Vorbis dev - Aug 2007 - Bounty for function to monitor an OGG-VORBIS stream

If this is the wrong place to post this, please forgive me, and PLEASE 
point me to the correct place.

I am looking to hire someone to write a function to scan a ogg-vorbis 
stream running on Debian etch and checking for two things. The first is 
static, the second is "dead-air". The function require no GUI, just
read
a configuration file for the url to scan and how many seconds of static 
to allow before flagging the static condition and how many seconds 
before flagging the "dead-air" condition. Once the condition is
flagged,
keep scanning until the condition stops unsetting the flag at that 
time.  If anyone is interested in picking up some cash for this please 
contact me off-list at this email address.

Thanks
George Baker
george@intelli-chopice.net

On Wed, Aug 22, 2007 at 08:52:33AM -0400, George Baker wrote:
> I am looking to hire someone to write a function to scan a ogg-vorbis
stream
> running on Debian etch and checking for two things. The first is static,
the
> second is "dead-air". The function require no GUI, just read a
configuration
> file for the url to scan and how many seconds of static to allow before 
> flagging the static condition and how many seconds before flagging the 
> "dead-air" condition. Once the condition is flagged, keep
scanning until the
> condition stops unsetting the flag at that time.  If anyone is interested
in
> picking up some cash for this please contact me off-list at this email
address.
You'd need to define "static". Do you mean the hiss that you get
when
an FM station is off-air? That's a tricky one. You're best turning it
into a dead air problem by enabling the receiver's squelch (mute)
circuit.

True dead air is a trivially easy problem. (Take the mean and
standard deviation of the lowest 99% of RMS samples over the given
period. If they're both nearly zero, you have dead air. Discarding
the top 1% removes any clicks and pops.)

-- 
Paul Martin <pm@nowster.org.uk>

On Tue, Aug 28, 2007 at 02:07:15PM -0400, xiphmont@xiph.org
wrote:
> On 8/24/07, Paul Martin <pm@nowster.org.uk> wrote:
> 
> > IIRC, Vorbis does encode noise (as in hiss) separately from tones
> > (fundamental, harmonics), so half the battle is won for you already.
> 
> No; vorbis does some implicit tonal estimation, but does not
> explicitly seperately handle tones/noise.  This is a fundamental
> limitation of the current generation of codecs we seek to improve in
> Ghost.
> 
> The rest of the reply is dead on; I like the insights provided about FM.
I sit corrected. :-)

-- 
Paul Martin <pm@nowster.org.uk>