[Pkg-icecast-devel] Bug#516305: Bug#516305: icecast2: Likes to disconnect sources streaming silence (extremely low bit rates)

[Jonas Smedegaard]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, Feb 24, 2009 at 06:19:22PM +0200, hhaamu at gmail.com wrote:
> On 22.2.2009, at 20.38, Jonas Smedegaard wrote:
>> On Fri, Feb 20, 2009 at 04:08:15PM +0200, hhaamu at gmail.com wrote:
>>>
>>> Possible work-arounds:
>>> 1) adjust source-timeout in icecast.xml
>>>
>>> 2) use encoder in managed bit rate mode (set a minimum bit rate)
>>
>> Which of the workarounds would you consider best?
>
> Difficult to say. I used the source-timeout adjustment for my local 
> uses (source is on the same box). I adjusted it to 30 minutes 
> (obscenely high), but I tested further and I believe 50-60 seconds 
> would be enough. The drawback: a timed-out source can't reconnect 
> during the timeout period.
>
> The managed bit rate work-around... legend says that it degrades the 
> audio quality slightly, the encoder uses more cpu, and it's a waste of 
> bandwidth.

Sounds like just setting a timeout is best approach, then.


> Feel free to experiment yourself:
> # create 10 minutes of silence
> dd if=/dev/zero bs=44100 count=2400 | oggenc -r -o silence.ogg -
> # stream it
> <silence.ogg oggfwd localhost 8000 hackme /test.ogg
>
> (Also play with the managed bit rate options of oggenc)

Thanks for the test.

I must admit that currently I have no icecast running myself. I maintain 
the package (i.e. make sure it conforms to Debian packaging standards) 
but would very much like more people to help with the actual icecast 
functionality.

That's why I asked if you perhaps would want to join our little team.

I guess your not commenting on that means that you are not interested. 
Fair enough.


>> Could you perhaps
>> provide a proposed patch?
>
> I believe that's a job for the upstream.
>
> Nevertheless, I created a patch. It doesn't disconnect if there are  
> still bytes left in the queue (I have *no* idea if this counter goes to 
> zero on a regular timeout). I don't know how it affects the timeout  
> behaviour of real-world situations.

Thanks. Your help with this is much appreciated!


Kind regards,

  - Jonas

- -- 
* Jonas Smedegaard - idealist og Internet-arkitekt
* Tlf.: +45 40843136  Website: http://dr.jones.dk/

  [x] quote me freely  [ ] ask before reusing  [ ] keep private
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAkmkJIMACgkQn7DbMsAkQLiuBQCdGOpr+DdZ8EENG6cilsph5VyE
GicAnAuZW4TODHTw+dJZvevObWQkNsd0
=kShn
-----END PGP SIGNATURE-----