UVI Falcon - v4 = 2026 released - rumors, ads, praise, kindergarden, auto-sampling and off-topic inside!

[ErikH]

- Falcon standalone
- Falcon Factory > Granular > Granular E-bow
- 3 voices
- 246ms grain size
- 288 latency
- Win7Pro, i7 2600K @ 3.4GHz

play C1
3 voices (in osc, bottom right indicates 2) no problems CPU use 9%
6 voices (in osc, does not update bottom-right) no problems CPU use 13%

no problems
Period or Grain size makes no difference

I tried more extreme settings of about everything and do not seem to be able to reproduce your problem.
Or I misunderstood something.

[c_voltage]

- Falcon standalone
- Falcon Factory > Granular > Granular E-bow
- 3 voices
- 246ms grain size
- 288 latency
- Win7Pro, i7 2600K @ 3.4GHz

play C1
3 voices (in osc, bottom right indicates 2) no problems CPU use 9%
6 voices (in osc, does not update bottom-right) no problems CPU use 13%

no problems
Period or Grain size makes no difference

I tried more extreme settings of about everything and do not seem to be able to reproduce your problem.
Or I misunderstood something.

Hm, yes, really, i tried just now, in standalone Falcon i get almost the same result as you describe (but for just in case - C-1 (C minus 1), not C1, well nevermind, not significant difference).
BUT despite native Falcon CPU meter in standalone mode are show those small values, nonetheless i hear the distinctive subtle crackling while click to note (in Granular E-bow)
And similar behavior and in Ableton BTW (during no large cpu usage amount in his native meter).

Well, then it turns out that needed user exactly with Studio One (any version), or at least need test this behavior only in plugin (not standalone).
Will need be test in Bidule for just in case.

Erik in which DAW you can test it ?

[newuser2014]

i hear the distinctive subtle crackling while click to note (in Granular E-bow)

There is a waveshaper in Granular e-Bow that will effect a definite crackle

[c_voltage]

i hear the distinctive subtle crackling while click to note (in Granular E-bow)

There is a waveshaper in Granular e-Bow that will effect a definite crackle

No, its exactly not fx-related crackles. At first because it is mostly reproduced exactly at the moment of clicking to note (i.e. coincides together with cpu activity moment), Initially i even thought that this something related with phase when start couple play markers together, or modulation for marker moments, but when i increased buffer size a much (to 1024 and next) then this "noise" begins to disappear.

[c_voltage]

I tried just now "Granular E-Bow" in Bidule, well occur exactly what about im talking - CPU usage bidule meter show about 9-12%, but in the same time show "CPU overload" message.
I.e. as i understand, occur very fast cpu peak, which not all cpu daw meters can catch for display.

[Sampleconstruct]

I tried just now "Granular E-Bow" in Bidule, well occur exactly what about im talking - CPU usage bidule meter show about 9-12%, but in the same time show "CPU overload" message.
I.e. as i understand, occur very fast cpu peak, which not all cpu daw meters can catch for display.

Well, no fast CPU spikes here, please accept that this issue is not a common one, but somehow related to your setup/computer/DAW/OS whatever.

[c_voltage]

I tried just now "Granular E-Bow" in Bidule, well occur exactly what about im talking - CPU usage bidule meter show about 9-12%, but in the same time show "CPU overload" message.
I.e. as i understand, occur very fast cpu peak, which not all cpu daw meters can catch for display.

Well, no fast CPU spikes here, please accept that this issue is not a common one, but somehow related to your setup/computer/DAW/OS whatever.

In general, yes, but if (if would) this symptom is confirmed by other users, then this is can be as certain good info for optimize Falcon device perfomance. In any case ofcourse I can not call it as a bug.

[evilantal]

I tried just now "Granular E-Bow" in Bidule, well occur exactly what about im talking - CPU usage bidule meter show about 9-12%, but in the same time show "CPU overload" message.
I.e. as i understand, occur very fast cpu peak, which not all cpu daw meters can catch for display.

Well, no fast CPU spikes here, please accept that this issue is not a common one, but somehow related to your setup/computer/DAW/OS whatever.

In general, yes, but if (if would) this symptom is confirmed by other users, then this is can be as certain good info for optimize Falcon device perfomance. In any case ofcourse I can not call it as a bug.

Are you noticing this as slight "glitches" when you strike notes?
In that case, I'm noticing that too with this patch. Earlier Simon thought it might be the drive/distortion I was hearing. But I don't think that's it. It sounds like an ever so slight click/glitch when striking notes. Not when they are sustaining.

I don't see any spikes in the CPU though.
Strange thing is. I only seem to hear it when playing it live. The clicking is not audible in a render (from Reaper). It also doesn't disappear when I increase the buffer size of my interface.

[c_voltage]

(I'm sorry, just now woke up)

Are you noticing this as slight "glitches" when you strike notes?

You mean, exactly in moment of click (i.e. press) on note ? Yes, i told it.

I don't see any spikes in the CPU though.Strange thing is. I only seem to hear it when playing it live. The clicking is not audible in a render (from Reaper). It also doesn't disappear when I increase the buffer size of my interface.

