I'm wondering if anyone has tried porting the Ogg Vorbis encoder to
ARM. Has anyone been successful? I understand that the current
implementation uses floating point but I was hoping to see if the
encoder would still be fast enough with fp software emulation. I
guess 1/2 real-time will be sufficently fast on an Xscale 400Mhz.
Any sugguestions will be appreciated.
I suggest that you rewrite the code into Integer mode, an
integer version of mp3 will run as cheaply as 40 Mhz at full
rate.
Will
The "Tremor" decoder library, an integer-only, fully Ogg Vorbis
compliant software decoder library is now available under a totally free
BSD-like free software license. Download complete source code as a a
tar.gz format archive, or check out module 'Tremor' from Xiph.Org CVS.
Klaus
Does it encode as well?
--
Grant Edwards grante Yow! Do you have exactly
at what I want in a plaid
visi.com poindexter bar bat??
Jimmy
"William Munns" <will_...@iname.com> 撰寫於郵件新聞
:b2fug2$1c07oi$1...@ID-177444.news.dfncis.de...
[Tremor]
> Does it encode as well?
Not last time I looked.
--
__ _______ ______ __
/ |/ / __/ ___/ /_/ / # Dan Maloney.
/ /|_/ / _// /__/ __ / # Disclaimer: Not my fault.
/_/ /_/___/\___/_/ /_/ # mailto:me...@toth.org.uk
> I'm wondering if anyone has tried porting the Ogg Vorbis encoder to
> ARM.
Yes.
> Has anyone been successful?
A bit. Sometimes it works and gave a working ogg, sometimes it
just crashes.
> I understand that the current implementation uses floating point
> but I was hoping to see if the encoder would still be fast enough
> with fp software emulation. I guess 1/2 real-time will be
> sufficently fast on an Xscale 400Mhz.
Even with a really really fast soft float library, I think
that would be pushing it.
> Any sugguestions will be appreciated.
Try it and see?
I only suceeded in compiling the encoder without libfloat. It takes 4
mins to encode a 10s 16bit 16kHz audio file on an Xscale PXA 200MHz.
Would anyone know what performance increase libfloat will have? The
oggenc binary just seg faults on me when compiled with -msoft-float.