Jump to content

M9 Banding and noise issues are gone!


Paul J

Recommended Posts

Advertisement (gone after registration)

This just gets funnier. Come on people - take a reality check!! or keep entertaining me this morning.

 

Not trying to be adversarial over this - but it't just utter poppycock.

 

I don't know where to begin! Philipp - In this context, film & tape are analogue media, I'm sure you remember ;-) It's complete nonsense to compare this with digital recording.

 

Take this to its logical conclusion. We will start seeing threads on how the colour is better on Lexars, Sandisk has better grain than Transcend and Calumet own brand has fantastic dynamic range for the price.

 

The card is not part of the optical process! It's just there recording binary digits after whatever A-D conversion and processing has already taken place.

 

You just cannot attribute aesthetic qualities to a digital data stream on a card, that's not even picture until it's processed again. It just works, or it doesn't. It's a file format. End.

 

Am I being had here ;-) Is there some forum conspiracy going on that I'm not part of !

Link to post
Share on other sites

  • Replies 113
  • Created
  • Last Reply

The card is not part of the optical process! It's just there recording binary digits after whatever A-D conversion and processing has already taken place.

 

This thread is about striped pictures. The stripes are not part or result of the optical process. The problem varies with the brand and type of card used. The card is not a mere "medium" but an active component within the entire chain. Its proper working is subject to a power supply kept within certain parameters, among other things. Being an electronic component which draws a non-trivial amount of current, it influences the systems's power supply in turn. It is quite conceivable that this affects other functions of the camera which are executed during the time the card is being written to. Leica have throttled the data rate with which the card is written to in order to reduce some problems in handling the cards. This has led to fewer problems with SDHC cards on one hand and to complaints about slower transfer of data to the cards on the other hand.

 

It's not magic. It's not imagined. It's, perhaps, not even unconnected with the fact that the new Leica M has a battery with double the capacity.

Link to post
Share on other sites

Now when you put it that way, it makes far more sense ;-)

 

The whole SD card debacle has been a real eye opener, (and I've experienced issues myself). Who'd have thought that innocuous little piece of plastic could cause so many random issues and thus headaches for Leica (and ourselves).

 

I wonder what the tolerances are between different manufacturer's cards that could have a significant influence on the whole A-D process within the camera, causing failed writes, and the introduction of noise in to the signal path.

 

Is the whole operation of the M9 running on such a knife edge with that little battery? Mine is stable now, since I switched to Lexar, seemingly regardless of the latest firmware. I had problems with Sandisk before.

 

I know I experienced write fails with fully charged batteries. It's just hard to conceive that one card draws that much more power than another, even with a full battery, that it affects the A-D conversion of what's falling on the sensor, as it's happening. I'm still not totally convinced. There's still a whiff of snake oil and gremlins about.,

 

Anyone know the sequence of events? Does the CCD output its sequential data, which is then converted and written to the card after the conversion. Or is it a real time process, where the writing happens as the image is being converted? Sounds like the latter if the card is to have any effect on the process.

Link to post
Share on other sites

Advertisement (gone after registration)

Damn!

 

If only we had invented an optical sensor device that collected light and then output 1's and 0's instead of an electric signal! All these problems with noisy High-ISO, blowing highlights, lost shadows, Banding, Red Lines, and everything else connected with converting an optical signal to an analog electric signal and THEN QUANTIZING it would be gone!

 

What were we thinking!

 

We can store the raw digital on an IBM1360 storage unit.

 

http://en.wikipedia.org/wiki/IBM_1360

 

To the OP- Thankyou for posting this thread and reminding everyone that there is no such thing as a "Digital Device". All devices that work directly with electric and optical signals are analog devices that store a "quantized" signal. Stuff happens. Congratulations on solving the problem. Now others will know what to look for, what to make of it, and get a good start on addressing it.

Link to post
Share on other sites

I'm sure those old IBM drives didn't affect any read/write operations merely by their presence ;-)

 

So the problems aren't card related directly.... (as in defective or low quality cards).

 

It's simply a power issue affecting performance of the A-D conversion, and possibly write reliability.......... on a random, unpredictable basis.

Link to post
Share on other sites

But, what does this have to do with why different cards would show more banding? If, you know the answer you don't have to make it a conundrum.

 

Yeah, it's quite baffling.

 

It could maybe that some cards have poor shielding, and it's interfering with the electronics, but it's quite a stretch of the imagination.

 

I've not had banding problems myself, so I'm not sure sure what the difference looks like.

 

Would be great if someone could post some goodcard / badcard shots for comparison.

