> It's not just the capture card, you can't trust the laser disc player output.
As long as the LDP has a RS 232 connection, we are able to capture each frame, can seek and jump to each frame and start with frame 0 or 1, which is crucial. Using one of the last produced LDPs, would assure better quality output, than a model from the eighties. So i can trust the LDP output and there is absolutely no reason why i should not (especially for some last generation LDP models), but the MAME devs cannot it seems. Your linked thread also does not prove your POV at all. Nowhere in the thread, talk someone about LDPs that are untrustable, only about the quality you can get out of them, which is a logical thing, considering a lifetime production span for nearly 2 decades. It is out of question, that a latest model, will output more decent quality, but the medium that is played is 90% junk in terms of quality. It is not a premium LD from Star Wars New Hope.
I quote some of the important stuff of this long 14pages thread from your link:
The circuitry in the LDV8000 introduces a moderate amount of noise, but given the **** quality of the FF disc, where multiple generation edits have been done with some seriously noisy gear, it doesn't really matter that much.
And Firefox, is one of the best looking games here. 90% of the LD games will look way more ugly. They where produced very fast and very cheap with the lowest eighties technology.
For playback on a PC, you'd most likely get better results by just using a capture card as it will have a TBC, 3d (hopefully) comb filter, etc. As I was ultimately going to play out composite, going through that process and then back to composite again would degrade the picture.
The Mame image, for instance, looks fine on a PC, but nasty when remodulated. The explosions, in particular, are covered in artefacts where the comb filter hasn't managed to get rid of the crosstalk between luma and chroma.
At the end of the day, thats what will be happening for LD games anyway. They will be played on a PC. Even if we want to be able, that the games can be "remodulated", we would need to keep the interlace/fields and would need realtime deinterlace shaders, so that it doesnt look ugly on a PC. There are systems already in MAME that sacrificed exactly these things. Why the devs cannot do the same here, i dont know. Losing the fields is a degrade process in itself and only CRTs are capable of displaying it properly.
Going the route of not designing any new hardware, using a digital i/o card would be doable, it's just a lot of cash to spend to then find yourself limited by the 8-bit ADC in the LDV8000. Better results could be had by using an A->D card and capturing the composite out from a player with an analogue-only path from disc to composite port. I looked into this when I started, but I couldn't find a cheap card which retained the composite signal without processing it.
Dave talked about cheap cards in a 4 years old thread. A Blackmagic is not even cheap today, but a lot more than 2010. More important, it doesnt process the composite signal, what comes in, comes out with interlace/fields or without (depending on input) and with VBI data.
Adding your own hardware increases the choices a bit. If I were planning to do lots more of these, I'd design an A->D board which could hook in as early in the signal path as possible and grab video + audio before they've been split. I'd have a digital signal coming out the back to be captured by any means necessary... FPGA dev board, my LD emulator PCB or one of Colin's I/O cards. Net result would be the same as hacking wires into the LDV8000, only with more bits coming out. The risk is that the additional circuitry introduces so much noise that the results aren't as good as a straightforward capture from the composite out.
Again, like Dave, annikom and me already said, the result is questionable, in terms of "will it be a better capture" or better "If good enough is not good enough, and you want an absolute perfect preservation of an analog medium, then of course there is no practical way of digitizing the signal. At that point the problem will be unsolvable.". Even with the intended rocket science setup, the result can differ from one laser to another laser, cables used etc.
Example: Its like listening to a classic concert, a record of it, will never be the same, like the real life experience. The difference here is, that the audio you listening to, is a cheap box radio mono tune, compared to the classic concert and you are trying to capture that with equipment and science that belongs more to the classic concert.
I doubt that anyone will see a difference in a capture done like you intend and a "normal" capture with good equipment coming from the last years of analog production lifetime. There will be no better equipment (for capturing) in the future, as this technology was buried in the past.
All i know is, that time for the LD medium is running out, day by day.
Grabbing a combined a/v signal seems like a good idea in theory, but isn't really necessary in practice. I sampled the Firefox audio twice, once with one channel hooked up to the composite out. Vertical sync falls within the audio band (composite is designed this way to allow sync separation with really noddy electronics) and can easily be seen on the captured waveform or detected by a bit of software.
Cutting the audio 'perfectly' doesn't give ideal results because there appears to be quite a bit of a/v sync variation in the master. You'll also never have a perfect cut because the audio is bandwidth-limited and can't instantaneously switch from one level to another. The original player would have lost the audio for approx 2 ms or so when doing fast seeks and I suspect the disc was mastered with the audio slipped back a fraction to move the transitions into the time period where the output would be muted. 2 ms is a guestimate based on Atari blanking 32 lines in each field to hide video disruption.
How to tell if an LD player has an analogue output path: an analogue player cannot freeze frame a CLV disc. Get a copy of the service manual for a candidate to be absolutely sure as there may be some players with no frame buffer, but a small digital path for noise processing and TBC. Don't know of any fitting into the latter category, just speculating.
Bits for sampling: I would use 10 or more, although that's just a rough estimate based on thinking the 8-bits in the LDV8000 is right on the edge of not being enough. Doesn't matter about overkill mind you because you can always use more and throw away what you don't use.
There are loads of players with digital frame stores. The Pioneer CLD-D925, for example, has an 8-bit digital signal path, just like the LDV8000.
The last Blackmagic card, that has composite input and does 525i and 525p with VBI capturing, samples with 10bit 4:2:2 which is more than enough for SD material. Also the Pioneer CLD-D925 (1996) has a way more decent signal path, judging by all reviews i have read about this model (it even has PAL/NTSC), it sadly has no RS 232 connection and so it falls out for my choice of capturing.