Just have a chance to get one, on behalf of my friend.

Enjoy it:

Noise test, 1D3 vs 1Ds2

Original


[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_1600_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_1600_100crop.jpg[/img[/url]]
[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_1600_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_1600_100crop.jpg[/img[/url]]

[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_3200_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_3200_100crop.jpg[/img[/url]]
[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_3200_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_3200_100crop.jpg[/img[/url]]

[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_6400_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1D3_6400_100crop.jpg[/img[/url]]
[url=http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_6400_100crop.jpg" target="_blank">http://www.kyphoto.hk/my_photos/1D3_test_shot/1D3_vs_1Ds2_noise/1Ds2_6400_100crop.jpg[/img[/url]]


Edited by Jeff on Jul 12, 2007 at 07:49 AM GMT (Reason: Un-embedded links to oversized images)

Holy cow! Too bad I can't afford a 1D3 (yet).

P.S. thanks kylee

(Space Cowboy's reply reminded me to be polite. I guess I was too distracted by the 1D3's excellent performance to rember to say thanks.)

Edited by jcolwell on May 19, 2007 at 04:11 PM GMT

Thank you Kylee!

Cheers.

Call me nuts but I like the 1ds2 shot better up to 3200. They seem to show much more detail and have better contrast. At 8x10 print size I have to believe they would be more impressive. Maybe too much nr in the 1D3?

twistedlim wrote:
Call me nuts but I like the 1ds2 shot better up to 3200. They seem to show much more detail and have better contrast. At 8x10 print size I have to believe they would be more impressive. Maybe too much nr in the 1D3?


No, you aren't nuts but it seems that the two cameras have very different strengths and weaknesses. The 1DMkIII has much better shadow detail but the 1DsMkII seems to have more detail in the brighter areas. Interesting.

Two years of development show the difference.

How can the 1D2 100% image be larger when the 1D3 has more MP?

dcmiller wrote:
How can the 1D2 100% image be larger when the 1D3 has more MP?

The 1D2 is not present in this test, but the 1D Mark III(10Mp) and 1Ds MKII(16.7Mp) are.

Space Cowboy wrote:
dcmiller wrote:
How can the 1D2 100% image be larger when the 1D3 has more MP?

The 1D2 is not present in this test, but the 1D Mark III(10Mp) and 1Ds MKII(16.7Mp) are.




I guess actually reading the labels instead of assuming can be good sometimes

The 1DsII may not have an NR applied by the user, but it certainly has NR in the red channel of the ISO 1600 crop. In some areas there's no red noise but has that weird NR look.

The difference in sharpness is the first thing I noticed. NR off, or not, it appears we're seeing quite a bit of default NR in the 1DIII image. I hope they reconsider in later firmware.

Interesting also, the clearly smaller DOF the FF has in the shots.

jray wrote:
The difference in sharpness is the first thing I noticed. NR off, or not, it appears we're seeing quite a bit of default NR in the 1DIII image. I hope they reconsider in later firmware.

I guess I don't see the tonality being better. The 1D3 just has more DOF.

Cancel all 1D3 orders (or perhaps Mr. Kylee is messing with us)

There appears to be a change in lighting between the two sets of shots. Extra lighting in the 1D3 shots to show the power line as well as better detail on the speakers. IF Av/Tv are identical then 1D3 has an advantage.

Please tell me if I'm wrong.

some points added

lighting the same, from a few bulbs on the celing

the focus do not rest on the same spot quite clearly ( I used the same AF point on the camera, only found this when they are enlarged on the monitor)

all shot in Av mode, no NR applied in camera or PP

I will try more samples tonight, may be at half time of Man U vs Chlsea :-)

no problem if anyone doesn't like the 1D3, it is not mine

they are not meant to show details, but noise

I will try later tonight, anyone has any idea on the improvement in methodology is clearly welcome :-)

I'm wondering, looking at pixel density from both sensors, it appears that the 1D3 has the same pixel density as the 1DsM2. This leads me to believe that the new camera is using a sensor based on the sensor in the 1DsM2. The improved noise is impressive as the 1DsM2 is no slouch with noise levels. Cutting the sensor dimensions should cut down on noise and the improved electronics (Digic III) processing and such just shows that the future is looking bright. I'm wondering if the 5D update (if and when it happens) will change the pixel density to get more on the same size chip and they use the enhanced processing power to reduce the noise levels even more. The 5D was the low noise champ until the MKIII if I'm guessing correctly. Need to see the full blown tests before we jump to conclusions. The MKIII may be losing detail in favor of noise reduction.

I have a potential issue with this test. The OP did not mention what lens and at what distance the pics were taken at. This leads me to believe that a method of "foot zooming" was used to get similar FOV. This creates a potential comparison issue.

To get a comparison of the same number of pixels, which I believe would be a better true test of noise, and especially since the photosites are nearly the same at about 7um, one would use the same lens at the same distance from the subject. In this way one would get the same number of pixels covering the same object. In this way it is possible that the noise difference may have been even more striking!

Im glad however someone is finally throwing us all a bone.

update ISO 1600 test, higher ISO test to come later

all jpg staight out, no pp/sharpening/NR

original:

suchs wrote:
I have a potential issue with this test. The OP did not mention what lens and at what distance the pics were taken at. This leads me to believe that a method of "foot zooming" was used to get similar FOV. This creates a potential comparison issue.



I used 70-200 2.8 non-IS

no zoom with lens and/or feet in the first set of test

Thanks for the great posts kylee. Any chance you can do a comparison in RAW? I assume the camera came with a version of DPP for RAW processing. Doing a comparison with RAW would let us (and you) determine how much of the noise reduction is on chip and how much is in the JPG conversion.

ISO 3200

ISO 3200

1D3 jpg


1D3 RAW from DPP

Remember that no sharpening or NR means that it's set at the lowest user setting. We don't know what the manufacturers do with these files. We don't want to pretend that there is a true neutral, unprocessed file because this encourages companies to increasingly mess with the files.

Also, looking at 100% crops puts the higher res camera at a disadvantage relative to results we would expect in the real world. Not that there is a single right way to do this. Just that attempts to be as objective as possible can lead to misleading results.

I think a true test (and he may have done this), would be to use the same lens, at the same distance with the same exact exposure settings (manually set) on both bodies with the same exact lighting. The difference in FOV should be apparent between the bodies and the images, that would be a more accurate representation of the noise relative to image size. The 1DsM2 is going to resolve more image, and the noise is really relative to what gets printed. Based on what I've seen here so far, I don't think you'd notice the slightest difference in a large print. For pixel peepers, you may be able to detect it on a monitor though. If you are going to move the cameras to compensate for FOV, you should do it both ways. Move the 1Ds closer AND move the 1DIII back. Show the effects of both actions.

Edited by radiodenver on May 19, 2007 at 09:27 AM GMT

ISO 6400

jpg