Valo - public beta demo, second attempt

[Leslie Sanford]

and 1/8t = 1/12 if I'm not mistaken?

Hmm, I think 1/8T is 1/3. You're dividing a beat into three parts.

I take 1/1 in fractional notation to mean one beat. Applying this resolution to an LFO would result in its frequency cycling once per beat. In time signature notation 1/1 translates to 1/4 (a quarter note).

But I dunno... could be that 1/1 is interpretted as 1 bar (as you seem to be doing), not beat.

Sorry. Don't mean to muddy the waters. This is something I've been a little fuzzy on myself.

[VitaminD]

I agree we still need more LFO routings but I suspect thats already in the works..

More routings in the sense of more routing slots (unfortunate unlikely to happen very soon, would need a mod-matrix for that) or in the sense of more routing targets in which case please please please suggest some, other than resonance (which was mentioned), I'm not sure what's missing, but would gladly see if things added whenever that's possible.

targets yes.. all of the delay and chorus parameters..esp the panning and wet/dry parameters.. oh and drive from the amp env..

[mystran]

targets yes.. all of the delay and chorus parameters..esp the panning and wet/dry parameters.. oh and drive from the amp env..

Hmmh, let's see... delay/chorus are insert-effects run globally for all the voices (it'd get a bit too CPU expensive to run them for each voice separately, I think) so in order to have those parameters available for modulation would basically require forcing the LFOs to global mode, and since there isn't one yet, it currently can't be done (though see below)...

... unless I can think of a magic trick that would allow the wet-levels (and delay pan, for each voice separately too, but won't work for the other parameters) but it gets a bit too expensive if not only chorus/delay but also reverb dry/wet modulation is desired... I have to think of it, but getting it to work could be a bit tricky... so don't expect that overnight in any case...

With global mode LFOs one could modulate other parameters as well I guess... I'll think about that as well once the LFO global mode is there.

As for drive... yeah I guess that would be a possible target and not that hard to add and shouldn't make much difference in performance. The distortion (or whatever you wanna call it) is run after the amplitude envelope (and any gains from modulations), so adding pure gain will add more distortion when drive is nonzero (so amplitude envelope will control distortion in a sense), but it's also true that the drive-knob does more than just act as a dry/wet control (the saturation gets shaped too) so I'll add that to the list of targets to be added.

[mystran]

and 1/8t = 1/12 if I'm not mistaken?

Hmm, I think 1/8T is 1/3. You're dividing a beat into three parts.

I take 1/1 in fractional notation to mean one beat. Applying this resolution to an LFO would result in its frequency cycling once per beat. In time signature notation 1/1 translates to 1/4 (a quarter note).

What?? Ofcourse it's the same as time-signatures. Why would it be in beats? So 1/1 translates to a semibreve (also known as a whole note) which in 4/4 time equals a bar. I never actually thought anyone would think of divisions in beat.

So I assume 1/8T is an individual note out of a triplet of eighths giving 1/8 * 2/3 = 2/24 = 1/12. This is how it seems to work where I've seen such notation used.

Anyway, theoretically speaking, I guess 7/16 would come out as 1/4DD (double-dotted quarter = 1/4 + 1/8 + 1/16 in normal musical notation) but that still doesn't work for 5/16, since there isn't a standard way to notate such a thing as a single note as far as I know. I could invent one though, call 'em 1/4HD for half-dotted, quarter and half of a dot = 1/4 + 1/2 * 1/8. No idea how to draw a half-dot though...

Would that be a sane way? I could put the full ratio in parenthesis in the menu then...

edit:

Actually... now that I think of it.. hmmh... 1/4DD, 1/4D, 1/4HD, 1/4... that does look kinda okayish list, definitely easier to read than the current one.

But mm.. 5/12 coming out as 1/2THD (= half-dotted triplet half-note) is kinda evil looking.

[VitaminD]

targets yes.. all of the delay and chorus parameters..esp the panning and wet/dry parameters.. oh and drive from the amp env..

