Another 5d classic shot. Pano shot with ND grad. Foreground shadows pushed a bit in LR. Would this be a problem with the 5dII?

But they aren't perfect, and the shadow noise does sometimes cause difficulties when I want to stretch dynamic range to the maximum.

Patient: "Doctor, it hurts whenever I do this."
Doctor: "Don't do that."

Doo-bop wrote:
5DMKII flaws, please post your best pictures which were sadly lost through pattern noise or whatever flaw the Canon has to offer. No test shots please.

People are invited to offer solutions to safe the images or tips how to avoid the problem in the futur.

I am eager to see lots of unique images which were rendered useless by the 5D

added:
The thread s purpose is that we have a place to discuss problems of an excellent camera and offer solutions where possible. I am sick of people who jump in all threads with always the same irrelevant test shots.
At the same time I think with a new tool you always have to learn about limits and problems.



lost is kinda of extreme

i have some shots where some more usable DR would've been nice, but they were not lost

occasionally the pattern banding is annoying, but mostly the pics looks amazing

michael49 wrote:
I used a 2 stop ND grad with in this shot, but I also pushed the shadows on the rock in the foreground without fear with my 5d classic. Would this be an issue with the 5dII?

That depends on how underexposed the rock was and how much you pushed the exposure there. But there are defintely some reds in there, so...

michael49 wrote:
Foreground shadows pushed a bit in LR. Would this be a problem with the 5dII?

Again, depends on how much is "a bit"...

michael49 wrote:
.... I use a 5d classic now and I shoot mostly landscapes and nature. I use ND grads and blends, but I also often push the shadows in PP as well. Thus, if this really problem showed up in my images with a 5DII then it would be an issue for me. I used a 2 stop ND grad with in this shot, but I also pushed the shadows on the rock in the foreground without fear with my 5d classic. Would this be an issue with the 5dII?

Hard to say for sure without knowing how much "push" we're talking about. But, overall ... and in general ... I'd say no problem.

In the shot below, the foreground (the side of the barn on the right) was very, very dark and near black for the most part. I pushed the shadow control in DPP then selectively brightened the area in Photoshop. Shot at ISO 400 w/24-105 f/4L.

David Baldwin wrote:
Dawei Ye,

If you read this please see how I have apologetically amended my previous post above.
Thanks David, no apology necessary you were absolutely right with your logic

M Vers wrote:
? The noise patterns are clearly different from example to example. Again, this looks like processing issues NOT sensor issues. All of this talk about gridded pattern noise and yet there is none in the above example, rather blotchy 'noise' or compression artifacts instead. Of course the best way to find out is to DL the original file to have a look for yourself.
---

corndog wrote:
Agreed, resize for web issue.
---

form wrote:
That kind of noise looks like JPG compression artifacts.
---

I can't really prove it's not JPG Compression artifacts. I guess you just have to take my word for it that what you see in those screenshots is visible on my screen in Canon DPP. I took great care to make sure the shot was not posterised in any way. As you see, the text on the EXIF is crystal clear, no halos or artifacts whatsoever...

The reason it doesn't look like the "pattern" in the other shots is because the area is quite dark and not the typical "grey/brown" type colour, and it has not been pushed yet to bring out the pattern (my photo had zero adjustment). If I do a "cubic function" curve to it, this is what it looks like:

Obviously no sane person does this, but this is just to show you that it is not JPG compression that is causing those strange markings...it is the underlying patterned noise at ISO 100...it is not JPG compression artifacts...

From what I can tell, the criss cross pattern/banding is most obvious when shooting in low light (especially under mixed yellow/red indoor lighting) and against dark/underexposed backgrounds. It seems to be primarily in the red RGB channel. You need to push exposure/shadows to make it visible.

Dawei Ye wrote:
Obviously no sane person does this, but this is just to show you that it is not JPG compression that is causing those strange markings...it is the underlying patterned noise at ISO 100...it is not JPG compression artifacts...


The noise pattern still looks entirely different from one example to the next. If you post the image at its original size somewhere you can see this. In the latest example the criss cross pattern is clearly there, while its not in the original where there are blotchy artifacts. Posting the original file, untouched, for people to view is the only way we will ever know.

Daan B wrote: That depends on how underexposed the rock was and how much you pushed the exposure there

I decided to test my recent 5DII. I shot a high dynamic range scene, by chance, at ISO 400: bright full sun on white stucco and beige paint, yet on the other side, a very dark garage with car and black tires. The histogram showed levels from slightly blown highlights to inky black shadows, and every tone in between. I processed it normally in DPP. It showed an excellent full dynamic range image.

