Tech Preview: Hive Wavetables

[nilhartman]

To be honest I just wish I could use cytospur's WT editor (but I'm on a Mac)... or have access to his brain directly (just the WT scripting part hehe).
Hive 2 is fantastic as is, now I just wish I'll learn how to translate ideas from say Bazille to H2's WTs. I'm still floored by the lack of any sort of artefacts with FM based uhm WTs.

[nichttuntun]

Evildragon,

it isn't. From your point of view. There are synth which do audio rate very CPU efficient. And if HIVE can't do it efficiently tippy do not need to use it. But I think it's great to be able to choose. Have joy and be sound.

[EvilDragon]

There are synth which do audio rate very CPU efficient.

But not at 16 voice unison per each oscillator and sub-oscillator, I'm pretty darn sure

[Funkybot's Evil Twin]

Not every synth needs to do everything. Hive has an excellent balance. Great sound, great workflow, moderate CPU, super easy to use... No offense to U-he, but I don't see how they could add audio rate without mucking up at least three of those four (wouldn't adversely impact great sound, but the rest, yes). I'd have no qualms about seeing audio rate+wavetables in Zebra3, but I totally get why Urs is reluctant to shoe horn that into Hive. Not every product needs to do everything. The best products are usually very limited IMO.

[zvenx]

Not every synth needs to do everything. Hive has an excellent balance. Great sound, great workflow, moderate CPU, super easy to use... No offense to U-he, but I don't see how they could add audio rate without mucking up at least three of those four (wouldn't adversely impact great sound, but the rest, yes). I'd have no qualms about seeing audio rate+wavetables in Zebra3, but I totally get why Urs is reluctant to shoe horn that into Hive. Not every product needs to do everything. The best products are usually very limited IMO.

This!!!

Well said.
rsp

[nilhartman]

Definitely.

[nichttuntun]

I do understand your points and needs. It's a save feel-good zone, like a well known home territory. Always sounding right. Working fast. No risk of entering new territory and maybe get lost or discover stranger things and become consumed by them. Yes please...do go on

[EvilDragon]

You can enter plenty of new territories with the additions in Hive 2 already.

[pdxindy]

Not every synth needs to do everything. Hive has an excellent balance. Great sound, great workflow, moderate CPU, super easy to use... No offense to U-he, but I don't see how they could add audio rate without mucking up at least three of those four (wouldn't adversely impact great sound, but the rest, yes). I'd have no qualms about seeing audio rate+wavetables in Zebra3, but I totally get why Urs is reluctant to shoe horn that into Hive. Not every product needs to do everything. The best products are usually very limited IMO.

+1

If it were left up to the users, every synth would become a disaster. The real irony is as people ask for more and more features, they only use about 15% of the capability that is already there.

[cytospur]

Here is a 2 operator FM script that wraps the normal FM formula with the arcsin trigonometric function. The arcsin of a sine wave is a triangle wave. This means that the normal output is 'triangularized'. This makes for a brighter sound.

Code: Select all

Info "Set Tables to 16\nFeedback FM with a twist\nBy Mark Holt\n"

NumFrames=256

Wave target=aux1 "pi*phase"

Wave target=aux2 "1"

//Set aux2 to 1 for sawtooth feedback and 2 to get Reface DX style square feedback.
//Higher values are also possible.

Wave "2/pi*asin(sin(2*aux1+2*((frame/16)%1)*sin(2*aux2*aux1)/(1-0.5*cos(2*aux2*aux1+sin(2*(floor(frame/16))*aux1)))))"

//Change the value of the multiplier of the cos term to affect feedback amount.
//0 is no feedback, any value less than, but not including 1, can be used.
 
Export Source=main "MH ASinFM.wav"

[Urs]

Wut?

[cytospur]

Sorry, it's a three operator script, which is maybe why it looks extra complicated

[KBSoundSmith]

I do understand your points and needs. It's a save feel-good zone, like a well known home territory. Always sounding right. Working fast. No risk of entering new territory and maybe get lost or discover stranger things and become consumed by them. Yes please...do go on

Having become an FM enthusiast myself, I understand where you're coming from... but I'd say a few things about the addition of FM to Hive.

