Speeding up Audio, When would I do it?

[Hittz]

Hello people, I would like to learn something to the point that I am confident about it, I have had cases where I have converted an avi to dvd, I would throw the whole avi into an encoding program like TmpgEnc, (which usually works well for me) resulting in an out of sync audio file contained in the dvd, after authoring it. Now, I understand some encoders dont handle vbr audio well at all but the case I am reffering to is this:

Lets say I have an avi which has a frame rate of 25, I want to create a dvd (NTSC), so, would I extract the audio file first to uncompressed wav (PCM) then change the framerate of the avi to 24, or (23.976) or go up to 29 ? Then should I speed up the audio, in say belight to match the video, by the same amount ? (video avi from 25 to 29, then speed up audio to 29 to match the increased frames in the video, and at the same time convert to ac3? Finally then author my project?
I want to learn so that all my encoding and conversions are synced , and to ensure my video plays as smooth as possible (without any choppyness, or that stop and go effect- assuming the video is of good quality)

I am trying to clearly understand this rule if it is such, ( the do's and donts, the whens and the why's, relating to this subject


Thanks Much
Hittz

[Alex_ander]

The only case when you need to change audio duration is when you go from 25 fps to 23.976fps in video by assuming a different framerate with the same number of frames. You can avoid spoiling the audio by slowing it down 4% if you use f/rate conversion method without changing video duration.
1.Resize video for NTSC (720x480)
2. Encode video as elementary progressive stream (video only) at original 25 fps.
3. Run the result through DGPulldown with 25=>29.97 setting
4.Process audio separately and use it at authoring with your video, it should be in sync (in some cases needs delay only adjustment).

[Hittz]

by assuming a different framerate with the same number of frames. You can avoid spoiling the audio by slowing it down 4% if you use f/rate conversion method without changing video duration.

thanks much for your response and guidance Alex_ander but can you clarify the above part,(I dont just want to say , ooooooh ok, And I really dont understand it )

[Alex_ander]

This means (with one of conversion methods) that if the frames of a video, originally encoded at 25 fps are shown at e.g. 23.976fps, this changes video duration proportionally. The corresponding AviSynth procedure used for this is:
AssumeFPS(23.976)
Then the video can be encoded at 23.976 with pulldown flags ('telecined' the standard way) making it generate extra frames (by combining fields of existing ones) at playback, thus producing 29.97fps. With this method you have to change audio length 4% to keep it in synch.
Instead of this you can apply a non-standard pulldown to a 25fps video offered by DGPulldown (after encoding; the video should be progressive and 'elementary stream'). Since pulldown doesn't change video length, you don't need processing audio.

[Hittz]

Im reading your answer along with the meanings of the words pulldown, and telecined, and the differences between various framerates, I should be able to then understand it. Thanks much Alex_ander for your time, and effort to help. I understand some now, Gonna sleep on it and study more tomorrow

Thanks Much

[Hittz]

Ok Alex_ander, I've tried to understand and read some of the definitions in the glossary section of Videohelp.com. Here is in example using a 10minute video sample of an avi that would be a pal 25fps film

Would I then set for example, Tmpgenc xpress 4, like this to create an mpeg2 (dvd) elimentary video file?



I would extract the audio (ac3) using virtualdub mod, as is, to be used when I author the project?
Are these the correct settings?

[Alex_ander]

The pictures show that you are trying to encode a PAL avi to standard NTSC DVD template. TMPGEncXpress will really do it keeping video length (and audio will be in sync). But the quality wouldn't be the best possible (problems with stutter-like motion reproduction at fast panning).
The better method suggested above needs encoding of an NTSC-sized video (720x480) at PAL framerate (25fps), so you have to make a step aside from that template by using 'MPEG output' button (on the right side of your picture). This (after a warning) will bring you to another page with some more settings, where you'll be able to change framerate to 25. Then open GOP tab (group of pictures) and change 'max GOP frame number' from 18 to 12 (it is needed for correct pulldown that will be applied). The rest settings (progressive, elementary) will come from the NTSC template you previously adjusted. Then encode and use DGPulldown software: set 25=>29.97 and process the video.
The direct encoding you've just tried is easier, but the second method should give better quality.
Now it's my turn to go to bed . Good luck!

[Hittz]

LOL. I like that last part
Thanks Alex_ander, and nighty night
Ill give it a try

Edited 12-16-07 Worked perfectly Alex_ander.