Thr fact that this cannot be reproduced in render wav (but hear only in realtime) just proves that this is cpu related behavior. Though i write probably obvious things, but let it be gor just in case.
So, about buffer increase, again - as i wrote above, for me it help, BUT while extremely increased bufer, just now test again for clarifi (standalone falcon, gran.e-bow) - for me this is value 1024, then without "start ctackles". But in 512 it already appear, although quite subtle.

[c_voltage]

Else one moment, about Start Phase knob in WT. Folks or Olivier, can you describe me any circumstance (example) in which changes of this Start Phase knob may produce an audible result ?

[Sampleconstruct]

(I'm sorry, just now woke up)

Are you noticing this as slight "glitches" when you strike notes?

You mean, exactly in moment of click (i.e. press) on note ? Yes, i told it.

I don't see any spikes in the CPU though.Strange thing is. I only seem to hear it when playing it live. The clicking is not audible in a render (from Reaper). It also doesn't disappear when I increase the buffer size of my interface.

Thr fact that this cannot be reproduced in render wav (but hear only in realtime) just proves that this is cpu related behavior. Though i write probably obvious things, but let it be gor just in case.
So, about buffer increase, again - as i wrote above, for me it help, BUT while extremely increased bufer, just now test again for clarifi (standalone falcon, gran.e-bow) - for me this is value 1024, then without "start ctackles". But in 512 it already appear, although quite subtle.

Ok guys, I have found the problem, the factory patch for Granular Ebow is simply wrong, I was testing my initial preset which I sent to UVI and did not load the preset from the factory bank. The problem is, that the layer is set to 2 unison voices, which is the reason for the CPU spikes. Falcon (and Mach Five) don't like being run with several unison voices in granular mode, especially not the new multi-granular engine, so the voices get doubled and the CPU spikes. Please set the layer to 1 unison voice and all will be good.

And it's actually not the same as using like e.g. 6 grain streams in the multi-granular engine, CPU gets even more than doubled by using 2 unison voices. Problem solved, I'm glad I checked again.

[ErikH]

Indeed, I see it is set to two unison voices. But I have no spikes in cpu use. And I do not hear a difference when I make it only one voice. Or three. At six voices, I get some audible overload-problems, and my cpu value (Falcon) rises from some 9 to some 12.

Yet, when I tried to switch from my latency setting, Falcon froze and cannot be started up again. The Ircam bug they are trying to nail down I guess.

[c_voltage]

(I'm sorry, just now woke up)

Are you noticing this as slight "glitches" when you strike notes?

You mean, exactly in moment of click (i.e. press) on note ? Yes, i told it.

I don't see any spikes in the CPU though.Strange thing is. I only seem to hear it when playing it live. The clicking is not audible in a render (from Reaper). It also doesn't disappear when I increase the buffer size of my interface.

Thr fact that this cannot be reproduced in render wav (but hear only in realtime) just proves that this is cpu related behavior. Though i write probably obvious things, but let it be gor just in case.
So, about buffer increase, again - as i wrote above, for me it help, BUT while extremely increased bufer, just now test again for clarifi (standalone falcon, gran.e-bow) - for me this is value 1024, then without "start ctackles". But in 512 it already appear, although quite subtle.

Ok guys, I have found the problem, the factory patch for Granular Ebow is simply wrong, I was testing my initial preset which I sent to UVI and did not load the preset from the factory bank. The problem is, that the layer is set to 2 unison voices, which is the reason for the CPU spikes. Falcon (and Mach Five) don't like being run with several unison voices in granular mode, especially not the new multi-granular engine, so the voices get doubled and the CPU spikes. Please set the layer to 1 unison voice and all will be good.

And it's actually not the same as using like e.g. 6 grain streams in the multi-granular engine, CPU gets even more than doubled by using 2 unison voices. Problem solved, I'm glad I checked again.

Hm, thanks, tried just now set to unisson 1, good finding, and really reduces cpu usage and for me, though cpu spike still occur, but a much less (not 100%), and no occur crackle, but unfortunately in my case it was be just coincidence that i mention about this preset, and he was not quite correctly configured.

I mean the problem with click \ cpu spike in multi granular, about what there is a speech (during specific config what i mentioned) are reproduced not only in "granular e-bow", but in most part of Granular factory patches (in the lower octaves). I.e. in presets in which settings coincide with those osc config what i mentioned.
(Well, this observation at least for me, again).

[Sampleconstruct]

Indeed, I see it is set to two unison voices. But I have no spikes in cpu use. And I do not hear a difference when I make it only one voice. Or three. At six voices, I get some audible overload-problems, and my cpu value (Falcon) rises from some 9 to some 12.

Yet, when I tried to switch from my latency setting, Falcon froze and cannot be started up again. The Ircam bug they are trying to nail down I guess.

I also get no spikes with 2 unison voices (or 3, but this mac is a monster), but I remember that on my old MacPro I struggled for months trying to find the issue until I found out that more than 1 unison voice will create CPU problems in relation to granular mode on older/slower computers. As far as I have understood, the IRCAM engines are already highly optimized and the CPU is what it is now, there is not much to expect on the optimization front, so getting a faster computer will be the best advice to give for Falconists struggling with the IRCAM engines.

[c_voltage]

IRCAM engines are already highly optimized

I would not say