Hmmh, let's see... delay/chorus are insert-effects run globally for all the voices (it'd get a bit too CPU expensive to run them for each voice separately, I think) so in order to have those parameters available for modulation would basically require forcing the LFOs to global mode, and since there isn't one yet, it currently can't be done (though see below)...

... unless I can think of a magic trick that would allow the wet-levels (and delay pan, for each voice separately too, but won't work for the other parameters) but it gets a bit too expensive if not only chorus/delay but also reverb dry/wet modulation is desired... I have to think of it, but getting it to work could be a bit tricky... so don't expect that overnight in any case...

With global mode LFOs one could modulate other parameters as well I guess... I'll think about that as well once the LFO global mode is there.

As for drive... yeah I guess that would be a possible target and not that hard to add and shouldn't make much difference in performance. The distortion (or whatever you wanna call it) is run after the amplitude envelope (and any gains from modulations), so adding pure gain will add more distortion when drive is nonzero (so amplitude envelope will control distortion in a sense), but it's also true that the drive-knob does more than just act as a dry/wet control (the saturation gets shaped too) so I'll add that to the list of targets to be added.

non-Global mode LFOs are very wonky to me.. I only prefer the global mode type..

[mystran]

We're going to experiment what happens when I put my all-so-great 1/8THD notation in there... that'll be in the next version (obviously with the raw ratio in parenthesis) and we'll see how many complaints we get that way.

[mystran]

We're going to experiment what happens when I put my all-so-great 1/8THD notation in there... that'll be in the next version (obviously with the raw ratio in parenthesis) and we'll see how many complaints we get that way.

No wait.. hmmh... 7/24? How does such a ratio come out as any sort of sane notation? I guess I'll leave the triplets alone, and use stuff like 7/16t to signify 7 times a 16th-triplet-note (2 full tripplets and one).

edit: rationale - anyone working with such division probably will be able to deal with the math anyway

[Leslie Sanford]

and 1/8t = 1/12 if I'm not mistaken?

Hmm, I think 1/8T is 1/3. You're dividing a beat into three parts.

I take 1/1 in fractional notation to mean one beat. Applying this resolution to an LFO would result in its frequency cycling once per beat. In time signature notation 1/1 translates to 1/4 (a quarter note).

What?? Ofcourse it's the same as time-signatures. Why would it be in beats? So 1/1 translates to a semibreve (also known as a whole note) which in 4/4 time equals a bar. I never actually thought anyone would think of divisions in beat.

That makes sense.

I think I got the 1/1 equals one beat from the value the VST host gives you; it's in pulses per quarter note. So if you get a value of 1.0 from a call to getTimeInfo, that's one quarter note, or beat, that's expired. If your sync resolution is set to 1/4 (a quarter note), no scaling is needed to use the ppqnPos value, hence the 1/1 scaler value.

Anyway, I'm apparently off on this (wouldn't be the first time). So I need to go do some studying!

Valo looks and sounds great, btw.

[mystran]

I think I got the 1/1 equals one beat from the value the VST host gives you; it's in pulses per quarter note. So if you get a value of 1.0 from a call to getTimeInfo, that's one quarter note, or beat, that's expired. If your sync resolution is set to 1/4 (a quarter note), no scaling is needed to use the ppqnPos value, hence the 1/1 scaler value.

Yeah that's right, VST uses quarter notes, but that doesn't mean I can't use signature-times in my GUI. I don't think a musician needs to know what VST does internally.

Valo looks and sounds great, btw.

Thanks.

[mystran]

daily update time, "pb2" changelog:
- slightly lightened default theme shading for darker gamma monitors
- added some EQ and slight degradation into feedback loops of delay
- limiter now has much better attack curve, being less clicky
- PWM nolonger causes "crackling" noise with narrow pulse-widths

Long versions:

There is now both high-pass and low-pass in the feedback loop of the delay, making the delay sound somewhat duller than dry sound. This makes it less of a mayhem on high-resonance sweeps, and hopefully keeps the low-end a bit more tight. However, this was tuned to what I'd normally use from a delay (and what I liked as a good compromise for the range of sounds I can manage to get from the synth), so I might have filtered too much (or too little) so opinions are welcome.

