U-he ACE!

Official support for: u-he.com
Post Reply New Topic
RELATED
PRODUCTS

Post

goldenanalog wrote:
pdxindy wrote:Use different stack settings will change that though...
Yes, but in my experience thus far, your 'voice' resources stay the same; ie (1) stack, 8-note poly; (2) stacked, 4-note poly, etc.

-goldenanalog

hmmm... that would mean that with a stack of 4 you could never get more than 4 notes polyphony even on many? That seems limited for pads with some release time when you have draft and standard quality settings to keep CPU in check...

I've never been happy with Zebra's few/medium/many options. I think Urs does not appreciate having to manage cpu resources on lesser machines.

String Studio has 2,3,4,6,8,10 etc and it is much easier to control cpu use. Zebra has 4,8,16 and it is too crude and easy to have too much voice stealing (4 or few) or too much cpu (8 or medium). It is annoying when I have to change what I play to suit that limitation (or thinking about it as I play instead of just playing) when having the option of 6 voices would solve that in most cases.

Fortunately I am getting a faster machine (new intel mac vs old G5) so that annoyance may be a thing of the past for me!

Post

pdxindy wrote:hmmm... that would mean that with a stack of 4 you could never get more than 4 notes polyphony even on many? That seems limited for pads with some release time when you have draft and standard quality settings to keep CPU in check...
I agree with you, pdxindy.
pdxindy wrote:I've never been happy with Zebra's few/medium/many options. I think Urs does not appreciate having to manage cpu resources on lesser machines.

String Studio has 2,3,4,6,8,10 etc and it is much easier to control cpu use. Zebra has 4,8,16 and it is too crude and easy to have too much voice stealing (4 or few) or too much cpu (8 or medium). It is annoying when I have to change what I play to suit that limitation (or thinking about it as I play instead of just playing) when having the option of 6 voices would solve that in most cases.
Again, I agree with you, from what you are saying....I personally don't have any experience with Zebra, but after brief exposure to ACE, I don't doubt that Urs' crude implementation of voice management caused you, and plenty of other regular users, problems.

But what Urs just posted makes sense...it's all math; if you have multiple 'engines' that concurrently 'subroutine' inside of a given vsti, they all have to 'add up' when it comes to available voicing resources coupled with event scheduling. And that can definitely get pretty complicated from a developer's standpoint.
pdxindy wrote:Fortunately I am getting a faster machine (new intel mac vs old G5) so that annoyance may be a thing of the past for me!
That, pdxindy, is the final solution, the last word! :tu: Good for you!

-goldenanalog

Post

thanks Urs

dave

Post

Hi Urs,

I've came across a weird behaviour with the Patch system. When I drag & drop a patch to the MIDI Programs folder sometimes it doesn't get copied, but moved instead :-o Happens e.g. for "TAS Bond" (TASMODIA/01 Bass). I'm on the Windows version of ACE.

While I got your attention: Is it possible to make your patch browsers respond to MIDI Program Change commands 36 & 37 in order to step through the list of patches via MIDI Keyboard?

Thanks!
bM3w

Post

bM3w wrote:Hi Urs,

I've came across a weird behaviour with the Patch system. When I drag & drop a patch to the MIDI Programs folder sometimes it doesn't get copied, but moved instead :-o Happens e.g. for "TAS Bond" (TASMODIA/01 Bass). I'm on the Windows version of ACE.
Normally things should only be moved unless you hold shift.
While I got your attention: Is it possible to make your patch browsers respond to MIDI Program Change commands 36 & 37 in order to step through the list of patches via MIDI Keyboard?
I would love to, but there's still that problem that a plugin must not access the hard drive upon a MIDI message. It would crash almsot certainly. That is why the presets that are in MIDI Programs are loaded at startup and kept in memory.

However, many hosts (especially simple ones like SaviHost) do not intercept all key strokes. Just click on the patch browser once to load a presets and use the up/down keys on your keyboard from then on (until you click something else). This doesn't work with all those greedy hosts though that don't give my plugins key strokes.

