Greetings again. Music is the topic today. I hope you get a bit of a lesson out of it. I've been wanting to create music for a few months. It started with Csound and has morphed from thing to thing since then. Not a single piece of decent music has come from my efforts. However, my efforts are not in vain. I'm learning how to make music and I'm thinking about very interesting things in the process. At the current point, I am happy where I am with Soundtracker for Linux. I'm past the learning curve for the tracker's basic functionality. Now I'm working on creating samples with Csound (I have a pretty good start) and putting them into a pattern. It hasn't created anything musical, so I'm still working, but I've gotten a pretty good start in three days. With a month of work, I can create bunch of decent songs for Hack Mars and have the skill to do as many more as I have patience for. I really like listening to modules. I have a collection of ~1600 minutes of electronica (most of it is: monotonik vol. 1-4, acid-100, and ). I have about 50 mods growing daily. I like about 80% of my electronica which gives me a large collection to listen to each day and not get repeats. Compared to my commercial collection of ~4500 minutes, it's taking over. I'm not buying anymore RIAA music. Since the electronica I get is free, it allows me to listen to all indy and support what I like. It's a downhill battle from here. No business can fight a downhill battle. How about you?

One thing I like about tracking is that it overcomes problems I've had in the past. It's like CSound, but doesn't require any algebra. It's like midi2 that midi2 will never become. Being able to use samples gives it flexibility that midi will never have. Being able to create in realtime gives it the usability that Csound lacks. I also like the fact that when the mod is finished, I can export it to Csound. How is that? Well... I was thinking about it. Tracking is just a subset of Csound. Csound support samples. It supports algebraic expressions also. So I use the tracker to make a song, let's say. Then when it is perfect, I manually type in the csound equivalent in a text editor. Then I can use some of Csound's features. Csound also is not limited to number of channels and real-time processing. Cool, huh? Another benefit of Csound is that it can create very intresting samples from scratch. That's cool because I don't have any samples of my own. So I take all the work that I have done in Csound (a small bit, ~10 attempts at songs) and I grab samples of each instrument that I've played in each song. Then I import them to the tracker. Then I can use the instrument in my pattern.

What is the lesson for today? It's about structure. Csound goes to great length to have a very open structure. A person can start with a wave, add exponential decay and linear envelope. A person can use grains, fm oscillators, all in series. Really a huge amount of algebraic functions for use in instruments. Then the song can play this at any point with any pitch, any random data anywhere. A person like me with very little understanding of music can make some ugly stuff. What is a C4 note? What does melody really mean? Harmony? Don't need that... Doh! A tracker on the other hand has structure. 1-8 channels, N instruments, M samples, bpm=140, tempo=6, pitch=A1-G5. There's no question of something being off beat, the BPM and tempo ensure that the instrument at place J is played at time t_j. Melody isn't especially hard and using samples is incredibly easy. There's no doubt that a sample will sound a certain way when played with a certain pitch.

I'm going to go back to read my Musical Acoustics book.