How to stretch audio track very slightly

[Anakunda]

Hello,

got this video composed from encoded video and audio tracks and doing AV sync check in different phases of the movie, there's a little speed issue, while the sync is perfect at the movie beginnings, the audio is approx. 5 frames ahead at the endings. It's not drastical but 5fr (200ms) is noticeable already. Is there a way to fix?
I convert from AC3(2ch)->AAC(2ch)
Now I have 2 choices, besweet and eac3to, is one of these tools capable doing such a precise stretch, eventually which one is more accurate, what command options should I give?
Thanks in adv.

[tebasuna51]

With eac3to you can only use changes between standard fps. Then the slightly method is:

eac3to input.ac3 output.m4a -24.000 -changeTo23.976

enlarge your audio by 1.001 factor, only if your input is 200 sec. you can finish with 200.2 like you want.

You can use BeLight-BeSweet or BeHappy-TimeStretch (Rate method recomendated).

[Anakunda]

Could you be more precise how to specify stretch factor by eac3to/besweet if I want to enter is as fraction or float number?

Audition offers these 3 stretch modes:
1.Time stretch (preserve pitch)
2.Pitch shift (preserve tempo)
3.Resample (preserves neither)
+
Preserve speech characteristics: ON

1stly, which is the mode I shall choose if I want to stretch audio track but don't want to deform voices any way?
I guess it's the 1st mode but not sure.
Furthermore, do the command line tools perform stretch better than Audition?

[Blue_MiSfit]

I'd imagine you could do this type of operating using SOX or Audacity.

Derek

[tebasuna51]

Could you be more precise how to specify stretch factor by eac3to/besweet if I want to enter is as fraction or float number?
Like I say before you can't use eac3to. Only can work if your sound have exactly 200 sec.

To be precise I need know the full duration of sound to convert, because the stretch factor is a percent value of duration source.

Audition offers these 3 stretch modes:
1.Time stretch (preserve pitch)
2.Pitch shift (preserve tempo)
3.Resample (preserves neither)
+
Preserve speech characteristics: ON

1stly, which is the mode I shall choose if I want to stretch audio track but don't want to deform voices any way?
I guess it's the 1st mode but not sure.
- Pitch shift (preserve tempo) can't do the job.

- Time stretch (preserve pitch) can do the job, but need a complex and not accurated algorithm. Recommended for high percent change when pitch change isn't admisible.

- Resample/Rate (preserves neither) can do the job with simple and accurated algorithm. Recommended for slight changes, like here, when the pitch change is unnoticeable.

Furthermore, do the command line tools perform stretch better than Audition?
BeSweet is a abandoned soft with very old routines, Audition is better.
Maybe Sox is the best option between free software.

[Anakunda]

thx for sox.
I've already done by audition, I used Time stretch with about 99.98% ratio and the sound doesnot sound anyway deformed.