Released: Free Synthesizer Odin 2 [Win / Mac / Linux]

[Erisian]

Awesome - thank you!

[exmatproton]

yeah.,.great update. Tested the OSC as mod to parameter and it works great! The dist per voice sounds very nice as well! It is also great to have 2 filters per voice and 1 master filter. Great choice

[tomx2188]

Awesome update! Dist per voice is great! Dist on the master can be achieved by adding fx after the plugin, so yeah, we don't really need that (imho).

Awesome update indeed!
By the way, you can still use filter 3 to slap some distortion on top of the combined voices Having a saturation knob for various filter models is really great

It's an amazing synth already!
I do have a simple feature request though, if that's okay:
The ability to set a unison count, so each note triggers a stack of N voices instead of always 1.
In addition to that, "stack index" as modulation source, which would assign each triggered voice a value in [-1,1] (maybe a unipolar variant too, in [0,1])

Since the amp is now polyphonic as well, one could easily use a bipolar stack index to pan voices across the stereo field .. opens up lots of other possibilities as well.
What do you think?

[SeBaer]

Hello Frederick,

thanks for the update, and also for abandoning the AVX dependency. Now I can get updates, too.
I have a minor issue with the Specdraw Oscillator. When drawing a spectral band, it's drawing to the right of the curser and not in the middle. It's not a problem, when you want to draw wide sweeps, but if you want to edit a specific band, it get very fiddly. Increasing the GUI-size doesn't help with the accuracy.
Wavedraw Oscillator and Chipdraw Oscillator work as expected.

Cheers!

[TheWaveWarden]

I do have a simple feature request though, if that's okay:
The ability to set a unison count, so each note triggers a stack of N voices instead of always 1.
In addition to that, "stack index" as modulation source, which would assign each triggered voice a value in [-1,1] (maybe a unipolar variant too, in [0,1])

Since the amp is now polyphonic as well, one could easily use a bipolar stack index to pan voices across the stereo field .. opens up lots of other possibilities as well.
What do you think?

This is a very interesting suggestion - not implementing unison and a stereo field for individual oscs are probably the biggest gripes I have in hindsight. Because implementing this would need to rework the entire synth structure.

However, I never considered using the existing voice system for this task. This is as simple as it is genious. (Other synths do it like this?) I will definitely put this on my todo list. I like the suggestion with the modulation source as well!

The downside of using the voicing structure would be that oscs are not usable in unison individually. Also the amount of voices would be limited by 12 / N. But more choice isn't bad I guess.

EDIT: Thinking about it I might even increase the number of voices... After all it wouldn't really hurt. Takes some more memory for a plugin instance, but no CPU-load increase - unless you use more voices that is

[TheWaveWarden]

Hello Frederick,

thanks for the update, and also for abandoning the AVX dependency. Now I can get updates, too.
I have a minor issue with the Specdraw Oscillator. When drawing a spectral band, it's drawing to the right of the curser and not in the middle. It's not a problem, when you want to draw wide sweeps, but if you want to edit a specific band, it get very fiddly. Increasing the GUI-size doesn't help with the accuracy.
Wavedraw Oscillator and Chipdraw Oscillator work as expected.

Cheers!

Could you provide me with your OS / DAW? Happens on both big and small GUIs? Feel free to pack all the information into a bug-report on my homepage, then I'll have all the info compiled there!
I noticed that the draw behaviour is different on every machine for some reason. I fix it on one, then it is off on another - I've worked over this more than once already

[BlitBit]

Note also that the amp-envelope application was moved after the distortion: This means the distortion is now evenly applied. Before it would be much stronger when the amp-env was "loud" and less strong or none at the beginning and end of the note.

I think distortion might be more interesting if it comes after the amp envelope. Consider a guitar amp that set to be "on the edge of breakup". This means that it's set up such that if notes or chords are played rather gentle (low volume) the amp will sound "cleanish". However, if you really dig into the strings (high volume) then the amp will sound distorted.

In a synth context you can simulate this by applying velocity sensitivity to the amp envelope. With such a setup you can then vary the amount of distortion by hitting the keys softer or harder.

Putting the distortion after the amp envelope would make the synth more expressive and patches more interesting. If you want everything to sound the same with that setup you could for example add a gain stage before the distortion so that the distortion can be driven hard regardless of the input.

[TheWaveWarden]

I think distortion might be more interesting if it comes after the amp envelope. Consider a guitar amp that set to be "on the edge of breakup". This means that it's set up such that if notes or chords are played rather gentle (low volume) the amp will sound "cleanish". However, if you really dig into the strings (high volume) then the amp will sound distorted.

I think there is some confusion here. The structure right now is:

voices->amp->distortion->amp envelope

So basically the amp-envelope is not applied inside the amp but after the distortion. I know this seems confusing.
What this means however is: you can totally control the distortion by hitting your keys harder: Just route MIDI-velocity to amp-gain or distortion-boost.

