Interview: Tobiah

Tobiah recently wrote to the SAOL users list about the joys of using SAOL. We thought it would be interesting to talk to him about SAOL, CSound and making music in general. Tobiah lives on the California coast, North of LA. He can be contacted at <toby@rcsreg.com>.

m-station: First, let's set the scene... What's your background in music? How long have you been using Linux for music related stuff?

Tobiah:
I started taking music courses in junior college, and became absorbed in them. I began to view the music that I had listened to up to that point as poorly conceived, and less interesting than I had once thought.

I started to play classical guitar instead of electric guitar, and went off to college to learn more.

I found out how to tune my guitar like a renaissance lute, and found that I preferred that literature. As the college did not support this avenue, I turned to composition as a way to finish school...

Just an aside, but what's a lute's tuning?

I tune a guitar's 3rd string (G) down a half-step to F#, and then put a capo at the second or third fret. The second fret sounds better, and according to some, more closely approximates what was actually being heard in the period. The third fret gives the pitches that were written into the music:

6 5 4 3 2 1
G C F A D G

With the lowest sounding string being 6. Actually most of the lute music was tablature anyway, and so there was not really a pitch reference, but if you look at vocal accompaniments and ensemble works involving the lute then the above is true. Through time, lutes evolved more bass strings, and they were tuned mostly diatonically down from the sixth, and so would often be retuned from piece to piece (They had no frets under them).

... I slowly began to understand a far broader meaning for the word 'music' than I had previously held. Finally I took a class in Cmusic. Since then my entire compositional energy has been channeled into a computer. I fell in love with Unix, and installed Csound on a home 486 running Linux in about 1995, and started honking and wheezing through a soundcard.

As you know, I recently found mp4-sa, and I find it quite liberating in comparison to any other tool that I have used.

Oddly, my formal output has been sparse. I often get too wrapped up in the process, and in building programming tools to the point that I blur the motivation for composition.

But then I have always thought that if one is absolutely faithful to the spirit of a composition, he has to relinquish control, perhaps to the point that he finds out that it wasn't a music piece at all! One has to let go of the selfish desire to be a composer in order to find the composition.

So, as a side effect, I took a great job as a computer programmer, and when I'm not working, I am poking around with score generation programs, and playing lute music.

heh heh, the list of people doing that is fairly long (wrapped up in the process...). What sort of genre has the stuff you have done fallen into?

I never have a good time answering that question. I don't try to emulate, so categorization is tough. I am drawn to creating textures with algorithmically generated bursts and sometimes rhythms. Another person would listen to it and call it 'That academic crap'. Like I said however, I can count my finished pieces on my elbows, so it is hard to say.

I just want pleasant sounds to happen. I don't want to write 'music', I just want to sculpt sounds. I do want to distribute the sounds, and have it happen that others reap enjoyment from them. In this way, it is music as I understand it.

And perhaps whatever is in there will get out untrammeled by what clothes you wear or who you voted for last - that's towards a Keatsian purity or classicism (of whatever it turns out to be!) ... or do you mean something else by that?

I don't know who Keats was, but this sounds like the Idea. It is like being in love with someone so intensely that you would be willing to follow the *desire of the love itself* so much that you would be willing to watch the union dissolve as it expressed itself in terms other than as "lovers". It really takes the same effort (or maybe lack of effort) that one expends in order not to "sin". Let's say that you are in a situation in which you know beyond doubt that you would not be discovered if you performed a slight transgression against an anonymous entity or society. How do some know that it is a bad idea? There is promise of immediate reward, with no apparent penalty, yet these people listen to the *desire of the composition* of life. They are able to squelch the overpowering and selfish wants of the frontal lobe. There is a great plan, and we get clues as we create art. I am just going for nice sounds. In the end, when all the chips have fallen, wouldn't it be best to have those around? I also think that it is important that a work be timeless, rather then relying on hooks that are installed in minds of the time. This is also close to "the plan". Was it Michelangelo that would brush away compliments and claim that the statue was already in the block of marble? All he did was to remove everything else.

