Subject: Re: [ecasound] picking up stattic with -b:64
From: Kai Vehmanen (email@example.com)
Date: Tue Jan 25 2000 - 17:57:04 EET
On Mon, 24 Jan 2000, Jeremy Hall wrote:
>>> ecasound -i /dev/dsp -o stdout -z:db -b:64 | msend 188.8.131.52/24430/7
>> trouble if you go under -b:128. In this case, I suspect msend is not
>> able to pipe data fast enough (meaning latency, not transfer rate).
> it does a read() on 1024 bytes, then does a write() of the same data. It
> is a very simple program, and has no disk access. It then multicasts this
It can still cause problems. Soundcard i/o consumes some of the
available time, so in practice msend has even less than 1.45ms (-b:64)
to do its stuff. And as far as I know, network cards (and TCP/IP stack
implementations) suffer from latency problems. But I'm no network-h/w
expert, so you probably know more about these...
Btw; -z:db (double-buffering) doesn't do any good in this case. It
currently only affects .wav and .raw inputs.
-- Kai Vehmanen <firstname.lastname@example.org> -------- CS, University of Turku, Finland . http://www.wakkanet.fi/ecasound/ - linux multitrack audio processing . http://www.wakkanet.fi/sculpscape/ - ambient-idm-rock-... mp3/ra/wav
This archive was generated by hypermail 2a24 : Tue Jan 25 2000 - 17:06:10 EET