Bytebeat
Jump to navigation
Jump to search
bytebeat in x86 assembler
; noise.asm -- just fooling around :-) ; mct and leif, ; Sun Feb 19 19:08:13 PST 2012 ; ; Build with: ; ; nasm -felf32 noise.asm ; ld -melf_i386 -o noise noise.o ; ; Run with: ; ; ./noise | aplay ; ; Equivalent to: ; python -c'import sys;[sys.stdout.write(chr((t>>10^t>>(t>>13)%4)%256)) for t in xrange(2**20)]' |aplay global _start section .data section .bss var: resb 1 ; one bye variable section .text ; f(x) = (x >> 10 ^ x>>((x>>13)%4)) % 256 _start: mov esi, 0 ;int 3 .loop: mov edi, esi shr edi, 13 ; edi := x >> 13 mov eax, edi cwd mov bx, 4 div bx ; dx := (x >> 13) % 4 mov eax, esi mov cx, dx shr eax, cl ; eax := x >> ((x >> 13) % 4) mov edi, esi shr edi, 10 ; edi := x >> 10 xor eax, edi mov [var], al mov eax, 4 ; sys_write mov ebx, 1 ; stdout mov ecx, var ; buffer mov edx, 1 ; length int 80h ; syscall inc esi jmp .loop
### Basic Sierpinski harmony bytebeat t & t >> 8 in as few bytes as possible: .globl _start _start: inc %ebx # output fd 1 in %ebx for write() inc %edx # byte count of 1 in %edx for write() loop: inc %eax # increment t each time through the loop push %eax # save it on the stack and %ah, %al # compute t & t >> 8, our output sample push %eax # store it on the stack for write() lea 3(%ebx), %eax # a three-byte way to set %eax to 4 (__NR_write) mov %esp, %ecx # pass sample pointer to write() (little-endian!) int $0x80 # invoke system call pop %eax # discard sample pop %eax # restore t into %eax jmp loop # and repeat ### Kragen Javier Sitaker ### 2012-01-10
Compile with gcc -m32 -nostdlib and pipe the output to aplay. After stripping, the ELF is 240 bytes; size(1) says it has 18 bytes of code.
bytebeat in <140-character command lines with python
From @wiretapped:
python -c'import sys;[sys.stdout.write(chr(( t^t>>(t>>11)%3^t>>(t>>12)%4 )%256)) for t in xrange(2**19)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t>>(t>>11)%4^(t>>10)*(t>>15) )%256)) for t in xrange(2**19)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t<<3+(t>>10)%3^t>>4+(t>>12)%4^t**(2+(t>>13)%8) )%256)) for t in xrange(2**18)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t^t>>4^(t*((t>>(11+(t>>16)%3))%16))^t*3 )%256)) for t in xrange(2**20)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t+(t>>(5+(t>>10)%4)^t) )%256)) for t in xrange(2**21)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t+(t>>(5+(t>>10)%8))^t|t>>12 )%256)) for t in xrange(2**21)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t|(((t>>1)%(8+(t>>14)%4))+(t>>6)) )%256)) for t in xrange(2**18)]'|aplay python -c'import sys;[sys.stdout.write(chr(( t*(3+(t>>10)%(4+(t>>11)%8))|(t>>5) )%256)) for t in xrange(2**20)]'|aplay
From @isislovecruft (#FYB):
python -c'import sys;[sys.stdout.write(chr(((~t>>2)*(2+(42&t*((7&t>>10)*2))<(24&t*((3&t>>14)+2))))%256))for t in xrange(2**19)]'|aplay python -c'import sys;[sys.stdout.write(chr((((t*5&t>>7|t*9&t>>4|t*18&t/1024)|((t|7)>>5|(t|4)>>9))) %256))for t in xrange(2**18)]'|aplay python -c'import sys;[sys.stdout.write(chr((((t*(t>>13|t>>8))|(t>>16)^t)-64)%256)) for t in xrange(2**18)]'|aplay
for OSX
brew install sox python -c'import sys; [sys.stdout.write(chr(((t*5&t>>7|t*9&t>>4|t*18&t/1024)|((t|7)>>5|(t|4)>>9))%256))for t in xrange(2**18)]' > out.raw; sox -r 8000 -c 1 -t u8 out.raw out.wav; open out.wav
an attempt with luajit
local bit = require'bit' local ffi = require'ffi' ffi.cdef[[ int putchar(int c); ]] local putc = ffi.C.putchar local rsh = bit.rshift local lsh = bit.lshift local band = bit.band local bor = bit.bor local bxor = bit.bxor for t=0, math.huge do putc( band(t, rsh(t, 8)) ) end
run it on osx:
luajit 8bit.lua | sox -r 8000 -c 1 -t u8 - -d
busybox awk
might work with gawk, dunno.
throw it to /dev/dsp
#!/usr/bin/awk -f BEGIN{ for (t=0;;t++){ printf("%c",and(t, rshift(t, 8))); } }
of course awk will eat your processor alive.
alternately from your shell
seq 11111111 | awk '{printf("%c",and($1,rshift($1,8)))}' > /dev/dsp
should also summon cthulhu
http://sebsauvage.net/paste/?75bee7fafb41fe22#QhBl0RBbPKZyp8OGyrozxdDFEjK7B1+5NCTWc7c0vXE=
With Haskell!
links
- Today's formulaic music. - Metafilter post about bytebeat with many links
- Recordings made at Noisebridge: [1] [2] [3] [4]
- https://github.com/kragen/pytebeat - Python Bytebeat livecoding software in SDL
- http://music.columbia.edu/~douglas/strange_things/?p=78 - Algorithmic beat rotation with the EchoNest API