Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix: don't export loudness info to Engine #11979

Merged
merged 1 commit into from
Sep 15, 2023

Conversation

mr-smidge
Copy link
Contributor

Exporting the ReplayGain ratio to use as track loudness info on Engine exports could result in scalars that can cause rendering errors on hardware players.

This PR changes the export process to play it safe and not provide any track loudness info. Basic experimental testing shows that this does not cause any waveform-rendering regressions, but does fix tracks where the waveform was not previously displayed correctly. Furthermore, this change does not affect the outputted audio.

@mr-smidge mr-smidge marked this pull request as ready for review September 13, 2023 15:05
// However, getting it wrong and accidentally scaling a waveform beyond a sensible maximum
// can result in no waveform being shown at all. In order to be safe, no loudness information
// is exported, resulting in waveforms being displayed as-is.
snapshot.average_loudness = 0;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Judging form t he variable name it needs to be_

Suggested change
snapshot.average_loudness = 0;
snapshot.average_loudness = 1/pTrack->getReplayGain().getRatio();

Because the ReplayGain value is used as extra gain to have all tracks with the same loudness.
This means for instance normalized track has an average loudness of -18 LUFS (on a dB scale)
The replay Gain will be 0 for this track.
If the same track has a doubled sample voltage it has an average loudness of -12 LUFS, and the replay gain will be 0.5.
Applied this, we are back to the desired -18 LUFS.

Then the question is: What is the scale of average_loudness. Is it a ratio or a dB value?
But does it mean compared to the ReplayGain2 LUFS?

Do you have a set of tracks where the average_loudness has been set before we have messed it up with Mixxx?
Maybe we can spot a rule how it compares the the Mixxx Replay Gain column.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@daschuer This is indeed the purpose of upstream issue xsco/libdjinterop#103 - the exact meaning of average_loudness isn't currently clear.

In the past, I had made an assumption that ReplayGain ratio was a good fit, but Denon players use the value to scale the waveforms, and the bug being fixed here is that some bad values cause the waveform to be scaled out of range, at which point it is no longer rendered.

I've been working through more methodically with an analysis suite of tracks to try and determine the exact meaning of the fields in Engine format. When we have confidence in the true meaning of the fields, we can come back to Mixxx and set the value properly.

Copy link
Member

@Swiftb0y Swiftb0y left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@whanake-music confirmed this to be solving the issue. So lets merge and postpone further experiments to another PR.

@daschuer
Copy link
Member

OK.

@daschuer daschuer merged commit 69c90a9 into mixxxdj:2.4 Sep 15, 2023
22 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants