Monday, April 30, 2012

Notes on Translection

So: in my previous post, I coined the word "translection" to describe what I thought might be a computer-native form of musical variation, which I had previously been referring to as "tag-shifting". I like "translection" better because it contains a word fragment ("lect") that  comes from legere, the Latin verb for "reading", which is the crucial operation here. The tags are not changed (as the word "tag-shifting" seems to imply). In translection, the way the tags are read is changed.

I might as well try to make a short definition for translection here. Literally, it means to change the way a signifier is mapped to an operation. Put mathematically, this is equivalent to changing the transfer function. To changing the algorithm by which one set of symbols is mapped to another set of symbols (or, in the case of a computer, to a set of operations). It's a case of remapping that specifically refers to remapping the data of time-based media. Translection differs from Translation and Interpretation in that it involves using clear and defined algorithms to change the meaning. Translation and Interpretation invoke much fuzzier forms of remapping. They invoke an art reliant on using judgement, rules of thumb, code-shifting between various mapping systems. Translection refers a more literal, more transparent form of remapping.

So why take such care in defining the word? I'm thinking it is already a way of talking about a specific kind of variation, and could be specifically useful in talking about/thinking of glitches as a source of  musical (and possibly extramusical?) variation.

 It occurs to me that playing a traditional score in a different key is an instance of translection.

I also realize that my sequencer's translective variations were the consequence of  a feature of MIDI data structure, specifically of  its feature of defining the note's duration with velocity (a note-on command consists of the note-number accompanied with a nonzero note-velocity, whereas a note-off command consists of the note-number accompanied by zero note-velocity). While traditional Western musical notation treated note duration as an atomic unity, MIDI grammar split note duration into note-on and note-off,  i.e. two grammatical units. Doubling the number of signifiers that defined duration opened the possibility of performing operations on the two signifiers which not only were previously impossible but literally unthinkable in terms of traditional notation,.

Friday, April 27, 2012

Translection: a computer-native form of musical variation

My collaborator Malek Lopez and I were playing around with some MIDI sequencer code. Based on his desire to have a rhythm generator, I kludged a 16-step sequencer that would spit out a sequence of 16 commands. For the sake of simplicity, let's say the sequencer did this by randomly choosing a number between 0 and 2 sixteen times, and putting its choices into a list.

So for instance, it might spit out: 0112 1020 2201 2021

The number sequence was then read as a sequence of tags/commands according to the following system:

 0= start a note;  1 = end a note;  and 2 = do nothing.

Now, if we assume that the numbers determine the gating of a single sustained pitch, then the sequencer would output a sound that could be represented as:

 Where a stretch of blue squares indicated a sustained pitch, and a stretch of white squares indicated silence.
After listening to the sequencer do its thing for a few hours. I realized that changing the way the sequencer interpreted the commands would create musical phrases that would differ from one another, and yet be related by rules of translation.

The most obvious variation would be produced simply by inverting the interpretation of the start and stop tags, ie

0 = stop note;    1 = start note;     2 = do nothing.

 Well, I said it was the obvious variation. It produces a negative of the previous sound, where previous tones are replaced with silences of equal length, and previous silences are replaced with tone.

However, if we use a different system of tag interpretation, say

0 = do nothing;   1 = start note;   2 = end note;  

 then we get something like this:

Which is a sonic product with a different and less obvious relationship.

I'm currently referring this kind of variation as Translection, as it consists of changing/shifting the way the tags are read ("lector" = reader, from the Latin verb legere:  "to read"). I find the idea of translective variation interesting because this kind of variation is native to music made with computer code. As far as I know, it is not a named, known or acknowledged form of musical variation. Still not sure where it goes from here, but tag-shifting functions will definitely be coded into the coming sequencers we'll be making..