Reply 59580 of 59600, by Ozzuneoj
- Rank
- l33t
BitWrangler wrote on 2026-08-25, 14:31:The point it claims, is so you can come back to the captured RF (Humungous file) when tools are better in a few years time, that […]
The point it claims, is so you can come back to the captured RF (Humungous file) when tools are better in a few years time, that implies that they are only good enough to show potential at the minute. That means for it to be worth the bother, you want to preserve max info and store the RF and the approx quarter terabyte a tape is after Flac by my reading.
Also the intermediate steps to Media Player playability, make huge files, it is sounding like you need just over terabyte of room to handle a 2hr tape. 100MB a second is max cited for the uncompressed full frame AV.
Then conversion speeds, rf capture at real time of tape play, then sampling of RF capture takes minimum ten times play length, then conversion to AV on a high end ryzen 9 takes double to four times play length, then if you are lucky you can convert it to a consumable format at about double speed. So 2 hours of tape play, 20 hours RF sampling, 8 hours AV convert, 1 hour format conversion. 31 hours per tape.
Not all of that is hands on time, you can automate it, but it's all you've got a machine doing for a day and a half, just for one tape. You would probably want to do multiple rf captures in one day though, which puts you at needing 2TB RF capture space, that you have to mirror to the RF sampling machine because you don't want it thrashing the same disk. then that needs 1TB to work with, then you need another big data bucket for the AV conversion, and finally you get something that's VideoCD to DVD sized depending on what final format you want it in. Anyway, lot of "in flight" storage even if you end up with a small file at the end... but does it look the same by now as something captured with a decent card with frame synch and stabilisation? The Canopus he's saying does 75% is well known to chuck out half the color info at the DAC, then it is also compressing down to a bitrate for the firewire. So basically we know that anything that doesn't do that is wayyy better.
Anyway, when consumable and conveniently stored format is the objective, is the 99% correct version really gonna be worth 10-15 times the time vs a 90% version when all the visible advantage is likely to disappear in the format conversion?
Yeah, it doesn't sound like a fast or convenient process no matter how easy they say it is to get started.
Still, I don't think the final FLAC compressed RF files are that large. There is a chart at the bottom of this page that says a 120 minute recording is around 39GB using FLAC compression:
https://github.com/oyvindln/vhs-decode/wiki/FAQ
It suggests using the compressed RF files for archival purposes since that still allows flexibility for higher quality conversion in the future if technology improves. So, I don't think there is any reason to keep the full size uncompressed files.
I am not saying this is what I or anyone else should be doing to convert whatever VHS tapes they have around. I haven't done it myself, so the process could be even worse and even less practical than it sounds on paper. It just seems like another option that uses more of a no-compromises approach for maximum archival quality.
I definitely don't need anything like that for my old home movies. I just haven't found any practical or affordable high quality options yet because every time someone discovers a method it eventually ends up on youtube and the value of the devices to do it go through the roof.
Now for some blitting from the back buffer.
