The 5D2 has strong pattern noise at ISO 100

bomzai wrote in post #8625998
Alright, let me ask you one thing. With your own eyes, do you see a beautiful picture, standing in a shadow of a highriser watching up to the sun lit tower? The one that you'd like to share with others?
My answer is - no, that just strains my eyes.

At the same rate - why do we use flashes when we want to photograph subjects against bright backgrounds? So silly of us. Let's just demand high DR and pull HDR out of our ... single shots.
Or more importantly why in a world people use flashes in such scenarios when they shoot film? Those have sufficient DR.

And the answer is simple - because no matter how much you try to correct the problem, it'll always be better to avoid problem/not make a mistake in the first place.

No one here is arguing about the composition or subject matter of the picture. The picture was chosen to demonstrate a scene with a large dynamic range. Using flash on backlit subjects is typically done, to the best of my knowledge, in order to reduce dynamic range. If you want a picture with reduced dynamic range, go ahead and use fill-flash.

As I understand it, this thread is about preserving dynamic range until the file is output.

TeamSpeed wrote in post #8626033
Yes you are off-topic, we are not talking about photographic composition and proper exposure on a certain subject, we are talking about the limitations of the sensor and DR in high contrast fields of view. If Daniel posted "hey what do you think of my shot of this church", then yes, you would be on-topic.

This is more of an engineering topic instead of an artistic topic.

Well, I saw this as a topic "engineering prevents me from achieving my goal"-kind. Hence my comments.

Yes I was under assumption that the "goal" here was not to prove that 5DII doesn't have 11.5 stop DR. Pure DR comparison was done many times in different tests, results of which are posted publicly on the Web.

Well at least these "patterns" might appeal to the quilting crowd...

I have sure noticed these patterns when pulling the exposure back from a heavily underexposed photo, of course due to my user error.

Then I moved on to the next photo instead of blaming the camera...

Imagine that?

bomzai wrote in post #8625870
Regarding original photo and all other samples that show this "problem". All I have to say - these images would fly into Recycle Bin as fast as my hard drive can handle. Regardless of the noise or banding.

Proper exposure is the King, not all this techie junk. If you have to pull exposure in RAW more than 3/4 of a stop, or use shadow/highlights (in one form or another) for more than 20% - YOU have failed when you released that shutter, not the camera.

Making something useful out of a crappy shot isn't really worthwhile time spending.

Master your composition. Master your light. That's the way to successful photos.

Are you saying that a lot of great photographers are really dimwits who can't even get a proper exposure for shooting film stock and making the best of the dynamic range of the film

Our eyes are logarithmic, not linear. The sensor is linear. A jpeg image has 8 bits for each color channel. We obviously want these 8 bits to represent as much of the information our eyes are able to see as possible. If our eyes can see the highlights and still see that shadowed building, why shouldn't we want our photos to show higlights and shadow too? But to do that, we really have to depend on the sensor correctly capturing data in the low bits too. Especially if the camera manufacturer claims that the camera should be able to do it.

nicksan wrote in post #8626192
a heavily underexposed photo, of course due to my user error.

Then I moved on to the next photo instead of blaming the camera...

Imagine that?

you always shoot in M, or Av and Tv at times too? Because heavily underexposed may or may not have happened with a better meter... If it's a metering issue which causes a sensor flaw to be seen, it's still a camera issue not a shooter issue even if it's avoidable by increasing the exposure. If it's an issue due to limited DR, you're still SOL because the sensor in the 5D CAN NOT capture the scene this thread is about.

So while I agree there are things you can do to avoid the banding, because this camera does it and others DO NOT, it's still a flaw in the camera canon needs to address. If not in a firmware update, then at least in new models (the 7D is looking like it doesn't band)...

if we use your approach every single OOF frame from a 1Diii is user error as are any soft images from any new lenses. While I'll agree the vast majority could be user errors, without the users bringing the issue to canon the 1Diii never would have been "fixed".

jacobsen1 wrote in post #8626411
you always shoot in M, or Av and Tv at times too? Because heavily underexposed may or may not have happened with a better meter... If it's a metering issue which causes a sensor flaw to be seen, it's still a camera issue not a shooter issue even if it's avoidable by increasing the exposure. If it's an issue due to limited DR, you're still SOL because the sensor in the 5D CAN NOT capture the scene this thread is about.

So while I agree there are things you can do to avoid the banding, because this camera does it and others DO NOT, it's still a flaw in the camera canon needs to address. If not in a firmware update, then at least in new models (the 7D is looking like it doesn't band)...

I shoot M 99% of the time. I'll use Av when I need to get something off quick and sure, the meter is fooled pretty easily. When I'm shooting quickly, I have no time to dial in EC, so I can see where pattern noise when recovering would be annoying. I already mentioned I have seen this in my photos.

This is a combination of user error and the frustration that the 5DMKII does not give you the margin of error that sometimes you might want when recovering photos.

I understand. Still, I don't have a problem with it and I ain't exactly Mr. Get the exposure right.:o

When you start relating this to a metering issue...that's just getting a little too loopy for me.

I can understand trying to preserve highlights and maximize DR, but not "oh damn, the camera made me underexpose 3 stops".

^ well, except for the fact the 5Dii is well know for underexposing by 1/2~2/3 of a stop anyway. No, that's not 3, but it's enough you're constantly pushing shadows around with a sensor that's known to band.....

you don't see a problem with that?

my c1 mode had it dialed in for it, for landscapes [c2] I know I'd need to chimp and adjust depending, for my studio [c3] it was a manual mode. So for over 75% of my shooting I had +EV dialed in.

jacobsen1 wrote in post #8626492
^ well, except for the fact the 5Dii is well know for underexposing by 1/2~2/3 of a stop anyway. No, that's not 3, but it's enough you're constantly pushing shadows around with a sensor that's known to band.....

you don't see a problem with that?

my c1 mode had it dialed in for it, for landscapes [c2] I know I'd need to chimp and adjust depending, for my studio [c3] it was a manual mode. So for over 75% of my shooting I had +EV dialed in.

