That is good news... please add 32 to all the U-he synths... the 16 is not enough - especially with unison in Bazille, Diva and AceUrs wrote:We're contemplating to allow for 32 or even 64 voice polyphony (not sure if our voice management code needs tweaking)
U-HE Hive
-
- KVRAF
- 26944 posts since 3 Feb, 2005 from in the wilds
-
- KVRAF
- 2990 posts since 13 Apr, 2008 from Charleston, SC
Getting excited here. I have a billion synths (ok just a few thousand, heh) but simple is always good. It is why we go back to 808 for our electronic drums. The sounds and thoughts behind it are timeless.
Will be on lookout for beta.
Jon
Will be on lookout for beta.
Jon
-
- KVRAF
- 1959 posts since 21 Sep, 2007 from The Infinite Void
So am i to understand that "fake unison" is when each oscillator can have multiple voices (eg: Sylenth, Spire) while "true unison" is when the entire synth has multiple stacked voices (eg: Massive) and that "true unison" not only involves x-multiples of every oscillator but also x-multiples of any filters used?
Seems to me then that "fake unison" is kinda better in many ways, aside from less CPU there's the ability to vary the amount of unison for each oscillator. For example in Spire I could make a lead-bass patch with the higher frequency oscillators heavily stacked and detuned but keep the low-end solid and tight with just one or two voices and only slight detune.
In contrast, i'd have to use two or more instances of Massive to achieve the same sort of thing.
Seems to me then that "fake unison" is kinda better in many ways, aside from less CPU there's the ability to vary the amount of unison for each oscillator. For example in Spire I could make a lead-bass patch with the higher frequency oscillators heavily stacked and detuned but keep the low-end solid and tight with just one or two voices and only slight detune.
In contrast, i'd have to use two or more instances of Massive to achieve the same sort of thing.
-
- u-he
- 30192 posts since 8 Aug, 2002 from Berlin
That's correct.mutantdog wrote:So am i to understand that "fake unison" is when each oscillator can have multiple voices (eg: Sylenth, Spire) while "true unison" is when the entire synth has multiple stacked voices (eg: Massive) and that "true unison" not only involves x-multiples of every oscillator but also x-multiples of any filters used?
Essentially "fake unison" saves a lot of resources. Only the oscillators need to be calculated so and so many times while envelopes, filters, modulators, amplifiers, mixers and whatever else there might be in a synth voice is calculated only once.
It's also a drawback because things like filter overdrive often sound harsher with complex inputs. In a true unison setting where the whole signal path is doubled/tripled/multifold you generally get better control over filter drive.
Cheers,
- Urs
-
- u-he
- 30192 posts since 8 Aug, 2002 from Berlin
I'm not sure how this would play out with the multicore option. We might need to change the multicore setup if we offer more than 16 voices. Therefore it wouldn't necessarily be an easy thing to do.pdxindy wrote:That is good news... please add 32 to all the U-he synths... the 16 is not enough - especially with unison in Bazille, Diva and AceUrs wrote:We're contemplating to allow for 32 or even 64 voice polyphony (not sure if our voice management code needs tweaking)
-
- u-he
- 30192 posts since 8 Aug, 2002 from Berlin
Hmmm, one can get unsion in Synth1 but I haven't seen an option to multiply the oscillators? As far as I can see there's ony ever a maximum of 32 oscillators in total, or am I missing something?hakey wrote:Synth1Urs wrote:synths that do super saw (fake unison) as well as voice stacking (true unison) are often considered CPU hogs.
-
- u-he
- 30192 posts since 8 Aug, 2002 from Berlin
Hmm, couldn't find that on the Mac beta… will give it another shot today.hakey wrote:Oscillator 1 does the supersaw (+supersine/triangle/pulse).
-
- KVRAF
- 5234 posts since 25 Feb, 2008
For my money, both S1 and Sylenth are close to the feature-set/simplicity/cpu sweet spot, but each lacks features that the other has - S1 could do with more modulation slots, a second filter, and Sylenth's sequencer and FX chain; Sylenth lacks PWM, oscFM and RM, and voice stacking.
