I wrote recently about the interesting work being done on the open source music visualizer projectM, and that progress is being made on modernizing the graphics capabilities of this library. Since it was first created, there have been changes made in the way OpenGL is to be used for widespread compatibility and support for GPU shader programs.
Now the efforts are in the home stretch; OpenGL ES support is close to complete, and there is support for handling a decent percentage of the shader programs found in the visualization presets. The vast majority of the work has been done (by @deltaoscarmike), and mostly bug fixing remains.
After setting up vertex buffers and attributes and reworking the texturing and sampling code, the final step for full modern compatibility was dealing with the HLSL shaders. MilkDrop is the visualizer for Milkdrop that projectM is built for compatibility with. It was made with DirectX and the shader language that it supported in presets is HLSL, the language used by DirectX. Since projectM is designed to be run on other systems besides Windows, some form of support for HLSL was needed. Previously it was handled by the Cg toolkit from nVIDIA, but this project has been long abandoned and it was an unpleasant and sizeable dependency. The only other solution was to convert the HLSL code into GLSL, the OpenGL shader language. Use is now being made of hlsltranslator to transpile HLSL to GLSL, which is no small feat. Additional support for a #define preprocessor directive needed to be added due to its liberal use in previous code and presets.
Default warp and composite fragment shaders were created, switching to the preset versions if they are present and compile without problems. The blur shader was reimplemented in GLSL, and some helper routines recreated in GLSL for the presets that expect things like computing angle or texture sampling. The combination of running preset shader code on the GPU with the recent intel SSE optimizations for CPU per-pixel math brings performance improvements as well.
The real amazing news is that projectM now builds on a raspberry pi. This is notable not just because of the popularity of them, including running media centers, but because if it builds on a raspberry pi it should (with some effort of course) eventually run anywhere. One of the biggest issues was that projectM could only be run on a desktop or laptop computer. But now by using modern graphics operations it can run on accelerated hardware on mobile and embedded devices via OpenGLES, opening up many new opportunities, and potentially making it much easier to re-integrate into downstream projects like Kodi and VLC. Running it on the web via Emscripten, compiling to wasm and using WebGL is doable again. It’s amazing what possibilities modernizing software can bring.
It should be mentioned that the vast majority of the OpenGL work was done by superstar deltaoscarmike who apparently really knows what he’s doing.
As I’ve ended up with de facto maintainership of the illustrious projectM open source music visualizer I’ve seen a fair bit of interest in the project. I think I at least owe a blog post to update folks on where it’s at, what needs working on, and how to help make it better.
What is projectM?
projectM is a music visualizer program. In short it makes cool animations that are synchronized and reactive to any music input. I say music and not audio because it includes beat detection for making interesting things happen on the beat.
Some of you may remember the old windows mp3 player WinAmp. It contained a supremely amazing and innovative music visualizer called Milkdrop written by a gentleman from nVidia named Ryan Geiss, known just as Geiss. The visualizer was not a single set of rules for visualizing audio but rather a mathematical interpreter that would read in “preset” files which were sets of equations. You can read the very illuminating description here of how the files are defined if you’re interested. In short there is a set of per-frame equations describing colors and FFT waveforms and simple transformations, and there is a set of per-vertex equations for more detailed transformations and deformations.
Due to the popularity of WinAmp and Milkdrop there have been many thousands of presets authored and shared with really stunning and innovative visual effects ranging from animated fractals to dancing stick figures to bizarre abstract soups. The files are often named things like:
shifter – cellular_Phat_YAK_Infusion_v2.milk
[dylan] cube in a room -no effects – code is very messy nz+ finally some serious stfu (loavthe).milk
NeW Adam Master Mashup FX 2 Zylot – In death there is life (Dancing Lights mix)+ Tumbling Cubes 3d.milk
suksma + aderassi geiss – the sick assumptions you make about my car [shifter’s esc shader] nz+.milk
flexi + cope – i blew you a soap bubble now what – feel the projection you are, connected to it all nz+ wrepwrimindloss w8.milk
And so on.
As I understand it, possibly incorrectly, there were two major problems with Milkdrop. First that it was implemented with DirectX, win32 APIs and assembler, and secondly that it was not open source (though it was made open source fairly recently). So some enterprising folks in 2003 created projectM as an open source reimplementation that would be Milkdrop preset-compatible.
