Page 1 of 1

Podcast Time Shrinking / Stretching?

Posted: Mon Jan 22, 2007 8:58 pm
by Pedro Palhoto
Hey there everyone :),

I've been searching the web but I haven't been able to find a decent app that takes in mp3s or oggs and spits out time shrinked files.

Time stretching is altering an audio signal's speed or duration without changing the pitch. MythTV does this for recordings. This would allow people to be able to listen to twice as many podcasts (yeah, I'm a cast junkie :P).

I was just wondering if anyone has a recommendation to shrink podcasts (and doesn't take forever - MythTV is pretty fast). Being able to pass files via the command line prompt (and not via a GUI one at a time) would be perfect for scripting. A time-stretching portable player could also be a possibility..

I have developed code for MythTV (not for audio processing), and could pick up the time stretch routines and make them into an app, but I'm just too busy at the moment for some weeks of FOSS development.

Thanks.

Posted: Tue Jan 23, 2007 6:19 am
by coolfrood
I know that mpg123, and possibly ogg123 have the option of doing time stretch on the command line. You can get their output on stdout and perhaps pipe it to lame to reencode it.

Posted: Tue Jan 23, 2007 6:42 am
by Pedro Palhoto
coolfrood wrote:I know that mpg123, and possibly ogg123 have the option of doing time stretch on the command line. You can get their output on stdout and perhaps pipe it to lame to reencode it.
Wasn't aware of it, thanks coolfrood.

Just tried ogg123 -x 2 <ogg_file>, it maintains the pitch alright, but as it skips every other block, constant fast pops are heard. There isn't a smooth transition (or "anti-aliasing") between blocks. This isn't suitable for prolonged listening :(.

EDIT: mpg123 works smoothly :). It sounds pretty nice with tllts, but manager-tools... sounds really weird.

Posted: Tue Jan 23, 2007 8:00 am
by Patrick
Interesting concept. The challenge is making it sound natural.

Posted: Tue Jan 23, 2007 9:17 am
by Pedro Palhoto
Patrick wrote:The challenge is making it sound natural.
Accelerated laughter is probably the only thing that will remain unnatural how well the signal is processed.

Posted: Tue Jan 23, 2007 12:27 pm
by snarkout
I imagine a lot depends on how the mp3s are encoded. Podcasts are notorious for using broken vbr. I use mplayer to get through cbt vids faster, and it works great. The more I dig into mplayer, the more amazed I am at what an amazing piece of software it is.