Capture & Trim with VD

[Tafflad]

Simple Q .... I have been capturing VHS tapes into DV format using WINDV as capture program.

This results in lead in lead out 'garbage' I have been using VD to trim this out and then setting "Direct Stream Coyp" on both Audio & Video .. using save as .avi
Ready for when I want to process ........ is this non destructive ? (I had assumed so) ... I am not transcoding or rescaling anything ...... I ask as it seems to take quite some time for VD to do the save as ... (10min plus for some) at 10.6GB video rate.
Sort of thought with only snipping off start and finish seconds 'save as' would be quick.

[DB83]

AFAIK VD is very fussy when it comes to stream copying. EVERTHING must be identical and I guess as you can not save in the codec originally used then re-encoding takes place.

You would be better off doing the TRIM as part of an avisynth script.

[Tafflad]

I looked at Media info on original and the trimmed file .... both are exactly the same so don't think its a codec issue.

If I used an aviscript ... wouldn't that just use VD just the same ? just driving it form a different point ? (or I may be totally wrong)

Looking at clip on VD I can get it to exact frame for cut point, with script how would you do that ... or do you mean find frame numbers first in VD and then enter them into script.

[DB83]

Looking at clip on VD I can get it to exact frame for cut point, with script how would you do that ... or do you mean find frame numbers first in VD and then enter them into script.

Unless the real experts have another way of precisely locating frame numbers.

Of course you can do this(the lossless trim) is Vegas but, there again, that is also very fussy.

[sanlyn]

Find frame numbers first in VD, then enter into a script. Remember that the first frame in a video is numbered 0, not 1. Example, if you want to delete frames 0-650, then Trim(651,0) will delete frames 0 to 650 and keep frames 651 to the end. The new start frame will then be numbered 0.

If you want to remove 650 start frames and the last 1050 frames of a video with 200,000 frames (the last frame number in the video will be 199,999), use Trim(651,198949). Use "fast recompress" in VDub to save the results.

Easier: open the video in VirtualDub, cut your frames from the start, then cut off the end if desired. Save with "direct stream copy" or "fast recompress".

[sanlyn]

Simple Q .... I have been capturing VHS tapes into DV format using WINDV as capture program.

This results in lead in lead out 'garbage' I have been using VD to trim this out and then setting "Direct Stream Coyp" on both Audio & Video .. using save as .avi
Ready for when I want to process ........ is this non destructive ? (I had assumed so) ... I am not transcoding or rescaling anything ...... I ask as it seems to take quite some time for VD to do the save as ... (10min plus for some) at 10.6GB video rate.
Sort of thought with only snipping off start and finish seconds 'save as' would be quick.

Keep in mind that you are doing it correctly with Direct Stream Copy or even with Fast Recompress, but remember: you're making a 10GB copy of a 10GB video. Try making your own copy from folder to folder of 10GB of data; it will take a lot more time than a blink of an eye. Don't hold your breath while waiting.

[Brad]

I ask as it seems to take quite some time for VD to do the save as ... (10min plus for some) at 10.6GB video rate.
Sort of thought with only snipping off start and finish seconds 'save as' would be quick.

It should take as long as copying and pasting a 10.6GB file. It will be faster if you save to a different drive than the source file is located on.

AFAIK VD is very fussy when it comes to stream copying. EVERTHING must be identical and I guess as you can not save in the codec originally used then re-encoding takes place.

There is no re-encoding involved with Direct Stream Copy.

You would be better off doing the TRIM as part of an avisynth script.

Avisynth will require recompressing.

Easier: open the video in VirtualDub, cut your frames from the start, then cut off the end if desired. Save with "direct stream copy" or "fast recompress".

Not good to pointlessly recompress DV (Smart Rendering isn't enabled by default).

[sanlyn]

I keep forgetting it's DV.

[jagabo]

AFAIK VD is very fussy when it comes to stream copying. EVERTHING must be identical

Since he's using "Video -> Direct Stream Copy" everything is identical, by definition.

and I guess as you can not save in the codec originally used then re-encoding takes place.

Direct Stream Copy allows no recompression. It's a remux. Only cuts/pastes are allowed, only on keyframes. With DV every frame is a keyframe. So you have frame accurate edits.

You would be better off doing the TRIM as part of an avisynth script.

Trim in AviSynth would require re-encoding.

As long as you use Direct Stream Copy mode (audio and video) VirtualDub is lossless for DV to DV editing.

[DB83]

Points taken.