I didn’t work on projectM originally and I am not responsible for the vast majority of it. However the previous authors and contributors have for whatever reason mostly abandoned the project so it was left to random people to make it work. The code is quite old although the core Milkdrop preset parsing, beat detection, most of the OpenGL (more on that later) calls, and rendering is in fine shape. projectM is really just a library though, designed to be used by applications. In the past there have been XMMS and VLC plugins, a Qt application, pulseaudio and jack-based applications, and more.
OSX iTunes Plugin
Not really having a good solution for OSX I went ahead and ported the ancient iTunes visualizer code to work on a then-modern version of iTunes and voila! projectM on OSX. Though I did have to deal with the very unfortunate Objective-C++ “language” to make it work. Not Objective-C, Objective-C++. No I didn’t know that existed either.
I tried to submit the plugin to the Mac App store as a free download. Not to make money or anything, just to make it easy for people to get it. The unpleasantness of this experience with Apple and their rejection is actually what spurred me to start this blog so I could complain about it.
I decided that what would be better is a cross-platform standalone application that simply listens to audio input and visualizes it. This dream was made possible by a very recent addition to the venerable cross-platform libsdl2 media library adding support for audio capture. I quickly hacked together a passable but very basic SDL2-based application that runs on Linux and macOS and in theory windows and other platforms as well. Some work needs to be done to add key commands, text overlays (preset name, help, etc), better fullscreen support and easy selection of which audio input device to use.
The main application code demonstrates how simple libprojectM is to use. All one must do is set up an OpenGL rendering context, set some configuration settings, and start feeding in audio PCM data to the projectM instance. It automatically performs beat detection and drawing to the current OpenGL context. It’s really ideal for being integrated into other applications and I hope people continue to do so.
You can obtain source, OSX and linux builds from the releases page. This is super crappy and experimental and needed some configuration tuning to make it look good, and you need to drop the presets folder in. But it’s a start.
In their infinite wisdom the original authors chose the cmake build system. After wasting many hours of my life I will not get back and almost giving up on the software profession altogether I decided it would be easier to switch to GNU autotools, the same build system almost all other open source projects use, than to deal with cmake’s bullshit. So now it uses autotools (aka the “./configure && make && make install” system everyone knows and loves).
This is where you come in. If you like music visualizers and want to help the software achieve greater things there is some work to be done modernizing it.
The most important task by far is getting rid of the OpenGL immediate-mode calls and replacing them with vertex buffer object instructions. VBO is a “new” (not new at all) way of doing things that involves creating a chunk of memory containing vertices and pushing it to the GPU so it can decide how and when to render your triangles. The old-school way was “immediate mode” where you would tell OpenGL things like glBegin(GL_QUADS) (“I’m going to give you a sequence of vertices for quadrilaterals”) and give it vertices one at a time. This is tremendously inefficient and slow so it isn’t supported on the newer OpenGL ES which is what any embedded device (like a phone or raspberry pi) supports, as well as WebGL.
Astute readers may note that there already are iOS and Android projectM apps. They are made by one of the old developers who has made the decision to not share his modern OpenGL modifications with the project because he makes money off of them.
Another similar effort is to replace the very old dependency on the nVidia Cg framework for enabling shaders. Cg was used because it matches Directx’s shader syntax. GLSL, the standard OpenGL shader language is not the same, and requires manual conversion of the shaders in each preset.
The Cg framework has been deprecated and unsupported for many years and work needs to be done to use the built-in GLSL compilation calls instead of Cg and convert the preset shaders. I already did some work on this but it’s far from finished.
The reason I’m writing this blog post is because of the community interest in the project. People do send pull requests and file issues, and we definitely could use more folks involved. I am busy with work and can’t spend time on it right now but I’m more than happy to guide and help out anyone wishing to contribute. We got an official IRC channel on irc.freenode.net #projectm so feel free to hang around there and ask any questions you have. Or just start making changes and send PRs.
