Since the release of the IDM development examples, I have been taking a new approach to instrument design. I have started algorithmically generating my instruments. What I mean by this is that I am letting Algorythm Radio design it’s own synth sounds to make music with! Just let that settle in for a minute… If after that minute you’re not overwhelmed with excitement, then I would move on, because that’s the excitement peak for my post today as far as I’m concerned. If you’re more excited by numbers, feel free to skip ahead to the 4th paragraph where I’ll discuss some big ones.
I have been batting the idea of randomly generated instruments around in my head for a while, but I only recently put the time into figuring out how to best weave the idea into the app. Needless to say, it is now woven in, and boy has it helped development!
If you have listened to the previous development examples, you will have noticed that the range of instruments used from song to song is pretty limited. That’s because it can take a long time to create a collection of nice synth sounds or find a range of good samples, fine tune them and then mix and master them all so that every single one works with every other one. It is a complex and time consuming process. Now, all I have to do is write a handful of instruments, each with a list of randomizable (but well refined) parameters, and suddenly I have all the instruments I need. It may surprise you to know, that each instrument generator takes almost the same amount of time to make as one normal instrument. That may sound crazy, but the time I used to spend tweaking parameters to make a precise sound, is now spent figuring out the acceptable boundaries of that parameter instead. There are also certain advantages presented by this process when it comes to mixing and mastering.
This is the point at which I would like to tell you the factor by which my efficiency in sound design has increased, or the exact number of different sounds available from each instrument generator… but I can’t. I can tell you that in both cases, the number is big, real big! The data type I am using ranges from 4.94 * 10^-324, to 1.79 * 10^308. I wouldn’t usually use scientific notation when trying to wow you, but if I didn’t, the screen would be half filled with digits, and I have much better numbers to fill your screens with in the near future. The range isn’t linear though, and I am not using the full range for each parameter, so it is hard to figure out, instrument to instrument, how many different combinations of parameters there are. One thing I do know though, is that each generator is definitely capable of 18446744073709551616 different instruments at the very least, which if you remember, is how many different tracks Algorythm Radio can produce per genre combination. Or at least, it was…
I have made some astonishing improvements to the random number generators, but I will talk about that in my next post. In the mean time, look out for some audio examples of the randomly generated instruments that I should be revealing in the coming days, and it won’t be long now before I post the next selection of genre examples. The next genres will be as follows:
– Drum and Bass
Thanks for reading!