one-line algorithmic music: Bytebeat

[William Croft]

OK, apologies, this is not Analog Box, however it does have a certain amazement factor for minimalist softsynth fans.

http://canonical.org/~kragen/bytebeat/   general survey article

http://countercomplex.blogspot.com/2011/10/algorithmic-symphonies-from-one-line-of.html

http://countercomplex.blogspot.com/2011/10/some-deep-analysis-of-one-line-music.html

[Andy Turner]

Neat use of LFSR.
<intimidating name>Linear Feedback Shift Register</intimidating name>
http://en.wikipedia.org/wiki/LFSR for some pictures and background theory.

I'm not sure ABox could do an LFSR directly as shown (at least not at
a rate faster than 23millisec), but maybe a combination of delay
objects and equations would have the same effect? Other alternatives
might include the diff-system and csv-reader/writer.

None-the-less, it's pretty cool that music-like phenomena can be found
in even the simplest places -- all we have to do is listen for it.

[John Paul Burtell]

what a coincidence! i was just going through a book on numerical methods that i picked up for a course i'm taking and i came across a section on the LFSR. i immediate thought it would be an interesting system to implement in ABox, but i drew the same conclusion that it couldn't be done at fast rates because of the feedback delay.

 

still, i am going to look into this further at how it might be done using your other suggestions.

 

thanks for posting this!

--
You received this message because you are subscribed to the Google Groups "Analog Box" group.

To post to this group, send email to analo...@googlegroups.com.
To unsubscribe from this group, send email to analog-box+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/analog-box?hl=en.

[Andy Turner]

a happy coincidence indeed :-) ... it might just be possible ...
they're using integers ... should be a way to use floating point ...
here's one way to pick out single bits from some signal 'a'

int( ( a%b ) / b )

where b is the fraction 2^(-bitnum) and bitnum=0 is the most
significant bit (corresponds to b=1), and 1 is the next msb
(b=1/2) ... etc down to bitnum=23 -- which is the limit of the
floating point precision. The above would take the place of the C code

( ( a >> ( 32-bitnum ) ) & 1 )

which I believe might let us duplicate the ByteBeat mechanism, at
least with 24 bits.

To get an actual LFSR, is it possible to determine the difference
bewteen between 2 consecutive LSFR outputs? If so, then the diff-
system object might be able to do the job ... albeit messily