The thing with this distortion in particular is: I implemented it in a very crude way, for example:
Signal surpasses a threshold and get's clipped. If the signal passes below the threshold now, there is an really awkward transition from no distortion at all to suddenly having distortion. This is circumvented now.

[tomx2188]

This is a very interesting suggestion - not implementing unison and a stereo field for individual oscs are probably the biggest gripes I have in hindsight. Because implementing this would need to rework the entire synth structure.

However, I never considered using the existing voice system for this task. This is as simple as it is genious. (Other synths do it like this?) I will definitely put this on my todo list. I like the suggestion with the modulation source as well!

The downside of using the voicing structure would be that oscs are not usable in unison individually. Also the amount of voices would be limited by 12 / N. But more choice isn't bad I guess.

Thanks for considering this!

There are some synths that do it pretty much like this, U-He Tyrell N6 for example.
That one is limited to max 8 voices total, so whether you run into trouble really depends on your patch and how you play it. It's perfectly fine for monophonic lines though, and also great for drums and percussion

When I started learning about synths, Tyrell N6 was among the first ones I tried. After that, synths that implemented unison on an oscillator basis struck me as a bit odd at first. Just acquired taste I guess

[Erisian]

Hello Frederick,

thanks for the update, and also for abandoning the AVX dependency. Now I can get updates, too.
I have a minor issue with the Specdraw Oscillator. When drawing a spectral band, it's drawing to the right of the curser and not in the middle. It's not a problem, when you want to draw wide sweeps, but if you want to edit a specific band, it get very fiddly. Increasing the GUI-size doesn't help with the accuracy.
Wavedraw Oscillator and Chipdraw Oscillator work as expected.

Cheers!

Could you provide me with your OS / DAW? Happens on both big and small GUIs? Feel free to pack all the information into a bug-report on my homepage, then I'll have all the info compiled there!
I noticed that the draw behaviour is different on every machine for some reason. I fix it on one, then it is off on another - I've worked over this more than once already

Both myself and SeBaer have this problem and we both couldn't run with AVX. Could there be a connection?

[BlitBit]

The thing with this distortion in particular is: I implemented it in a very crude way, for example:
Signal surpasses a threshold and get's clipped. If the signal passes below the threshold now, there is an really awkward transition from no distortion at all to suddenly having distortion. This is circumvented now.

If hard clipping is the only type of distortion that you have right now then I recommend watching this video.

[SeBaer]

Could you provide me with your OS / DAW? Happens on both big and small GUIs? Feel free to pack all the information into a bug-report on my homepage, then I'll have all the info compiled there!

Just submitted. Unfortunatetly, now revitings this thread, I just realized I forgot to mention that changing GUI size makes no difference...

[TheWaveWarden]

If hard clipping is the only type of distortion that you have right now then I recommend watching this video.

I have some more, but they're all threshold based. I don't know this video though, I'll have a look at it thanks

[TheWaveWarden]

Just submitted. Unfortunatetly, now revitings this thread, I just realized I forgot to mention that changing GUI size makes no difference...

I'll keep it in mind, no worries!

we both couldn't run with AVX. Could there be a connection?

I'd be really surprised if there was
EDIT: Actually, I think the issue was on all systems... I was just careless with the biggui update
Fix will be coming with 2.0.19

[Tj Shredder]

I think distortion might be more interesting if it comes after the amp envelope. Consider a guitar amp that set to be "on the edge of breakup". This means that it's set up such that if notes or chords are played rather gentle (low volume) the amp will sound "cleanish". However, if you really dig into the strings (high volume) then the amp will sound distorted.

I think there is some confusion here. The structure right now is:

voices->amp->distortion->amp envelope

So basically the amp-envelope is not applied inside the amp but after the distortion. I know this seems confusing.
What this means however is: you can totally control the distortion by hitting your keys harder: Just route MIDI-velocity to amp-gain or distortion-boost.

The thing with this distortion in particular is: I implemented it in a very crude way, for example:
Signal surpasses a threshold and get's clipped. If the signal passes below the threshold now, there is an really awkward transition from no distortion at all to suddenly having distortion. This is circumvented now.

Usually distortion is applied by waveshaping. For example an arctan function if you want the most simple one (clipping is even more simple but a bit too nasty). Anyway, if you apply it before the envelope, its just the same as a different but static waveform. If you apply it after the envelope, you get a waveform which varies with the envelope. If you apply it after summing of the voices, you get all these difference and summed frequencies we know from guitar distortion. There isn't really a problem if we leave that guitar type distortion to external fx though.
Very interesting waveshaping functions are actually chebyshev functions, as you can control single overtones with it. Haven't seen a synth with this yet though...