My 5D2 Has Low Iso Banding

Csae wrote in post #11154037
* * *

I am not sure how you say there is only 1 stop of DR unless i apply some sort of gamma in PP, would you be willing to elaborate a bit more ?

He means that the basic RAW data is very "compressed," and a curve is applied to make it look "natural" when you view in in the RAW converter. Here's a screenshot "properly" exposed (actually, it was exposed to the right and looks "over" exposed before exposure/contrast adjustment in normal post processing) RAW image with "linear" checked in the DPP curves box, to give you a visual idea of what he meant. You can see how the shadows would need to be opened up ("pushed" to the right in exposure) to get a natural looking image from this.

HOSTED PHOTO
please log in to view hosted photos in full size.

OK, one of the earlier threads had a poser who had a current version of rawnalyze - well, the original site is down as the owner apparently passed away.

Through luck I found that the software has been given a new home:
http://dave-anderson-photo.com …bor-rawnalyze-author-rip/

And using that I just decided to throw it at a long exposure at ISO 100 (pure black, looking for hot pixles for a fellow poster) - and well, to be honest - I cannot see any banding in my camera on the red channel.

HTTP response: NOT FOUND | MIME changed to 'image/png'

DetlevCM wrote in post #11155786
- and well, to be honest - I cannot see any banding in my camera on the red channel.

I see it.

Daniel Browning wrote in post #11156470
I see it.

Do you see it because it is there or because your brain is looking for patterns?

Csae wrote in post #11154037
When you say, push the shadows are you talking about making them darker ? Or pulling them back, making them Lighter ?

No, when I say "push", I mean to make it brighter.

Csae wrote in post #11154037
The car analogy had to do with what you DO to your film, not how many stops you are using, most multi-million dollar features are still using film, so i reflect you back to your very own first argument.

Are you saying that since digital is linear, we should not try to use 12-13 stops of dynamic range, like we can with film?

Csae wrote in post #11154037
Digital is linear right? So wouldn't that mean that even if you are using GNDs you would still only be stretching your existing DR ?

Light is linear and digital is linear. When you use GND, you change the dynamic range of the scene.

Csae wrote in post #11154037
I am not sure how you say there is only 1 stop of DR unless i apply some sort of gamma in PP, would you be willing to elaborate a bit more ?

As other folks explained, the raw converter applies the gamma for you by default. The gamma is an extreme push of the shadows.

DetlevCM wrote in post #11156488
Do you see it because it is there or because your brain is looking for patterns?

Because it is there. It's not at all like random noise, which I find pleasing enough to *add* to images that don't have enough of it.

Daniel Browning wrote in post #11156518
Because it is there. It's not at all like random noise, which I find pleasing enough to *add* to images that don't have enough of it.

If you think it's there.

To be honest I have a feeling it's more my brain finding patterns where there are none.

It would really need mathematical analysis.

DetlevCM wrote in post #11155786
And using that I just decided to throw it at a long exposure at ISO 100 (pure black, looking for hot pixles for a fellow poster) - and well, to be honest - I cannot see any banding in my camera on the red channel.

All I can say is, "wow".

Banding is the most visually intense noise there.

It doesn't matter that the standard deviation of the banding, in isolation, is only about 10% of the rest of the noise. The RAW banding is usually the same in all channels. Red is often mentioned, because in all common light sources except incandescent/halogen ones, red is weak, *plus* the camera is least sensitive to red (a stop less sensitive than green). When you shoot at ISO 100 in daylight, the red channel is really at ISO 200, and the blue channel at ISO 140. To add insult to injury, the blue and red channels only have 1/2 as many pixels as the green, making their noise even worse than the real ISO sensitivities for the channels would suggest. The stated ISO is only for the green channel. In tungsten lighting, the blue channel is at about 3.5x to 4x the stated ISO. No, this is far worse than actually shooting those channels with those higher ISO settings, because ISO 100 to 400 have almost the same read noise levels, and therefore, the blue channel in a tungsten shot at "ISO 100" can have as much read noise as the green channel at about ISO 6400. Thanks to human perception, this is not as bad in the red and blue channels as it is in the green, but it still presents extra color noise.

rx7speed wrote in post #11148409
I'm make it simple. is this statement true or false

the 5dII can show banding when the shadows to a brighter level when used at lower iso levels?

Yes. No evidence has ever been provided that there are a majority or even any 5D2 cameras that do not have banding when you pull up the shadows at ISO 100. Most attempted "counterproofs" have been based on lame little pushes, or subject matter known to be good at masking banding.

If you want lots of DR in one exposure, the 5D2 isn't going to do it. Any recent Sony, Pentax, or Nikon DSLR, or even a good P&S with RAW, can generally give better results pulling up shadows, even if the midtones are noisier in the case of P&S and APS-C.

John Sheehy wrote in post #11157373
Yes. No evidence has ever been provided that there are a majority or even any 5D2 cameras that do not have banding when you pull up the shadows at ISO 100. Most attempted "counterproofs" have been based on lame little pushes, or subject matter known to be good at masking banding.

That is the most subjective assessment I have read.
Proof that what you say is the utmost truth?
Have you got any shots you took yourself that show both points of view?
I'd be interested as I can easily show banding at nearly any iso if i push the shadows hard enough.
I want to see what you mean by masking the shadows by you showing us your personal examples so I can mask my shadows next time.

Nikolas wrote in post #11157513
That is the most subjective assessment I have read.
Proof that what you say is the utmost truth?
Have you got any shots you took yourself that show both points of view?
I'd be interested as I can easily show banding at nearly any iso if i push the shadows hard enough.
I want to see what you mean by masking the shadows by you showing us your personal examples so I can mask my shadows next time.

I can't make any sense out of your reply, but I'll say it a different way, so that maybe you'll understand what I say well enough to respond to it.

No one has *EVER* provided a 5D2 image or RAW which is capable of showing banding, which has not shown it. IOW, all the qualifying samples provided point to the probability that *all* 5D2 cameras are afflicted.

By "qualifying samples" I don't mean samples that show banding; I mean conditions and exposure capable of showing it.

There are some very vocal people on the websites who say "not mine", but none have ever delivered the goods. A blackframe from RawAnalyze should do the trick.

DetlevCM wrote in post #11155786
And using that I just decided to throw it at a long exposure at ISO 100 (pure black, looking for hot pixles for a fellow poster) - and well, to be honest - I cannot see any banding in my camera on the red channel.

HTTP response: 502 | MIME changed to 'text/html'

I think the problem here is that you're looking strictly for banding, when what you should be looking for is pattern noise. Pattern noise is any noise that is not strictly random.

This is what strictly random noise looks like:

I see it too. a very distinct cross hatch pattern. hell there is a horizontal line running right through the middle of the image that is plain as day to see to me atleast.

I see Mickey Mouse in the second one

bohdank wrote in post #11158301
I see Mickey Mouse in the second one

Me too! That must have been taken with a Mickey Mouse camera.