4.0d capture glitch

johnmeyer wrote on 8/22/2003, 2:09 AM
I just tried to re-capture several videos that I made nine months ago. At various points in the capture, about 1-3 seconds of video and audio were completely missing.

I went back and recaptured a short segment just before one of the glitches, and captured until ten seconds after the glitch. The exact same segment was missing. However, I could see and hear the video and audio on the monitor and from my camera during the capture, and nothing was missing. The tape plays fine.

I did the above capture and subsequent test with Vegas 4.0d on a WinXP Pro SP1 computer, capturing from a Sony TRV-11 NTSC camera. I've done lots of captures in the three years I've owned this camera, and not had anything like this.

I then tried the same test on my old Win98 SE computer that is running VideoFactory 2.0c. I got the same exact glitch.

But wait, I then captured on my old Win98 SE computer using Studio DV. No glitch! It captured perfectly.

I then captured using SCLive on my Win98 SE computer. Perfect capture. Nothing missing.

Being a glutton for punishment, I reconnected to my WinXP machine, rebooted (just for fun) and tried capturing again in Vegas 4.0d. Nope, the problem was back exactly as before, and just like with VideoFactory. I should also note that both Vegas and VideoFactory capture reported zero frames dropped.

I then fired up the latest version of SCLive (Scenalyzer) on my WinXP machine and captured the same footage. Pefect capture.

Thus, the same tape, across multiple computers and O/S configurations exhibits gaps when captured by Sonic Foundry products, but not when captured by other programs.

This is the first bug of this magnitude I have found in Vegas. This is nasty. As it turns out, I captured almost four hours, which I then converted into 2 DVDs for a client (which was the whole point of re-capturing the finished tapes). I didn't find the glitch until I played the first few mintues of one of the DVDs. It took one whole day to capture the four hours of video, render to MPEG, and burn in DVD-A. This looks like a major bug, and one that must have been there for some time, since it shows up in VideoFactory as well.

I have virtually nothing running in the background on either machine (not even anti-virus software). No Windows Messenger, etc. Not that would cause the problem to occur at the same point every time.

I will hang onto the tape in case is could be useful to diagnose the problem.

Comments

Grazie wrote on 8/22/2003, 3:12 AM
John, excellent analysis. Well thought through. Yes, I've downloaded the latest 4.0d, I'm not installing until I'm more than certain that basic stuff like Capture over my suite of SoFo products is 100%. I mean it.

Again thanks for the "controlled" experiments. I for one DO appreciate such intelligent observations. S@D's Law states that someone else here will report 100% success with Capture - just you wait! Yeah?

Grazie
SonyEPM wrote on 8/22/2003, 8:51 AM
Is this happening with more than one tape?

fyi: Video Capture has not been touched in any 4.0 revision, and only minor changes were made to Video capture between 3.0c and 4.0 (outside of the new vidcap media pool stuff).

johnmeyer wrote on 8/22/2003, 12:04 PM
In answer to SonicEPM:

Is this happening with more than one tape?
I don't know. I have over 50-70 tapes at this point. I have never seen this in any tape that was created directly in the camera (that is, a tape that was made by pointing the camera at something and pressing the record button). The tape that causes the problem is one that I created in VideoFactory last December, and then printed back to my camera. I then re-captured this tape back into the computer yesterday in order to create a DVD.

fyi: Video Capture has not been touched in any 4.0 revision ...
I'm sure it hasn't. In fact, as my long initial post pointed out, my attempts to capture in VideoFactory 2.0c (Build 125) also failed. That capture code is probably a year old.

I realize that thousands of people have captured hundreds of thousands of hours of video with SoFo video products, myself included. Until yesterday, I thought the capture worked fine. Many people here have reported (in these forums) problems with capturing, but usually those are attributable to no disk DMA, faulty 1394 boards, bad BIOS settings, background programs, viruses, etc. Many of those things could apply to me as well, except for the fact that the problem happens on two completely different machines; I have been doing this for over three years (which doesn't make me infallible, but at least means the equipment has been working most of the time); and of course the fact that two other programs (Studio DV and Scenalyzer) capture the tape perfectly, one of them on both machines.

I strongly suspect that there is some sort of timecode interruption or anamoly that got put on this tape which Vegas/VideoFactory does not handle gracefully. If this is the case, one question to be asked is how that glitch happened. I cannot answer this (and of course, this may not be a correct identification of the problem).

The much bigger question is, why do the other programs deal just fine with this tape, but SoFo capture cannot? SoFo, please feel free to email me or call me if you'd like to discuss this further. I'd be happy to help assist in any troubleshooting required.
Jsnkc wrote on 8/22/2003, 12:06 PM
I've been using 4.0d since it came out pretty much every day on 2 diffrent computers capturing from a variety of diffrent tapes and I haven't had any problems with what you are seeing. To me it doesn't sound like a software problem or everyone would be reporting it (well at least more than 1 person anyways)
johnmeyer wrote on 8/22/2003, 1:55 PM
... doesn't sound like a software problem or everyone would be reporting it ...

I agree up to a point. Generally, capture has worked fine. However, I have a tape that was the result of a print to tape function, and now I'm recapturing it. I know for a fact that tapes resulting from computer renders can have problems when re-captured because Pinnacle's StudioDV had exactly this problem two years ago, and it took them two separate bug fix releases before they fully cured the problem. I suspect something similar is going on here. The Pinnacle problem was traced to an interaction between their software and Microsoft's qdv.dll file which is part of DirectX.
granmadave wrote on 8/26/2003, 10:29 AM
Hey, whadda-ya-know? You're not alone! I get that same error!

I have received that same error on certain tapes since 3.0C! As you mentioned, Pinnacle Studio and others will capture without errors. In Vegas, the audio tends to "skip" or "Stutter" where the source tape has an edit or a glitch, regardless of the intensity of the glitch in the source tape. I update through WindowsUpdate almost weekly, and currently have AutoUpdate, so all my drivers and system files are up-to-date.

I have captured through USB via a Dazzle DVC-80 in 3.0, and now a PyroDV 1394 card with a Dazzle Hollywood DV-Bridge in 4.0. The error, as you mentioned, does not seem to be hardware-related, due to the success of capture through other software.

It seems you have researched this a lot, and found more success than myself. I recently transferred a 26-hour, 26-Volume VHS set to 12 DVDs for a customer, wherein I had to capture through Studio8; Vegas kept getting an audio skip. It was more time consuming, for I had to then import the clips, wait a year-and-a-half for the audio peaks (yes, i know they're not necessary, but I like having them for editing purposes), before I could go on to setting the regions, and running the "Batch Render GUI" for the M2Vs and AIFs when I go home for the night.

As you said, it's possible to circumvent this problem, but it shouldn't BE a problem.