newsAdvanced Gtk+ Sequencer - News


GSequencer development release 1.0

Item posted by Joël Krähemann <jkraehemann> on Sat 01 Jul 2017 02:49:39 PM UTC.

My personal opinion is that current state of usability is given for Advanced Gtk+ Sequencer, as version 0.8.8. I am happy about Standard MIDI import/export functions and the new editing tools to move or crop notes.

The 0.8.x branch will be stalled. Only important changes are going to be applied.

Extending existing test-cases

The hunt for bugs just started. To extend the test-suit is mandatory to ensure future compatibility and reliability. The works is going to be done on 1.0 branch. And is going to be back-ported as needed.

The unit test coverage goal

Like every project that thinks about quality wants 95 % unit test coverage. There is where I would like to go. But still no Test Driven Development. Recently the got reorganized and the unit tests got its very own automake file. This makes the huge amount of coming unit tests better findable.

Functional testing of the automation editor

It is a fact that the automation editor has no functional tests. This is going to be changed. The functional test shall cover all aspects. Add/remove machines, enable/disable ports as well editing automation.

Audio raw-data and input support

Yay, one more DAW. With all whistles and bells.


This is one of my favored coming features. It could be used as scripting engine and remote control. Some prototypes are already available on 1.0 branch.

Additional computing power

The promise to support even more threads to compute the audio-data tree. At my opinion every one should have to facility to use High Performance Computing software. But I have real concerns that nowadays CPUs will just fail to give the needed performance.

Since gsequencer uses synchronized threads to compute in parallel the audio tree. There is some overhead but I feel like to gain the 1000 tics per second. Currently the thread maximum precision JIFFIE was throttled to 250.

The need for optimization and case study

Yes, it was a pain to implement the synchronized tree. The solution asked to be fearless. And after some kernel panics and some disturbed email complaints. Everything seems to work.

Anyhow, I have no idea why. But if they moved the problems to an upper limit. I just will figure it out.

by Joël

Back to the top

Powered by Savane 3.13-3cd0.
Corresponding source code