evaporated water vapor in the air, the sky color often changes very rapidly as the very warm (low color temperatrure expressed in Kelvin degrees) source of light from the low-sky direct sunlight may be revealed as it peeks out between low hanging cumulus clouds, or is revealed as the clouds covering it are thinned out or disappear entirely. If you look at the reflection in the 5D3 windshield, one can see this effect in the reflected clouds which have obviously been lit by much warmer light from the setting sun. This is all very clear from the examples.

As to the color noise, it was also very much affected by the color temperature. The "grain" size and frequency of the noise was obviously affected by the warmer color light; because the subject was much warmer (redder) in the 5D3 shot, it was recorded much more heavily in the red channel, the channel which always reproduces much more noise than either the blue or green channels. This is the same reason why images lit by low light tungsten lighting are always noisier than those lit by low light daylight-colored light sources. You can easily see this by the vastly greater number of red noise clumps in the 5D3 shot, making the image look more "noisy" than the image from the cooler (less red, more blue and green) lit 1DX shot.

To summarize, the color rendition issue is probably due entirely to ambient color temperatures at the time of each shot, and the noise differences are probably almost as much due to the same influence; without the color change, the 1DX probably would still have had less noise, but only a tiny fraction less, and the noise evident on the 5D3 would have been of the same general color range as visible in the shot taken with the 1DX.

Regards,
David
Thank you for this analysis
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: stewy on September 22, 2012, 05:02:17 PM
Ummm... wasn't this already known from the start. From the initial batch of high ISO 5D3 pictures I've seen, there was always a color shift when going from 12800 to 25600. If the photo was taken with the 5D3 at 12800 then there wouldn't be a color shift. In either case, the color shift can be corrected in post and is the least of your worries.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: bdunbar79 on September 22, 2012, 05:34:58 PM
The metering difference I don't get why people are using as an argument when it comes to iso evaluation. First off, which metering is more accurate? Absolutely no contest ... and in the X you can offset your 0 ev, mine is at +5/8. I've shot 20k shots on the 5d3 and 11k on the 1d X and the difference is huge in metering precision and noiseperformance. I always need to to adjust the metering back and forth on the 5d to get what I want which really slows down shooting, is very annoying and makes me miss moments. The 1d is consistent no matter what light it's crazy, i hardly ever touch the ev scale.  I guess all the 5d owners don't know or don't want to know and needs to defend their purchase. I too defend the 1d X over the 5d, and i have used them both to the limits. And it just isn't right the difference is minor and not worth the price. It might not be worth it to some people, but the 1d is twice the camera.

I have both and use both heavily.  There's not a bit damn difference that makes any difference in metering between the two cameras.  We're talking 1/3 of a stop.  It doesn't cost me any photos and noise cleanup at ISO 12,800 is similar, they're both highly useable and I've printed 8 x 10's from BOTH at that ISO level and they look great.  I suppose I'm just defending my purchase of the 5D3 and justifying it?  Well yeah, with facts and everyday useage.  It seems to me you're doing the same thing with the 1DX that you accuse 5D3 owners of doing, aren't you?  No offense meant, however.  It just seems you're doing the exact same thing about the 1DX as supposedly 5D3 owners are doing.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Viggo on September 23, 2012, 03:20:24 AM
Guess my reply was censored, well that pretty much sums it up...
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: NotABunny on September 23, 2012, 11:53:08 AM
Maybe the photo from 5D3 should be processed with another RAW processor. Maybe there is a problem during decoding.



Regardless of how good the 1DX sensor is, it is still slightly behind the Nikon D4:
http://www.juzaphoto.com/article.php?l=en&article=113 (http://www.juzaphoto.com/article.php?l=en&article=113)

"The performance of the two cameras is similar, but from ISO 6400 upwards the Nikon D4 is the winner by a little margin: I'd say that it has about 0.5 stop better high ISO performance than the 1DX; the D4 shows a little more noise and it has a more easy to remove, fine granulated noise."

He still doesn't understand the difference between noise per pixel and noise per image. He's also not saying how he got the same image resolution and field of view for different sensor resolutions. Were the cameras at different distances? Did he resize the images?

Nevertheless, I appreciate his effort in getting the gear and making the images. The 1DX's noise is blotchier.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: NotABunny on September 23, 2012, 12:00:48 PM
The metering difference I don't get why people are using as an argument when it comes to iso evaluation.

Because the amount of light that gets on the sensor affects the noise level. The ISO doesn't change the amount of light that gets on the sensor. Therefore, identical optical exposures are mandatory. For a scientific test, the ISO has to be changed so that the brightness of the images from both sensors is the same, which means that the displayed ISO value may be different.
Title: Re: A Quick Color & Noise Comparison: Canon EOS-1D X vs Canon EOS 5D Mark III
Post by: Viggo on September 23, 2012, 02:05:15 PM
The metering difference I don't get why people are using as an argument when it comes to iso evaluation.

Because the amount of light that gets on the sensor affects the noise level. The ISO doesn't change the amount of light that gets on the sensor. Therefore, identical optical exposures are mandatory. For a scientific test, the ISO has to be changed so that the brightness of the images from both sensors is the same, which means that the displayed ISO value may be different.

Yeah, for a scientific test point. But when metering and shooting a scene it doesn't matter. Less light more noise, but if one camera always meter under it creates noise. And when things happen fast and the metering is off, like with all three 5d's, that underexposes the problem isn't the displayed iso's. The difference is lab tests and how the camera creates the picture.