Bad Vhs transfer in digital as source

[mammo1789]

Hi guys, I appreciated all the help that i get from this community and i need your help once again.

I have some videos of fighting that were transferred poorly from Vhs source and the guy made dvix and mpeg files out of it. What do you think is smart think to do. The videos suffer from noise ( to bad they were not denoised before encoded to divx) and i assume were recorded on a lousy vcr without tbc. I want to make it watchable because now when i watch them on my dvd player the picture jitters a lot and is muled also the resolution is low 352 240 ( will you think the http://www.infognition.com/VideoEnhancer/ super resolution can help) Thanks guys i will post pictures and upload short clip so you can see what im talking about.

[mammo1789]

http://www.mediafire.com/?v85eh0d9ka32s4z

http://www.mediafire.com/?77sa3c57k7ex67j

The video are 1 mb so you can see divx files

[mammo1789]

First atempt. Video enhancer super resolution 704 480 on highest (slowest) quality setting+ msu denoiser + msu smart deblocking 0.8 and msu smart sharpen

http://www.mediafire.com/?9uk8277232xgf92

[2Bdecided]

There's an AVIsynth-based attempt at a TBC in software - I'd start with that.

Cheers,
David.

[sanlyn]

There's an AVIsynth-based attempt at a TBC in software - I'd start with that.

Cheers,
David.

From all I've heard (and seen) here and in doom9, the software TBC doesn't work.

https://forum.videohelp.com/threads/338581-Avisynth-function-to-TBC-old-VHS-video?p=210...=1#post2104100

[poisondeathray]

There's an AVIsynth-based attempt at a TBC in software - I'd start with that.

Cheers,
David.

From all I've heard (and seen) here and in doom9, the software TBC doesn't work.

https://forum.videohelp.com/threads/338581-Avisynth-function-to-TBC-old-VHS-video?p=210...=1#post2104100

But there has been some recent developments by jmac and others

http://screenshotcomparison.com/comparison/88810

http://forum.doom9.org/showthread.php?t=162726



Still work in progress , very slow, and only works in some limited conditions, but holds some promise

[sanlyn]

Saw those posts recently. Amazing, so far. Relieved this project isn't mine!

[mammo1789]

From all I've heard (and seen) here and in doom9, the software TBC doesn't work.

Me too but i like what I'm seeing on doom9 site. I didn't understand is this the same tbc avysinth plugin or is something more. I will have to ask jmac for that script (and plugins needed to work) he was very helpful

[Cherbette]

I've been reading up on this over at doom9 as well and am testing it out. From what I understand it doesn't work 100% properly as of yet on "2nd generation tapes, etc"??

[jmac698]

Oh, it does work in some conditions, Cherbette can atest to that
http://screenshotcomparison.com/comparison/91853

I can see I'm having "public relations issues" hahha. I'm talking about a lot of technical stuff and jumping around a lot of related ideas, and most people can't follow that. Maybe I need a blog or something where I can relay news in small, focussed chunks for the public.

Btw, anything I said about 2nd generation tapes, I was probably referring to making my own test of a 2nd generation tape. However what I have working so far has no problem on any generation tape, like I've said before I'm just lining up black borders, it's very simple
and I don't know why no one has done it before.

Anyhow I would be curious to know why anyone thought it would never work? I can't understand why it's supposed to be impossible?

[Cherbette]

I know now that it can work Very promising results so far.

[edDV]

First you model a mechanical VCR in software.......... down to pimch roller wobble.......

The rest should follow.

[sanlyn]

First you model a mechanical VCR in software.......... down to pimch roller wobble.......

The rest should follow.

[jmac698]

That can be done, you wouldn't care about the physical model, just assuming the equation for intertia/friction is polynomial say, just fit a polynomial to the resulting jitter and you have your model. This is the black box approach.

[mammo1789]

That can be done, you wouldn't care about the physical model, just assuming the equation for intertia/friction is polynomial say, just fit a polynomial to the resulting jitter and you have your model. This is the black box approach.

I totally agree almost everybody here svear in mpeg2 software encoding right ? Which was unthinkable some years ago you had to have hardware mpeg2 encoder or similar machine. Now even below average pc can encode high quality mpeg2 files ( even h264 which still there are not analogies in hardware domain like dvd recorders for mpeg2).

We encourage you to keep up the good work.
Now can I ask you to process the file that I posted trough you script and see the results
Thanks

[jmac698]

Instructions are posted with my script. I think it's more important that I spend my time improving the script instead of user support of a non-finished, experimental script. Thanks.