In search of better softsyths

by MrMessiah

Hello! First of all some disclaimers. This isn't another salvo in the long-since-boring MP3 vs Trackers debate. It's my opinion as a maker of electronic music and as an observer of the scene for a long time now. Although I put in my profile that I am a coder/graphician, which I also am, first and foremost I consider myself a musician, because I have been making music with technology since I was 6. (With that you may take my lack of having released anything in the scene yet as a measure of the relevence of my opinions if it helps you! :) )

I see a lot of talk, some of it quite reasoned, some of it very lame, about the merits of tracked music in demos, and particularly what makes me sad is the feeling that tracked music is in any way "dying out" as I saw in the headline of a new Musicdisk released recently. Partly because this isn't necessarily true. People react to statements like that in a number of ways: some people get together a load of their friends and make a kickass musicdisk, which is always cool. Some people react negatively and bitch about MP3's again.

What I think more and more in this debate is to remember why people started tracking in the first place. On the one hand you had totally generated sound, which was expensive to get any kind of good result, and if it became too much to code around, you'd get relegated to sound of such low quality it was very unexpressive as a musician. On the other hand, you could get a huge great big .WAV file and play it back, which would be great because you could make any sound you wanted in that, but sucked because of the file size, and the fact it may have been realtime when you recorded it, but it sure as hell wasn't afterward! :)

Tracked music though was a decent middle-ground solution. It provided a framework you could work within to produce outstanding, expressive, music-like results more easily, and was reasonably cheap on the CPU and small on the disk. There was a flurry of activity as formats came and went, and features were crammed into the trackers to help musical expression, and over a period of years, it more or less boiled down to a couple of really useful formats that everybody knew how to write for and code players for. This was very much the elegant solution of its time.

Now that is not to denigrate the tracked music scene in any way, and I myself have built up a lot of love for the skill involved in producing something. But if you look at the latter half of the tracking scene, it has shifted its mindset very clearly from its origins. What the hell am I on about? Well the thing with trackers used to be that they were a liberating musical tool, that you could produce tracks that were as good as if not better than the stuff you heard on the radio, and certainly a lot better than by using some other means of generating music on your computer. But more recently, when the power of computers has increased so much, and also just in terms of the people making the music, the tracker formats are the opposite, they are more of a restriction to work within.

For some people, that is the whole point: that working with the limited palette you have and hand-tweaking everything to get some kind of sound that is more than the some of its parts is the very ethos of the scene to start off with. Which if you think about it, is a very good argument, and there's no way you would ever want to argue with that. These people are like master-craftsmen who know how to use the restriction and the tools to their advantage and still produce kick-ass results. I have a lot of respect for the people that can do this, and that choose to do this, particularly as I find that my own skills in this field are lacking. I can spend hours getting to a fairly good result, but at the end of it I'm reaching for my studio equipment, or Buzz, or somesuch, just because I know I will have a musical idea float around my head for so long, and the quicker I can get that idea out into some more permanent form the better. :)

My problem with this for is that it is no longer about pushing the limits of the machine but about pushing the limits of an arbitrary format that has diminishing relevance when you look at what the modern PC hardware can do. I'm not just talking about CPU power either, the soundcard hardware is getting phenominally powerful, with professional grade synthesis and DSP capabilities just waiting to be tapped.

Even if you "just" consider softsynth trackers like Buzz, (which I'm no particular evangelist of, I can think of so many ways it could be made better), you have a range of expression in there that, I'm sorry to say, far exceeds what you can get out of a tracker, and as a musician (or one kind of musician, maybe) that just tempts me away every time.

If you look at the demos that do use MP3 playback, you can bet that a lot of these are sourced from the so called "next-gen" trackers like Buzz, and hard-disk-recorded that way. Or, pulled in from some professional musical setup. You see, what MP3 has given us is a new middle-ground technology, that enables you to again draw a line somewhere between pure synthesis and pure playback, write one library to replay it and boom, an easy solution.

Now, this is where I become unhappy with MP3 and share the frustrations of some of the tracker officionados. Because it is such a shame to rely on a technology like MP3 if the only reason it was used was to bypass the restrictions of oldskool trackers and as a stopgap. And as the move into Windows demos on the PC platform continues, I can only see this getting worse if something is not done.

I think that what is needed is some new way, a new framework if you like, that can be used to get this stuff into demos, in a synthetic format, rather than as MP3, in a similar way that OpenGL or Direct3D is currently cutting out the donkey -work of 3D coding for graphics coders. Somewhere between the two camps, that facilitates the writing of in-demo soft- synth systems, that provide you with the full range of expression you need as musicians.

Because another problem has been that the provision of easy one-shot libraries for replaying tracker formats and MP3 formats has had a more pernicious effect: isolating musicians from the coding process. This I feel is the biggest disappointment of all. Maybe that's just because I happen to like playing both roles of musician and coder! :) But I know I'm not alone in this, and there are still a few notable exceptions where a demo group has created a cool softsynth system. How about bringing these techniques into the open?

I know to some extent there is a growing feeling of "make what you like and forget everything else" which believe me is a good ideal, but I think that there is so much scope in the field of sound and music processing thats going untapped. Musicians' input shouldn't be relegated to "...and the music was OK as well". Synchro shouldn't be relegated to "... and the screen flashes white every kick drum. Coool!".

Demos of the past 3 years have blown us away with visuals, now how about blowing us away with sound?