Anyone here use Hive?

[fluffy_little_something]

Just downloaded the Hive demo again, seems there has been an update since the last time I had it on my computer.

Sent them both through a spectrum analyzer (primitive single-saw init patch, filter fully open) and the Hive spectrum seems to drop off a couple thousand Hz earlier than with Sylenth (for instance 16 kHz vs 20 kHz roughly). Maybe that is why I perceive Sylenth as brighter and edgier, others might say more digital.
https://app.box.com/s/vroe6ur2ki0ena56uaf4gba075hl4w89

On the other hand, with Hive the spectrum rolls off in a gentle curve from about 10 kHz on, whereas on Sylenth the spectrum often simply ends in a vertical drop higher up The difference doesn't look like much on the diagram, but since it is a logarithmic scale, it amounts to a difference of about 4 kHz at the top end, sometimes even more depending on the note.
However, I assume Hive's spectrum is more like that on hardware, right?

I don't know much about those things, but I found it interesting

[pdxindy]

1. (and the most important) I cannot stand Hive´s envelope attack... sounds to me like a heavy negative logarythmic curve (dunno, if this is correct ... I mean slow start- fast end)... with attack times about 300-600 ms there´s first nothing (or very little), then comes the sounds very abrupt ... I personaly find negative logarythmic (if this term is correct) very ugly for attack times about these values I posted... it´s just sounding shitty (sorry )

No sorry needed... you are welcome to like what you like

Me, I like the Hive envelope attack shape for some sounds and less for others... there is not a single envelope attack shape that is 'right' for all sounds... so I value that I can change the shape of just the attack, or decay or release in Hive

[drakmaniso]

1. (and the most important) I cannot stand Hive´s envelope attack... sounds to me like a heavy negative logarythmic curve (dunno, if this is correct ... I mean slow start- fast end)... with attack times about 300-600 ms there´s first nothing (or very little), then comes the sounds very abrupt ... I personaly find negative logarythmic (if this term is correct) very ugly for attack times about these values I posted... it´s just sounding shitty (sorry )

Hive uses different envelope shapes in each mode. In fact it also uses different detuning laws:

http://www.kvraudio.com/forum/viewtopic ... 1&t=427890

Personally I wish we could select each option independently rather than having a global mode, but that's probably not the spirit of the synth. I like the s-shaped envelopes (the ones you don't like) for very short attacks, but otherwise I agree I usually prefer the exponential envelopes used in "Dirty" mode.

As for the percentage based GUI, I think U-He is slowly moving away from this (or at least starting to provide more traditional values) in their latest plugins (e.g. Presswerk). So maybe Hive will evolve on this point too?

[Trancit]

Hive uses different envelope shapes in each mode. In fact it also uses different detuning laws:

http://www.kvraudio.com/forum/viewtopic ... 1&t=427890

Personally I wish we could select each option independently rather than having a global mode, but that's probably not the spirit of the synth. I like the s-shaped envelopes (the ones you don't like) for very short attacks, but otherwise I agree I usually prefer the exponential envelopes used in "Dirty" mode.

As for the percentage based GUI, I think U-He is slowly moving away from this (or at least starting to provide more traditional values) in their latest plugins (e.g. Presswerk). So maybe Hive will evolve on this point too?

Thanx for the link...didn´t know about this...but it´s exactly like I thought... S-type/linear type attack in the 2 mode I personaly would consider to use (I don´t like the sound of the dirty mode at all)...

It´s a pity, that things sometimes be other way round as you would like them to have... Is it such a big deal from a developing point of view to make envelopes more flexible (like so many synths have already)????

Sorry again, but I see this as wrong decision to dictate/restrict the user in a new developed synth... if you want this shape you have to use this mode... you don´t like the filter there???... tough luck... deal with it or find something else... we build in all you need, but we got the great idea to restrict you, how we think you have to use it...

[Urs]

Thanx for the link...didn´t know about this...but it´s exactly like I thought... S-type/linear type attack in the 2 mode I personaly would consider to use (I don´t like the sound of the dirty mode at all)...

We didn't use anything out of the ordinary. Here's where you find surprisingly similar envelope curves:

A tad S-shaped curve for VCA Attack with exponential shape for Mod Attack and all other stages: Sylenth1

Linear Attack + exponential Decay/Release: Virus

All exponential: Analogue stuff

However, unlike Sylenth1 and much like a Virus one can modulate an envelope's Attack parameter by the envelope itself. Thus one can make it as exponential or logarithmic as one could wish for.

[Dasheesh]

You have to ask yourself. Do I want a standard sound that is the norm or something with personality and character? What is your goal? Hive is not a do all instrument and it's not an iPhone sound that is so popular right now.

[Urs]

As for the percentage based GUI, I think U-He is slowly moving away from this (or at least starting to provide more traditional values) in their latest plugins (e.g. Presswerk). So maybe Hive will evolve on this point too?