The slight degradation means it no longer has perfect signal-to-noise which shouldn't make much difference with "normal" use, but will make it slowly accumulate some noisy character if you set feedback at maximum. I think it goes well with the EQ and all. Comments welcome though.

Oh, and the limiter... there was a bug (well let's say misimplementation) that caused the attack curve to get faster when it was already reducing gain (with more reduction->faster curve). This no longer happens, so it now always takes advantage of the full lookahead, greatly softening the clicks (making me wonder why some limiters with more lookahead actually click harder).

And well.. a minor tweak to the pulse-wave generation.. pulse-width modulation now let's you go all the way to 0% pulse in both direction in a clean way.

[bwwd]

How about this color schme for Your synth?

Gray colours are depressing for me and so uninteresting.
Maybe more oscillators ? different types and mixed ones.Or better... ability to load oscillators.

[mystran]

How about this color schme for Your synth?

I'll add some more colorschemes at some point, sure, as soon as I'm feeling artistic (and too lazy to write code). There's currently the red one (which is what I usually use myself) if you want something that isn't gray, but yeah, I can make a bright one too at some point. Will keep the default one to a more neutral one I think, but it's on TODO list to allow saving your preferred scheme as a preference.

Maybe more oscillators ? different types and mixed ones.Or better... ability to load oscillators.

More oscillators is unlikely, since this was specifically designed to have two. I've got a plan to make a more modular synth in the future, but it'll probably take several months before I even start that project, 'cos I wanna concentrate on making this one really good first.

As for different types, yeah that's been the plan for a long time, but that will happen after full version is available (but maybe before "1.0" release, don't know), and after more critical stuff is there. I don't wanna go too deeply on what I got on the drawing board, since I don't even have the math sorted out (and definitely no working prototypes) yet, so I don't know if that stuff is workable, but some sort of wave-table stuff anyway (or maybe grain-table, or maybe something in between, we'll see). Maybe some morphs to the existing saw/triangle/sine as well. However, doing new oscillator types properly isn't an overnight job, and there's quite a few more important things first.

[streifentier]

Hello, for anyone who's interested: you can find 64 patches
for Valo here.
No guarantee for quality or whatsoever... Have fun!

[mystran]

Btw, coming in next version (in a couple of days I hope) some MIDI improvements, like the ability to change patches with MIDI program change (I got tired of having to find the mouse to swap a patch, hehe).

[mystran]

This took quite a bit of time, and I'm not yet happy, and not completely sure this new version is stable so it'll have a new separate download right now:

http://mystran.googlepages.com/Valo-unstable.zip

In any case, feel free to test, and report any problems.

Changelog is basically:
- MIDI program change support
- won't calculate LFOs when they aren't used for anything (help CPU a bit)
- interpolation for wheel/touch (was temporarily disabled)
- takes advantage of SSE for mixing mods and audio, and tries to speed up filter
- half-sync mode (only sync every other cycle of master, essentially syncing to octave lower)
- multi-saw "sync" (that does tricky stuff.. basicly sync it like if it was a single morphing wave)
- driftyness retuned (high notes should be cleaner)

Short version: it should run a bit faster (if not, please report, especially if it runs slower) and there's a couple of minor tweaks to the oscillators.

The download is a bit bigger, and there's some internal changes such that it can now run registered version as well (once I get my license-key generator working). I'd like to hear if anyone has any trouble with this version.

I've also switched to proper version numbering now...

If there aren't big problems with this one, I'll be doing some bigger feature additions next. MIDI learn/LFO improvements/Sequencer/Osc-types on TODO (probably in order).