[vox-tech] Mp3 Stereo to mono

[Alex Mandel]

Micah J. Cowan wrote:
> On Sat, Jun 17, 2006 at 10:24:37PM -0700, Jeff Newmiller wrote:
> 
>>Alex Mandel wrote:
>>
>>[...]
>>
>>
>>>Any suggestions on this are quite welcome, keep in mind that these audio 
>>>files are for analysis of bird songs in a scientific setting so any 
>>>alteration of the actual original audio is unacceptable as we might lose 
>>>valuable information from filters and adjustments.
>>
>>If authenticity is so important, why are you using MP3 in the first place?
> 
> 
> MP3 or Ogg, or any other compression method that is lossy and based on
> throwing out information based on how the human brain processes audio
> information, does seem /completely/ inappropriate as a storage format
> from which to draw scientific conclusions...
> 

Those are some great points, I'll have to take up with the researcher 
who is developing the audio pattern matching software. I think we 
defaulted to mp3 since it was the only format that we could find 
off-the-shelf drives for... I would have picked FLAC personally, and if 
anyone knows of a portable media device that records to line-in FLAC I 
would love to hear about it. I've suggested an embedded linux board but 
that's out for this year since the devices are in the field this week.

As it is this whole thing is kinda out of hand, with an expected 3.2TB 
of mp3s I don't even want to think about how impossible it would be to 
store 10x that and analyize it.
I've been pressing for timers so we don't have to record 24/7 but once 
again I think that will have to wait till I find a linux board so I can 
run cron jobs at specified times.

So I guess, minimize the loss would be what I'm looking for.
Lame is actually the encoder used by Audacity that's why I tried that, I 
  just wasn't convinced that it was dropping a channel, the help said 
something about adding the channels together. I guess if we're pattern 
matching and not necessarily frequency matching and have a margin of 
error around each part of the pattern we should still get the matches. 
There's bound to be natural variablity between individuals of a species 
anyways.

Thanks for the help - Alex