[MPlayer-dev-eng] real mmap-support in ao_alsa

[Arpi]

Hi,

> what do you mean with float?? i guess int to float conversion has to
> be done earlier (in some decoder?). so far i get data only in plain
> ints, i see no reason to do int -> float in ao_alsa. what would be the
> improvements to do so?

i think he meant something like AFMT_FLOAT, and if there is an ao supporting
that, he could add float output support to filters/codecs

> i actually started an ao_jack, i did already a kind of 'sounding'
what is 'jack'  ?

> prototype :) but sound is horrible till now. since jackd recommends to 
> have data in float i think i have to make a conversion here. also
> buffering is a problem, cause jack has a kind of poll-mechanism and
> mplayer uses push&write method i started to implement another
> buffer. im stucking now on this (double)-buffer thing. its somehow
> also against jacks-philosophy, but it's the only way to get jack
> working with push&write-apps.
see ao_sdl, it uses such buffer thing (sdl is also polling stuff using
callback)

> > I would appreciate such a switch, make it generic cause It would be an
> > interesting feature for OSS as well. I want to be able to mix the data
> > stream with other types of input in a future karaoke mode, and that
> > requires short latency. It is also interesting for a possible VJ
> > option.
> 
> i don't know so much about OSS, sorry, and oss_plugin is arpi's domain
> ;). what i saw in ao_oss is that it trys to figure out the
you can set fragment size and number of fragment, so you can set up very low
delay stream too (down to 128 byte buffer size) - but mplayer - due to
push&play philosophy - doesn't like <32k buffers.

> max-available-buffersize and simply sets the max_buffersize then. -abs has no
> effect, but maybe i'm wrong?
-abs is to override detection of buffersize, but it doesn't override the
actual buffersize!

A'rpi / Astral & ESP-team

--
Developer of MPlayer, the Movie Player for Linux - http://www.MPlayerHQ.hu