I have decided to revert back to Jitsi. The Odysee live stream didn’t go as well as I hoped. People couldn’t leave messages unless they had an Odysee account, and in the end I didn’t manage to save the live stream. I know the latter problem is my own fault but for the sake of simplicity I’m reverting back to Jitsi.
Thus, for the time being, the appointment should be:
Time: Monday, 5pm UTC
Place: Jitsi Meet
See you there!
Hi, I am another so far silent reader, who would really love to see the tracker interface in Ardour.
Unfortunately I can’t make to the Jitsi meetings most of the time, but since you were mentioning recordings: Are these available somewhere to watch later or did I get it wrong. From skimming through the thread I found the one video from end of January, but nothing newer. I didn’t look very thoroughly though, my apologies, but I would be glad if someone could (re)post the link to past recordings in case they do exist.
Please keep up the great work @ngeiswei !
Hi @jmaibaum,
thanks for the words of encouragement!
Unfortunately I did not record any call since Ardour Tracker Interface Demo (a call for help)
Also these days I am making a small break from the Tracker Editor as I (and Rob) are focusing our attentions to complete Zynayumi GitHub - zynayumi/zynayumi: VST/DSSI/LV2 plugin based on ayumi, the highly precise emulator of AY-8910 and YM2149 http://sovietov.com/app/ayumi/ayumi.html.. I will ping you here as soon as I refocus on Ardour.
Hello, I am pleased to announce I have finally fixed the biggest remaining feature of the Tracker Editor which is the full support of automation of plugins.
I thought it would take months, it took years. The reason it took so long, beside the fact that I have little time to work on this project, is that it required a cascade of changes up to almost the very heart of the code base. The good part is that it has resulted in a nicer code. Yet, if I had known it would take so long I might have given up thinking “after a few years, an AI will probably be able to write that code in an instant, so why bother”. Well, I don’t think an AI can write that code today, let alone in an instant, so it was good to believe that it would take a few months.
So, the Ardour Tracker Editor is getting close to what I consider “The Tracker of my Dreams”, but that isn’t to say that there is nothing left to do (and that I do not need help), quite the contrary. Here is a sample of what remains to be done:
Fix limitation of max 8 columns for notes and automation. This limitation has never been a problem for me because it is a per track limit, which I have never reached. Nonetheless it should be addressed, especially for those into the occult art of black MIDI magic.
Add shortcut keys. Except for entering notes and automation, everything else is done with the mouse, this is not the tracker way! Changing octave, step size, etc, should have shortcut keys.
Optimize! It should be as responsive as Protracker running on an Amiga, which is far from the case. Fortunately I have refactored much of the code to facilitate such optimizations.
Improve stability. Well, it is stable for me but as soon as you guys use it, you are probably going to crash it.
Implement and fix the unknown unknowns. That, as everyone knows, or perhaps doesn’t know, is the most consuming part.
I believe the code is getting close to be mergeable upstream. I need to do more testing and cleaning before that. In the meantime, if you wish to compile it you may follow the instructions below
I will keep you up to date on the progress. Now, go track some music!
I’m not interessed but it’s a nice addition ! Keep it up !
Trackers are cool. It’s invaluable to have a set of different good tools to get things done. Thanks for your great work, I saw the video and really look forward to the addition!
I will be giving this a try when I have some time.
Thank you!
So great to see you keep chipping away at this.
Keep it up!
Just wanted to provide some updates. I consider that I have reached a major milestone in the development of the Ardour Tracker Editor, and I now need to thoroughly test it before I allow myself to submit it to the Ardour team. To that end, and because it has been a major motivator for me to make music, I have resurrected the Libre Music Challenge and you are of course very welcome to participate.
After this period of testing and fixing, I intend to create a pull request with the hope to get it merged upstream. Then there will still be a bunch of work needed for polishing and optimizing.
Hey Nil,
i’m glad to hear back from you. Great job on the effort, please let us know how can we contribute - if there is a way. Thanks!
Looking forward to it.
Thanks, dspasic, in terms of help, I think testing it and reporting bugs and crashes would be super helpful.
User experience is less important as this stage because it has not been the focus so far, for instance there are almost no shortcut keys, it is slow, lacks contextual menu, it is not particularly pretty, etc. I was trying to get all the functions in there first, which I think I have basically accomplished. So now it is the testing phase of these functions.
