Noise About Noise

[Panopeeper]

Mark,

I sent you a direct email through the forum's system.

Gabor

[bjanes]

Bill,

ISO 100 is the key. I do not have any other ISO 100 shot. I verified shots wityh perhaps all other ISOs, and all get +0.5.

I suggest you to make two shots of the same scenery (perhaps the Stouffer), with the same exposure, but one with ISO 100, the other with ISO 200, and compare their histograms. There should be some clipping as well.
Gabor
[a href=\"index.php?act=findpost&pid=200150\"][{POST_SNAPBACK}][/a]

Gabor,

I went through my test shots and found what you want. The exposure is the same for both shots, 1/15 s, f/8. As the histograms show, the same values are written to the raw file for both exposures, but they are processed differently by ACR according to the BaselineExposure value, which is -0.5 for the ISO 100 and +0.5 for the ISO 200. If I cancel out these values by adding +0.5 EV in ACR for the ISO 100 and -0.5 EV for the ISO 200, the results are more or less the same.

BTW, Rawanalyze reports ISO 200 for the ISO 100 exposure.

Bill

Histograms, 100 left, 200 right:
[attachment=6958:attachment][attachment=6959:attachment]

ACR defaults, 100 left, 200 right:
[attachment=6954:attachment][attachment=6955:attachment]

ACR Plus 0.5 EV left, Minus 0.5 EV right:
[attachment=6956:attachment][attachment=6957:attachment]

[Panopeeper]

Bill,

this is now clear; the D3 does not have any ISO 100 gain. I find it incredible, that this will not be documented; if you don't know this and meter for ISO 100, you get everything burned out. Don't you think you should post it on DPReview?

Re the ISO display by Rawnalyze: Nikon plays a dirty game. In this case there is no ISO spec in the Exif section. There is a proprietory tag in MakerNote for ISO, that is invalid *in this case*. Then there is another proprietory tag, which specifies ISO 200 in this case, but there is something else in that tag, which I have not interpreted yet. Anyway, I will include this if not otherwise then hard wired. The way this appears with Canons (the Highlight Tone Priority changes the ISO internally): ISO 100/200, indicating the nominal and factual ISO.

[gr82bart]

So is this a site about photography or camera equipment engineering? Or boys who think they are the latter?

Regards, Art.

[Tim Gray]

So is this a site about photography or camera equipment engineering? Or boys who think they are the latter?

Regards, Art.
[a href=\"index.php?act=findpost&pid=200294\"][{POST_SNAPBACK}][/a]

The site is largely un-moderated, so it's about whatever people post about.

I like that.

[Mark D Segal]

So is this a site about photography or camera equipment engineering? Or boys who think they are the latter?

Regards, Art.
[a href=\"index.php?act=findpost&pid=200294\"][{POST_SNAPBACK}][/a]

This thread results from an article I contributed about noise and resolution, which are important aspects of image quality. The people contributing to this discussion are approaching it from the perspective of the underlying fundamentals. If that's of no interest to you, fine, but it's not a reason to dump on those for whom it is.

[Panopeeper]

I just love, when people - usually anaphabetes of the digital aspects of photography - are complaining about openly discussing subjects they don't understand.

[Panopeeper]

Bill,

I found the "fine print" to the "Lo" ISO setting, it is specified as Lo 0.3, Lo 0.5, Lo 0.7 and Lo 1.0. I have to translate these in ISO numbers.

I guess, the third stops are 125 and 160 on Nikon too. But how does the half-stop ISO between 100 and 200 appear? 140? I'm pretty sure it is not 141.

Do you mind checking these out? Then I can display the ISO properly, like 125/200 or 100/200.

Thanks
Gabor

[Panopeeper]

Mark,

the shot I received has not been created by a neutral WB (UniWB); this is apparent by looking at it in ACR: it appears in the right colors. With a neutral WB it would appear rather greenish. Thus it is not surprizing, that the in-camera histogram does not reflect the raw data.

However, I don't understand, why the clipping is not indicated on the in-camera display. I extracted the preview JPEG, and it is properly white balanced, and it is clearly blown. For example the top right window: there is some sheet (I think originally bluish) hanging on the glass pane inside; the green is blown overall and the blue on most of it. This sheet appears in the preview with RGB=(255,255,255).

