Home Page Forums MAME MAME Discussing about the dumping of analog media - VHS

I wonder if somebody will ever be able to come up with a way of properly combining the data gathered out of many identical VHS tapes into one better "verified" dump.

I mean, just capturing one of them would show the same artifacts again and again, something that would differ from any other existant tapes of the same kind, making the single dump "incorrect" by default.

Of course people could capture an MPG out of it and "clean it", but that would be like guessing the missing bits out of a bad ROM dump.

chroma and luma aren't separate on VHS or S-VHS, you would probably want to store them separately as if they were though.


Laser discs have the same problem, sometimes you just have to accept you're not going to get broadcast quality master tapes for anything.

Just in case it can shed some light on this subject, here's a related document:
http://catalogs.infocommiq.com/AVCAT/images/documents/pdfs/TT189%20-%204611.pdf

Also:
http://cnyack.homestead.com/files/modulation/tapeav.htm

Yes they are. The chroma and luma are split, the luma is frequency-modulated, and the chroma is upconverted, then they're recombined and recorded with the same head. OK, they're on the same helical tracks and read with the same head, but then they have to be separated, the luma demodulated, the chroma downconverted, and then fed to S-video out as separate signals, or recombined for composite output.

Probably the most practical way of digitising VHS tapes would be to sample the composite video signal at 4fsc (i.e. 4 times the subcarrier frequency). The professional D-2 and D-3 video tape formats store a 4fsc-sampled composite video signal. There's probably a corresponding SMPTE standard for digital transport of 4fsc-sampled video. That should capture all data in the vertical blanking period.

Unfortunately I don't know of any hardware (consumer or professional) which can sample a composite video signal at 4fsc into a computer. Typical video capture hardware decodes the composite signal into component (YPbPr) first, then samples luma at 13.5MHz, chroma at half that rate. For preservation purposes, not decoding the composite signal before sampling would be better in my opinion.

Some professional (S)VHS decks might be able to output separated luma and chroma when playing VHS tapes. If you could sample the luma at 4fsc (and chroma at 2fsc maybe), that could give a better quality capture.

No consensus to how to do it, make me think if the already preserved media don't will "look so good for the project" in a medium term