If you remember the old windows music player Winamp, it came with an amazing visualizer named Milkdrop written by a guy at nVidia named Geiss. This plugin performed beat detection and splitting the music into frequency buckets with an FFT and then fed that info into a randomly-selected “preset.” The presets are equations and parameters controlling waveform equations, colors, shapes, shaders,”per-pixel” equations (not actually per-screen-pixel, rather a smaller mesh that is interpolated) and more.
Most of the preset files have ridiculous names like:
“suksma + aderassi geiss – the sick assumptions you make about my car [shifter’s esc shader] nz+.milk”
“lit claw (explorers grid) – i don’t have either a belfry or bats bitch.milk”
“Goody + martin – crystal palace – Schizotoxin – The Wild Iris Bloom – mess2 nz+ i have no character and feel entitled to one.milk”
Milkdrop was originally only for windows and was not open-source, so a few very smart folks got together and re-implemented Milkdrop in C++ under the LGPL license. The project created plugins to visualize Winamp, XMMS, iTunes, Jack, Pulseaudio, ALSA audio. Pretty awesome stuff.
This was a while ago, but recently I wanted to try it out on OSX. I quickly realized that the original iTunes plugin code was out of date by about 10 major versions and wasn’t even remotely interested in compiling, not to mention lacking a bunch of dependencies built for OSX.
So I went ahead and updated the iTunes plugin code, mostly in a zany language called Objective-C++ which combines C++ and Objective-C. It’s a little messed up but I guess it works for this particular case. I grabbed the dependencies and built them by hand, including static versions for OSX in the repository to make it much easier for others to build it (and myself).
Getting it to build was no small feat either. Someone made the unfortunate decision to use cmake instead of autotools. I can understand the hope and desire to use something better than autotools, but cmake ain’t it. Everything is written in some ungodly undocumented DSL that is unlike any other language you’ve used and it makes a giant mess all over your project folders like an un-housebroken puppy fed a laxative. I have great hope that the new Meson build system will be awesome and let us all put these miserable systems out to pasture. We’ll see.
Long story short after a bunch of wrangling I got this all building as a native OSX iTunes plugin. With a bit of tweaking and tossing in the nVidia Cg library I got the quality and rendering speed to be top-notch and was able to reduce the latency between the audio and rendering, although I think there’s still a few frames of delay I’d like to figure out how to reduce.
I wanted to share my plugin with Mac users, so I tried putting it in the Mac App Store. What resulted was a big fat rejection from Apple because I guess they don’t want to release plugins via the app store. You can read about those travails here. I think that unpleasant experience is what got me to start this blog so I could publicly announce my extreme displeasure with Apple’s policies towards developers trying to contribute to their ecosystem.
After trying and failing to release via the app store I put the plugin up on my GitHub, along with a bunch of the improvements I made. I forked the SourceForge version, because SourceForge can go wither and die for all I care.
I ended up trying to get it running in a web page with Emscripten and on an embedded linux device (raspberry pi). Both of these efforts required getting it to compile with the embedded spec for OpenGL, GLES. Mostly I accomplished this by #ifdef’ing out immediate-mode GL calls like glRect(). After a lot more ferocious battling with cmake I got it running in SDL2 on Linux on a Raspberry Pi. Except it goes about 1/5fps, lol. Need to spend some time profiling to see if that can be sped up.
I also contacted a couple of the previous developers and the maintainers on SourceForge. They were helpful and gave me commit access to SF, one said he was hoarding his GLES modifications for the iOS and Android versions. Fair enough I guess.
Now we’re going to try fully getting rid of the crufty old SourceForge repo, moving everything to GitHub. We got a snazzy new GitHub homepage and even our first pull request!
My future dreams for this project would be to make an embedded Linux device that has an audio input jack and outputs visualizations via HDMI, possibly a raspberry pi, maybe something beefier. Apparently some crazy mad genius implemented this mostly in a FPGA but has stopped producing the boards, I don’t know if I’m hardcore enough to go that route. Probably not.
In conclusion it’s been nice to be able to take a nifty library and update it, improve it, put out a release that people can use and enjoy, and work with other contributors to make software for making pretty animations out of music. Hopefully with our fresh new homepage and an official GitHub repo we will start getting more contributors.
I recorded a crappy demo video. The actual visualizer is going 60fps and looks very smooth, but the desktop video recorder I used failed to capture at this rate so it looks really jumpy. It’s not actually like that.