Not that i've tried this mind you. but the differents
appears to be where the output is going. when the
output is going to a audio device that plays music
at a fixed rate, it blocks on the write operation before
going and decoding another block of mp3 data....
whereas when you're writing to a network, you *can*
write as fast as your network will buffer it, no blocking
and all of a sudden, you've got it decodeing mp3 blocks
one right after the other, causing a nice little spike
'till the songs finished decoding. not playing mind you,
decoding into some network buffer.
this is not the desired behaviour, but i've used other
radio-streaming software that used lame as the decoder
and the re-encoder that would perform the same cpu spike,
but ONLY when it had an error connecting to the network
streaming server, and in effect, was writing out to the
network equivilent of /dev/null.
-answer
On Wed, 20 Sep 2000, Todd Goodman wrote:
> Date: Wed, 20 Sep 2000 10:38:57 -0400
> From: Todd Goodman <[EMAIL PROTECTED]>
> Reply-To: [EMAIL PROTECTED]
> To: Mark Powell <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED]
> Subject: Re: [MP3 ENCODER] Live-XMMS 1.0.0 and Lame 3.87
>
> * Mark Powell <[EMAIL PROTECTED]> [000920 10:29]:
> > I don't really understand where the problem is. A process like LAME is CPU
> > bound. i.e it does very little IO, but simply uses all the CPU it can get
> > to process data. If LAME is the only process on a box then it will use
> > around 100% of CPU. If you start up something else on that machine which
> > is also CPU bound then both will get around of 50% each. Whether you have
> > a 1MHz 8088 or a PIII 1.2GHz, LAME is still going to use ~100% of
> > CPU. With a faster CPU it will simply use 100% of CPU for a shorter time.
>
> OK. I understand that. What seems to be the problem is that when I run
> XMMS with the live-xmms plugin and it starts lame to encode it starves
> out the rest of the XMMS player (so that playback skips).
>
> However, if I run XMMS and playback and then run lame from a terminal
> window, it doesn't end up starving the XMMS player (playback is fine).
>
> > > Is this a known problem? Is it expected?
> >
> > I hope I explained what I *think* you were asking.
>
> Thank you. I think the answer is that noone here has run live-xmms and
> lame together and it's likely a problem with live-xmms or XMMS
> playback/plugin prioritization.
>
> Thanks,
>
> Todd
> --
> MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
>
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
