Re: [solved] Convert multichannel raw file to wav

On 2024-04-23 08:58, Rodolfo Medina wrote:

> Now, using sox I split each of them in four single channels

So... you've now got eight single-channel files?

> and observe,
> or hear,

When you say "observe" what do you mean?

Are you visually comparing waveforms in some GUI tool?

If you /randomly/ play one or other of each of the channels' files can 
you
tell which is which without already knowing?

(You might be able to do that with a script that picks one or other at 
random,
o, if you can't program a suitable script, make lots of 
differently-named
shortcuts to one or other file & double-click those), put them all in a
single folder away from the source files, then click lots of shortcuts
paying as little attention to their names as possible, then note which
source you thought each was, then work out afterwards if you were right
or not.)


> that in the first method those single channels are better separate:

What do you mean?  Nothing in what you've described should have mixed 
any
of the sounds of one channel with any other.  They should all be 
"equally
separate".  When listening to one of them there should be no other sound
being played that would (possibly) be muddled with it.

One cannot judge "separation" of a MONO sound.



> I can
> say so because one of my microphones is an ugly 6-dollar bad thing I 
> bought
> just for experiment and I can recognize it: it's a sensation but very 
> clear
> but difficult here to explain:

OK, it sounds poor.  We get that it won't sound as good as the same 
source
sound made directly in front of some other mic (assuming the mics are 
all
somewhat directional).



> however pacat alerts us with its message:
> 
>  Warning: failed to write channel map to file.
> 
> so the second methods is not good unless we manage to eliminate that 
> message,

Who knows what it means?  If it's just a list that says that the 
original
three or four sources are placed in the file in such-and-such an order 
(eg
source channels 1 2 3 4 are in the file in order 2 4 1 3) it should have
nothing to do with sound quality, especially after sox has extracted the
channels to separate mono files.


I think you should run the "stat" & "stats" effects against all eight of
the single-channel files.  You might find that one set of four are all
louder than the other set (which could easily make the louder ones seem
clearer, or "better separated").


-- 
Jeremy Nicoll - my opinions are my own