I need help writing a conkyrc file. It (should) call a lua script, which reads a FIFO pipe of PCM data generated by MPD. This FIFO is used to generate visualization in NCMPCPP but I'd rather it go to a bar graph on my desktop. And yeah, I'm aware that it would need to be updated pretty fast in order to look good. I'm hoping that instead of a full spectral analysis, I can do a simpler visualization that doesn't require an expensive FFT or wavlets... just something I can pipe into a graph showing activity of music at the moment.
[edit] Progress made, albeit with a lot of fudge factor...
my conkyrc file
lua_load /home/adam/Programming/myWork/conky/mpd.lua
update_interval .05
TEXT
HERP DEE DERP DEE DERP DEE DUUUR
${lua_bar fifo_func}
My lua file
do
-- configuration
local interval = 5
-- local variables protected from the evil outside world
local next_update
local buf
local int = 0
local colour = 0
local function update_buf()
buf = os.time()
end
local f = assert(io.open("/tmp/mpd.fifo", "rb"))
local block = 2048 * 2 --2048 samples, 2 bytes per sample
local list = {}
function conky_fifo_func()
local bytes = f:read(block) --read a sample of block bytes
local power = 0
for i = 0, 2047 do
--j = string.byte(bytes, 2*i, 2*i+1) --extract 2 bytes
j = string.format("%u", string.byte(bytes, i*2, i*2+1))
power = power + math.abs(j-(256/2))
--io.write(j..'\n')
end
r=((power/10000)-20) * 15
io.write(r .. '\n')
return r
end
-- returns a percentage value that loops around
function conky_int_func()
int = int + 1
return int % 100
end
end
Based on the NCMPCPP source code, the FIFO is an array of 2048 16bit ints