Bug in Autodj Transition

[ikr83]

Bug Description

Hello,

Since the update to version 2.3.5, there has been a regression in the use of the autodj function. Instead of aligning the transition of the "intro start marker" of track 2 with the "outro START marker" of track 1, Autodj sometimes aligns the transition of the "intro start marker" of track 2 with the "outro END marker" of track 1. The bug seems to occur more frequently on one of the two main decks than the other. It appears that the bug is related to certain "intro start markers." On the affected tracks, deleting the marker and repositioning it slightly seems to resolve the issue. However, without knowing which tracks are affected, it's difficult to fix everything.

In the attached video, there are two transitions (from the same track), and the bug occurs on the first transition but not on the second one.

Example.mp4

Thanx!

:)

Version

2.3.5

OS

Window 11

[daschuer]

The "Full Intro + Outro mode" aligns either Outro End with the Intro End or Outro Start with Intro Start in a way that both is fully played:

        // In the diagrams below,
        // - is part of a track outside the outro/intro,
        // o is part of the outro
        // i is part of the intro
        // | marks the boundaries of the transition
        //
        // When outro > intro:
        // ------ooo|ooo|
        //          |iii|------
        //
        // When outro < intro:
        // ------|ooo|
        //       |iii|iii-----

If no intro is set, it should use the outro time. It looks like Mixxx assumes a zero lenght intro in the first case which causes the crossfader jump. I will investigate how this can happen.

[daschuer]

I can confirm the issue. It happens when you seek the next track before the intro start. This can be a micro seek but this is obviously a bug. Introduced here: 7c039ad

[ikr83]

Thank you! I have installed Mixxx using the downloadable installer from the website. How do I proceed to correct Mixxx with the autodjprocessor.cpp file that you modified? Should I wait for the next update? Or should I reinstall Mixxx from the source code with the updated autodjprocessor.cpp file located at src/library/autodj/autodjprocessor.cpp?

[daschuer]

I have just added some improvement that limit the intro length when it is in the middle of the track an you have re-cued the track at the beginning. Please check if it feels natural.
You can install the GitHub artefact that will appear in a few minutes at the bottom of:
https://github.com/mixxxdj/mixxx/actions/runs/5175300378?pr=11629

If you are satisfied (or not) please comment at the Pull Request. Than it wil lol be merged and becomes part of our next release version.

---

I am happy to hear that you are able to build from the source code. Do you have interest to join us and contribute new features or bug-fixes? A fist thing could be to check the first time build experience. How does it work for you and what can be improved.

[daschuer]

Now it's here:
https://github.com/mixxxdj/mixxx/actions/runs/5175462455?pr=11629

[ikr83]

:) Two points now

First point, I don't know how to use ...

https://github.com/mixxxdj/mixxx/actions/runs/5175462455?pr=11629

... but I will !

Second point
I haven't been able to compile successfully yet. I'm on Windows. I've installed Visual Studio Community 2019 with C++ and CMake, and I ran 'tools\windows_buildenv.bat'. However, when I run CMake, I encounter some errors... (missing SScache + missing pthread.h + Performing Test HAVE_GNU_POSIX_REGEX -- failed to compile + Chromaprint).

CMakeOutput.log
CMakeError.log

[daschuer]

The windows installer is here: https://github.com/mixxxdj/mixxx/actions/runs/5175462456

[ikr83]

It's perfect !!!

[ikr83]

Thanx!

[daschuer]

We close issues when the fix has been merged: #11629

[daschuer]

Merged