Re: tone detection

On 2021-05-29 07:17, Robert Jeffares wrote:
> I have an audio file which is recorded from a stream which contains
> time 'pips' I wish to use as a cue point.
> 
> The head can be anything from a few seconds to 2 minutes depending on
> the transmission delay time.
> 
> I can detect silence in the file which I can use to trim the audio but
> the silence based script fails when the head audio is longer then 30
> seconds or so.
> 
> Using the 1kHz tones as a marker will give me a more accurate edit 
> point.
> 
> Is there something in the Sox arsenal which can detect this tone?

I don't know.

A while back I wanted to find out if the overall unequal channel balance
on a recording (of a choir singing) was consistent all the way through.
I suspected that it would be; the Sops & Tenors were tending to sing
towards the mics whereas Altos+Basses somewhat across them.  (Microphone
placement was far from ideal).

Of course sox's STATS effect could tell me levels for the whole 
recording
or, after I'd divided it up into different pieces of music, for each 
one.


But I wanted a closer look, so wrote a script to hack a track file into
an arbitrary number of short slices of sound and run STATS on each one,
then tabulate the data.

I think you could possibly do something similar; chop up a copy of such
a file into lots of pieces each eg half a second long (assuming that the
pips last for more than half a second).  Then run one sox STATS against
each slice and extract from its output the min/max level information.
It should be possible subsequently to decide reasonably accurately where
each pip is, and also where the proper content starts.

It might be you'd need to use many more shorter files, but I'm sure it
would be possible, provided you can program well enough to write code
to do it.


My script was written in a programming language named ooREXX, to run
on Windows.  It probably wouldn't be a great deal of help to you, but
if you want it you can have a copy.

-- 
Jeremy Nicoll - my opinions are my own