wslam Posted November 7, 2006 Share #21 Posted November 7, 2006 Advertisement (gone after registration) The more samples I look at from the M8, the more they resembled the original Canon 1D (CCD, not CMOS). The D200 had a different banding problem which apparently has to do with the slight mismatch of the quad-channel amplifiers of the sensor data. The 1D, similar to the M8, has more random 'streaks' that is apparently also a result from the channels being read out 'too fast'. This is why it seems strange because the 1D, with its fast fps... But the M8 is nowhere near that. Why would they have such a similar issue? The original 1D was never fixed, but the banding was only minimized with firmware upgrade. Link to post Share on other sites More sharing options...
Advertisement Posted November 7, 2006 Posted November 7, 2006 Hi wslam, Take a look here Weirdest thing about banding. I'm sure you'll find what you were looking for!
rvaubel Posted November 7, 2006 Share #22 Posted November 7, 2006 I wish you did as well, I think you'd be able to provide a useful extra pair of eyes on the analysis... As regards exposure, I measured the light with a spot meter and correctly exposed, it's fine. Even +1 EV seems to show some banding. Mark I just had an idea about the RAW image of the man in the cafe. If you are familar with that picture, it had a shot of the same subject and lighting @ ISO 160, 320, 640, 1250, and 2500. The shot at 1250 showed all the artifacts that I have got use to predicting i.e. banding, dark mirror image, etc. I had never noticed all the features before. But they are all there. However, the same shot at 2500 shows NO banding or other artifacts. Very strange, is it not? I actually have an explanation that fits into my model. That is the +4 F stop model that I mentioned above. Believe me it fits the model but its complicated to explain. I'm pretty jazzed about this because every picture I look at fits perfectly. I hate to be a party pooper when I get so close but its almost 2:00 AM Nite Nite Rex ....ARF! Link to post Share on other sites More sharing options...
scott kirkpatrick Posted November 7, 2006 Share #23 Posted November 7, 2006 The shot at 1250 showed all the artifacts that I have got use to predicting i.e. banding, dark mirror image, etc. I had never noticed all the features before. But they are all there. However, the same shot at 2500 shows NO banding or other artifacts. Very strange, is it not? I actually have an explanation ... I pressed send on the same observation (about the 1250 image) a few seconds before you did. It is already three or four posts back. Don't leave us in suspense about your theory. scott Link to post Share on other sites More sharing options...
rvaubel Posted November 7, 2006 Share #24 Posted November 7, 2006 I pressed send on the same observation (about the 1250 image) a few seconds before you did. It is already three or four posts back. Don't leave us in suspense about your theory. scott Oh Christ, are you guys going to keep me up! Well I'll try to explain what I've found but bear with me for a minute. You probably have to have the RAW plug for CS2 to duplicate what I found. And you definately need the RAW file of the man in the cafe ISO 1250 and 2500. I'm going to the bathroom. If anyone is awake when I get back I will continue. Rex ..arf (tired doggy) Link to post Share on other sites More sharing options...
wattsy Posted November 7, 2006 Share #25 Posted November 7, 2006 Capture One also has a "pattern noise removal" action which you can check. It appears to be used for files coming from chips in the KAF family, such as PhaseOne backs, Some Olympus products, and the Leica DMR. I am familiar with "pattern noise" from my E-1 in its early days. Nobody is reporting this artifact for the M8. Is 'pattern noise' the same thing as moire? If so, the M8 is probably as susceptible as any 10MP camera without an AA filter. Frankly, I'd be surprised if the M8 didn't show moire rainbow effects on occasions. See the 100% crop from a shot below (which also, incidentally, demonstrates that the default Capture One sharpening settings are rather generous for the already sharp combination of 35/F1.4 ASPH and M8): Link to post Share on other sites More sharing options...
scott kirkpatrick Posted November 7, 2006 Share #26 Posted November 7, 2006 Oh Christ, are you guys going to keep me up! If anyone is awake when I get back I will continue.) Awake? You're keeping me from lunch! regards, scott Link to post Share on other sites More sharing options...
scott kirkpatrick Posted November 7, 2006 Share #27 Posted November 7, 2006 Advertisement (gone after registration) Is 'pattern noise' the same thing as moire? If so, the M8 is probably as susceptible as any 10MP camera without an AA filter. No, they are different. For Moire, COne provides a PS plug-in. Sean Reid showed an example of its use. You can treat the whole image, or make a layer and brush the correction across only where it is needed. "Pattern noise" typically in the strong reds or yellows, is a maze-like pattern of short strips of white in an area of strong saturated color. I've never learned where it comes from, because those who create it would rather not tell us. scott Link to post Share on other sites More sharing options...
rvaubel Posted November 7, 2006 Share #28 Posted November 7, 2006 Awake? You're keeping me from lunch! regards, scott Oh crip. I hoped you were gone. I can't go into the whole thing in detail but briefly (using CS2 RAW plug in) open the 1250 image *set the exposure slider to 0 *check the blown highlight box. The windows will go red *move the exposure slider down all the way to -4 Notice that some of the blown highlights are recovered. Also notice that the highlights that aren't recovered correspond to the banding. Cool HUH!! OK now go to the same guy in the cafe @ ISO 2500 * go thru the same procedure * but notice this time all the highlights are recovered. Also notice that there is NO BANDING Cool HUH!! Thats about it. I think it means that the blown areas must be blown by more than a certain amount or no banding happens. Also it explains why the bands are always narrower than the apparent blown highlight. Also look for the mirrored dark image of the window. Its there. Also notice that the banding stops at the centerline of the picture. Finally notice that the windows on the left never band because they are not blown enough i.e. you can always recover all the highlights. Very Cool This same analysis works on all the RAW files I have, which are very few. I want my M8!! Rex going beddy-by Bezerkeley by the Bay Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.