Archiving Laserdiscs without any player processing

[NJRoadfan]

Came across this neat project.

https://www.domesday86.com/?page_id=978

Basically it samples a Laserdisc's signal using the RF output directly from the laser, completely bypassing all of the player's processing (TBC, dropout compensation, NR, etc.). From there, the RF signal is decoded by software to actual video.

They have some test card samples here: https://www.domesday86.com/?page_id=1332

Looks like the software processing needs some work, particularly the comb filter (its hard to tell if this is artifacts from the actual decoding or "baked into" the source), but in theory this could lead to a much better restoration down the line as the software decoder improves. This cuts out variables like player and capture card quality entirely. I'm curious if this is theoretically possible with a VCR. One would have to demod the "color under" signal on home formats like VHS if the signal is tapped from the heads, or tap the signal further down the line.

More background on the ld-decode software: http://forum.lddb.com/viewtopic.php?t=2671

[hodgey]

There is already a video on youtube of someone that did something similar with a VHS VCR so it's very much possible in theory.
https://www.youtube.com/watch?v=t0eLxQY6-RI

It's interesting that they did this using a pretty standard video capture card with modded drivers rather than some expensive oscilloscope. Maybe it would be possible to decode a composite or s-video signal in software using hacked drivers as well which would give some extended opportunities without having to tinker with VCR internals.

[metaleonid]

So what's the progress with LDs? I am soon to be ready to re-capture with Analog ADV 8742.

[hodgey]

So they've actually gotten quite far with this project, even designing a custom capture device using an off the shelf A/D converter and other components, and they keep improving it. Maybe it would be an idea to try to collaborate with them and look into developing something similar that could work on video tape output.

I actually decided to buy one of these cx2388x based cards they initially used for prototyping, as a cheap way of capturing raw signal data with a custom linux driver. (they are very cheap, so in the worst case it's not a lot of money wasted.) I'll probably just try it on a composite signal at first (as it's a bit simpler than messing with the raw rf output) and see if I can do some rudimentary software decoding as a hobby project, but maybe I'll look into doing more complex stuff later.

[lordsmurf]

I think my interest in this would come down to signal timing, and ability to augment embedded timing errors as well as direct (with the latter being what a TBC fixes).

But I'm just an advanced user, not somebody that can engineer the hardware.
I can look at hope, help spread the word, sometimes pony up a few bucks, be a cheerleader, but that's it.

[juhok]

It is possible for VCR. There was an engineer in Finland who built custom DSP for direct signal from the video heads of FS-200 if I recall correctly. The gear he used cost something like 10k€. Results looked pretty good, better than I could ever do with normal process. He quit years ago but I still got a sample if someones interested.

edit: Oops, it seems I deleted the samples along with everything else while I quit a while ago.. Sorry