alsa-lib resamping error

alsa-lib resamping error

[Paul Davis]

if you try to play a 22.05kHz stream on hardware that supports 32kHz
and 44.1kHz but not 22.05kHz, alsa-lib's plughw layer chooses to use
32kHz and resamples. this seems like an error to me: it would be much
better to choose 44.1kHz and do simple integer resampling. that is:
the search for the best rate should first try integer multiples and
divisors of the desired rate, not the "closest" rate.

--p


-------------------------------------------------------
This SF.net email is sponsored by: Scholarships for Techies!
Can't afford IT training? All 2003 ictp students receive scholarships.
Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more.
www.ictp.com/training/sourceforge.asp

Re: alsa-lib resamping error

[Takashi Iwai]

At Wed, 22 Jan 2003 12:38:31 -0500,
Paul Davis wrote:
> 
> if you try to play a 22.05kHz stream on hardware that supports 32kHz
> and 44.1kHz but not 22.05kHz, alsa-lib's plughw layer chooses to use
> 32kHz and resamples. this seems like an error to me: it would be much
> better to choose 44.1kHz and do simple integer resampling. that is:
> the search for the best rate should first try integer multiples and
> divisors of the desired rate, not the "closest" rate.

agreed, it would be better.

but... errm...  changing this in alsa-lib is like going into a big
labyrinth for me :)


Takashi


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: alsa-lib resamping error

[Takashi Iwai]

At Thu, 23 Jan 2003 17:20:02 +0100,
I wrote:
> 
> At Wed, 22 Jan 2003 12:38:31 -0500,
> Paul Davis wrote:
> > 
> > if you try to play a 22.05kHz stream on hardware that supports 32kHz
> > and 44.1kHz but not 22.05kHz, alsa-lib's plughw layer chooses to use
> > 32kHz and resamples. this seems like an error to me: it would be much
> > better to choose 44.1kHz and do simple integer resampling. that is:
> > the search for the best rate should first try integer multiples and
> > divisors of the desired rate, not the "closest" rate.
> 
> agreed, it would be better.
> 
> but... errm...  changing this in alsa-lib is like going into a big
> labyrinth for me :)

and after a long wandering in the labyrinth, the changes are on cvs
now (both for alsa-driver and alsa-lib).
the new code will try the integer multiples at first.
i hope this works correctly...


Takashi


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com