jui. we got a hacked version of pd outputting to stdout as well as to dev/dsp so redirected the pcm into 'encoder', the only mp3encoder aprt from fraunhofer stuff, that reads stdin, and haad the encoder write on a named pipe. but stil, we couldnt manage to finally stream directly to the icecast server, because the streaming utilites coming with icecast, cant really cope with the pipe stuff .. but once that works out, it would be easy to encode from a named pipe written into by an 'sfread'. (and then start the encoder with an 'exec' obj on demand).
there also other problems: -for example that the encoder isn tfast enough on a p233. ;) (or: why isnt there a -speedratherthanquality option to the encoder(s)) -also, pd still depends on an existing /dev/dsp to ever output any signal, and be it 'only' through streamout or sfwrite.?
(guenther: i have still many jodels in the box, so i can easily afford one just to cheers you up ;) and of course, if its not longer than one mailscreen, its not a todolist .. fifo!
-- bzorstogipptude
17.hzV.tRL.478 discourseth:
jui. we got a hacked version of pd outputting to stdout as well as to dev/dsp
hmm..could you post a patch for this?
so redirected the pcm into 'encoder', the only mp3encoder aprt from fraunhofer stuff, that reads stdin
which encoder is that?
Thanks, :)
Eric
|-< |-<hmm..could you post a patch for this? |-<
its only one line in s_linux.c line 422 or so: int sys_dacs_write(int fd,void* buf,long bytes) { here-> write(1, buf, bytes); return write(fd, buf,bytes); }
|-<> so redirected the pcm into 'encoder', the only mp3encoder aprt from |-<> fraunhofer stuff, that reads stdin |-< |-<which encoder is that?
ftp://wopr.campus.luth.se:/pub/mpeg_layer_3
but apparently lame seems to be pretty cool, although i havent got it run yet. especially you can turn off the psychoacoustic model, so its gaining in speed, which is cool for liveencing ..
gluk.so