High-speed continuous data streaming to PC (analog video)

[NiHaoMike]

I'm not sure if it's possible to use the MSI.SDR as plain ADCs, many SDRs do a final stage of frequency domain shifting and filtering in DSP.

Save the raw "I/Q" data to disk, then you have basically unlimited tries to tweak the extraction algorithms to get the best results.

[woodchips]

Just looking at archiving VHS tapes and saw this thread.

My experience is decades old, but moved colour PAL video to a computer by simply sampling the analogue signal and storing it to disc. This was the days of 6 bit flash converters and IDE discs. Worked. Was from a camera rather than a video player, perhaps that is the difference.

Might be useful to use an 8 bit ADC but why not just do that? I am unsure as to what the fancy splitters to extract the chrominance info do for you?

Once on disc, can then process to your hearts content.

[Novgorod]

I'm not sure if it's possible to use the MSI.SDR as plain ADCs, many SDRs do a final stage of frequency domain shifting and filtering in DSP.

Save the raw "I/Q" data to disk, then you have basically unlimited tries to tweak the extraction algorithms to get the best results.

I guess we'll find out in 2-6 weeks - I ordered one from China to play around with .. I hope all the processing is bypassable to get access to the raw ADC data. And of course I would record the raw ADC stream to disk to develop the processing tools first, but I'm still skeptical that the final result would be any better than using a dedicated chip which has a much higher internal sampling rate and all the "fine-tuned" processing.

[Novgorod]

My experience is decades old, but moved colour PAL video to a computer by simply sampling the analogue signal and storing it to disc. This was the days of 6 bit flash converters and IDE discs. Worked. Was from a camera rather than a video player, perhaps that is the difference.

Might be useful to use an 8 bit ADC but why not just do that? I am unsure as to what the fancy splitters to extract the chrominance info do for you?

Once on disc, can then process to your hearts content.

Even with 6bit instead of 8 it's tens of MBytes of data per second to store a full-bandwidth composite signal. What kind of ADC and interface did you use?

I don't know if I got your question right, but if it's about using S-video vs. the composite signal, it would (theoretically) improve the quality a bit when capturing from tape. Luma and chroma are demodulated from the tape signal and then processed separately by the tape player before being combined again into a composite signal. If it has S-video, you can get the separated components before they are combined into composite, so you avoid one filter step and crosstalk degradation. Of course VHS is already a kind of composite signal, so if you capture it directly from the tape head (as suggested by the other guy), you bypass the whole demodulation process inside the player and can do it all digitally. But if you only have access to the video outputs on the player, then S-video should be the better choice.

[Novgorod]

Just a final note on this (for now): After a bit of digging, I found the "Domesday project", which is very close to what I'm looking for. They designed a 10bit 40Ms/s digitizer for continuous streaming (including a FPGA and USB3.0 board) with the main purpose to digitize the RF signal directly from a laser disc head and do the entire RF to Y/C to RGB decoding in software. The software project has a fork for decoding VHS RF (called VHS-decode) which produces very decent results already, including color, and is in active development, so I don't have to do it from scratch. It seems to be the superior approach compared with digitizing the composite or S-video signal, so I will give it a shot sometime when I'm free enough. There is no Hi8 support yet though (which would be more important than VHS), but that's a matter of software adaptation once the RF is recorded from the tape.