I just see that step a bit of a waste of time if the intention is still to filter in avisynth whether from lagarith (better) or DV.

Of course, converting the DV to lagarith will also take time so its a bit of a no-win situation.

[Tafflad]

Easier: open the video in VirtualDub, cut your frames from the start, then cut off the end if desired. Save with "direct stream copy" or "fast recompress".

That is the way I use it .. just wanted to check it was not doing any degradation ... using 'direct stream copy' as it was taking a long time.

Thanks for confirmation

[sanlyn]

Riht. As jagabo noted, for this particular task of just making simple unfiltered edits, use Direct stream copy.

[2Bdecided]

I just see that step a bit of a waste

Of course it's not. No one should throw away their original captures, but there's no point keeping blank space. DirectStreamCopy in VirtualDUB performs a lossless copy of the audio and video data giving you the part of the capture that's worth saving, without further loss. Using Trim in AVIsynth forces re-compression (=filesize bloat for no reason if re-saving to lossless, or loss if re-encoding to DV). Going via AVIsynth also hard-codes any DV decoder errors into the file, and (in the case of stuff transferred from actual DV tape) strips all the shooting metadata, timecode, time and date information, and error flags. VirtualDUB preserves all this.

Even though VirtualDUB is lossless, unless there's minutes of blank space, I wouldn't bother trimming it. I usually do bother because I leave the thing copying and end up with minutes (sometimes hours!) of nothing at the end of the VHS>DV capture. I never bother trimming DV-AVI transfers of actual miniDV tapes though.

Cheers,
David.

[Tafflad]

Same reason, short blank section at front, . and about 10 mins' of overrun

[PuzZLeR]

Will also confirm - been using VirtualDub's "Direct stream copy" for over a decade now (wow, can't believe it's been that long) and it has never re-encoded/re-compressed anything - 100% non-destructive.

It wasn't frame accurate for MPEG-4 formats like DivX, Xvid, but that's because the key frames are further apart on those. With DV being intra-frame, accurate cutting is assured (and will not need any smart rendering).

So go ahead and use it to trim ends on your captures or make cuts, and even lossless joins as well with "File Append AVI segment".

[PuzZLeR]

Even though VirtualDUB is lossless, unless there's minutes of blank space, I wouldn't bother trimming it. I usually do bother because I leave the thing copying and end up with minutes (sometimes hours!) of nothing at the end of the VHS>DV capture. I never bother trimming DV-AVI transfers of actual miniDV tapes though

Interesting you brought this up.

For camera content/tapes I can agree, since it's what's been shot, and the true digital source - unless you purposely want to edit it out (and that's a whole other debate and philosophy on the integrity of capture, etc, blah, blah, blah...).

I will comment on those captures from VHS tapes though, since it's not a direct transfer of any digital source from the tape (and much of it may be from TV, or second gen stuff, etc.), only a digital representation of it.

Honestly, I find it much less time consuming to just let it run hours on end. I just feed the machine, tape by tape, but do other stuff while it's working, in or even out of the house, then come back when it's done (or even long after). I've even fallen asleep while forgetting a capture is running too.

But that's fine. I could care less about the junk it may be capturing. (And always have good stop conditions, and it's on a dedicated capture box, anyway.)

When done, I brush through the video and cut out what I want to cut out with direct stream copy and delete the original capture string. Sometimes I find I only want, like 20 minutes, from a multi-hour capture. (Using "Job Control" is very helpful if you cut out many pieces at a time.)

I've saved so much time doing it this way, even if it was for only 20 minutes of content, as this takes minutes to cut out, instead of babysitting each capture, one by one - after hundreds of tapes, even labelled some 10-20 years ago, I still don't know for sure what's on there.

My point is that Direct Stream Copy is wonderful, and should be highly encouraged. I wish we had something like this during the old days of VHS recording (while we actually had to be attentive to each session - recording/editing/stopping/starting/etc - ugghhhh).

[DB83]

Pedantic mode on:

My friend, in his OP, said "Ready for when I want to process"

Agreed that he has a head and a tail - not my friend unless he is a dragon in disguise - but his WinDv cap.
Agreed that if he trims this in VD that does not alter the original (unless he then deletes that)

So the 'processing' ,however it is being, done will involve re-encoding. If the trim is done in avisynth then it still becomes one pass as part of the 'process'. Only if there is no intention to do ANY filtering with the cap if there real benefit in doing this in VD

Pedantic mode of:

[This Dragon will now retire to his cave for this topic ]

[Tafflad]

Well I thought it was a simple question when I asked it.