Michael, please discuss (shadow) noise in 1Ds-III

[Brad_Stiritz]

"Banding destroys this camera's viability"

An unsubstantiated statement by an unknown person. Why even give it the time of day?

Anyone can write any rubbish on the net, with any motive.

Michael

[billh]

If possible, I'd like to download one or more CR2 files from a 1Ds-III, so that I can check them out carefully in Camera Raw 4.3 and PS CS3. If anyone can point me in the right direction, I'd really appreciate it. (I did check out Ted Pedersen's test images on dpreview.com, but only saw JPGs posted, not raw captures.)


Brad
[{POST_SNAPBACK}][/a]

Here is a link to new 1DS MKIII ISO 200 and ISO 1600 Raw-samples:

[a href=\"http://www.walther-fineart.de/1dsmkIII.htm]http://www.walther-fineart.de/1dsmkIII.htm[/url]

[Brad_Stiritz]

Here is a link to new 1DS MKIII ISO 200 and ISO 1600 Raw-samples:

http://www.walther-fineart.de/1dsmkIII.htm
[a href=\"index.php?act=findpost&pid=158023\"][{POST_SNAPBACK}][/a]

Danke, Bill!

[Wayne Fox]

"Banding destroys this camera's viability"
http://reviews.cnet.com/digital-cameras/ca...56&cval=2457556


[a href=\"index.php?act=findpost&pid=158013\"][{POST_SNAPBACK}][/a]

You have any other information regarding banding?  

A single post about a camera thousands are using ... on Cnet no less.  Personally I don't see any value to that review at all.

Certainly evaluating files is important, but it seems that after 5000 images Michael would have piced up on any banding problems.

[Brad_Stiritz]

Michael, Wayne,

Regarding the anonymous claim of banding in a prior Canon 1D camera model:

"Banding destroys this camera's viability"

Michael, you wrote:
>An unsubstantiated statement by an unknown person. Why even give it the time of day? Anyone can write any rubbish on the net, with any motive.

Wayne, you wrote:
>You have any other information regarding banding? A single post about a camera thousands are using ... on Cnet no less. Personally I don't see any value to that review at all.

Very good questions. I was remiss in my original (now redacted) posting not to have mentioned that I also had done my own quick investigation *before* reading that anonymous claim.

* I found a 10.1MB CR2 file that s/o had posted on the web, downloaded it, and examined in CS3 Camera Raw 4.3.
* Though there was no camera info metadata in the file, I presumed the image was from a late-model Canon 1D camera. I could be completely wrong here, I admit.
* At high magnification, this one image clearly shows vertical banding in the deep shadow regions, similar in appearance to the banding I see occasionally in scans of underexposed film (Nikon 9000 scanner).
* Sample size = 1 for this investigation. I know, I know, not very respectable...

I admit, it was sloppy of me not to have taken the time to have included all the above info in my original posting. I apologize.

I did also carefully examine the two 1Ds Mark III raw files (that billh was very kind to have posted a link to  and did not find any vertical banding. So it does seem fair to state (very provisionally) that banding may well have been a problem in at least one earlier model, and that the problem may well not exist in the current TOTL model.

Again, if anyone knows where I can find more 1Ds-III CR2 files on the net, that would be super-helpful, thanks!

Brad

[Seth Honeyman]

Hi Brad

Rob Galbraith (the same fella that broke the news about the 1D Mk111's dodgy AF performance) was very complementary about the camera's overall image quality.  In his review, he said:

"...... Of the two dominant digital SLR brands - Canon and Nikon - it's Canon that already produces cameras with the best high-ISO image quality and the most natural-looking darker tones. At ISO 800 and above it's no contest really, and the EOS-1D Mark III only widens the gap further. This camera produces images with the smoothest shadow gradations we've seen in a digital SLR from the two makers. Shadow naturalness is still decent at ISO 3200, which is a first to our eyes. Combine that with lower noise levels at all ISOs, and especially those at the top of the camera's range, and the result is a noticeably better image file whether you shoot at ISO 100 or ISO 3200 all day long.
 
Comparing photos of the same scenes from the EOS-1D Mark III and EOS-1D Mark II N, both on-screen and in 16 x 24 inch prints, we see almost a full stop improvement in noise levels at the upper ISO settings. But if you factor in shadow naturalness - which impacts the quality of high-ISO photos almost as much as noise when it's time to print them - plus, the fact that the EOS-1D Mark III doesn't seem to show the same faint bands that sometimes crop up at higher ISOs with the EOS-1D Mark II N, and the improvement is probably more like 1.5 stops."

  (The review was written in April - before release of the D3)

Regards- Seth

[Brad_Stiritz]

Hi Michael, Hi SDH,

Michael, thanks for linking to Phil Holland's 1Ds-III review from What's New. Phil has a couple of CR2 files there for downloading, which have been very interesting to check out in CameraRaw. I'm just getting into CR for the first time, so I can't say too too much, but the CR2s are very encouraging.

SDH, thanks for posting Rob Galbraith's comments. Now that I've personally examined several CR2's from the 1Ds-III, I definitely agree: Canon looks to have really gotten noise under firm control with this camera! This is really exciting territory for me  I might be able to finally leave the film world behind  

Brad

[geronimo]

"Banding destroys this camera's viability"

An unsubstantiated statement by an unknown person. Why even give it the time of day?

Anyone can write any rubbish on the net, with any motive.

Michael
[a href=\"index.php?act=findpost&pid=158013\"][{POST_SNAPBACK}][/a]

Michael, ISO 3200 is more than usable with 1Ds III, but ... banding noise really seems to be an issue in some cases especially if you're going to print large size. Of course you could always lower down the ISO (I havent seen banding at all @ iso1600) and use a tripod, but it depends on the the type of photography you're doing.

I could post a link to some raw files if you need real example of my experience
~Cheers~

[Panopeeper]

ISO 3200 is more than usable with 1Ds III

I think ISO 3200 is fake, i.e. it is a numerical derivative of ISO 1600. In other words, the gain is the same as with ISO 1600, but the dynamic range is reduced by one stop. This may be good for JPEG, but when you are recording raw data, you can do the same in the raw processing without reducing the dynamic range.

Addendum

I do not *know* if ISO 3200 is a fake or not; I only suspect it, because the specification mentiones this as "enhanced ISO". However, a raw file reveals the truth. The 1DMkIII has true ISO up to 6400, but its pixel sites are 32% larger than those of the 1DsMkIII.

banding noise really seems to be an issue in some cases

Do you mind posting examples for that (raw file)? Then I can demonstrate how ISO 3200 "works", and I would try to reduce the banding.