Then in PS, I went into levels, and pulled the 0 levels up to 50 (out of 255), rendering the blacks to a very uniform but ugly gray. That is +4 stops? I examined the tire and shadows, at 100% pixels, and to the most hyper-critical observer, there was no banding whatsoever.

What is the formula that gives perfect banding?

Don Clary wrote:
Daan B wrote: That depends on how underexposed the rock was and how much you pushed the exposure there

I decided to test my recent 5DII. I shot a high dynamic range scene, by chance, at ISO 400: bright full sun on white stucco and beige paint, yet on the other side, a very dark garage with car and black tires. The histogram showed levels from slightly blown highlights to inky black shadows, and every tone in between. I processed it normally in DPP. It showed an excellent full dynamic range image.

Then in PS, I went into levels, and pulled the 0 levels up to 50 (out of 255), rendering the blacks to a very uniform but ugly gray. That is +4 stops? I examined the tire and shadows, at 100% pixels, and to the most hyper-critical observer, there was no banding whatsoever.

What is the formula that gives perfect banding?

Hysterical question! . Thanks for the levity.

- Steve

Don Clary wrote:
What is the formula that gives perfect banding?

Pushing the shadow slider to the maximum in DPP...

I can't seem to replicate the phenomenon in any of my 5D Mark II images in CS4 or LR2 until ISO 3200 or higher, so it's pretty much a non-issue for me.

Once upon a time we had different films, and these films could have enormously different characteristics. Dynamic range was among the characteristics that could vary dramatically between films.

Interestingly, it was very difficult to find a film that had both the highest resolution and finest grain and the longest dynamic range--resolution and dynamic range tended to vary inversely with film.

If you wanted extreme resolution and extreme dynamic range at the same time, you could get it at a price--just go medium format. But that price was not only a much more expensive camera, but also a camera that was much more difficult and slower to handle.

It seems going digital hasn't changed the fundamental choices. There still ain't no such thing as a free lunch.

It seems to me that DPP seems to be the main problem.
I haven't seen these sort of issues with acr on cs4.
But them again I haven't pushed my images to the obscene levels I have seen here, any badly underexposed shots get binned.

Don Clary wrote:
What is the formula that gives perfect banding?

That is something I try determine myself... because if I know that, I can work around it.

In the meantime, here are some observations I made:

1) The banding shows regardless of ISO values. I can get banding to show in ISO100 files, but in ISO3200 as well.

2) The quickest way (for me) to reveal any kind of banding is pushing the LR2.5 fill light slider to +100 (maximum). Even though I would never use such a drastic measure for real... only for testing purposes.

3) There seems to be a direct relation between the amount of underexposure (for either the whole image or certain parts of it) and how much banding will be visible in the shadows after pushing like descibed under #2 -in high contrast scenes that exceed the DR of the sensor excessively. But I haven't determined a very precise threshold yet. In any case, it is all related to how far you will exceed the DR of the 5D2's sensor if banding will show when you push the shadows. And that differs from scene to scene... To avoid or lessen banding, "exposing to the right" (or even slighty overexposing) is still the way to go

I'm considering to get a 5DII and therefore thought this thread might give me clues.
What I don't understand is how people can (technically) judge a picture on a screen.

The picture that hits the eye is processed by a couple of programs in series, starting with the amplification in the camera and finished by the display driver. Everything in the chain is adding, changing and removing things and consequently has effects on the result that hits the eye.
The camera has a limited DR and any software to push the result beyond the DR limits is showing the limits. Why would we complain about that? Spend some bucks to get a wider DR if you need that, or do you think I'm misunderstanding the "problem"?

Perceptor wrote:
The camera has a limited DR and any software to push the result beyond the DR limits is showing the limits. Why would we complain about that?

We don't... at least I don't.

Spend some bucks to get a wider DR if you need that, or do you think I'm misunderstanding the "problem"?

Yes.

Yes.


Yes what?

Perceptor wrote:


Yes.


Yes what?

Yes, I think you are misunderstanding the "problem"

Seems to be a few camps here on FM - "There is no problem if you use the camera properly", and "this is a horrible problem." When I got my 5dii, i pushed it hard to try to see the banding problem, and i could see it in a few examples. It didn't bother me because It wasn't real shooting.
Every once in a while though, it rears it's head when I don't expect it. Look at this shot, it is a throw-away because I under-exposed. background is a resampled web size shot, with a 100% crop of the center. In the 100% crop, there are 2 sides, one with exposure "pushed" one half stop. Shot data: RAW, iso200, no sharpening, no noise reduction. I can see the pattern before the exposure is even pushed.