Hive has 10% scaling, just like Presswerk. The reason is, we lay our UIs out on a 10 pixel raster at 100% so that even if we scale the UI we do not need to resample a knob more than once. That means, we can scale the knob graphics once and directly copy them into the screen buffer without further need for interpolation.

It's a compromise between having a resizable interface and sticking to bitmap graphics.

OTOH Sizes from 70% to 200% in steps of 10% should pretty much fit anyone's bill (or screen, that is)

[drakmaniso]

Hive has 10% scaling, just like Presswerk.

Sorry, I made that quote out of context: we were speaking about the unit used for displaying parameter values, not the scaling!

If I'm not mistaken, you mentioned on the forum you were considering displaying more "musical/real life" units for some paramaters of the synths, in a similar way that Presswerk already does.

[mutantdog]

Yeah, you are referring to the parameters such as filter where its 1 - 100. Its actually based on semitones or something i believe. U-he have held onto that one as a statement that its actually more logical (in many ways it is) but I've never really got my head around it either.

One thing i'd like them to add to the preferences would be the option to change these to the more traditional frequency based scale.

[drakmaniso]

Yeah, you are referring to the parameters such as filter where its 1 - 100. Its actually based on semitones or something i believe. U-he have held onto that one as a statement that its actually more logical (in many ways it is) but I've never really got my head around it either.

Actually, I like the 0-100 scale for most of the tasks: it makes modulation easier, especially when the same source is used for multiple targets. I just wish there was an additional display of a more traditional value (e.g. note names for cutoff frequencies, or dB for the volume).

[Urs]

Hive has 10% scaling, just like Presswerk.

Sorry, I made that quote out of context: we were speaking about the unit used for displaying parameter values, not the scaling!

If I'm not mistaken, you mentioned on the forum you were considering displaying more "musical/real life" units for some paramaters of the synths, in a similar way that Presswerk already does.

Ah, yes, of course.

Yep, it's one of our ongoing projects to improve the "units" of parameters, wherever it makes sense.

[Trancit]

We didn't use anything out of the ordinary. Here's where you find surprisingly similar envelope curves:

A tad S-shaped curve for VCA Attack with exponential shape for Mod Attack and all other stages: Sylenth1

Linear Attack + exponential Decay/Release: Virus

All exponential: Analogue stuff

Yes I think I was wrong here... Indeed it sounds very similar to Sylenth...

However, unlike Sylenth1 and much like a Virus one can modulate an envelope's Attack parameter by the envelope itself. Thus one can make it as exponential or logarithmic as one could wish for.

Nevertheless, this is a very abstract method and a bit hard to get one´s head around...the envelopes of i.e. Synthmaster are an (for me) ideal example of how it should be...

About CPU usage:
Did a little test here (all with the same settings as far as possible... because of no having proper value readout in Hive, it´s hard to match exactly for the envelope times)

Overall: Padsound, 16 Voices on all Synths enabled
OSC 1: Saw, 7 unison voices, detuned a bit, Stereowidth at 50%
OSC 2: Square one octave lower, 7 voices, detuned a bit, Stereowidth at 0%
Amp envelope: Attack at about 400ms, Sustain 100%, Release about 2.5 sec
In Hive: Attack modulated by Mod 1, OSC 1 detune modulated by LFO 1 (that it not sounds too ugly )
No Filter engaged
A bit of internal EQ in all
Just compared quickly with Reapers internal CPU readouts and a little sequence:

Sylenth1: highest peak at 0.6%
Synthmaster: same at 1%
Hive: highest peak at 2.1%

nearly 4 times more than Sylenth and nearly the double of Synthmaster...

[Urs]

16 Voices on all Synths enabled

Did you multiply that with the number of oscillator unison set in Sylenth? A common mistake is to think that voices in Sylenth equal voices in Hive, but in Sylenth each single oscillator counts as a voice.

Also, maybe post the patches so we can run our own tests - if it's true, we'll figure out a way to make Hive faster.

[tonedef71]

If it's true, we'll figure out a way to make Hive faster.

As a paying customer, I gotta appreciate that.

[Urs]

Well, thinking about it, there may indeed be cases where Hive may be slower than others - foremost with "simple" sounds. If only one filter is used for instance, both are calculated. That's because they're written in SSE assembly, which means we calculate 4 filters at once (2 VCF * 2 channels stereo) at the price of one. This way we managed to get a oversampled filters to run as fast as the non-oversampled filters of some other product, but only if two a used. Making it faster for just one VCF used we would have to drop the oversampling. Alternatively, the "Clean" filters aren't oversampled because they don't distort.

Likewise, the oscillators calculate in badges of 4x unison. If it's set to only 5, Hive will still calculate 8 oscillators - for the same reason as the filters - we get 4 at once for just little more than the price of one, at a really good quality.

Hence, the more you have, the lighter on CPU it gets in comparison. And that, is all that counts. Because for simple sounds Hive is still CPU friendly enough anyway.