Can Anyone Try This On There Fusion?

General questions about the fusion belong here.

Moderators: Jesse, parametric, Jesse, parametric

degaffman
General
General
Posts: 132
Joined: Wed Sep 14, 2005 10:43 pm
Location: Mobile, AL

Voice Limitation

Post by degaffman »

I can easily reproduce the problem on my Fusion. Why if we're all running the same software rev can't this problem be reproduced on all Fusions?

Anyone have any idea why the setting the number of voices to 24 (Program/Pitch/Number of Voices) doesn't help in this case? I assumed that this parameter would cause voice stealing, but it doesn't ... the polyphony still rises to about 48, then the CPU gets overloaded (48 voices * 3 ocsillators per voice * mod routings = CPU overload).

Also, I still don't quite understand what the overlap release parameter is? It seems to me that would be if the same note sounds, to cut off the release of the prior note until that much time expires. Thanks!


degaffman
Drocket
General
General
Posts: 57
Joined: Tue Apr 18, 2006 11:33 pm
Location: NSW, Australia

Post by Drocket »

I can reproduce it at will (never tested this patch in this way before). Pretty much happens as everyone describes.

I have a sustain pedal connected but didn't use it and managed to get the noise (actually sounds like static). I tried it with the pedal disconnected same problem.

The patch has not been edited so has the default values.

Sustain Decay: 9.5 s
Release: 158.375 ms
Overlap Release: 619.637 ms

OS: 1.22

Drocket.
User avatar
SEA
Advanced
Advanced
Posts: 455
Joined: Tue Jun 06, 2006 8:36 am

Post by SEA »

Drocket wrote:I can reproduce it at will (never tested this patch in this way before). Pretty much happens as everyone describes.

I have a sustain pedal connected but didn't use it and managed to get the noise (actually sounds like static). I tried it with the pedal disconnected same problem.

The patch has not been edited so has the default values.

Sustain Decay: 9.5 s
Release: 158.375 ms
Overlap Release: 619.637 ms

OS: 1.22

Drocket.
Thanks for testing it out Drocket!

One thing that works very well is when you play a power chord in your right hand like an A major chord right above middle C and the in the left had play the bass octave spread and then with both hands (in a rock style) simply bang away on the keyboard like 12 to 15 hits like 1/8 notes at 140 bpm.

Now as you're doing this you should see the CPU level rise all the way to 95% where you should hear all the clicks & pops like static from the CPU overloading. You don't even need a sustain pedal for it.

So for those who haven't been able to get the CPU to overload on this patch this might help.

Once again the patch is located in [HD: PRESET 5] and it's A-3 (002) Piano & Analog Mood.

The tech guys at Alesis were also able to reproduce it but the first time they said they couldn't, but they had the wrong patch.

Let me know if anyone else can reproduce it. It should be inherent to all Fusions.

SEA 8)
Dare To Be Different... FUSION Image
gwenhwyfaer
Advanced
Advanced
Posts: 387
Joined: Fri Mar 10, 2006 9:19 pm
Location: Not here any more

Post by gwenhwyfaer »

OK... now I can reproduce it... weird.

Even weirder - after I stop "randomly bashing keys", the Fusion continues playing notes for a little while, almost as though it lost track of all the notes I was playing and is going to squeeze them in now whatever happens - or that somehow the main CPU has "got behind" and has dumped all the notes into a queue which it proceeds to service as quickly as it can...

That being the case, any sound with similar characteristics ought to produce similar results...

Also - switching to a fixed number of voices changes nothing. But switching the patch to a monophonic one and playing the same way (think "monkey on speed") has the most curious effect of all... it "loses" the piano sound (although it doesn't pop or crackle at all). And a moment's thought reveals that this is because the sample isn't retriggering... so it's playing out to its full length of 30 seconds, regardless of what the envelopes are doing...

.....

HAH! GOT IT!!!!

Add a modulation route from env 3 to Volume, move it to the top of the list, and that solves the problem! The problem being that without a global Volume parameter, the Fusion is apaprently having some difficulty determining when the voice is "off", and therefore when it's OK to steal it. Clamping the Volume just makes sure.

Does everyone else find this solves the problem?
User avatar
jaggeddoctrine
General
General
Posts: 71
Joined: Mon Apr 24, 2006 8:53 pm

Post by jaggeddoctrine »

I was able to reproduce it. I'm on a 6HD, 75 GB HD, no memory expansion.

Gary
User avatar
GuyDenruyter
Mega
Mega
Posts: 1018
Joined: Fri Jan 13, 2006 9:11 am
Location: Belgium

Post by GuyDenruyter »

gwenhwyfaer wrote:
HAH! GOT IT!!!!

Add a modulation route from env 3 to Volume, move it to the top of the list, and that solves the problem! The problem being that without a global Volume parameter, the Fusion is apaprently having some difficulty determining when the voice is "off", and therefore when it's OK to steal it. Clamping the Volume just makes sure.

