@dfkettle said:

So I tried a three-way comparison. I played the file back in Audacity, my patch and WIndows Media Player. None of them matched...

It seems, if you want them to match, that using a different operating system may be the only solution.

I know, that's a terrible answer... but "terrible" is also a kind word to describe the utter disaster that is professional audio in Windows.

How about this: Record your Pd patch to a disk file, and measure the number of samples between soundfile attacks (or phasor jumps). This will either match the sample rate (assuming the 1 Hz cycle) or it won't.

If it does match (give or take one sample, due to fp rounding), then you know definitively that Pd is producing a correct sequence of samples and the situation in your patch cannot be improved. At that point, it's irrelevant that Audacity and WMP are wrong -- evaluating the worth of a correct result, against other software's incorrect behavior, is not useful.

If it doesn't match, then it would be a bug report. But I'm quite confident that won't be the case.

hjh

@whale-av said:

@dfkettle That's unfortunate.
Maybe try Asio4All...... https://www.asio4all.org/....... as a universal driver for Pd and Audacity.
It will have a settings window...... individual for each application..... so try to keep them as identical as possible.
It will give you lower latency and allow you to use multiple soundcards so it can be useful anyway.
David.

I don't think that's an option for Audacity, at least not under Windows. It uses PortAudio for all three "hosts" (MME, Windows DirectSound or WASAPI), as you can see in the screen capture above. I have to confess that I don't really understand the difference between audio "host" and "driver", though, as they're called in Audacity. Maybe there's a way to tell Audacity to use ASIO4All, but I don't know how. I already have ASIO4All installed, I use it with other programs.

@dfkettle Sorry...... I see Audacity has to be recompiled to use ASIO.
That's unfortunate too...
David.

possibly related? idk: https://github.com/pure-data/pure-data/issues/1998