Archiving Laserdiscs to DVD

jgourd wrote on 5/18/2002, 9:52 AM
I have been on a very long project of archving my collection of Laserdiscs to DVD. Many of my discs are not already on DVD and it is not expected that they ever will since I have been collecting obscure import releases for years.

At first I used a Dazzle DVC II which was fast, but the picture isn't great. I have now honed the process to where I get amazing results.

I capture the S-Video out and analog audio through an analog to DV box (camcorder with pass-throug or Dazzle DV bridge). At the same time I record the SPDIF out of the player to a stereo track in VV3. I do this so that the sound track I work with is uncompressed and not re-digitized by the DV device.

In VV3 I sync the two sound tracks. This turned out to be easy because the DV device and the SPIF in on the sound card (RME Hammerfall) are getting their clock from the same source - the player. Once I have captured both sides and used VV3 to re-assemle the whole moive, I render both a WAV and DV avi file.

I take the rendered avi and pass it through a few filters in Virtual Dub so that video noise from the laser disc is minimized. I then make a new project in VV3, bringin the wav, and the picture from the filtered avi and render to an MPEG-2 DVD file whos average bit rate has been calculated to make a 4 gig file from the project. I have a spread sheet that can predict the total file size given a 15 second render that tells me what the average bit rate should be.

Overall, the DVDs look better than the original Laserdisc!

Comments

johnmeyer wrote on 5/18/2002, 10:51 AM
Two questions:

1. Its a lot of extra work to bring in the audio separately and then sync it back up with the video. Are you really able to tell the difference?

2. What filters, and what settings for those filters, are you using in VirtualDub. As far as the video is concerned, that seems to be the only thing you are doing to modify the video prior to encoding your MPEG-2. If you are geating superior results, it must be from this pre-processing, and I'd be curious to know what it is.

jgourd wrote on 5/18/2002, 12:06 PM
In virtual dub I use chroma noise reduction and temporal cleaner. Both of these are 3rd party filters I downloaded.

In many cases where the original sound track is poor to begin with it really makes no audible difference. In cases where the sound track is excellent like concert videos (SRV Live at Buddakan for example) it does make a difference. You want the sound track as close the orignal bitstream as possible. Re-digitizing can intoduce quantization errors and audible artifacts that audiophiles find unpleasent.

In some cases I use a Mastering plugin like Ozone to improve the sound where appropriate. This is a time consuming process for sure, but if you do it this way, you won't wish you had did a better job a few years down the line when all you have is the DVD to watch.

The spreadsheet I use to predict the size of the final MPEG-2 render is available for download at: http://www.lamoateffe.com/MPEGFilePredictionCalculator.xls

Enjoy
johnmeyer wrote on 5/21/2002, 1:05 PM
Thanks for the info on VirtualDub filters. Very helpful.

John
altphase wrote on 5/22/2002, 8:59 AM
I used a similar method to convert a movie on LD to a DVD. I recorded the movie through a SONY media converter and later exported the wav file and DV avi as separate streams. The problem is that the movie on the LD had a 3:2 pulldown added to it, so I really wanted to do a reverse telecine to be able to encode a final 24p mpeg-2 file to use on the DVD. Unfortuately I found out that there are no great tools for doing this on the PC (not cheap at least). Out of all the software I tried only the VirtualDub IVTC filter worked well, but it still had trouble with some scenes and ended up mixing some interlaced frames in with the 24p avi. Overall the result was very good. I used TmpegEnc to encode the 24fps file with the 3:2 pulldown on playback flag and was able to burn a DVD that looks almost better that the original LD (aside from the occasional motion jitter caused by the imperfect IVTC process).
johnmeyer wrote on 5/22/2002, 10:39 PM
I've used the IVTC in TMPGEnc and found that it works very well. You have to remember to tell it to look for the frames to throw away by selecting Auto-Setting within the Inverse Telecine option (you have to double-click on the Inverse Telecine option to get to the settings, something that is not at all obvious).

Once you select the Auto-Settings option, select 24 fps (Flicker prioritized), set Deinterlace to None, and then click on Enable when encoding.
waypt wrote on 6/14/2004, 4:52 PM
John
I'm starting to do this Laserdisk backup to DVD and am wondering if Vegas 5 can do real Inverse Telecine yet? If not, then I guess I'll use decomb with AVISynth. Does that sound OK?

Running through my mind is this question --

If I capture an LD that was originally shot on film and has been telecined to 29.97fps on the LD, then place it into a 24p project, can this avi be considered as 24p to Vegas, and can Vegas then remove the duplicate fields (IVTC)? I think it should, from all I've read, but I can't find a concrete answer to this and the sequence of steps to accomplish this.
johnmeyer wrote on 6/14/2004, 6:15 PM
wondering if Vegas 5 can do real Inverse Telecine yet?

I don't think so, but perhaps someone else knows of a way.