Does everyone else find this solves the problem?
Very interesting diagnosis! This would mean that the Fusion does not look at the real amplitude, but rather at the envelope. I imagine this is easier since no need for real-time analysis. I will try myself later today.
User avatar
jaggeddoctrine
General
General
Posts: 71
Joined: Mon Apr 24, 2006 8:53 pm

Post by jaggeddoctrine »

gwenhwyfaer wrote:
Does everyone else find this solves the problem?
Just tried your solution..and it works for me. I couldn't get my CPU above the high 80s. Nice job!

Gary
User avatar
SEA
Advanced
Advanced
Posts: 455
Joined: Tue Jun 06, 2006 8:36 am

Post by SEA »

jaggeddoctrine wrote:
gwenhwyfaer wrote:
Does everyone else find this solves the problem?
Just tried your solution..and it works for me. I couldn't get my CPU above the high 80s. Nice job!

Gary
I'll try it right now (at 7:32 AM EST) and will let everyone know!

SEA
Dare To Be Different... FUSION Image
gwenhwyfaer
Advanced
Advanced
Posts: 387
Joined: Fri Mar 10, 2006 9:19 pm
Location: Not here any more

Post by gwenhwyfaer »

GuyDenruyter wrote:Very interesting diagnosis! This would mean that the Fusion does not look at the real amplitude, but rather at the envelope.
Quite the opposite; it IS looking at the amplitude, otherwise it wouldn't know when to stop the sound at all. But when there's no overall volume control, it has to try and determine it from three individual amplitudes - and I'll put money on there being a "nearly 0 but not quite" issue when adding their current amplitudes together.

Anyway, hopefully we can now lay that particular demon to rest, and it's a useful additional tip for sound designers - "always include a Volume envelope; don't rely on the individual oscillators".
User avatar
SEA
Advanced
Advanced
Posts: 455
Joined: Tue Jun 06, 2006 8:36 am

Post by SEA »

Hey Guys!

I gotta split.. I'll be back in an hour or so and try to do the fix!

I'll keep you posted!

SEA 8)
Dare To Be Different... FUSION Image
User avatar
Hollow Sun
Extreme
Extreme
Posts: 944
Joined: Sat Feb 11, 2006 10:57 am
Location: Wales,UK
Contact:

Post by Hollow Sun »

Well spotted Gwen :)

And you are absolutely correct in what you say - Fusion has a final amplifier that MUST be controlled by an envelope (typically Envelope 1) else ... well - we've seen/heard the result! What's happening in this patch is that some voices are just never getting shut off and the only way they get cut is when other voices steal them (rather inelegantly), hence the clicks.

Of course, that particular patch is compounded by the obscenely long decays and sustain releases and 'ordinary' releases on some of the 'aux' envelopes - reducing those helped but your fix is definitely the one (although simply changing the first mod assignment from Env1 > Osc1 Vol to Env1 > (master) Volume will also fix it effectively).

To be honest, my other suggestion for a fix for this would be ... ermmm ... don't use/play that particular sound!!!!

I've not experienced this coz I only installed Preset 5 for the first time this morning to check it.
User avatar
Luc Henrion
Extreme
Extreme
Posts: 937
Joined: Fri Nov 25, 2005 8:21 am
Location: Belgium
Contact:

Post by Luc Henrion »

congrats guys, this is what I call an instructive forum :wink:
User avatar
Roberto
Extreme
Extreme
Posts: 530
Joined: Sun Feb 05, 2006 11:38 pm
Location: Suffolk, UK
Contact:

Post by Roberto »

Top work Gwen :)
Rob
Hollow Sun
Fusion 6HD|Akai S6000-Z8-MPC1000-MPC500-MPK49|Korg Prophecy|Yamaha CS1x-DX27-RM1x-SU10-DTXpress|Roland SC88-PMA5|Elka X605|Mattell Synsonics|Event EZbus|AKG C444
User avatar
SEA
Advanced
Advanced
Posts: 455
Joined: Tue Jun 06, 2006 8:36 am

Post by SEA »

Hollow Sun wrote: I've not experienced this coz I only installed Preset 5 for the first time this morning to check it.
No wonder you couldn't duplicate it... you didn't even have the patch!! :lol:

Well at least we all know now and I'll definitely contact Steve at Alesis about it soon!

Thanks again Gwen for the fix! :wink:

SEA 8)
Dare To Be Different... FUSION Image
User avatar
SEA
Advanced
Advanced
Posts: 455
Joined: Tue Jun 06, 2006 8:36 am

Post by SEA »

gwenhwyfaer wrote: Add a modulation route from env 3 to Volume, move it to the top of the list, and that solves the problem! The problem being that without a global Volume parameter, the Fusion is apaprently having some difficulty determining when the voice is "off", and therefore when it's OK to steal it. Clamping the Volume just makes sure.

Does everyone else find this solves the problem?
I DID IT!!!! Image

SHE'S-A-FIXED!!!!! Image

THANKS GWEN!

SEA Image
Dare To Be Different... FUSION Image
Post Reply
  • Similar Topics
    Replies
    Views
    Last post