I’ll see if I can find some time to test this out. Really interested in this.
Many thanks for the (years of) ongoing effort!
Is this tied to the version of Ardour you started to work on it at that time?
It is (re)based on the development version of Ardour (Ardour9). Don’t hesitate to ask if you have difficulties compiling it.
I might not be able to help at all , but since i’m into tracker this is a very happy “news” for me!
I’ve got some major news, good, bad and neutral.
Good
The tracker editor, although not completely finished, has reached the point of usability that I feel I can recommend it to others. I will go as far as saying that, as of now, it is the best tracker in the world! According to my own metric anyway.
Bad
I cannot offer to the Ardour team to merge the tracker editor. You will probably guess why, this last month especially it has been coded with the help of a coding assistant, and due to the Ardour team’s policy on machine generated code, they will almost surely not accept my contribution. Let me say that I find their policy very reasonable but I learned about it only recently (this week-end). On the other hand I cannot say that I regret because I now finally have the tracker of my dreams! And that very much thanks to the help of a coding assistant.
Neutral
I suppose I probably need to express a bit my view on this whole AI coding assistant debate.
First, I am not against AI in principle, I am an AI researcher myself, even though my research (on formal reasoning) is somewhat at the opposite of what we now call AI. Ever since this LLM based coding assistant technology has appeared I have been resistant to it. For a long time I dismissed it as being too stupid to be helpful, viewing LLMs as knowledgeable intellectual idiots. I still somewhat view them in this way, but let’s face it, they have reached the point of competency that makes them extremely useful if you know how to drive them intelligently and responsibly.
Second, I obviously dislike the lack of transparency of big AI service providers, the imbalance of power it creates, the shackles and all the things that made so many of us move to FOSS. Even open weight models lack transparency. The fact that the training data and process is not disclosed means that we do not know if we are complicit of things we may find unethical. We cannot exercise informed consent because we are not informed. It is not necessarily profoundly different than the lack of transparency on other goods and services the free market offers, but it is still disturbing especially given how big it is. To reduce my negative footprint I only use FOSS coding assistant software paired with open weight models, locally or remotely. I know it is not perfect but it is what it is.
Weighing the good and the bad: like every technology there is potential for good and bad. Personally my main concern it that people progressively lose their intellectual autonomy. If we stop thinking with the idea that the AI is always going to do the thinking for us, then I think (while I still can) that we are in trouble, especially the day something goes wrong. But I will admit, programming the Tracker Editor while being assisted has been an utter joy for me. First, I have a coder companion, that’s pretty cool by itself, and since I know the code very well (I still wrote 90% of it) I can bounce ideas back and forth with the coding assistant, and let it do the laborious work and then enjoy the fruit at unprecedented speed and ease.
If you disagree with the path I’m taking you are welcome to express your opinions, just please try to remain terse as I suppose it would be best if this thread remains about the Tracker Editor.
You can find the code at the usual place
It is not a hard fork and is planned to remain rebased on the main branch indefinitely, I will keep it in sync over time. The documentation is currently very partial but it will progressively get better. I’m obviously happy to get you started if you need help. You still need to build it yourself though. I could provide binaries but that would create a competition with the mother project, so I don’t know what to do about that for now. Or maybe I could provide binaries while strongly inviting users to donate to the Ardour project. I probably need to get in touch with the Ardour team.
That’s it for me. Any feedback is welcome and if you like music tracking give it a try, I hope you will love it as much as I do.
Could you share a screenshot with us ? I’m curios of how it looks.
I’ve added a screenshot in the README.md.
As you may see it does not look much different than before. It is a sober albeit hopefully functional design. I’m not a graphical designer but I welcome contributions on that front.
However it is now:
There are still missing things like:
I’ve just gone through about a third of this mega thread about Ardour and AI and based on what Paul and Robin said I think they may possibly accept my contribution.
First, I believe I can confidently claim copyright of this work, without of course having to dissimulate the use of a coding assistant, given how enormously involved I have been in the coding process.
Second, the tracker editor lives in its own library, virtually does not touch the rest of Ardour and can easily be disabled at compile time.
Third, I have already contributed to Ardour in the past, so they know I am a legit programmer.
So my plan is to test it and review it as thoroughly as I can, harden anything that can be hardened, then I’ll make a PR and we’ll see how it goes!