If you have access to the camera again, you should put back this image on a card and check it out again.

Gabor

[Mark D Segal]

I just love, when people - usually anaphabetes of the digital aspects of photography - are complaining about openly discussing subjects they don't understand.
[{POST_SNAPBACK}][/a]

Gabor,

The word is "analphabet" : [a href=\"http://www.smba.nl/en/newsletters/n-84-analphabet/]http://www.smba.nl/en/newsletters/n-84-analphabet/[/url]  

[Mark D Segal]

Mark,

the shot I received has not been created by a neutral WB (UniWB); this is apparent by looking at it in ACR: it appears in the right colors. With a neutral WB it would appear rather greenish. Thus it is not surprizing, that the in-camera histogram does not reflect the raw data.

However, I don't understand, why the clipping is not indicated on the in-camera display. I extracted the preview JPEG, and it is properly white balanced, and it is clearly blown. For example the top right window: there is some sheet (I think originally bluish) hanging on the glass pane inside; the green is blown overall and the blue on most of it. This sheet appears in the preview with RGB=(255,255,255).

If you have access to the camera again, you should put back this image on a card and check it out again.

Gabor
[a href=\"index.php?act=findpost&pid=200317\"][{POST_SNAPBACK}][/a]

I believe the D3 shots were made using the camera's Auto WB. Looking at the file in Rawnalyze, I see that there is moderate red and blue channel clipping. This is what did not show on the camera's aggregate histogram. The stuff blown in the window should be considered specular highlights - white paper against glass with direct sunlight bouncing off of it. I doubt the card still contains those images. They'd have to be reloaded onto a card and reviewed in the camera. I'm not sure I'll be able to get that done, but if I can I'll report back. At this stage, let me just say that two of us obsrved the same phenominon.

[Panopeeper]

The word is "analphabet"

Thanks; actually, I happen to know that. However, my typing is getting worse with my new keyboard (formerly I have used the same IBM keyboard 14 years long), and these minisculous fonts on the forum don't help my aging eyes.

two of us obsrved the same phenominon

Right, but my issue was and is, why the camera did not indicate the clipping (that was my understanding).

Btw, the word is "observed", the other one is "phenomenon"   (but I don't intend to get into a spelling war anyway; English is my third language).

[Mark D Segal]

Thanks; actually, I happen to know that. However, my typing is getting worse with my new keyboard (formerly I have used the same IBM keyboard 14 years long), and these minisculous fonts on the forum don't help my aging eyes.
Right, but my issue was and is, why the camera did not indicate the clipping (that was my understanding).

Btw, the word is "observed", the other one is "phenomenon"   (but I don't intend to get into a spelling war anyway; English is my third language).
[a href=\"index.php?act=findpost&pid=200324\"][{POST_SNAPBACK}][/a]

Touché - but didn't you like the content of the address I provided?  
I had no doubt you knew the word, but I couldn't resist. And yes I need to use the spell check - going too fast for my own good.

Your understanding is correct - the camera did not indicate the clipping. It surprised us to see it as well when we opened the files in ACR. I can only think it is something to do with the JPEG settings.

[bjanes]

Bill,

I found the "fine print" to the "Lo" ISO setting, it is specified as Lo 0.3, Lo 0.5, Lo 0.7 and Lo 1.0. I have to translate these in ISO numbers.

I guess, the third stops are 125 and 160 on Nikon too. But how does the half-stop ISO between 100 and 200 appear? 140? I'm pretty sure it is not 141.

Do you mind checking these out? Then I can display the ISO properly, like 125/200 or 100/200.

Thanks
Gabor
[a href=\"index.php?act=findpost&pid=200314\"][{POST_SNAPBACK}][/a]

Gabor,

Using half stop increments, I was unable to determine the ISO for Lo 0.5. ViewNX reports the ISO as "0.5 EV under 200". ACR and Bridge leave the ISO blank and the BaselineExposure in the DNG is 50/100.

For third stop increments, ACR reports the ISOs as 125 and 160. The DNG BaselineExposure values are -17/100 and 17/100 respectively.

Bill