Page 1 of 2

[Unison stereo spread DONE in 5.5] VA-Beast features.

Posted: Sun Mar 19, 2017 4:55 pm
by UrsaMajor
Master tune
Noise as separate osc with freq/filter/level sliders
Unison stereo spread amount (0 = mono)
Filter (pre) input gain to avoid distortion when res is high.
Mod env ADR mode (continue looping after release)
Mod env time sync mode
Exp+log decay curve for all envelopes

Re: VA-Beast features.

Posted: Tue Mar 21, 2017 6:41 am
by planet-h
Welcome to the forum, UrsaMajor.
And thanks for your message.

UrsaMajor wrote:Master tune

Interesting point, indeed. I'll consider that.

UrsaMajor wrote:Noise as separate osc with freq/filter/level sliders

You already have that.
Use Noise as waveform in OSC2 and you can control the noise frequency and filter type using the tune sliders at the right side.

UrsaMajor wrote:Unison stereo spread amount (0 = mono)

Just turn UNISON off and you have spread = 0.

UrsaMajor wrote:Filter (pre) input gain to avoid distortion when res is high.

Good point, I'll consider that.

UrsaMajor wrote:Mod env ADR mode (continue looping after release)

That's what the AD Repeat mode is for.
If you need the looping after the release of a key, use an LFO (saw waveform) in KB sync mode instead.

UrsaMajor wrote:Mod env time sync mode

Please use a synced LFO (saw waveform) in KB sync mode instead.
That does exactly what you're looking for.

UrsaMajor wrote:Exp+log decay curve for all envelopes

All decay (and release) curves in the VA-Beast synth are exponential.
A logarithmic decay would result in a very hard transition from decay to sustain. What would be the benefit of that?

Re: VA-Beast features.

Posted: Thu Apr 06, 2017 2:26 pm
by UrsaMajor
Noise as separate osc
I of course mean using both osc1 + osc AND Noise (with filter options and so on)

planet-h wrote:You already have that.



Unison
Sometimes you want unison ON, but the option to set the amount of STEREO spread,
as well as DETUNE amount.
Many times you may want a thicker sound, but still in mono.

planet-h wrote:Just turn UNISON off and you have spread = 0.



Mod env ADR mode
The AD Repeat mode stops the looping after release.
Yes you could use an LFO, but many times you are already using all 3 LFO's for other stuff.
This would basically give you a total of 6 LFO's.

planet-h wrote:use an LFO



Mod env time sync mode
Many times when doing long drones or pads, you want the parameters to return to the original positions just in time for the pattern to end.
Same as above; many times you are already using all 3 LFO's for other stuff.

planet-h wrote:Please use a synced LFO



Envelope curves
You are of correct, my bad. I meant option for exp and linear curves for the decay stage. Especially good for longer drones and pad-sounds.

Re: VA-Beast features.

Posted: Thu Apr 06, 2017 4:36 pm
by planet-h
UrsaMajor wrote:Noise as separate osc
I of course mean using both osc1 + osc AND Noise (with filter options and so on)

I cannot give you a promise here, as it is unclear if this would fit into the existing UI.
What you're actually asking for is a third oscillator.

UrsaMajor wrote:Unison
Sometimes you want unison ON, but the option to set the amount of STEREO spread,
as well as DETUNE amount.
Many times you may want a thicker sound, but still in mono.

OK, got it.
You can achieve that by detuning the oscillators against each other.
But I understand that this is not exactly what you want.

The probably easiest way to achieve mono detuning is to apply a chorus/flanger as insert effect.
The chorus/flanger does exactly that, but always with two layers.

What would you say about a simple mono/stereo switch in the unison section. Would that cover your needs?

UrsaMajor wrote:Mod env ADR mode
The AD Repeat mode stops the looping after release.
Yes you could use an LFO, but many times you are already using all 3 LFO's for other stuff.
This would basically give you a total of 6 LFO's.

Mod env time sync mode
Many times when doing long drones or pads, you want the parameters to return to the original positions just in time for the pattern to end.
Same as above; many times you are already using all 3 LFO's for other stuff.

