There has been discussion here recently about early 1Ds bodies needing to have their sensors replaced, due to shadow banding and colour mottling issues in underexposed areas of an image. I searched, but cannot find this thread. Jeff (our moderator) responded on that thread, so he may chime in here, too. I got rid of my 1Ds on account of this problem.
Does anyone know from what serial number Canon fitted the replacement sensors to the roll-out of the 1Ds product run, or if in fact this happened?
Or was it that Canon replaced faulty sensors only, which may have occurred any time in the production run?
Finally, short of testing an actual body, how can one know that a camera that is described as having had a sensor replaced really has had the sensor replaced?
Don't want to hijack an important question, just want to add that I'd bought a new 5D in Feb. and feel that color noise in the shadow areas, even at the lowest ISOs is worse than previous D60 and 20D. I'll be interested in knowing if anyone has seen similar problems in the 1Ds or any other Canon digital SLR?
Believe I'll send the the 5D back to Canon soon to see if they agree.
Appreciate anyone out there with information on out of calibration sensors?
So, to recap from the last thread, my 1Ds (purchased new 3/03) exhibited banding in the 1/4-tones in properly exposed images; that is, the 'macro-scale' banding showed up in darker tones, like navy suits. For example (and this is what I sent to Canon):
The sensor was replaced by Canon, under warranty, at about 11 months and 2 weeks of age. It did take a bit of going around with them to finally show them that it wasn't an underexposed image that I was over-tweaking the gamma, but I did convince them. Secondarily, if you do have an underexposed image, or if you severely open up the shadows, from my experience any 1Ds will exhibit similar behavior, and it's mostly (or all?) in the red channel. This does not mean that there is anything wrong with the camera, just that there is something wrong with the exposure, and you shouldn't be pushing it that far. It's when this behavior manifests itself in the darker tones of a properly-exposed image that you have a problem.
In summary, I'm nearly certain that Canon never owned up to any large-scale problem with the 1Ds production line (nor do I think they should have, from what little I've seen of the 'issue'), and am equally 'certain' that there is no way to identify any serial number run or timeline to indicate that a given camera should/should not have the problem. I believe that it was more of a random occurrence (production-wise), and that other than testing a given body, there's no other way to predict whether that body will or will not exhibit the behavior, (other than the fact that it is statistically unlikely). What Canon knows of this issue is anybody's guess. RD Grier's 5D chroma noise issue (above) is more endemic to the Canon CMOS sensor, is easily rectified in post, and is most certainly not what is illustrated in the example image above, even though it is admittedly 'color noise in the shadows'. I've only seen this 'macro-scale banding' from 1Ds MkI models, though other cameras may exhibit such garbage data when the image is pushed too far.
As far as knowing if the sensor was replaced, I have no idea whether even Canon keeps such records long-term, and absent the paperwork showing that the sensor assembly was indeed replaced, I doubt there's any way to know by examining the camera. Even so, does it really matter? If the sensor doesn't exhibit the phenomenon, I don't think I'd care one way or the other if the sensor was 'original' or not. As long as it worked properly.
For overseas purchasers of these bodies, the only way to be sure is to ask the owner to shoot something that has significant shadow areas, and ask to look at the files. hth, kl
Hi,
I know that this information was detailed in the old Rob Galbraith forums, which I used to read prior to becoming a member here. While the old forum is gone some of the information can still be searched in the new forum website. I read this information very carefully when i bought my 1Ds and recall that the body's with serial number like 8300 and lower were the ones with banding problems. I may be totally off since my memory cells have been wiped since having kids. I do know that my old camera which was serial ending 8644 did not have the problem unless the images were seriously underexposed and adjusted in post.
Don't want to hijack an important question, just want to add that I'd bought a new 5D in Feb. and feel that color noise in the shadow areas, even at the lowest ISOs is worse than previous D60 and 20D. I'll be interested in knowing if anyone has seen similar problems in the 1Ds or any other Canon digital SLR?
Believe I'll send the the 5D back to Canon soon to see if they agree.
Appreciate anyone out there with information on out of calibration sensors?
Bob
My 5D shows no banding in the shadows or noise. Certainly not compared to my 20D which was worse than my 10D especially if the image was even slightly underexposed. The 5D is the best quality sensor I've ever used. So either you have a bad sensor and quite rightfully Canon should replace it. Or you're slightly underexposing your images. You should expect excellent images and dynamic range from your 5D.
I am lucky. I pushed +2 stops but the noise in the shadow areas is random. My serial number is 113204. (Have a hard time believing they sold this many, though.)
Mine's 113579 and I don't seem to have the banding problem described in the previous page either. I'm guessing it was just on the first batch of 1Ds bodies released and canon fixed the issue either through firmware or sensor replacement.
What Jeff showed is exactly what I saw and had to deal with on my 1Ds, underexpose by a stop and your blacks would start becoming unuseable due to the red/green banding. I rented another one to see if it was just mine but it was exactly the same. When I asked about it people told me 'yup that's the 1Ds'.
For all the AF problems I'm having at the moment with my 5D's, I wouldn't take a 1Ds back if you paid me, it was really hard to deal with.
For a 7K camera it should not have a banding issue if underexposed by only a few stops.
It wasn't only a problem of underexposing but a problem taking images with lots of dark areas, even if properly exposed such as using a black backdrop with studio flash.
When I did send my 1Ds in for a checkup about the problem they more or less fixed the problem by increasing the blur to the image. It was still there but not quite as noticeable, unless you turned off the default NR in the RAW converter.
It was a 7K camera for its time, you should have seen the loops that PJ's had to deal with shooting the very first DSLR's that cost a hell of a lot more than that..
The 1Ds noise performance was never the best (the original, MK I) but it wasn't bad. I did notice some banding in the samples I used from time to time but nothing majorly objectionable and nothing that kept me from using them on multiple occasions until I finally got my 1D MK II.
I purchased a 1Ds in April 2003 and then sent it in for a sensor cleaning just about one year later. Canon replaced the sensor with no explanation. I noitced that the dark areas were better, with no banding. I hope this helps
cyberstudio wrote:
I am lucky. I pushed +2 stops but the noise in the shadow areas is random. My serial number is 113204. (Have a hard time believing they sold this many, though.)
Mine was a 108xxx. I suspect the first digit (1) may be a placeholder.
I think the anti-aliasing filter of this camera is rather weak and I can get it to moire rather easily.
I read dpreview again, and quite interestingly this camera has worse low-ISO noise with a 8.8 micron pixel pitch than the older D60 with a 7.4 micron pixel pitch. Actually I was quite relieved upon reading this as this implied Canon beefed up the color filter of the sensor for better color separation at the expensive of noise. Of course a modern camera such as the 5D can achieve both excellent color separation and low noise at the same time, but still, the 1Ds has better AF accuracy. There is a reason why the two of them cost approximately the same on the market. The market cannot be mistaken. I like the 1Ds, except after holding it for an hour. I am not sure if I would change my mind after using a 5D, but I probably won't have the chance to find out.
