"duration" is an integer, should be double

[Chloé AVRILLON]

Summary :
The Mixxx control value "duration" gives a rounded integer value

We need :
the actual "duration" value as a real number (400 milliseconds difference is 1 beat !)

"duration should be equal to =track_samples/track_sample_rate/2 (real value) "/2" because stereo

[Daniel Schürmann (daschuer)]

As a workaround, you can use
"track_samples" and "track_samplerate"

[Daniel Schürmann (daschuer)]

If you like, you can fix it your self.
https://github.com/mixxxdj/mixxx/blob/0774d77a3bf03a520a207e4b3c9be1dd83e221be/src/trackinfoobject.h#L74
you can make it a double, all ControlObjects are double anyway.
But you need to track where this value is used and if the fractions of a second didn’t do bad things.

Our general bugfix workflow is described here:
http://www.mixxx.org/wiki/doku.php/bugfix_workflow

[Chloé AVRILLON (chloe-avrillon)]

hum, something clear is missing here daniel :
http://www.mixxx.org/wiki/doku.php/bugfix_workflow

are you sure it is the good link ?

[Chloé AVRILLON (chloe-avrillon)]

Majors problems for me :
1) a bit lost in the Mixx code
2) I did not program in C since ages. I am a Delphi developper girl
3) I am not used to gitub

Hey, I have some limits :/

Anyway, I will try to understand everything, but the bugfix workflow page is gone somewhere in the Twilight Zone ! :D

[Chloé AVRILLON (chloe-avrillon)]

hannnnnn.... I understood something reading the source code !!!!
The duration given is the duration given by the tags of the file , wich is an integer.

Soooo it's better to leave as it is.

But what is interesting and not dociumented is that we have a property available : "Channels", that provide us the info to know if our track is Stereo or not.

Therefore we can return a real value for the duration like that :
real_duration==track_samples/track_sample_rate/channels

[Daniel Schürmann (daschuer)]

The page was hijacked, now the original content is back.

> Majors problems for me ....

If you still have fun to take these challenges, go ahead, I will try to give you a help if yo ask.

[Chloé AVRILLON (chloe-avrillon)]

Thank you Daniel, but they are not priorities since they can be get around without problem.
In my list of bug I have one very serious, and 3 that are a bit annoying :

The very serious one is this one (and I think it is urgent and critical since it can resolve a bunch of issues that might be related to it) :
https://bugs.launchpad.net/mixxx/+bug/1499738

The two annoying :
https://bugs.launchpad.net/mixxx/+bug/1498868 (those two permitted me to discover the serious one)
https://bugs.launchpad.net/mixxx/+bug/1498960

https://bugs.launchpad.net/mixxx/+bug/1498134 (this one is diferent, but can get around)

[Daniel Schürmann (daschuer)]

After the track is loaded, to the player, TrackInfo::m_duration is updated with the real value, which is sample precise. The "duration" ControlObject is set from this real value. A ControlObject is always of type double, so it might make sense to use double precision here. But as said, we need to make sure not to introduce any regression.
And if you have an other workaround ....

[Chloé AVRILLON (chloe-avrillon)]

Yes, the workaround is that in midi scripting we have accesss to the track_samples and track_sample_rate value

We could expose the Channels value without any drawback for everything else, so that we could get in MIDI scripting our value like that :

___________

MyController.getduration = function(group) {
       var track_samples = engine.getValue(group,"track_samples");
       var track_sample_rate = engine.getValue(group,"track_samplerate");
       var track_channels = engine.getValue(group,"channels");
 return track_samples/track_sample_rate/channels;
};

___________

[Daniel Schürmann (daschuer)]

Yes, this will work for me as well. We should call it "track_channels", since a deck has currently always two channels.
And even without the channel value, you should be able to guess the value of channels.

[Chloé AVRILLON (chloe-avrillon)]

"And even without the channel value, you should be able to guess the value of channels."

Ho yes we can guess the value of channel, 99% of the time it will be Stereo.
But for the 1% left, the Midi script is not able to guess that the provided track is mono.

For mp3, I don't know if Mono audio exist (I think not), but for PCM wav file, yes, for sure.

[Daniel Schürmann (daschuer)]

If "duration" is 3:30 and "track_samples / track_sample_rate / 2" is 1:15:3 we have a mono track. :-)

[Chloé AVRILLON (chloe-avrillon)]

This is good! Nice trick!

[Daniel Schürmann (daschuer)]

Taglib 1.10 supports finer duration values. We should update Mixxx accordingly.

[Uwe Klotz (uklotzde-deactivatedaccount)]

https://github.com/mixxxdj/mixxx/pull/970

[Swiftb0y (swiftb0y)]

Mixxx now uses GitHub for bug tracking. This bug has been migrated to:
https://github.com/mixxxdj/mixxx/issues/8229