bavc
9th and 10th bit Error Detection
A few of us (@kieranjol @bturkus) have been discussing signal chain errors that cause the 9th and 10th bit to either blank (padded with zeros), or incorrect (containing false data, some sort of interpolation, but not truly a 10 bit signal).
After speaking with a number of analog video people, it seems that some gear causes this issue by converting poorly between the analog and digital domain. Also, it seems that this error is very difficult to detect with older equipment, and that QCTools is really the only tool in which this error can be easily and clearly detected.
With that in mind, I'd like to request a graph that graphs this error. Scale from 0 to 1. 0 meaning the error is not present at all, 1 meaning that the error is present. If the 9th and 10th bits are all zero, then the value would be one. I would imagine this portion would be fairly simple to implement?
The more difficult part, I imagine, would be detecting that the data in the 9th and 10th bits is interpolated, or otherwise incorrect and not truly a 10 bit signal. In all of the examples I've come across, the bit-plane slices of the 9th and 10th bits seem far less noisy than they would potentially be. This suggests that by comparing the entropy of the 9th and 10th bits to that of the 7th and 8th bits, a determination could be made as to whether the issue is present.
"it seems that some gear causes this issue by converting poorly between the analog and digital domain" ping @kierank
I'm also confused by what you are requesting: the percentage of 0's per bitplane per frame. You should review and comment on this PR, https://github.com/bavc/qctools/pull/198