Cheers,

;) Urs

Post

Urs wrote:
bM3w wrote:Hi Urs,

I've came across a weird behaviour with the Patch system. When I drag & drop a patch to the MIDI Programs folder sometimes it doesn't get copied, but moved instead :-o Happens e.g. for "TAS Bond" (TASMODIA/01 Bass). I'm on the Windows version of ACE.
Normally things should only be moved unless you hold shift.
Yep, but it does it here without the SHIFT key on random patches :? Can you check on that and confirm?
Urs wrote:
bM3w wrote:While I got your attention: Is it possible to make your patch browsers respond to MIDI Program Change commands 36 & 37 in order to step through the list of patches via MIDI Keyboard?
I would love to, but there's still that problem that a plugin must not access the hard drive upon a MIDI message. It would crash almsot certainly. That is why the presets that are in MIDI Programs are loaded at startup and kept in memory.

However, many hosts (especially simple ones like SaviHost) do not intercept all key strokes. Just click on the patch browser once to load a presets and use the up/down keys on your keyboard from then on (until you click something else). This doesn't work with all those greedy hosts though that don't give my plugins key strokes.
OK, I guess that's the way to go then.

Thanks for your quick reply!
bM3w

Post

bM3w wrote:
Urs wrote:
bM3w wrote:Hi Urs,

I've came across a weird behaviour with the Patch system. When I drag & drop a patch to the MIDI Programs folder sometimes it doesn't get copied, but moved instead :-o Happens e.g. for "TAS Bond" (TASMODIA/01 Bass). I'm on the Windows version of ACE.
Normally things should only be moved unless you hold shift.
Yep, but it does it here without the SHIFT key on random patches :? Can you check on that and confirm?

Thanks for your quick reply!
bM3w
Yes, I reported a very similar thing a couple pages earlier in this thread.

Also, does ACE support arrow keys for patch browsing? They don't work in REAPER, even enabling send of all keys to plugin...

- Mario

Post

Yeah, there seems to be an issue... I'll work on it. Thankfully it isn't too serious though :oops:
mabian wrote:Also, does ACE support arrow keys for patch browsing? They don't work in REAPER, even enabling send of all keys to plugin...
Well, it works in other hosts, so I guess Reaper isn't sending all keys then... try using it in that other separate Window mode...

Post

Hi Urs,
in the past I had some semimodular synths from John Bowen for the Pulsar card.
He has on/off switch for every synth module in his synths.
For example when I only need one VCO then I switch the others off.
So I could reduce the cpu usage.
Would that be an idea for ACE to reduce the cpu usage?

Post

Hello Urs,
in the past I had some semimodular synths from John Bowen for the Pulsar card.
He has on/off switch for every synth module in his synths.
For example when I only need one VCO then I switch the others off.
So I could reduce the cpu usage.
Would that be an idea for ACE to reduce the cpu usage?

Post

Urs wrote:Yeah, there seems to be an issue... I'll work on it. Thankfully it isn't too serious though :oops:
Danke :)

Post

@ Urs

is there a new version that sorts the tap mod bug whilst using multiple instances yet?
im nearing completion of the extras in the factory bank..im just typing out the descriptions today..
if there's a new version then just email it to me..if this is the release version then it's fine im nearly done..


cheers

Rob

Post

Gucky wrote:Hi Urs,
in the past I had some semimodular synths from John Bowen for the Pulsar card.
He has on/off switch for every synth module in his synths.
For example when I only need one VCO then I switch the others off.
So I could reduce the cpu usage.
Would that be an idea for ACE to reduce the cpu usage?
Unfortunately this isn't viable. 90% of the cpu are required for the oscillators and the filters. The oscillators are processed in parallel as far as possible while filters are completely parallelised. Switching one filter off would only improve things if both filters were turned off, and switching an oscillator off or two wouldn't have much of an effect either. In fact I think the option to switch something off would cost more cpu than it saves.

I would rather spend some time in finding other areas of optimisation, I might be able to gain some cycles if I hand code certain areas in assembly, which I might do at some point (most of the code is written using compiler SSE/AltiVec intrinsics, which is something compilers don't optimise so well)

;) Urs

Post

Rob, I'm still working on a fix.

The feedback is just so overwhelming that it's hard to concentrate on coding... I must still have 50+ emails to reply to...

Post

Urs wrote:Rob, I'm still working on a fix.

The feedback is just so overwhelming that it's hard to concentrate on coding... I must still have 50+ emails to reply to...
Ok Urs no probs i'll work with what i have :) not a big issue anyways,just making sure i had the same version these guys have.

cheers

Rob

Post Reply

Return to “u-he”