hi all,
I am pretty new in the Axoloti-scene and right now in the process of building a delay -
is there any patch which could be used for like delay times up to 16, 20 seconds (in stereo) ?
any link or help is much appreciated =)
thanks so much in advance!
this is not possible on the axoloti, since you don't have enough ram to store so much audio.
what?
this guy https://www.youtube.com/watch?v=2Q1fnBYn8pQ
wrote me:
""85 seconds in mono into the 8Mb of SD-Ram (in 48kHz/16bit), but the biggest delay time I seem to be able to use is half of that (43 seconds)"
The maximum delay time in mono is 43.7s
here is the "long delay" patch
multiply it by 2 and reduce the size a notch (to 21s) and you have a nice long stereo delay
we could actually make a bit longer delay if we also make tables in normal memory and read these as being one big table. Though, it won't be much..
Also, creating 8bit delays with pre&post filtering is also possible, this will allow delays twice the size.
and top it all off..
I just found out a bool array can't get longer than an 8bit array
Does anyone know why you can't make longer delays when using a boolean array instead of 8bit? To me it would seem that you'ld save 7 bits with every slot of the array...
oeh.. as now I think about it.. I could use the 8bit array as 8 1bit arrays...
so I could make a 1bit version of 8x86 seconds of length or having 8 delay channels of 86s each... XD
proof of concept (1bit PWM, bool array):
I must say, even though it's 1bit PWM synthesis.. it's not even that bad..
this could actually reproduce the stereo 86 seconds (LP-filtered) delay....
proof of concept 8bit length
whahahaha, this is sweet! XD
So, the longest time possible is about 86 seconds for an 8bit array.
I'm using the first 4 bits for the left signal and 4 bits for the right signal.
So... I've got a 350s stereo 1bit delay XD
I was thinking.. 1bit.. surely that's going to sound TERRIBLE...
But you could use it for a trigger-pulse delay. Not sure how useful that would be, mind you..
a|x
not necessarily.
I compare the input with a filtered signal. Whenever the input is higher then the filter signal, it generates a gate high and whenever it is lower, it generates a gate low. This gate high/low signal is then used as the filter-input.
This way the pwm signal is based on the input and when using filters over the delay-output, it gives you back the input signal, though dampened.
Oh, OK. But the Axoloti audio system works at a fixed 48kHZ. Doesn't PWM rely on having a clock running at a large multiple of the carrier frequency. 8bit PWM signals require a clock running at 256x the carrier frequency in order to generate 0-255 possible levels, for each carrier cycle. Or am I completely misunderstanding..(highly likely)?
a|x
If you want to build an extremely long stereo delay you should run the audio through a compressor to decrease its dynamic range, then dither the signal, so that you can decrease the number of bits as much as possible while still retaining whatever minimum standard of quality. After that, you might be able to use some lightweight data compression algorithms (here I mean compression like zip, not nonlinear amplitude response as earlier) to keep the recorded audio a little smaller, though you'd probably have to include some lookahead for the decompression during playback.
I was thinking it would be cool to have say, 4 channels for looping. That would give you 5.25s for 4 stereo channels.
a|x
cool.
