Re: [solved] Re: Live recording
Jeremy Nicoll <firstname.lastname@example.org> writes:
> On Thu, 17 Aug 2017, at 08:57, Rodolfo Medina wrote:
>> Jeremy Nicoll <email@example.com> writes:
>> > You do realise that merging files, adjusting balance etc are all possible
>> > with sox?
>> > One reason I do that sort of thing with sox is that by keeping note of the
>> > commands I used to do each stage, I automatically document how I
>> > manipulated a set of audio files. Documenting anything done via a GUI as
>> > in Audacity is a great deal more difficult.
>> > It's worth documenting what you do so that if necessary you can exactly
>> > repeat the process at a later date.
>> Thanks... In fact, what you suggest is exactly what I wanted to do but
>> didn't manage to... In fact, suppose I have two files: piano.wav and
>> voice.wav, created by sox in recording piano and voice respectively,
>> simultaneously. piano.wav is a stereo file, with the two channels, left and
>> right, inside it; whereas voice.wav is a monophonic file. When I put them
>> together to create the final, say, result.wav, I must properly allign them.
>> Now, Audacity makes this job easy thanks to the graphical possibility of
>> seeing the waving forms and magnifying them. Instead, with sox, I tried to
>> use the `delay' option with no success. Maybe can you suggest a better and
>> proper way to do that...?
> I've never tried that. The sox mail list is a good place to ask for sox
> One thing I'd suggest (since sox is potentially complicated) is that instead
> of trying to use sox with one command that does everything, you try to do it
> in stages. That way you can check that what you get at the end of each stage
> is what you expect. Sox has effects that will tell you about what is in a
> I think first I'd check simply that I could create a file containing all
> three channels, eg by:
> sox -V2 --no-clobber -M piano.wav voice.wav threech.wav
> The -V2 sets a fairly high level of verbosity; you might need -V3
> The --no-clobber prevents sox from overwriting a file. It's easy to get a
> command wrong and accidentally destroy an original file; personally I work on
> copies of files AND use --no-clobber.
> -M tells sox to merge the files. You should end up with a 3 channel file,
> with the piano audio on channels 1 and 2 and voice on 3.
> If that works (and ironically, Audacity might be an easy way to make sure) I
> think I'd next try to make a file with the voice channel offset from the
> others, perhaps by (guessing again!):
> sox -V2 --no-clobber threech.wav delay 0 0 2t voicedelayed.wav
> For an experiment, it's worth making the delay substantial, for example the
> two second value (2t) here. It'll be easy to tell if that worked. Later you
> can try different values to get the audio aligned where you need it to be.
> (Also note that there's an effect named 'silence' that can be used to chop
> unwanted 'silent' or at least quieter than the wanted sound audio out of a
> track, so a different approach that might help would be to make sure that
> none of the tracks have much silence at the start, before you merge them.)
> Note that so far, none of the three channels have had their relative volumes
> To do that, and create a stereo file from the three, you'd need to use the
> remix effect.
$ sox -M piano.wav voice.wav voicedelayed.wav delay 0.3 0.3 0
to delay piano of 0.3 seconds (`delay' at the end of the command line), and
$ sox voicedelayed.wav final_result.wav remix -m 1p-5,3p-5 2p+5,3p-5
to decrease the volume of the left channel of piano of 5db, increase the volume
of the right channel of piano of 5db and decrease voice's volume of 5db. It
seems to be working fine. Thanks. Then I'll be using sox instead of Audacity
to manipulate recorded tracks...
>> One more thing: I remember, Jeremy, your suggestion of purchasing a
>> multi-channel audio interface, to be preferred to a mixer, and will do in
>> future. But do you think that the solution I'm adopting for now: different
>> files on different single-channel audio cards and then merging them - dou
>> you think this is a good solution as well...? What difference between this
>> cheap one and more or less expensive multi channel interface...?
> Your current solution is the cheapest I can think of, and no doubt you'll
> learn quite a lot doing this.
The reason why I'm starting with that cheap solution is not only its cheapness
but that I first want to learn with it, and also because it suits my basic
needs for now.