semorg wrote:
It's funny you mentioned that. Because that's what I tried doing. Adding grain seem to fix the problem, but it's so lame to have to do that. Another thing I found out with adding grain, is your intended print size. I found out, I need to use different size grains for different print sizes. But in general, I've never been 100% happy creating B&W images out of digital images, however the D3X seem to look like a great camera. Just waiting to see if they would put that electronics into the D700 replacement and if so, I'll probably move to Nikon. ...Show more →
I hear ya about B&W from digital and not being happy with the results. It's very challenging. I have to say that so far I'm most happy with the Nik Color Efex Pro conversions - that's why I may give Silver Efex a try. I'm bummed you're not having more success with it tho. As for changing systems, I'm also open to that come spring when i need to decide how to upgrade my whale/critter DSLR. If you're thinking of dropping a large sum of cash on a Nikon D3X, it might be worth your while to try a 1Ds3 first since that sensor is supposedly of a higher quality and produces fewer artifacts.
Too bad this thread is AGAIN turning into a "the problem doesn't exist" vs "the problem does exist" debate... This is just sad
Can we just stick to finding ways to work around the "problem" - for those who are bothered by it anyway, PLEASE?
Pardon me, but I do beleive i'm one of the few who have actually posted some examples. My comments were not to cheapen the purity of this banding problem, merely a cul-de-sac rambling. I know it exists, i've also made numerous comments wishing the low ISO hadn't been sacarified at the expense of high ISO. Untill we have a real solution however (which is certainly not going to come from anyone on this forum) ramblings are all we have.
Pardon me, but I do beleive i'm one of the few who have actually posted some examples. My comments were not to cheapen the purity of this banding problem, merely a cul-de-sac rambling. I know it exists, i've also made numerous comments wishing the low ISO hadn't been sacarified at the expense of high ISO. Untill we have a real solution however (which is certainly not going to come from anyone on this forum) ramblings are all we have.
I think I was perhaps sloppy with my quote in my post - I was responding to the lines near the end of the quote and not to what you wrote.
I apologize for the confusion, and I have edited my original post to try to make this clearer.
Well it looks like Im going stick with what I have. I've been eyeing up this camera for so long and I can officially say that the examples shown here are reason enough to back away. This really blows.
I'm sure some people in this forum will find problems with what you have...
Chefdaniel wrote:
Well it looks like Im going stick with what I have. I've been eyeing up this camera for so long and I can officially say that the examples shown here are reason enough to back away. This really blows.
I've been interested in this thread. It would be nice if people could seek information, honestly, not frivolously, without having their inquiries given short shrift by people whose experience happens to be different.
My (second) 5D II went into Canon three times, starting within a few weeks after I bought it. After the reliability and clarity of the 5D, the 5D II was rather a shock. I thought there might be some kind of electrical malfunction. Focus was highly erratic (I use center spot), sharpness hit or miss, there was what I think of as scaling or luminance haze in the darks, and not only the darks, and I wasn't able to set exposure comp with the Quick Control Dial.
The last time I got it back, the report said, "We have examined the product (THE CAMERA!) according to your request, and upon close inspection the exact cause could not be identified but it was found that the dial could not operate properly from time to time. The internal component (WHAT INTERNAL COMPONENT?) was replaced. Other electrical adjustments, inspection and cleaning and parts replacements (WHAT . . . ?)were carried out."
I called and was told that 'probably the whole front mechanism was replaced'. What did that mean? The guy I spoke to at the repair center couldn't be more specific.
The camera is better overall (though setting exposure comp more than one way still isn't fixed), but I think it's still far off the 5D's standard in producing images that hold up at high magnification.
gdanmitchell wrote:
Haven't been to this thread for a few days but, my God!, it looks like a few others are actually talking sensibly about this "problem." Thanks! :-)
Dan
Not to pick on you in particular, Dan, but I do not "get" this attitude at all. You put "problem" in quotation marks as if we're making up this stuff. Sorry, it's not a "problem," it IS a problem.
I photographed a jazz group in a dimly lit club (biased heavily to red, even worse) a couple of weeks back using the 5D2. I bought this camera specifically for its low-light capabilities, to be able to produce technically good photographs in conditions where my 40D, and then 5D and 1D MkIII could not do the job. I was shocked when I opened the files in Lightroom 2.5. The most minute tweak to white balance or to open up the shadows resulted in awful pattern banding, obvious when viewing the full image within Lightroom on a 24" monitor, not at 100% or higher magnification. I had never seen anything like this. Quite frankly, iso 1600 images from the 40D have looked much better than these 5D2 images at iso 3200.
