[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: how to record sound to mp3 [wav, for those who can]



On Thu 25 Mar 2021 at 23:22:35 (+0100), Nicolas George wrote:
> David Wright (12021-03-25):
> > > > $ arecord -d 10 -f cd -v -v -v -D plughw:0,0 /tmp/audiofile.wav
> > > This command does not record the sound being played.
> > … on your machine.
> 
> On no machine, unless specifically configured, which is not trivial at
> all.

I'm afraid it's my PCs that are making a liar of you.

> It would be helpful if people around here learned to read carefully the
> questions before trying to answer them. If they did, they would have
> noticed that the question was not to record the ambient sound but the
> sound BEING PLAYED.

What, you think that I left the speakers running so that the
microphone could record them?

Or did you think that the sound of Thursday evening traffic in
St John's Wood would carry across six timezones?

Or you think that's it's impossible that I should have been able to
record sound the computer is playing (from whatever source, external
or internal) on an OOTB PC for over twenty years?

Is that why you're shouting?

Anyway, back to talking about PCs.

> To achieve it requires either a hardware connection
> between the output and the input of the sound controller

That's my understanding. Using the terminology of the High Definition
Audio Specification (Revision 1.0a June 17, 2010), there is presumably
a link from the § 7.2.3.4 Mixer (Summing Amp) Widget output to the
§ 7.2.3.5 Selector (Multiplexer) Widget. This link in inside the
Widget Interconnection "Cloud" of Figure 49, Module-Based Codec
Architecture.

"The exact number of possible inputs to each widget is determined by
design;" (§ 7.1.1), which is why you can't just conjure up any facility
on any PC.

But the machine is no more "specifically" configured than any PC which
has HDA and a mobo: the vendor (Intel on my old one, Dell on the new)
decides how much of the architecture they will implement. And the
modern way seems to be to go cheap, particularly with consumer-grade.
OK, this one's a decent machine, but it's still a 10-yr old cast off.

Perhaps take a look at the specification and see how much is left open
to the vendor. Hence the need for scripts like alsa-info to tell you
exactly what you've bought with any given "sound card".

This PC was not cheap when it was bought, largely because it's
supposed to be fast: it was bought for students to run geophysics
programs on. You can now pick them up for just over $100. If I ever
have to hand it back, I might just do that. It'll be the first
computer I've ever bought.

> or the
> collaboration of the sound driver.

I assume by this that you're talking about pulseaudio. That's why
I've mentioned it each and every time. (This is the third—should
I put it in my signature?) I can't advise how the OP might use it,
because *I* don't¹. But perhaps that's not expected here—one just
replies   "pulseaudio"   like the people saying   "audacity"
or   "sox"   or whatever.

Finally, who's the audio expert round here? I posted what I think
is a determining factor for just vanilla ALSA and the card to work
with my command line. Presumably there's a definitive answer to
this? Do *you* have it?

¹ self-imposed simplicity: no PA, no OSS.

Cheers,
David.


Reply to: