simioliolio

sorry if this is a repeat of a previous problem.

i am making a live input for my patch dj64, a djing utility for the monome 64. the patch contains some time-based effects which can be applied to audio loaded into the patch (stutter, granular speed and pitch shifting, random granulator and scratch emulation). i have re-written these effects so they can be applied to live incoming audio (being careful to avoid trying to obtain audio that has not occured yet!)

the effects rely on audio being played using the wavetable ramping method (the playback position being a value between 0 and 1, and obtaining a value from a 'wavetable' every audio clock tick using an accumulator/ramp)

incoming audio is written to a buffer~ using record~, and using the record~ sync output, audio is simultaneously read from a point in the buffer~ 3 samples behind the record position (to ensure the audio data is written before it is read).

i noticed when comparing the incoming audio to the audio received from the live input wavetable, it seems there is a slight bit of audio degradation when the audio undergoes this process. in theory, there should be no change in the audio signal, it is written to a wavetable and read back from the wavetable, and although the two processes occur in quick succession, there is a slight fuzziness at HF, and a slight loss of detail to the audio. the audio is made ever so slightly fuzzy.

i would like to know 1) if i am going mad, and there is no difference 2) does my machine not like this process? 3) is this a good method, and if not, is there a better method?

the example patch was made in max 4.6.3. thankyou for reading this post.

i would like to know 1) if i am going mad, and there is no difference 2) does my machine not like this process? 3) is this a good method, and if not, is there a better method? 

write-and-simultaneous-read-to-buffer-causes-audio-degradation

sorry about lack of agate, wasn't thinking. a zip of the replacement patch is attached. i have used max 5 this time.

i chose 3 samples because i tried 1, and got problems, and tried 2, and got problems, then tried 3, and all was fine. the effects i am using throw the playhead around quite a bit, but 3 samples delay was enough to prevent any rounding errors etc from pushing the playhead too far forwards and getting a nasty cranky noise.

i have tried the interpolation settings on wave~, and they make no difference to my problem.

stefantiedje suggested i used elay~, which i completely overlooked. this solves the problem of reading/writing in the same signal vector.

i have replicated my problem even if there is a long delay (ie 1 second), so i don't think my problem has anything to do with this. i recorded my output from groove using record~, then played it back using another groove object along side the current groove box playing the original file, and there was little if no difference, so i think the problem lies with either record~'s sync output, or wave~'s inability to replay sound from a buffer~ correctly. ??

@Patrick
i chose 3 samples because i tried 1, and got problems, and tried 2, and got problems, then tried 3, and all was fine. the effects i am using throw the playhead around quite a bit, but 3 samples delay was enough to prevent any rounding errors etc from pushing the playhead too far forwards and getting a nasty cranky noise.
i have tried the interpolation settings on wave~, and they make no difference to my problem.
stefantiedje suggested i used elay~, which i completely overlooked. this solves the problem of reading/writing in the same signal vector.

@stefantiedje
thanks for the tip on delay~. 

i have replicated my problem even if there is a long delay (ie 1 second), so i don't think my problem has anything to do with this. i recorded my output from groove using record~, then played it back using another groove object along side the current groove box playing the original file, and there was little if no difference, so i think the problem lies with either record~'s sync output, or wave~'s inability to replay sound from a buffer~ correctly.  ??


"i noticed when comparing the incoming audio to the audio received from the live input wavetable, it seems there is a slight bit of audio degradation when the audio undergoes this process. in theory, there should be no change in the audio signal, it is written to a wavetable and read back from the wavetable, and although the two processes occur in quick succession, there is a slight fuzziness at HF, and a slight loss of detail to the audio. the audio is made ever so slightly fuzzy."

play a tune that you know which has a lot of detail at HF, and you will notice there is a slight loss of detail. if you dont notice anything, then there is a problem my end. if you do notice something, then is it just a slight problem with the wave~ / record~ object?

write and simultaneous read to buffer causes audio degradation