It's probably worth saying that working with something like SAOL you don't actually need a whole bunch of other stuff - just your computer, a sound card and speakers... but some people might add to that. Do you use anything else?

I agree, and that is certainly one of the most refreshing aspects of generating sounds with a computer.

At one time, during my electric guitar stage, I began to seek variety in sounds and composition techniques through the then budding market of electronic noisemakers that were available at the local music shop. I remember getting a small voltage control style monophonic synth, and thinking that that was going to give me any sound that I wasn't able to make with what I had. I figured that that would be enough equipment for me for quite some time.

Then I realized how important it was to be able to play more than one note at at time, and so I got a synth that would do that. Then I realized how blind I had been, and that the synth needed to be able to play not only more than one note at a time, but more than one timber at a time. So I got one of those. Then came sampling keyboards, and so on.

I went to school about that time, and was so poor that the point was of no consequence. Now I find that there are no more attractive electronic whammy boxes at the music shop that I am interested in. I can do it all at home in my computer with far greater flexibility (no more, "Gee I wish that this synth had a slider for this function") and fidelity.

So, as for my tools, I have a classical guitar, a set of wooden recorders, a midi piano, a couple of percussion instruments, a pair of mics, and a small mixer. I also have a laptop that I can use to harvest sounds from nature. There are a few good sounds on the internet as well.

All that, combined with the things that I can do to those sounds in the computer provide plenty of palette for me to work with. Then there are the classic forms of synthesis which are also available through software.

And, once in a great while, when I am feeling particularly ambitious, I will pull out a pencil and staff paper.

Just to go back a bit, I wonder what your working method with SAOL is ... do you "hear" what you'd like to do and then program it or is it a more iterative process where you generate something and then keep adding and subtracting until you like it.

I am still learning SAOL, and so have only tried to compose with Csound.

I wish that I could say that I methodically constructed a piece mentally, and then maybe made sketches, and then, finally programming what I had invented so that it could be heard by others.

That is what I look forward to being able to refine my technique to, but that is not really what actually takes place right now. I did one piece that started with an inspiration brought about by a sound that I created through pure experimentation.

I was playing around with a short sample of a cello playing a low note. I started writing cscore programs that would generate scores for it. I began to feel my algorithmic oats, and tried to see if there was a limit to what my computer could do. There wasn't a limit really. So I started generating tones by mixing thousands of windowed sections of the cello sound together at very slightly varying pitches, and at staggered start times. The result was a haunting low pitched howl that made me decide to put a piece around it.

Later, when the csound 'grain' opcode came out, I learned that I had not invented anything new, save that of using the score to implement the granular synthesis, but I'm sure that many people have done that as well.

I continued to add sections to the piece, overlaying 'tracks' until I had an all around texture that I was happy with. This was really just a collage, and so I am not as proud of it as I might be if I had really executed a brilliant idea. Oh well, next time.

By the way, when making a large piece, I use dozens of orchestra and score pairs, which each render by themselves in a reasonable amount of time. I arrange these units of sound in a logical hierarchy, where the deeper levels are individual sound nuggets, and the shallow areas are culminations of the smaller parts, and the root of the structure is where the final 'tracks' or sections are mixed together to create the final piece. Each directory or node contains it's own makefile that checks it's subdirectories for sounds that are out of date with respect to the orc, sco or score gen program or anything else that is needed to create that sound. The end result, is that I can listen to a 5 minute piece, and decide that the third bass note of the fourth minute should be a few dB higher. I would then make the change in the score for that note or phrase, and then go back to the root of the piece and type 'make'.

Just as with a large C program, this monster composition engine would travel down to the out of date section, render it, and then climb back up to the top, along a line of simple mixers that very quickly patch the entire piece together again. This saves countless hours which would otherwise be devoted to rendering parts of the piece that hadn't changed. You can also just dwell in one of the deeper sections of the tree in order to focus on one sound or group of sounds without hearing the rest of the piece along with it.

This method takes far more disk space, as there is a lot of redundancy between the levels. Fortunately, disk space is very abundant and inexpensive these days, so it does not pose a problem.

Thanks a lot Tobiah.

home  music  news  opinion  software  tips  email