pouët.net

thanks guys :) here's one more:
t>>4+!(-t>>13&7)+2*!(t>>17)|t*t*(t>>(t>>12^t>>11)%3+10)/(7+(t>>10&t>>14&3))*!(t&512)<<3+(t>>14&1)

late night noize chillout: L: (t*0.01)^0x6f|(t^43)-(t^62)*1.77 | R: ((t&(t*0.01)^0x2f)^0x8f)*1.55

v2: L: (t*0.01)^0x6f|(t^43)-(t^62)*0.0000097*t | R: ((t&(t*0.01)^0x2f)^0x8f)*0.0000095*t

a bare-bones version of the previous one to see how small it can be:
t>>4|t*t*(t>>6&8^8)*(t>>11^t/3>>12)/(7+(t>>10&t>>14&3))

nice!

sounds like a souped up version of one of glooms work lol.

My No 4:
((v=t/4)>>v)/(1+(v>>v>>7))+((u=v^31)>>u)/(1+(u>>u>>7))&t/32|(t>>11)%8%3*t*t*(t>>9&2^2)&31

I still think there should be a more standardized instruction set, a few rules and a defined syntax not to mention some tools for wav/exe/playback/codefordemo and online playback tool.

Has anyone been working on something like this?

I really like xpansive's piece :)

Got my No 4 smaller without really making it sound worse :)
((v=t/2^(t&64?63:0))>>v)/(1+(v>>v>>7))&t/32|(t>>11)%8%3*t*t&15

Oopsy. Sorry, some debug crud in that link, please use this instead:
((v=t/2^(t&64?63:0))>>v)/(1+(v>>v>>7))&t/32|(t>>11)%8%3*t*t&15

I've posted longer lines before:
L: o=(t>>14),s=t*1.14*(o<8),a=4096/((s+(s>>12)^(s>>3))&4095)<<5,l=o-10,t-=0x13fff,p=(7-l)*(l<7),u=(10^(t>>(15-p)&7))+3,y=(t>>(11+p)&7)/u,f=t*8*y,z=(t*16/u)&99|f|f*1.01,z*=(o>3),a+z*2 | R: a*1.2+z*4

And here are the two parts seperated:

1. s=t*1.14,a=4096/((s+(s>>12)^(s>>3))&4095)<<5

2. L: u=10^(t>>15&7),u+=3,y=(t>>11&7)/u,f=t*8*y,z=(t*16/u)&0x63|f|f*1.01,z*2 | R: z*4

That's sweeeeeeeeeeeeet, mu6k! The detuned sound makes a great job at disguising any inexact note frequencies.

yehar and mu6k's pieces togheter: L: s=t*1.14,a=4096/((s+(s>>12)^(s>>3))&4095)<<5 | R: ((v=t/2^(t&64?63:0))>>v)/(1+(v>>v>>7))&t/32|(t>> 11)%8%3*t*t&15

xpansive, thats freaking unbelievable.
I think it's the one of the most impressive things since the beginning, it's in the spirit of the first ones, ultra simple formula, incredibly complex and long tune.
It's obviously not planned.

I feel there is some mathematical things to be discovered about this, maybe some kind of theory, like fractals. This should be able to lead to graphics too.

mine, yehar & mu6k pieced together: L: s=t*1.14,a=4096/((s+(s>>12)^(s>>3))&4095)<<5 | R: ((v=t/2^(t&64?63:0))>>v)/(1+(v>>v>>7))&t/32|(t>> 11)%8%3*t*t&15||(t*2)^t|(t&(t*0.75))-(t*0.99)^(t*6)

Replying to xpansive's with ONE MILLION alarm clocks:
t*t/(t>>12&t>>8)<<7

Thanks guys :)
Here's another one, quite a bit more deranged, though...
t*(t/((t>>9|t>>8)))

No 5:
(u=3*t>>t/4096%4&-t%(t>>16|16)*t>>14&8191)/(u>>6|1)*4

Playing around some more with the t%t/ stuff...
(t%(t>>(t>>7&t>>2))<<2)|(t%(t/(t>>15&t>>6))>>4)

great

http://www.youtube.com/watch?v=NPD9zCOpmuc

I think this C64 is creating something similar by mistake, by accidentally writing stuff to the SID registers because of some glitch. (Maybe dirty pin contacts on the 6502 or something.)