@bocanegra said:

Sooo, best way is to add the offset as a fraction (offset in samples / total samples) then run the ramp through a [wrap~] before you scale it up with the multiplier. Like this:

[phasor~][/~] <- offset/length
|/
[+~]

Hi @bocanegra - just wondering, what kind of a number would i be putting into [/~] ? if the array is 65536, then should i divide (and it appears to be an audio-divide) the incoming 0...1 numberspace (coming from my midicontroller knob)?

this is the bit that i'm most confused about.

@whale-av said:

It is very hard for us to test, communicate or make corrections unless you upload your patch...... and extremely hard to spot any errors.

hi!
plom.pd

here is the current state. if i have a slider going from 0 to 1 (replacing [r p$1-offset], and twist a knob (on midicontroller) or insert a slider there, i will hear loud sounds. you can load a sample by clicking on the (L) and then set the numberbox underneath it to be 1 (or 2 or whatever). then just replace anything after [*~ 1] as [dac~ 1 2]
anything to the right can be fully ignored.

@esaruoho Just change [line] to [line~] in your patch (post 18 of this thread I think), and set all the timings to 5ms, and remove the 1 from the [*~ 1] which is just a volume control at the end of the sampler that will duck the audio as the click occurs.
The [1 118( message is far too long, and that is why you hear the sound drop out.
If you upload plom.pd here I can correct it and post it again for you.
David.

@whale-av said:

@esaruoho Just change [line] to [line~] in your patch (post 18 of this thread I think), and set all the timings to 5ms, and remove the 1 from the [*~ 1] which is just a volume control at the end of the sampler that will duck the audio as the click occurs.
The [1 118( message is far too long, and that is why you hear the sound drop out.
David.

ok, i moved to below 10ms but this is what i hear
resample 0008 [2018-10-14 184954].mp3

and patch looks like

@esaruoho The output from [r p$1-offset] is going to the [+~] after [phasor]........ before [line~] takes the volume down. You need to delay that data with [pipe 4] so that the volume is at zero as the change is made.
David.

@whale-av said:

@esaruoho The output from [r p$1-offset] is going to the [+~] after [phasor] before line takes the volume down. You need to delay that data with [pipe 4] so that the volume is at zero as the change is made.

hi! i added the [pipe 4] and before that i was messing around with modifying the message data amounts and delay amounts until i got something that felt like there was no dropouts:

so this was with a regular signal going into 0 1( and a delay 10 signal going into 1 30(

i added pipe 4 and it seemed to help. if i take 0 1( and increase it to 0 5( i'll start hearing audible clicks. hmm. this is really getting there, so thanks for your help!

looks like delay 15 to 1 5( and regular signal going into 0 1( and the actual offset-data going through pipe 4 helps. I'll see how this sounds like. i'm hearing drops, which is fine, but no crackles, which is great.

i'm not sure how to rejig this so that the drop would be as minimal as possible, but there would be no stuttery clicks. what's the recipe, please?

@esaruoho Yes, if you go to 0 in 5 ((... [0 5(.... )) with a [pipe 4] then the offset message is sent after 4 msec..... and the [line~] needs another 1msec to get to zero........ because you sent it the message... go to 0 in 5msecs.
David.

@whale-av said:

remove the 1 from the [*~ 1] which is just a volume control at the end of the sampler that will duck the audio as the click occurs.

thing is, if i don't set it to 1, the sampler will not play until a knob has been turned.. so i'm trying to figure out how to force-set that to 1 without using *~ 1 which i normally use for setting volume to 1 by default. i even tried loadbang->bng->1( -> *~ input 2 but that didn't seem to work either

@esaruoho I didn't think of that.....!......
Put the loadbang to a message [1( sent into the left inlet of the [line~]

Or maybe you can leave the 1 in your version of Pd. In Pd Vanilla the [*~ 1] will not let you connect an audio signal to the right inlet, only a control signal.
David.

@whale-av said:

@esaruoho I didn't think of that.....!......
Put the loadbang to a message [1( sent into the left inlet of the [line~]

Or maybe you can leave the 1 in your version of Pd. In Pd Vanilla the [*~ 1] will not let you connect an audio signal to the right inlet, only a control signal.
David.

if i have [*~ 1] on purrdata, i cannot connect a line~ there. i can, however, connect a line there.maybe i should try regular non~ line again?

@esaruoho Sorry, long phone call.

That's what I thought (about [*~ 1]. You could try [line 1 1] which will start with a value of 1, and when you send it a message it will "jump" every ms...... but I think you will still get clicks as the level change will be a "staircase" with 5 steps.
[line~] and the loadbang will be better.
David.

@esaruoho said:

Hi @bocanegra - just wondering, what kind of a number would i be putting into [/~] ? if the array is 65536, then should i divide (and it appears to be an audio-divide) the incoming 0...1 numberspace (coming from my midicontroller knob)?

this is the bit that i'm most confused about.

Me I'd use a controller ranging from 0 to full sample length, so I can pick the exact sample offset, then divide by full sample length (to get a float between zero and one) before adding it to the phasors signal. But if you're fine with a 0 to 1 controller just forget about the [/ ] object