Hi, I'm wondering how to go about doing this:

1. When .pd script is loaded, a bunch of waveloops (4/4, 126bpm) are loaded into the memory and their playback is started
2. at the same time, a stepsequencer is started that starts playing a 4/4 126bpm (119.048ms metronome) 32 step sequence (in pd) with a bassdrum on every 4th.

Should I have the second step be delayed by the amount of time it takes to load the samples in (is there a way to calculate that in milliseconds), or, should I instead actually toggle the sample loop to play by a pd-stepsequencer, within pd, so that they will "line up"? There's a whole trainwreck going on with this one and I'm not sure how to go about solving it.

Any advice would be much appreciated!
(vanillaPD because libpd because iOS..)

how are you loading the samples? if you're using [soundfiler], then the outlet will give you the arraysize when the file is loaded. You can ignore the value and use it to trigger a bang (with [t b] or whatever), if you want.

@mod said:

how are you loading the samples? if you're using [soundfiler], then the outlet will give you the arraysize when the file is loaded. You can ignore the value and use it to trigger a bang (with [t b] or whatever), if you want.

I'm loading 7 loops of audio, 667200 is the maximum filesize of them (5 of them are 667200, the 2 others are 33600 (according to soundfiler).
I guess instead of using loadbang to start the step-sequencer, I could have a 667200 loader do the [t b] and have that start the stepsequencer. would that make sense?

Yeah, [t b] from the outlet of the last file you load.

@mod said:

Yeah, [t b] from the outlet of the last file you load.

Hi, if they are of the same length, should I somehow actually make it so that the loading of the files is chained together? i.e., that once one wavefile is loaded, that triggers the loading of the next? Because, they're loaded with the same loadbang at the start..

there are dozens of ways in which you could enforce order here.... i assume you're already looking into it and have found one.

let us know which one you found.

if you get really stuck, i'll tell you what i'd do, but i kind of feel in this situation, you're best off looking for the solution yourself. You'll probably get some 'aha' moment that will help you from now on with pd, too.

let us know how you go

@mod said:

there are dozens of ways in which you could enforce order here.... i assume you're already looking into it and have found one.

let us know which one you found.

if you get really stuck, i'll tell you what i'd do, but i kind of feel in this situation, you're best off looking for the solution yourself. You'll probably get some 'aha' moment that will help you from now on with pd, too.

let us know how you go

I think that so far I'm not entirely sure how to go about doing this - for, you see, every wavefile that I load, I load with a specific .pd file. such as:

msx~ wave.wav
msx~ wave3.wav
sfx~ wave4.wav

so since they are all the same pd script, I'm not entirely certain how to have soundfiler deliver me the results that I need to load a second one in. It'd be extremely easy (err, well, compared to this current setup) to do this if I just had a singular pd script that loaded the wavefiles into specific arrays, instead of the way I have it set up right now.

...ahh

are you using [loadbang] inside those abstractions? in that case, here's a simple workaround... remove the loadbangs, and chain the abstractions in series:

[loadbang]
|
[msx~ wave.wav]
|
[msx~ wave3.wav]
|
[sfx~ wave4.wav]

then, inside each abstraction, you have something like this:

[inlet]
|
[symbol $1]
|
[read -resize $1 arrayname(
|
[soundfiler]
|
[t b f]
| |
| [(outputs array size)
|
[outlet]

that way you enforce the order and also keep your preferred method

@mod said:

...ahh
are you using [loadbang] inside those abstractions?

I was. I now modified sfx~.pd and msx~.pd to have a third inlet, which receives bangs to cause them to load the audio.

I started with a
loadbang
|
s startload

then I went through 6 different pd abstractions and modified them to talk to eachother.
So, basically

r startload
|
s bangfromsfx1
|
r bangfromsfx1
|
s bangfromsfx2

and so on until it was time to start up the stepsequencer. I think now the stepsequencer only starts playing when the samples have been loaded. However, seems that the way the msx loader is done, it starts playback immediately. I've attached msx~.pd as an example. (All 7 instances of msx~.pd are put to play immediately, so can switch between the 7 "versions" of the same song without the tempo being hindered).
Interestingly, the 7 loops still play "on time", but the stepsequencer doesn't start when "it's supposed to". I guess this is related to the music-players starting playback immediately on initialize.

http://www.pdpatchrepo.info/hurleur/msx~.pd

Should I make it so that "r forward" is only banged once all the samples are loaded? is that it?

yes, BUT you also have that [/ 1e+07] object coming straight from the soundfiler into the hot inlet of that number box, so that will send that value directly, even without the trigger from [r forward]. so, you should replace that number box with a [f ] object, and send the output of [/ 1e+07] to the cold inlet, so it gets stored until the [r forward] receives a bang.

also, generally it's a bad idea to use GUI objects (number boxes, sliders, GUI arrays, etc) in abstractions unless they are explicitly needed for visual feedback. You have all those number boxes and stuff as part of your patch, and that's not a real good idea. It slows things down, wastes cpu...and also in this case, you actually NEED proper objects with cold inlets.

hey, if ALL your files are going to play at the same speed, then here's my suggestion:

only run ONE phasor~, outside of all the individual abstractions, to a [s~ myphasor] object.

Then, inside each abstraction, use [r~ myphasor] before the [+~ ] scaling.

that way, you only have ONE phasor~ to worry about syncing, and as long as you trigger its frequency after all files are loaded, you should be good to go

sup, i wound up switching to having all the audiofiles be toggled by stepsequencers - switching from using msx~.pd to sfx2~.pd made the audiofiles play at the exact right tempo (instead of slightly, subtly detuned lower, which was (i'm guessing) the probable reason for the stepsequenced bassdrums starting to fall out of sync all the time).

anyway, now it works and since i chained all the loading of the loops and the drums together and for the stepsequencers to start playback only after all the wavefiles were loaded, it stays in sync too. so far this seems like the working solutions.

thanks for all the suggestions, mod! i think it was the audio-loop-player that was the main issue, but after i replaced it with a regular one-shot-player, all of your advice came to play and everything turned out alright!