U-HE Hive

[tehlord]

So far you are winning.

[ddeez]

Ok. I really hope the thing sounds very "hifi"

I havent used zebra recently because i feel the overall sound quality of it isnt really competitive with whats on the market now, especially diva and bazille, spire etc. hoping the sound quality is there, more then happy to use a lot of cpu for that purpose

[News From The Sky]

We use a technique that will have some aliasing, but not in the audible spectrum. I.e. aliasing stops above 18+ kHz. There might (might!) be a mode that introduces more aliasing as a feature (!), but not of the metallic low dumbing type.

Hey Urs, just curious: do any of the U-He synths use dynamic sample rates (ie, a very high integer multiple of the fundamental frequency of the note being played)? Or is it possibly something you're considering for a future product? Seems like an interesting way to deal with digital wave generation (since any aliasing will be harmonic, at least), but maybe for some reason it's not practical in software.

[News From The Sky]

I wouldn't expect Hive to sound as good as Diva, since one of the stated design goals was to make it relatively light on the CPU (compared to other U-He synths anyway). It might still sound better than Zebra if the fixed audio path allows for optimizations.

[Kriminal]

Good? Dive is an analogue emu, hive isnt. Both should sound 'good' but diva would naturally be a bigger cpu hit.

[GeorgeZ]

^+1

[xh3rv]

We use a technique that will have some aliasing, but not in the audible spectrum. I.e. aliasing stops above 18+ kHz. There might (might!) be a mode that introduces more aliasing as a feature (!), but not of the metallic low dumbing type.

Hey Urs, just curious: do any of the U-He synths use dynamic sample rates (ie, a very high integer multiple of the fundamental frequency of the note being played)? Or is it possibly something you're considering for a future product? Seems like an interesting way to deal with digital wave generation (since any aliasing will be harmonic, at least), but maybe for some reason it's not practical in software.

This technique is accomplished by manipulating the clock rate of digital components, which is to say when the samples occur in continuous time. It's really part of a functional process that's discrete-to-continuous / digital-to-analog. Within software there are whole families of DSP algorithms that are in some way analogous but functionally only discrete-to-discrete is possible.

[Urs]

We use a technique that will have some aliasing, but not in the audible spectrum. I.e. aliasing stops above 18+ kHz. There might (might!) be a mode that introduces more aliasing as a feature (!), but not of the metallic low dumbing type.

Hey Urs, just curious: do any of the U-He synths use dynamic sample rates (ie, a very high integer multiple of the fundamental frequency of the note being played)? Or is it possibly something you're considering for a future product? Seems like an interesting way to deal with digital wave generation (since any aliasing will be harmonic, at least), but maybe for some reason it's not practical in software.

This technique is accomplished by manipulating the clock rate of digital components, which is to say when the samples occur in continuous time. It's really part of a functional process that's discrete-to-continuous / digital-to-analog. Within software there are whole families of DSP algorithms that are in some way analogous but functionally only discrete-to-discrete is possible.

One can accomplish this though. One can run an oscillator at a constant rate, say, 100 samples per cycle. Afterwards one would downsample (or upsample) to the desired frequency. So one would always render the oscillator ahead a bit, to get enough supply for the current block of samples. One can also use a variable integer cycle size, but switching between those due to pitch/frequency modulation will introduce aliasing nonetheless.

IIRC Vemberaudio's Surge uses this technique (render at integer sample cycle, then downsample to whatever is desired)

[Urs]

Good? Dive is an analogue emu, hive isnt. Both should sound 'good' but diva would naturally be a bigger cpu hit.

Yep!

[xh3rv]

IIRC Vemberaudio's Surge uses this technique (render at integer sample cycle, then downsample to whatever is desired)

I stand corrected! Makes me sort of think, continuous-to-discrete, which seems logically possible in digital. Maybe not quite. I suppose oscillators built with piecewise, functionally described curves might make more sense as continuous-to-discrete

[News From The Sky]

Good? Dive is an analogue emu, hive isnt. Both should sound 'good' but diva would naturally be a bigger cpu hit.

Right - because Hive isn't going to be emulating analog circuit behavior with all the CPU-expensive quirks that make Diva sound so rich.

I'm not worried about Hive not sounding good. I'm sure it will be lovely. I was just responding to this:

Ok. I really hope the thing sounds very "hifi"

I havent used zebra recently because i feel the overall sound quality of it isnt really competitive with whats on the market now, especially diva and bazille, spire etc. hoping the sound quality is there, more then happy to use a lot of cpu for that purpose

Hive isn't going to sound like Diva, but it doesn't need to in order to sound very good.

[ddeez]

I hope it sounds better then zebra ^

[zvenx]

I hope it sounds better then zebra ^

Ouch.. So you know, some of us think zebra is one of the best sounding software synths.
Rsp

[pdxindy]

I hope it sounds better then zebra ^

Ouch.. So you know, some of us think zebra is one of the best sounding software synths.
Rsp

That's why if it sounds better then Zebra it'll be rather amazing!

[KingTuck]

I mean the public beta should be up soon enough that all this speculation is a bit unnecessary right? I'm sure it will sound good, u-he have a good track record, but if you're not gonna let their past products necessarily decide the quality of their future products than just wait like a month. Soon we can all be testing and helping improve the Hive !