I don't think it's a good idea to mangle the envelopes to make them act as LFOs.
What you actually want are more LFOs, right?
That is definitely something to consider.

UrsaMajor wrote:Envelope curves
You are of correct, my bad. I meant option for exp and linear curves for the decay stage. Especially good for longer drones and pad-sounds.

I'll think about that.
Do you specifically mean the decay stage only, or also the release stage?

Re: VA-Beast features.

Posted: Fri Apr 07, 2017 6:06 am
by UrsaMajor
planet-h wrote:What would you say about a simple mono/stereo switch in the unison section.


A width slider would be preferred, because it gives you the option of spreading it a litte bit.
But a mono/stereo switch would work great as well.
Or the button could have several modes like 100%, 75, 50, 25, and 0%.
That way it would still give you the option to separate the sound in the mix a little bit.

I have noticed that the more you spread a sound left and right, the wider the sound of course, but often you lose the punch and impact of it, it becomes more diffuse.

planet-h wrote:What you actually want are more LFOs, right?


Yes, give us more LFO's - you can never have enough LFO's haha :D
It may also be interesting to throw in a few more waveforms in the LFO's.
Perhaps add some sort of shaping parameter of the waveform.
And also 1-shot mode ;)

planet-h wrote:Do you specifically mean the decay stage only, or also the release stage?


Exp/Lin switch for all stages would be great.

Re: VA-Beast features.

Posted: Fri Apr 07, 2017 7:46 am
by planet-h
UrsaMajor wrote:A width slider would be preferred, because it gives you the option of spreading it a litte bit.
But a mono/stereo switch would work great as well.
Or the button could have several modes like 100%, 75, 50, 25, and 0%.
That way it would still give you the option to separate the sound in the mix a little bit.

I have noticed that the more you spread a sound left and right, the wider the sound of course, but often you lose the punch and impact of it, it becomes more diffuse.

OK, now we get to the core of the problem.
The fact that the sound loses punch if with higher unison depth is not mainly because of the stereo spread.
Actually, the stereo spread has only a minor effect to the punch of the sound.
The punch is lost because of the detuning, as the detuned waves might meet each other at -1 (wave1) and +1 (wave2).

As you see in the docs, the stereo spread already depends on the number of layers. The more layers, the more spread.

The User Manual wrote:The VA-Beast Unison offers 3 different signal multiplication levels:
2x (pans 1 signal slightly to the left of the stereo field, and 1 to the right)
3x (pans 1 signal fully to the left of the stereo field, 1 to the right, and keeps 1 in the center)
5x (pans 2 signals fully to the left of the stereo field, 2 to the right, and keeps 1 in the center)

If you want to keep the punch, reduce the layers, and even more important, reduce the unison depth.
A good amount for fat bass sounds with unison is 2x and depth = 6 - max 10.


UrsaMajor wrote:Yes, give us more LFO's - you can never have enough LFO's haha :D
It may also be interesting to throw in a few more waveforms in the LFO's.
Perhaps add some sort of shaping parameter of the waveform.
And also 1-shot mode ;)

I'll see how much the performance impact of 6 LFOs would be (I cannot give you a promise yet, but I agree that more LFOs would be nice).
A shape slider might take place (no idea where exactly yet) with a similar shape functionality like the oscillators have (e.g. sine morphs to saw approximation).
All VA-Beast LFOs already have a one shot (env) mode.


UrsaMajor wrote:Exp/Lin switch for all stages would be great.

I'll definitely consider that.

Thanks a lot for your very well thought out suggestions.
I really like most of them.

Btw. are you already member of the beta community?
And if not, are you willing to join?
Just visit https://groups.google.com/d/forum/g-stomper-public-beta-group and click "Apply for membership".
Would be great to have you aboard;)

Re: VA-Beast features.

Posted: Fri Apr 07, 2017 4:39 pm
by UrsaMajor
planet-h wrote:The punch is lost because of the detuning,

Yes i know what you mean. When you duplicate the voices in unison mode the sound becomes thicker, true. And you will also get som beating between the waves when they cancel each other out and so on. I fully agree with this. However, this is not what i am talking about.