1) It's already possible via the .uhm scripting language. Obviously it has the disadvantage of needing to learn a scripting language (which I haven't even begun to do ), but...
2) it has the advantage of off-loading CPU strain
3) ease of use once converted into a wavetable (it doesn't become unpredictable/unwieldy to small parameter changes like a true FM synth, so those who aren't good with FM can still benefit)
4) can be used with the unison engine without weird things happening (try FM on a Zebra oscillator using unison)
5) maintains the desired philosophical simplicity of the synth... FM can be time-consuming to program, due to the sharp economy of the synthesis method.
6) Hive can already get into "new" territory. The included wavetables are good and quite powerful, and they reward careful exploration and clever manipulation. And this is without considering the possibility of adding new wavetables

Regarding FM in general, I think it has a bit of mystique about it that causes people to either be disproportionately repulsed or attracted to the method. It isn't the unfathomable monster some imagine it to be, but neither is it fairy dust which makes every piece of music a sonic unicorn. I've dived pretty deep, and I've found that some of the sounds which can be obtained on a truly deep dive can be achieved faster/easier with... drum roll... wavetables

I don't want to really discuss other developer's synths here, but since it's topical, I've tried Phase Plant recently. It allows extremely flexible FM routing, and Wavetables, Samples, Noise, and Classic Wave Shapes ("Analog") can all be used as operators. I've found the results to be... samey. It's actually a whole lot less interesting than you'd think.

The quality of a sample being used for FM really matters -- bad samples don't sound less bad when used as FM operator -- they often just sound worse. The predictable results are from the sustained portions of sound...which are often sawtooth/square wave-ish already. Factor in bad sample looping control, why not just actual sawtooth or square waves for the FM (or triangle waves... more useful than squares, since it is usually Phase Modulation in actual use... triangles are an extremely useful waveform for FM, IMO...kind of why I don't understand why a ModMap has to be used up in order to use them in Bazille)? The transients are what make most samples unique sounding, and they have to be treated gently/carefully with FM I've found, in order to sound interesting/good -- otherwise you get pretty samey noise. I've also tried various filter types with samples too... example, bandpass filters. BP filters have a tendency to sound Formant like on samples...so why not use a formant waveform included with a lot of FM synths, rather than fiddle with samples/filters and walk on egg-shells? Complex time-varying waveforms in FM are pretty overrated... the FM tends to just sound like a distortion module, rather than cleanly generating new waveforms. I found FM with Percussive samples most interesting, using it as distortion.

Wavetables definitely are far more interesting for FM (to me at least), since the time-varying content of the operator is more controllable, plus you can also use them as static waveforms. The juggling act here is the inherent complexity of the waveform being used. My FM observation: beyond a certain threshold, the more complex a waveform becomes, the less interesting/useful the waveform becomes for FM (the point of FM is to take simple waveforms to generate complex waveforms...if the waveforms are already complex...). FM will sound more like a distortion method rather than a synthesis method after a certain point... low FM amounts and fewer operators become necessary as the operator waveform complexity increases. Otherwise you get samey noise very fast. You likely will be better off running the wavetable through a distortion unit, or just by itself entirely, unless you specifically want the FM distortion.

Now, having gotten the taste of FM with wavetables, I think it would be a great feature for Zebra 3. I'm not so convinced of it for Hive 2. Again, as operator waveform complexity increases, the usefulness of heavy FM amounts and multiple operators per stack decreases. To get interesting, controlled, and varied sounds, multiple carriers strike me as more interesting, since they can be used as layers for sculpting the sound. Obviously, Zebra is better equipped for this approach. Plus, I strongly prefer proper MSEGs for working with FM.

[pdxindy]

I do understand your points and needs. It's a save feel-good zone, like a well known home territory. Always sounding right. Working fast. No risk of entering new territory and maybe get lost or discover stranger things and become consumed by them. Yes please...do go on

Hive 2 already offers lots of new territory and strange things to be discovered. What's holding you back?

[Howard]

Here is a 2 operator FM script that wraps the normal FM formula with the arcsin trigonometric function.