Sampler issue (audio crack) at loop point

[La Tenaille]

Hi,
When I record a drone on the fly with a sampler, I've an audio crack at the loop point. I've tried the zero crossing option but it doesn't solve the problem (what does this option ? I haven't found it in the documentation)
I can imagine why it cracks but is there a way to avoid this ?
Thanks

[La Tenaille]

Up

[nay-seven]

this option is to avoid clicks when the start & end parameters are modified.you can turn it off when recording.
About drone, not so easy to automatically find the zero, if you know the length of your drone, you can first record one, zoom on the start and end of the waveform and create fades with the volume curve, now save this patch to record your drones ?

[La Tenaille]

Thanks Yann,
I'm not using the sampler that way. Music is recorded "live" and samplers give the sync. I don't have time to modify the sample waveform while playing live !
I've read that in Mobius looper soft, the sampler pre and post record a few msec, then apply a crossfade on both to have a smooth - imperceptible - connection.
I imagine all the live loopers tools have a trick to avoid cracks at the loop point...

[nay-seven]

yes, i ve understand the live option, that why i suggest to prepare a patch, recorder with the volume curve prepared that you can overwrite in Live
but i admit it's tricky

[La Tenaille]

Do you think there's a tool in Usine to create a short audio buffer ?
I've read a few forum messages on looper sites, and they all talk about buffering audio to create seamless audio loops.
I wonder if I could create a workaround, but I can't think about a solution without buffer...

[23fx23]

you could maybe just tigg a one buffer fade (ie multiplying audio by the output of ramp module) at the concerned time

or if wanna perfect crossfade withoud fadeout think would work to store the audio array when clic rec (just pass it to an array via a pass module), do the same for the ending array on end rec button clic that will also trigg the to crossfade from current buffer to first stored one, and will do so at each loop pulse.
crossfade buffer being (ramp*bufferStart) + ((1-ramp)*CurrentBufferAtEnd))

[La Tenaille]

Perfect crossfade ? Of course I want
Your suggestion is above my skills, I'm not familiar with arrays, I didn't even know I could send audio to an array. Which array module would you use to store audio ? display or set ?

[23fx23]

yes in fact audio is an array of asio buffer size, yes i was thinking of display or set. i was trying to patch that to give ya a made one but i still get a click for now mmm i ll let ya now if i manage to get a workable solution.
what is your record way, freeloop or on sync?

[oli_lab]

could it be something like that ?

http://www.sensomusic.com/forums/upload ... er%201.wkp

[23fx23]

yes does the trick, nice one oli!

[La Tenaille]

Thanks for the patch oli. I've tried it with an oscillator, but the "pop" is still here at the loop point. It seems the only solution would be a crossfade starting before the end of the master loop.

23fx23 : My loops are freeloops, but in sync : I set a master sampler that trigs the master sync (tempo calculated from its duration), and the other samplers are then synced with quantize modules (rec, play, mute, restart...)

[oli_lab]

I suggest that the pop thing comes from the way you sync your slave samplers : forcing them no matter what to restart will create pop

just maybe

I be without internet an Usine for the next 45 days, so no more answer from me until then.

cheers

olivar

[La Tenaille]

lucky you are
I tried with your wkp unchanged.

[oli_lab]

hum, do you mean you get clicks even with my patch ?
could it be some tweaking to do with your setting ?
bigger buffer like 256kB
faster (or slower) refresh rate
the famous IRQ priority trick http://www.sensomusic.org/wiki3/doku.ph ... q_priority

[La Tenaille]

Yes : opening, sending oscillator from device.
I'll try again tomorrow but I've never had troubles with my gear since years...

[sephult]

So as far as you loop popping, does this happen in every instance of recording? Like everytime you record or only in specific instrumentation?
Do you have a recorded file, that I can load into one of these loopers to see if I hear the popping here as well?
I can also examine the waveform in something else like Wavelab and see if there is a huge difference that would cause the pop in the wav.

This has me curious. I was messing around with my SLR8x the other day and was expecting to hear loop pops, but I was not hearing them.

[sephult]

Now, with the crossfade idea, couldn't an ADSR method be used?
Like trigger the attack with the sample to minimize a pop.

-s