Hi all,

I built a patch that uses phasor to read from a table - quite simple really.
Everything works fine when I "bang" it myself - all notes sound the same, phasor syncs to my bang. If I connect a [key] to the bang I play the patch from my keyboard and it's working: all notes sound the same, phasor syncs no matter how fast I bang it.

BUT - when connected to a [metro], each note sounds different: it seems like phasor is syncing a bit differently each time!

Got into it a little bit, and here's what I found:

There is a correlation between the frequency of the [metro] and the number of different sounds I'm getting. 500ms produces two alternating sounds. 450ms produces a series of 13(?) different sounds that loop over and over in a cycles. In fact, each [metro] value produces its own repetitive series of sounds.

Strange!

I managed to find a [metro] value that seems to produce a stable sound: 476.01 ms.
Then I found out that half that value and twice that value work as well.
Changing the sample rate to 48KHz creates a dramatic change: every [metro] value that is a multiple of 4 would stabilize the sound.
And at sample rate of 24Khz, only [metro] values that are multiples of 8 produced a steady syncing of [phasor].

OK... So can someone tell me what's going on?

And... is there a better way to achieve a reliable synced reading of waveforms in PD?

Thanks,

ith

well I'm not sure how you're syncing it but it probably has to do with the fact that when you hit a graphical bang the bang only hits on the next dsp block, and when run from a metro they hit in between dsp blocks. Therefore you'll get a form of aliasing at the block size. it's because phasor only allows the phase to be set at block boundaries as well, so what is happening is that when the metro happens to be at an integer multiple of the block size (in ms) everything works out but when it's not it will sync the phase at times that sound odd.
mmb made a phasor that does sync in the middle of a block https://github.com/dotmmb/mmb

Hi Sebfumaster, thanks for your reply!

Your analysis was correct, so I downloaded mmb's [vphsor~] and replaced it in the patch - and it worked!

Thanks so much,

ith

I think you should just reset the phase upon banging, that way phasor really starts at the beginning of it's waveform. Bang a message with a zero into it's second inlet and that should be fine.

@seb-harmonik.ar @ith I actually posted a more accurate [vphasor~] external here if you're interested. My original abstraction was sample-accurate (and vanilla), which is probably fine for most uses. But given that [metro] is sub-sample accurate (meaning it will bang between samples), if you really want to nail the reset perfectly, the external is better. At really high frequencies the difference between the two will likely be a bit more noticeable. At lower frequencies it probably doesn't matter so much, though.