Full replay gain support for track metadata, file tags and library

[uklotzde]

Sorry, I couldn't resist ;) Another PR for master.

This PR completes the replay gain support for track metadata (including file tags) and the library (database) by properly reading/writing the peak amplitude of a track.

The peak amplitude is currently not computed by the RG1 analyser, but could be added easily. The RG2 analyser seems to perform this analysis on the fly.

[daschuer]

An additional hint that strPeak = "1.0" represents the digital full scale could be useful.

[uklotzde]

I also thought about adding a static assert, because we implicitly assume
here that CSAMPLE_PEAK = 1.0

On 11/03/2015 08:15 AM, Daniel Schürmann wrote:

In src/util/replaygain.h
#766 (comment):

@@ -56,6 +69,18 @@ class ReplayGain {
m_peak = CSAMPLE_PEAK;
}

• // Parse and format the peak value metadata according to the ReplayGain
• // 1.0/2.0 specification.
• // http://wiki.hydrogenaud.io/index.php?title=ReplayGain_1.0_specification
• // http://wiki.hydrogenaud.io/index.php?title=ReplayGain_2.0_specification
• static CSAMPLE parsePeak(QString strPeak, bool* pValid = 0);

An additional hint that strPeak = "1.0" represents the digital full scale
could be useful.

—
Reply to this email directly or view it on GitHub
https://github.com/mixxxdj/mixxx/pull/766/files#r43721107.

[daschuer]

Thank you for the PR. I have only commented some naming nits.

[uklotzde]

Added column (read-only) for replay gain in table views.

[daschuer]

This PR is growing. When will it be ready for merge?
I have already tested and reviewed it until 10b31ed and I can say LGTM up to that.
So if you mind, you can do separate PR to merge this now.

I will now read though the remaining commits.

[uklotzde]

I don't plan any further additions.

The GUI extensions are just the icing on the cake that I need to test your
RG2 / EBU R 128 stuff ;)

On 11/06/2015 10:43 PM, Daniel Schürmann wrote:

This PR is growing. When will it be ready for merge?
I have already tested and reviewed it until 10b31ed
10b31ed
and I can say LGTM up to that.
So if you mind, you can do separate PR to merge this now.

I will now read though the remaining commits.

—
Reply to this email directly or view it on GitHub
#766 (comment).

[daschuer]

I don't know why, but now there is something broken.
The track info dialog is empty for every track.
The history does not track files.
This also happens in current master... strange.

[uklotzde]

I've tested it only with a fresh Mixxx library/config. I will test again
with an existing Mixxx library/config.

On 11/06/2015 11:25 PM, Daniel Schürmann wrote:

I don't know why, but now there is something broken.
The track info dialog is empty for every track.
The history does not track files.
This also happens in current master... strange.

—
Reply to this email directly or view it on GitHub
#766 (comment).

[uklotzde]

My tests with both an empty and an existing Mixxx config were successful.
Also the history view works as expected.

Did you try a clean rebuild? Sometimes SCons does not reliably switch
between different branches.

[daschuer]

Strange! It seems I had messed up my working directory. Now the issue is like that:
Plain recent master: no issue
Merged with this branch scons -c and rm -R lin64_build: no track info content.
I will now debug why.

[daschuer]

My error: [ no such column: replaygain_peak ]

The library predict that is was already at 25, but that was wrong. After changing current the schema version back to 24, it works as desired.

LGTM

[uklotzde]

Great! Thank you for reviewing and merging, Daniel.

You can now rebase your PR "Replaygain 2.0" on master for a fresh start. Since no one has actually reviewed the code this is safer and clearer than merging.