Output Stereo

[ytsrt66589]

Thanks for your amazing package!

I'm currently trying to use this package to align my multi-tracks audio with its mixture. However, I face the problem that the output file for mixture is always mono. Is that possible to keep the mixture be stereo?

Thanks a lot.

[iver56]

Here's a workaround: If you install fast-align-audio and get the offset (which denotes the amount of samples the delayed_signal is delayed compared to the reference signal) from audalign, you can align your mixture (and keep it in stereo) like this:

aligned_audio, gaps = fast_align_audio.align_delayed_signal_with_reference(
    reference, delayed, offset=offset
)

[benfmiller]

Thank you @ytsrt66589 for raising this issue, and thank you @iver56 for the workaround!

This can also be done through the _write_shifted_files method https://github.com/benfmiller/audalign/blob/main/audalign/__init__.py#L624 .

I think this functionality should also be present to the align methods, so I'll add support for the unprocessed flag there

[mmathis76]

Hi Ben,

Thank you for Audalign which seems to be a great help in aligning tracks recorded with multiple recorders (thus having to deal with alignment and clock drift issues).

I have the same issue with my stereo input files which are aligned correctly but also down-mixed to mono in the process.

Let's assume this is my script. How would I have to use _write_shifted_files to preserve stereo output?

import audalign as ad

fingerprint_rec = ad.FingerprintRecognizer()

fingerprint_rec.config.set_accuracy(3)

results = ad.align("target/folder/", recognizer=fingerprint_rec)

results = ad.align_files(
    "bak/sample1.wav",
    "bak/sample2.wav",
    destination_path="destination/",
    recognizer=fingerprint_rec
)

Thanks in advance for your help. This looks like a fantastic tool for live archivists and field recordists.