The ideal Swiss Army knife synth would combine these features. The talk of Hive as a Sylenth-killer had me thinking this might be that synth...
The ideal Swiss Army knife synth would combine these features. The talk of Hive as a Sylenth-killer had me thinking this might be that synth...
-
- KVRAF
- 5234 posts since 25 Feb, 2008
I posted audio (now deleted) of Synth1 doing a JP800 impression in the supersaw thread - some seemed to think it sounded close.Urs wrote:Hmm, couldn't find that on the Mac beta… will give it another shot today.hakey wrote:Oscillator 1 does the supersaw (+supersine/triangle/pulse).
-
- u-he
- 30192 posts since 8 Aug, 2002 from Berlin
Well, Hive is foremost supposed to be ultimately CPU efficient. It has PWM, but it doesn't have RM or FM. Audiorate modulation requires oversampling, and that automatically ruins CPU efficiency.hakey wrote:For my money, both S1 and Sylenth are close to the feature-set/simplicity/cpu sweet spot, but each lacks features that the other has - S1 could do with more modulation slots, a second filter, and Sylenth's sequencer and FX chain; Sylenth lacks PWM, oscFM and RM, and voice stacking.
The ideal Swiss Army knife synth would combine these features. The talk of Hive as a Sylenth-killer had me thinking this might be that synth...
As Hive doesn't sound exactly like Sylenth I don't think it's a "Sylenth killer". It covers many similar grounds, but one feature or two might make 1:1 patch conversion impossible. Such as, gain compensated filter resonance, pitch-based detune attenuation and possibly a dozen other features.
-
- KVRist
- 71 posts since 3 Oct, 2014
Urs,
I know you want it to be CPU efficient, but that was only one factor in Sylenth's success. The main thing was that all the sounds were great and it generally sounded great. So I hope sound quality isn't compromised in the name of CPU efficiency. I'd rather have CPU that was a bit higher than Sylenth if:
- it sounded as good or better,
- looked better,
- had better file management,
- had a clearer layout/functionality.
But it absolutely must sound as good (or better), otherwise those other factors become irrelevant. Many synth have the features above, but I won't consider them cos I don't like the sound. CPU usage only annoys me if extreme.
Excited to hear it.
I know you want it to be CPU efficient, but that was only one factor in Sylenth's success. The main thing was that all the sounds were great and it generally sounded great. So I hope sound quality isn't compromised in the name of CPU efficiency. I'd rather have CPU that was a bit higher than Sylenth if:
- it sounded as good or better,
- looked better,
- had better file management,
- had a clearer layout/functionality.
But it absolutely must sound as good (or better), otherwise those other factors become irrelevant. Many synth have the features above, but I won't consider them cos I don't like the sound. CPU usage only annoys me if extreme.
Excited to hear it.
Last edited by martingifford on Mon Nov 03, 2014 9:37 am, edited 1 time in total.
-
- KVRAF
- 1924 posts since 15 Oct, 2008 from Germany
What about oscillator sync?Urs wrote:Well, Hive is foremost supposed to be ultimately CPU efficient. It has PWM, but it doesn't have RM or FM. Audiorate modulation requires oversampling, and that automatically ruins CPU efficiency.hakey wrote:For my money, both S1 and Sylenth are close to the feature-set/simplicity/cpu sweet spot, but each lacks features that the other has - S1 could do with more modulation slots, a second filter, and Sylenth's sequencer and FX chain; Sylenth lacks PWM, oscFM and RM, and voice stacking.
The ideal Swiss Army knife synth would combine these features. The talk of Hive as a Sylenth-killer had me thinking this might be that synth...
-
- KVRAF
- 24411 posts since 7 Jan, 2009 from Croatia
That too needs oversampling since hardsync does create an awful lot of harmonic content.