 

- Steve

Link to post
Share on other sites

I'm not trying to make this thread go on but IMHO those two images of lens caps does not show me that the fault is in the card.

A bad SD card shot and identical tethered shot, normal scene, light and deep shadows from a tripod. Thanks!

 

Please, try it out yourself and let us know what you find. Thanks.

Link to post
Share on other sites

Maybe someone can tell me whether any fine debris on gold contact points affect images?

(even minimal) ??

 

It's in the realm of anything is possible. Old-time computer users remember "Reseat the Card" fixes. Dirty contacts can change the electrical levels. At some point, that can cause problems.

 

These are complex "computers" with "Optical sensing devices" built in as peripherals.

Link to post
Share on other sites

The OP has successfully corrected the problem that occurs in his camera by changing brand of SD card used. If others get such problems occurring, this will be a good place to start looking. The problem is solved, what else needs to be done?

 

Try sticking a Wi-Fi card in the M8u- see if that produces a banding problem. If it does, remove the card and see if it stops.

Link to post
Share on other sites

However, there are Wi-Fi cards which are said to damage the socket in the camera.

 

 

AFAIK the warning was issued for the M9.

I have not seen it for the M8 though, but I didn't search for it excessively.

 

So, if you have to use that memory card, it is at your own risk.

Link to post
Share on other sites

Paul,

 

I finally had the chance and bought two of the Transcend 323GB Class 10 cards today and just ran some tests that I felt could induce banding. I tried ISO 160 and 640, single frame captures on S advance without review and let the file write to card before the next exposure, but also on C advance, shooting an entire 7 frame DNG uncompressed sequence and immediately reviewing an image, zooming in on it and panning around it while the buffer was writing other images to the card. I should note that I always have the camera set to not automatically display the image after capture.

 

I repeated the test images with one card from my normal stock of Transcend 16GB Class 10 cards and also one Transcend 16GB UHS1 card.

 

Most of the test frames were 1/4000 second black frames, boosted +5 in Lightroom, but I also did some normal captures around the house, mostly at about a stop under so I could boost them a bit in LR, which I felt was reasonable because there have been many times I've deliberately underexposed M9 files to preserve highlights with the aim of opening darker tones in post.

 

This is what I discovered with the pushed black frames:

 

I could induce banding on the 32GB cards. It was more obvious at ISO 160 because at 640 the noise was much stronger and disguised some of the finer banding. Banding was typical in images made during a continuous sequence until the buffer was full and images were reviewed while the buffer wrote to the card. It was not typical for S advance mode images where the camera was left to write the image to the card before the next capture, but did appear in S advance images where each was made about one second after another. Here the banding was stronger in the 3rd and subsequent images and absent in the first two.

 

The UHS1 card, with which I had immediate visible banding from the first set of normal images I shot with it a couple months ago, and immediately retired from the M9 rotation, exhibited much, much worse banding than the regular 16 and 32GB Class 10 cards in the C advance black frame tests. It was OK in the S advance single frames where I waited for the buffer to clear before the next exposure.

 

My understanding is many UHS1 cards don't play well with the M9, so I wasn't too surprised about this. I had hoped to use these in order to reduce card to computer transfer times, but I guess this will have to wait until the new M.

 

While the banding was fairly noticeable in the 32GB card black frames pushed +5 (and pretty much identical to the 16GB Class 10 card), one has to keep in mind how unrealistic this situation is, which is why I also did normal photos around the house, both on S and C advance modes. While I only shot a limited number of normal images on the 32GB cards, none exhibited visible banding at ISO 160, even with about +1.5 push in LR. From a sequence at ISO 640, one out of seven had a single visible line near the top of the frame (which is often where I see banding with my M9).

 

Therefore, I will add the 32GB cards to the rotation to get a better longterm feeling about them during normal use.

 

My feeling is the problem lies somewhere in the circuitry of the M9, between the sensor itself, and the card. As speculated, it's possible some cards generate disturbances that affect the camera's circuitry. Otherwise, why would a card such as the UHS1 sample consistently record images with considerably worse banding than the regular Class 10 cards?

 

IMO the final solution to this problem, for me, will be the new M. I really do like the M9's image quality, when it works, but having been a long time Canon 1D series user (and still am), where I can immediately review images and never induce banding, etc., using the M9 is truly stepping back in time to the circa 2002 digital camera era. Considering that the M9's 'guts' are essentially transplanted from the M8 and carried over by the same Jenoptik team, I don't think I'm that far off in considering it to be a nearly 10 year old digital camera.

 

I'm looking forward to moving up to circa 2010. :)

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...