What i am referring to i think, is more of a psychoacoustic phenomenon, not related to a specific effect or DAW. I think it has to do with how the brain interprets the location of the sound. When the sound is in mono in the mix, even if you use more than one voice, the sound will have a distinct position. You can even pan the channel to the left or right, it will still have a distinct position. If the voices of the unison is panned left and right, i think the brain interprets this as two separate sounds coming from two differend directions at the same time. It just gives it a less defined character. The more the stereo-spread, the more diffuse. Another example but same principle: sometimes I like to use chorus to thicken a sound slightly, but then sum the channel to mono to give it a more distinct position in the mix.

planet-h wrote:If you want to keep the punch, reduce the layers, and even more important, reduce the unison depth.

Yes, but the VOICE DETUNE, and the VOICE STEREO SPREAD is not the same thing. For example, look at Korgs emulation of the Mono/Poly. It has "Voice Count", "Detune" and "Spread"
Image

planet-h wrote:All VA-Beast LFOs already have a one shot (env) mode.

Excellent! Thanks.

planet-h wrote:are you already member of the beta community?

Thanks, I will check it out soon.

Re: VA-Beast features.

Posted: Fri Apr 07, 2017 4:59 pm
by planet-h
Yes, you're right.
Stereo spread just controls the panning of the layered voices, which is fixed in the current setup.
And yes, I agree, that a stereo spread ctrl (either button or slider) would be very useful.
Thanks again, I'll see what I can do here. As always, the biggest challenge is to fit it into the existing UI. The technical implementation is pretty simple;)

Re: VA-Beast features.

Posted: Sat May 06, 2017 10:25 am
by planet-h
Wile creating the "Stereo Tool", I did some experiments with the Stereo Spread of the Unison in the VA-Beast.
Just to be sure that we're talking about the same thing here...

With "Stereo Spread" I mean the panning of the layered 2, 3 or 5 Unison voices. Is it this what you had in mind?

Narrowing the Unison in the stereo field can be achieved by different methods:

1) The proper and real method:
2 voices: both will be panned against each other based on the stereo spread
3 voices: 2 will be panned against each other based on the stereo spread, 1 remains in the center
2 voices: 4 will be panned against each other based on the stereo spread, 1 remains in the center

This is the real thing, as you'd have all layers at both channels if the width would be set to zero.
The downside is, that this would have a serious performance impact, as the left and right have to be accessed (which uses CPU time), ... and it has to be done per voice and Unison layer. For the records: each OSC produces a true stereo signal, as it can be generated waveforms and/or PCM based multi samples (which can be stereo as well).
In the current setup (without a spread setting), the stereo spread is fixed at 100%. This means only the center layer uses both channels, the panned layers only use the left or right channel (which significantly reduces the CPU load).

2) Narrow the already spread Unison signal:
Simply said, right at the output of the Would a MID/SIDES processor take place to narrow or even enhance the stereo width.

This would of course sound slightly different, as the mono signal would be built from the fully spread unison signal.
In other words, the center layer would be twice as loud as the other (panned layers) after summing them together to get the MID channel.
The benefit would be a much lower CPU cost, as this could be done once per voice or even for all voices together.

You can test and compare the two methods in Ableton for example (with the Utility device and the Monopoly Synth).
I'd be interested to hear what you think about that.

Re: VA-Beast features.

Posted: Sat May 06, 2017 3:22 pm
by UrsaMajor
planet-h wrote:2) Narrow the already spread Unison signal... ...In other words, the center layer would be twice as loud as the other (panned layers) after summing them together to get the MID channel.

With option 2 - would it be possible to compensate for the volume difference? Increasing the level of the panned layers in relation to the (decreased) spread. Or maybe that is a "dirty" solution, programming wise?

planet-h wrote:For the records: each OSC produces a true stereo signal, as it can be generated waveforms and/or PCM based multi samples (which can be stereo as well).

Ok, just so I understand - does this mean that stereo PCM samples are summed to mono first, then panned to the left and right (for the panned unision layers)? For example, when using 3 layers of unison.