Polyglot tracker module data decrunching, processing and crunching
I was sorting out a bunch of old digital files and folders and was quite amused at the realisation that it seems like I have written some sort of tracker module converter or reader in pretty much every programming language I've ever used.
NOTE: a tracker is a music creation program, and a module is how their files are called---they're tightly optimised compressed binary files and use all sorts of tricks to save data.
Other people write To-Do apps to learn new languages, and I guess I write tracker module conversion tools? 🙃
So, in order of creation, as far as I can remember:
2005: C++: IT to events (in XML)
I wrote a utility in C++ to pre-process and parse pattern data from Impulse Tracker's IT files, which would then be output as an XML file with event data, and then the C++ executable would use tinyxml to load the file. I am not entirely sure of why I didn't just incorporate this event processing and reading directly into the executable instead of doing that at build time—I think I either didn't have enough time to finish the exporter and I ended up generating the important events with Audacity or something like that, or my parser was very slow and I didn't want the demo to take ages to load.I can't find this code, but I remember it was very buggy and I wasn't progressing very fast. I wrote about it back then! There was a lot of confusion and non-written knowledge around the tracking format that I finally understood asking other people for help (e.g. how to calculate how much time do each row and each 'tick' take). Also, lots of segmentation faults and time spent thinking about memory and string allocations! And I was using the wrong tool for the job.
GitHub and Google Code didn't even exist back then and SourceForge was terrible, so this is probably collecting digital dust in some folder.
2007: Ruby: XM to events
Apparently I also wrote a Fast Tracker's XM to timed event utility in Ruby because I was going to work with a musician that used XM (hi JosSs!) and I was so into Ruby as server side and general utility language back then—the demo was going to be built in Flash with probably ActionScript 2.I actually didn't finish the demo, therefore the exporter is quite incomplete and I haven't bothered publishing it, but it was interesting to use Ruby to process binary data from a file in a sort of stream like way.
In absence of a demo, here's my favourite song from JosSs:
https://soundcloud.com/josss-1/josss-pixeltropia
2008: PHP: IT to events (in a .h file)
Another iteration of this event list idea was to parse the data with PHP and generate a song.h file containing a big data array to be used in 64k intros, like this one.I can't find the code, but if I remember correctly, the nicest aspect of building this was that I could use some sort of text template to generate the .h file, and since string manipulations are way easier with PHP than with C, I got to the point where I wanted to be really quickly and without tons of segfaults.
2009: PHP: XRNS to MOD
I found another twist on my data conversion efforts: xrns2mod takes a Renoise song and converts it into an Amiga MOD file.This is not a trivial conversion, as Renoise files are XML based and the player is significantly evolved compared to the MOD format: it can have a lot of channels, and more than one track per channel, plus arbitrary length patterns, plus a lot of instruments, and more than one sample per instrument, etc, whereas the most basic MOD files only have 4 tracks, use a limited note range, different base frequencies, less sample effects, and lesser quality samples, etc. So it's not just moving numbers from one place to another.
I think I wanted to enter some MOD contest but I found all the MOD trackers quite uncomfortable compared to Renoise, and I wanted to have the best of both worlds: ease of use and also entering the contest. It's incomplete and quite hardcoded in places, but it was already converting files quite decently the last time I ran it, so I put it on GitHub, just in case it maybe helps someone. Maybe.
Of course, minutes after I published it I found XRNS2XMOD, a similar tool which seems more complete as it also exports to XM in addition of MOD, but it requires Mono and what not.
~2010: Python + Lua: XRNS to events (in a .lua file)
I wrote an exporter in Python for generating a list of timed XRNS events and use them in a demo forThis was used in this demo -- as you can see the visuals are tightly related to the music:
I later tried to port this demo to WebGL, but I never finished it... perhaps I should just share the code for whoever was interested. It was a bit challenging to migrate because I was using OpenGL immediate mode in the Luisita version, so I needed to change my mental paradigms to WebGL/three.js's and also write a shader or two to make it look as I wanted it to look... also those were a ton of events that I wanted to send to JavaScript, so my naive initial attempts were making the Garbage Collector work a lot more than it should have worked.
Funnily enough I remember being frustrated with not having tweening in the Lua version, but it could be 'easy' to add it to the web version using tween.js. But it's not going to happen.
2013: node.js: XRNS to JSON
I also wrote a node.js module for extracting XRNS data into something the browser could use more easily (i.e. JSON). I guess not many people are using it or are interested in it, because 3 years later I haven't heard of anyone improving or using it, but if you want you can install it with npm:
npm install renoise
I used this to generate the events and note data for the slides for my JSConf.EU 2013 talk in which I programmed virtual instruments to play a tracked song in the browser, built their interfaces with Web Components and also used hardware controlled/read with Open Sound Control and node.js! Yay to mixing everything together! 8-)
Here's the video starting with the actual demo:
https://youtu.be/5s7sJBQd6es?t=19m19s
Well, I hope you have enjoyed this tour, and remember to always use the best tool for the job ;)