This thread comes up in searches for software TBC, so hello internet! And I want to update you on the state of the art in software TBC.
First of all, you need access to the video signal in some type of raw form. There are two raw forms, the composite output from the laserdisc/VCR, or the RF signal from the optical pickup/ VCR head.
The next step is to digitize that signal. The two approaches most common so far, are special drivers for a CX or BT based capture card, or an oscilloscope. The capture card method requires special drivers, which are very rare and experimental in windows, but much more developed in Linux, especially recently as there is a new hobby called Software Defined Radio, and people from that area are using the old video capture cards for this new purpose, so that development work is indirectly helping the efforts of software TBC also.
The last step is simply software: understand the signal as video, and apply processing to improve it in various ways which are not well addressed by even the best VCRs/Laserdiscs/hardware TBCs. Now I can tell you the final conclusion here: the video doesn't look better than good analog equipment under ideal conditions. The real potential of software decoding is to avoid a) damages from poor analog electronics (not really an issue, if you just used better hardware) b) perfect TBC, less noise, less dropouts, this is the real need of restoration. And it's done (as proof of concept), and it works. What is needed now is to make this technology useable in the hands of the archivists.
So I say this is done, and here's the videos to prove it. First, this is what I used to consider the ultimate idea possible: pure software decoding of VHS, directly from the RF signal of the VCR head. This is as direct and raw as it gets, and due to all the decoding being in software, can extract the maximum theoretical quality ever possible. And the result is a bit disapointing, like I say it's not much better than a high-end VCR under ideal conditions - but bear in mind, this is a consistent level of quality, and could work even on currently "broken" tapes.
software decoding of raw RF signal from VCR head
And here is software TBCd VHS video:
you can see that it's perfectly stable, as would be expected.
Now here is video decoded by a portable USB oscilloscope, which is convenient since no special drivers are needed:
If you're tired of seeing samples in grey, here is a fully colour sample:
laserdisc decoded in a fairly advanced software. In this case, certain problems of ghosting, noise, dropout, and jitter are solved by using this technology.
I hope this overview gives you some idea of the potential of the technology/concept. Make no doubt however, software TBC is 100% possible, all the impressions you've had from hardware boxes and their poor working are obsolete. Of course, this still needs to be demonstrated in a very clear form before people will realize it.
I should also note that you won't find these samples when searching for 'software TBC', because it's redundant to call it by this term - software decoding of raw video, is by it's very nature, a type of software TBC; it's just inherent to the process. If software decoding ever became fully developed, the whole concept of a 'TBC' would become obsolete, because the decoding would always just work, and none of the problems of today would even be an issue anymore.
Until there is easily usable software and equipment for this technology, it remains in the hands of a very few.Eventually it would be possible to buy an off-the-shelf USB oscilloscope of the necessary specs, and run a program, and out pops the result with not much fuss. Every once in a while some new update would come out, and you could re-run the decoding to fix even more problems, like 3d comb filtering or other quality improvements, until at some point most video problems are 99% solved and there would be nothing more to do.
Btw, here is a fully software 3d comb filtering, which is pretty much perfect in the static scenes:
software 3d comb filtering
-- merged --
There's not much left for me to do in the software TBC from raw capture, if someone really needs to restore some samples, what I could do is help you get set up with the existing software to do this. You need a CX based capture card or digital oscilloscope, possibly and install of Linux, and I can direct you through all the other steps. The problem with baked-in jitter is still an open problem, and so far as I can tell I'm still the only person on the internet with an idea to solve this. Unfortunately I lost all my work in several hard drive crashes, and have to start all over again.