I wound up converting the whole shoot using Breezebrowser Pro, which is based on some Canon software (for raw conversions, anyway.) This resulted in far less banding - which tells me that some of the problem is in the converter. Since then, I've been consciously increasing my in-camera exposure to minimize the need to lift shadows if at all possible in pp. This requires a change in my exposures from what has always worked before, and takes away some of the purported high iso advantage of the 5D2. The 5D2 is extremely unforgiving of any underexposure. In the venues where I work, there are going to be shadow areas falling off into the dark. Having burlap-sack type pattern noise show up all the time is not what I bargained for when I shelled out $2500 for this camera body. I expected a camera that offered another full stop of useful range above iso 1600. I didn't expect that extra stop to come with a form of noise that we haven't seen to anywhere near this extent from Canon bodies before this.
It's not just a "gear-head" obsession to talk about this problem, especially not in a forum like this one. It's harder to produce an emotional photograph when your intent does not include a bunch of noise. You want a cross-hatch pattern on your pictures? Add it yourself in post. Don't expect all of us to like it in our own photographs, even when we don't want it!
Something else: the boundaries of the technically acceptable change over time. You try to submit a batch of photos to a stock agency that are marred by the cross-hatch banding, and see how many of them are accepted. Photographs that got published in major magazines twenty years ago wouldn't make it across an editors desk now, because the technical capabilities of readily available equipment are so much better today. The art world may be different, but that's not the world in which I function.
brainiac wrote:
OK - this sounds great. I'm going to send in both my 5D2's and see if I can pry a fix out of Canon UK.
I think that is the way to go after reading into the 5d2 White Paper:
"14 bits per color A/D conversion in DIGIC 4 Image Processor yields finer gradations,
more accurate colors, and improved highlight and shadow details" (compared to 5D1)
This is not the case as shadow details can get spoiled by noise patterns under ISO 400. I stil think the issue gets too much attention on FM as in real live it is very unlikely to come across the porblem
As for pattern noise at ISO 3200, I can not reproduce this, maybe people could support their post with raw files. at high ISO the 5d2 is much better than any other camera from Canon including the 1ds3 IMO
Jim Levitt said, "The 5D2 is extremely unforgiving of any underexposure". That's pretty much it in a nutshell. Great camera, but, aside from pattern noise, which I don't have but don't doubt is an issue for some, it's finicky about light levels: I think you have less room for error than with the 5D.
My understanding is that the problem is in the electronics that reads out the sensor. If the problem is more apparent at low ISO, then it is in the electronics (such as the ADC) in the signal processing chain after ISO amplification takes place; if the problem is roughly independent of ISO, then it is the electronics (such as the sensor) located before ISO amplification takes place. Reason being that ISO amplification boosts the signal, so if noise is after amplification the S/N is improved by raising ISO at fixed sensor illumination, while if noise is before amplification then raising ISO boosts signal and noise at the same rate and has little effect.
I'm no electronics expert, but an educated guess is that the line noise may be radio-frequency noise in the electronics. The pixels are read out a column at a time, at a rate of several tens of MHz (since there are 21 mega-pixels read out in a fraction of a second). Any RF noise would oscillate up and down, whichever direction the noise is on a particular half-cycle, the pixels in that column will all be shifted that way.
Why it appears in some shooting conditions and not others for a given boost of the exposure in a converter is very likely related to the color temperature. Reds and blues are boosted by large factors depending on the color temp of the lighting, so are virtually always "underexposed" relative to the optimal sensor operating point, which is typically (but not always) set by the green channel which is closest to the luminance data. You can see how much the R and B are boosted by white balance if you look in the exif data using exiftool or a similar metadata reader. For instance I have a 7D file to hand for which the relevant line in the metadata is (cloudy WB)
Measured RGGB : 518 1024 1024 743
and another for which it is (tungsten WB)
Measured RGGB : 1038 1024 1024 314
In the cloudy case, the R channel has half the exposure (518) as green (1024); and blue is about a half-stop underexposed (743 vs 1024). In the tungsten case, red and green are approximately balanced, but blue is about 5/3 stop underexposed. In each case, to get white balance the R and B channels must be boosted in the converter software by a factor 1024/518, 1024/743, or 1024/314, respectively. So you already start off with a stop or more of "lifting shadows" in R and/or B even before you touch the exposure slider.
HTP also intentionally underexposes the image by using 1 stop lower ISO internally in the camera, and the converter internally has to boost the shadows by another stop.
So I would recommend anyone interested in seeing the boundaries of how to expose without generating issues with their 5D2, to look at a sampling of shots to see how much lifting of shadows is needed to show problems for their camera at various ISO, and then look in the exif data to see what the white balance multipliers are. That will give a rough guide to how much amplification of the color channels is taking place. It should be that the problem appears consistently when the combination of exposure compensation and WB multipliers exceeds some threshold. And by examining how light color temp relates to WB amplification of R and B, you will know in a particular shooting situation how much leeway you have in exposure -- how much of the total compensation before pattern noise becomes a problem remains, because some has already been used up by the WB amplification that will be necessary in converting the file.