I've got to agree with you in that the metering in the 5DMKII leaves me a little confused, especially since I really never have these problems with my 1DMKIII.

This leaves me chimping a whole lot more than what I am used to...yes, kinda annoying.

jacobsen1 wrote in post #8626492
^ well, except for the fact the 5Dii is well know for underexposing by 1/2~2/3 of a stop anyway. No, that's not 3, but it's enough you're constantly pushing shadows around with a sensor that's known to band.....

you don't see a problem with that?

my c1 mode had it dialed in for it, for landscapes [c2] I know I'd need to chimp and adjust depending, for my studio [c3] it was a manual mode. So for over 75% of my shooting I had +EV dialed in.

I've found that when I point the meter at a grey card it's accurate, and I'm familiar enough with the metering system to that I have no problem compensating with EC. +/- EC, doesn't really make a difference to me as long as I get the exposure that I want (which for me, generally means clipping some highlights).

That doesn't mean I don't want to be able to push the shadows around without issue...but I'm fine with the performance of the camera for my uses. I guess I'm one of those that bought the camera for high ISO use...so, the pattern noise isn't as much of an issue for me...but it is still an issue to be dealt with at times.

nicksan wrote in post #8626498
I've got to agree with you in that the metering in the 5DMKII leaves me a little confused, especially since I really never have these problems with my 1DMKIII.

yep, same for me coming from my old 5D...

bacchanal wrote in post #8626566
I've found that when I point the meter at a grey card it's accurate, and I'm familiar enough with the metering system to that I have no problem compensating with EC. +/- EC, doesn't really make a difference to me as long as I get the exposure that I want (which for me, generally means clipping some highlights).

that's my point though, it's EVERY SHOT, consistently off... So if I forget to dial it in (it was in my custom mode thank god) EVERY shot needed to be bumped in post. Which led to noise in the shadows to some extent (these were all 3200+ shots).

I used mine 75% of the time at high ISOs. At high ISOs it was nail the exposure or get banding in post for me. I'd err on the side of over exposure (hampseter or where Jake's calling it now) to be safe as I'd rather blow a highlight than deal with banding. But I was used to the banding from my 40D and 5D so it wasn't new. The banding at low ISOs is similar, but it's more of an issue BECAUSE the sensor has limited DR you're forced to push exposures to get those details back. Once you do that you bring in noise at base ISOs which isn't ideal there either.

I saw MUCH LESS banding at low ISOs, just noise in general, but for 100 that was annoying. At high ISOs I saw banding whenever I bumped my shadows, but considering this was at 3200 or 6400 I didn't complain much because it's pretty incredible when you think about it.

Like you said, the camera did/does what I needed for the most part, especially at high ISOs. But the low end noise (in general as well as banding) was a more annoying issue for me. I get especially frustrated with it because I assumed it was my processing.

jacobsen1 wrote in post #8626746
I used mine 75% of the time at high ISOs. At high ISOs it was nail the exposure or get banding in post for me. I'd err on the side of over exposure (hampseter or where Jake's calling it now) to be safe as I'd rather blow a highlight than deal with banding. But I was used to the banding from my 40D and 5D so it wasn't new. The banding at low ISOs is similar, but it's more of an issue BECAUSE the sensor has limited DR you're forced to push exposures to get those details back. Once you do that you bring in noise at base ISOs which isn't ideal there either.

I saw MUCH LESS banding at low ISOs, just noise in general, but for 100 that was annoying. At high ISOs I saw banding whenever I bumped my shadows, but considering this was at 3200 or 6400 I didn't complain much because it's pretty incredible when you think about it.

Like you said, the camera did/does what I needed for the most part, especially at high ISOs. But the low end noise (in general as well as banding) was a more annoying issue for me. I get especially frustrated with it because I assumed it was my processing.

Sucks that all those photos you took in the past year have all these issues.

jacobsen1 wrote in post #8626746
that's my point though, it's EVERY SHOT, consistently off... So if I forget to dial it in (it was in my custom mode thank god) EVERY shot needed to be bumped in post. Which led to noise in the shadows to some extent (these were all 3200+ shots).

If the meter performs consistently and your shots are under exposed, I just don't see that being a meter issue, it's user error. If the 0 point of Nikon's matrix metering minimizes user error for you then that is great, I don't think that necessarily means that Canon's metering is off or broken.

Actually, now that you mention it (and this is getting way off topic)...one of the things that I really like about the D700/D3 system is that it can be set up to be much more "point and click" than the 5DII (or any canon). Usable auto ISO and spot metering tied to the active AF point can really free up mental resources for things like composition and timing...granted the exposure isn't going to be perfect every time, but it will probably be close enough.

bacchanal wrote in post #8626897
If the meter performs consistently and your shots are under exposed, I just don't see that being a meter issue, it's user error. If the 0 point of Nikon's matrix metering minimizes user error for you then that is great, I don't think that necessarily means that Canon's metering is off or broken.

How is someone not using EC on each and every shot considered user error? Shouldn't the camera be calibrated to get a proper exposure on most ordinary scenes, with EC only needed for difficult lighting? At least, isn't that the theory?

I love that you posted this topic. What is the trend of low dynamic range media though? This topic reminds me of the loudness war.