Re: Banding caused by slow SD card

[Mac]

if you leave your SD card in the camera whilst downloading data to the computer then this would cause horizontal banding (I have the explanation filed away somewhere)

I would be interested if you could dig out that info, I cant believe that copying over a file from a camera would introduce banding, That would mean that the file is being changed (corrupted) whilst it is being transferred, if this were the case then the error checking should capture this when copying over the file as the check sum would be different.

This would also mean that any file transferred to a camera potentially has the risk of corruption, if this was the case then there should be a huge amount of firmware upgrade failures as the transferred firmware file would be corrupt when copied to the camera.

Mac.

[Mac]

Actually thinking about it it would be very easy to prove if this is true or the original finder of this problem has had squiffed results.

If you take say 1000 images and then copy them from the camera to the computer the probability of banding introduced banding MUST appear in at least one of these images by the fact that someone has said that this is the case (not very logical i know but not knowing the probability 1/1000 seems a reasonable starting place), also copying over the images from the camera should not alter any file data of the original file.

If you then remove the memory card and copy over the same images via a card reader you remove and chance for this banding to be introduced.
Then use a piece of software to do a byte by byte file comparison until you find a file that has changed.

Another way is to copy over 1 file a thousand time and do the same checks

#its sort of a hijack but in a friendly way#

Mac.

[Carole]

I'll see if I can find the thread/info.

Carole

[Carole]

Having a struggle trying to find the thread, but reading later comments it seems it is something to do with old SD cards:

the slower/older/failing cards can cause "bumpy" transfers to your pc /laptop

Still trying to find the original thread, it was a few years ago so might have been deleted.

Carole

[Carole]

Right - found the thread, there is no technical explanation but comments on people who have been advised to remove the card and actually tried it, these are various comments on that thread:

Interesting..Historically I have had a lot of issues with my Canon 30D where with long exposures I get horizontal lines. These are beyond just gradient and appear no matter what I was imaging and irrespective of moon etc.

So i whipped out the memory card..no more issues.

Can only assume there is stuff happening in between the physical process of writing an image to the card and pushing it through to the PC. Somewhere, electrical charge is being introduced into the image file

The banding happens because the camera is trying to capture to the card and the outside device at the same time and the onboard processor just isn't up to it.

I had the problem 4yrs ago whilst using the canon software so my images uploaded direct to the pc.
Then a photographer friend told me that I have to remove the card.
Strange thing is it doesn't happen to all canon D20's so it must be to do with the processor type?

I started on the horsehead after removing the card and used Flats, Darks and Bias frames then stretched the hell out of it and no banding evident.

[Rick]

The sensor is a sensitive dynamic device. Until each pixel is read (and digitised) its value can change. The primary way pixel values chage is when photons hit them. However, other electronic disturbance can also affect them. If the circuitry is insufficiently de-coupled/isolated then the signals being generated during the process of writing the data (to memory, SD card, USB, or wherever) might affect the values of the pixels yet to be read, and give rise to odd effects...

[Rick]

The banding happens because the camera is trying to capture to the card and the outside device at the same time and the onboard processor just isn't up to it.

Right... So the added load slows something down, and that affects the values being read out of the sensor.

[Rick]

But thats  different issue compared to noise being introduced whilst copying from sd card to pc.
I agree that until the data is pulled off of the chip and saved to the card then there is a possibility of corruption, but that corruption is not the fault of being saved to the sd card
that is the due process of reading and storing the data, which is a different cause. I found the original post on astroshed which dosent go in to any detail except one persons view that it might have been the slow cards and that was just a hunch.

The banding has to be happening at sensor read time. A slow SD card might well cause the sensor read-out to be done slowly or less evenly as it is read from the sensor and written to the card. Put a faster SD card in, or cut the SD card out of the loop altogether, and the sensor read out won't suffer those extra delays.

[RobertM]

I can't really see how that can happen Rick as the sensor is read to an internal buffer before being flushed to SD Card.

[Rick]

Point is the banding must be introduced during the sensor read stage, because that's the only stage where some disruption will only cause relatively subtle changes, as it's the only stage where there's analogue-to-digital activity. Any corruption happening in purely digital stages would cause far more dramatic and random damage to the image.

If the SD card is directly involved in the read loop then timing effects are an obvious possible cause.

If the SD card is not touched at all until after all the sensor reading is done then there must be something else going on as a result of having the SD card there that is affecting the analogue-to-digital conversion. Maybe there's some electronics associated with the SD card that is putting a ripple on a signal line somewhere, or causing a thermal ripple.

My guess is that, though there is some internal buffering, they're actually starting to write data to the SD card before they've finished reading it from the sensor (because that's the sort of over-lapping you do if you want to try to get things to seem faster than they actually are).

[MarkS]

Memory cards may or may not cause interference.  But the main cause of banding is in the sensor readout circuits. It's low level banding at a level less than the least significant bit of the AD converter.

Mark

[RobertM]

Do you get the same banding with the battery as the step down PSU - the regulator circuit uses a 1MHz switching frequency (it's the same regulator that's in my Starlight-Xpress camera).  Ripple should be less than 30mV but even so there could be an outside chance of it affecting the camera.

Robert