Finally, conventional NR won't deal with this issue well, nor will downsampling help. NR filters are designed to preserve edges and texture as much as possible, and pattern noise is texture. Downsampling has little effect until the factor of the downsample exceeds the periodicity of the pattern, and most people won't want to be turning their 21MP camera into a 1MP camera.
ejmartin wrote:
My understanding is that the problem is in the electronics that reads out the sensor. If the problem is more apparent at low ISO, then it is in the electronics (such as the ADC) in the signal processing chain after ISO amplification takes place; if the problem is roughly independent of ISO, then it is the electronics (such as the sensor) located before ISO amplification takes place. Reason being that ISO amplification boosts the signal, so if noise is after amplification the S/N is improved by raising ISO at fixed sensor illumination, while if noise is before amplification then raising ISO boosts signal and noise at the same rate and has little effect.
I'm no electronics expert, but an educated guess is that the line noise may be radio-frequency noise in the electronics. The pixels are read out a column at a time, at a rate of several tens of MHz (since there are 21 mega-pixels read out in a fraction of a second). Any RF noise would oscillate up and down, whichever direction the noise is on a particular half-cycle, the pixels in that column will all be shifted that way.
Why it appears in some shooting conditions and not others for a given boost of the exposure in a converter is very likely related to the color temperature. Reds and blues are boosted by large factors depending on the color temp of the lighting, so are virtually always "underexposed" relative to the optimal sensor operating point, which is typically (but not always) set by the green channel which is closest to the luminance data. You can see how much the R and B are boosted by white balance if you look in the exif data using exiftool or a similar metadata reader. For instance I have a 7D file to hand for which the relevant line in the metadata is (cloudy WB)
Measured RGGB : 518 1024 1024 743
and another for which it is (tungsten WB)
Measured RGGB : 1038 1024 1024 314
In the cloudy case, the R channel has half the exposure (518) as green (1024); and blue is about a half-stop underexposed (743 vs 1024). In the tungsten case, red and green are approximately balanced, but blue is about 5/3 stop underexposed. In each case, to get white balance the R and B channels must be boosted in the converter software by a factor 1024/518, 1024/743, or 1024/314, respectively. So you already start off with a stop or more of "lifting shadows" in R and/or B even before you touch the exposure slider.
HTP also intentionally underexposes the image by using 1 stop lower ISO internally in the camera, and the converter internally has to boost the shadows by another stop.
So I would recommend anyone interested in seeing the boundaries of how to expose without generating issues with their 5D2, to look at a sampling of shots to see how much lifting of shadows is needed to show problems for their camera at various ISO, and then look in the exif data to see what the white balance multipliers are. That will give a rough guide to how much amplification of the color channels is taking place. It should be that the problem appears consistently when the combination of exposure compensation and WB multipliers exceeds some threshold. And by examining how light color temp relates to WB amplification of R and B, you will know in a particular shooting situation how much leeway you have in exposure -- how much of the total compensation before pattern noise becomes a problem remains, because some has already been used up by the WB amplification that will be necessary in converting the file. ...Show more →
That all makes sense to me, in that it conforms with my experience of the camera's performance over many tens of thousands of images in varied conditions.
It should be that the problem appears consistently when the combination of exposure compensation and WB multipliers exceeds some threshold. And by examining how light color temp relates to WB amplification of R and B, you will know in a particular shooting situation how much leeway you have in exposure -- how much of the total compensation before pattern noise becomes a problem remains, because some has already been used up by the WB amplification that will be necessary in converting the file.
This makes a lot of sense, and it should be fairly easy to test as well.
Above 1ds2, 1ds3 & 5d2 measurements (RGGB ratio) are very close through out iso 100-3200 range under the controlled test light. 5D2 RED was always measured less than half of 1Ds2 & 1Ds3 RED compared with their GREEN channels.
This is the most interesting RGGB logic - above are some black frame RGGB numbers and 5D2's Red & Blue are in reverse compared w/ 1ds3's while all 1ds2's are zero.
Look like there is some shifting going on here...with 5D2 & 1Ds3...
3) This is the case of RGGB data in banding/pattern noise (1 stops push). I still can't correlate the banding noise issue with below RGGB number - in this case, WB blue coeff = 4.2 and WB red coeff = 1.3 - thought
All three cases tell me - 5D2 red channel is different from 1ds2 and 1ds3 design - Sure is, 5D2 is seeing less ( or more if reverse) "RED" than any other canon bodies. It is good to understand how the issues is caused by 5D2's RGB over amplification but it is not easy to guess/avoid the problem based on what we try to do here. So, correct exposed theory won't help. It's all depending on the colours spectrum that 5D2 seeing.
...A quick check this morning, so far 5D2 measured RED is not the same as 1ds2, 1ds3, 40D, 1D2, 1D3, 5D based on my old files...Thought
