Re: edirol r09 and video

Jim Z wrote:

I am somewhat shocked at the fact the
clocks on these devices are not very accurate.

Oh, they're VERY accurate, which is why you can usually get away with "wild" (= no) sync for a reasonable length scene. But they're not perfect. Most of them use a crystal originally designed for TV sets that costs just a few pennies because they make so many of them. Still, no two are exactly alike, and in an hour, even if they're different by just a few hertz, the error accumulates enough so notice.

I thought electronics has advanced to the point
where frames per second video and audio recording
could be done in perfect sync from device to device.

"Sync" means that there's something tying them together so that they run off the same clock. When cameras were driven by spring-wound motors and recorder speed was determined by a DC motor and friction between the tape and capstan, you needed a common time reference for both. We graduated to SMPTE time code, and now we have word clock. Each system provided more accurate synchronization.

However, what sort of program could be used to
"tweak" the tracks to stay in sync ? There must be a
way to stretch the time frame of the video with some sort
of markers that will ensure perfect timing with audio.

Most A/V workstation programs have time stretch functions. Vegas is a popular one on a PC. However, this may seem like the layz way to accomplish synchronization but it's not a very good way. Just because the beginning and end line up doesn't mean that every point in the middle will line up. Generally the way these things go, you'll want to do some editing and you'll be chopping up scenes, so you chop up the audio along with the video. Because the speed is sufficiently accurate, if you line them up at the beginning, in the 30 seconds to five minutes before the next edit, the tracking will be close enough not to worry about it. But if the video runs continuously for an hour, you'll probably need to adjust the audio at a couple of points if synchronization is important. You need to look for "hit" points and line those up. For instance if you're doing a video of a band, when there's a close-up of the drummer, you want to hear the drum beat when you see the stick hit the drum. But if it's a long shot, you don't care about such precise accuracy.




--
If you e-mail me and it bounces, use your secret decoder ring and reach me here:
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers (mriv...@xxxxxxxxxxx)
.

Relevant Pages

  • Re: Custom ReferenceClock to slow down video but leave audio alone?
    ... > I have tried a ton of things to slow down / speed up video frames ... > So I have created my own Reference Clock based on the CBaseReferenceClock ... > The problem is that the audio render follows my custom clock too. ... > audio will be out of sync with the video, but that's actually what I want. ...
    (microsoft.public.win32.programmer.directx.video)
  • Re: Custom ReferenceClock to slow down video but leave audio alone?
    ... Then, in theory, if you have regular clock - audio will be streamed to BASS ... at the regular speed while video speed can be adjusted by "slowing" the ...
    (microsoft.public.win32.programmer.directx.video)
  • Re: Unsychronous recording from DV cameras
    ... This inevitably produces some delay. ... dv _includes_ audio this delay normally is no problem and there is no need ... However I also agree with you: Audio normally comes with the video in the DV ... If they don't use the same clock, ...
    (microsoft.public.win32.programmer.directx.video)
  • Custom ReferenceClock to slow down video but leave audio alone?
    ... I have tried a ton of things to slow down / speed up video frames including ... So I have created my own Reference Clock based on the CBaseReferenceClock ... The problem is that the audio render follows my custom clock too. ...
    (microsoft.public.win32.programmer.directx.video)
  • Re: AVI viewer for Palm?
    ... Better synchronization than MMPlayer ... for the file I tried (between audio and video, ...
    (comp.sys.palmtops.pilot)