TAL Mod

[ToguAudioLine]

It took me some time to find the "issue". It also happen when you use a rectangle LFO and modulate the filter 1.

After some research i found out that this is a normal behaviour under these conditions:

- Filter in serial mode.
- Play a sound > 1000Hz (example) with ritch harmonics (saw, pulse...)
- Filter 1 with rectangle LFO that modulates the cutoff and goes down to 0 Hz or below 1000Hz (huge range).
- Filter 2 with a cutoff a bit higher 1000Hz that filters out some harmonics of the sound. Or choose 200Hz to filter out the whole sound to get an extreme effect where you only hear the pops.


Now you can hear that low frequency pop sounds at the transition points. Especially when the filter closes (from 20000Hz to 0Hz for example). This is because the filter needs some time to make the frequency jump. What you hear is some type of fast filter sweep from 20000Hz to 0Hz. This is totally natural.

The solution is easy. Just switch the filters and filter the high frequncy with the first one and do the filter modulations (evelope, rect LFO) with the second filter.

Not sure I understand your explanation. But it results from the dual filter architecture, right?
In serial mode there are three cutoff knobs involved since one can't just bypass either filter. Maybe adding a Bypass option in the filter mode drop-down menu would help. Oops, forget it, that's what the parallel mode is for...

Just loaded that patch again and tested stuff.
I still think it has to do with the envelopes or voice stealing as I had originally suspected. From my patch, when I reduce the filter env attack to zero, oddly the clicks go away. With other synths it is the other way round
When the filter env attack is slower than the amp env attack, many synths make the typical synth brass bump, but it seems to be something different.
And as I said earlier, it goes away when I add more voices. With two voices it is still there, but more rarely. With 4 voices I no longer hear the clicks.

Thanks for the additional information

The synth does some fades when he steals a voice, but this only affects the amp.

You can only hear it when the synth re-uses a voice, because it has to make that frequency jump (go down to the attack start point of the filter ADSR). A fresh note (especially when it has filter TRIG enabled) starts with fresh initialized values and don't have to make that jump. It just starts at the start frequency and does not have to jump to the attack start frequency. It also does not have older values of the higher frequency in his history. Maybe all values are zero.

IIR filters remember old values for some time and the filter needs it's time until he forgets the older values and is completely closed.

[ToguAudioLine]

It took me some time to find the "issue". It also happen when you use a rectangle LFO and modulate the filter 1.

After some research i found out that this is a normal behaviour under these conditions:

- Filter in serial mode.
- Play a sound > 1000Hz (example) with ritch harmonics (saw, pulse...)
- Filter 1 with rectangle LFO that modulates the cutoff and goes down to 0 Hz or below 1000Hz (huge range).
- Filter 2 with a cutoff a bit higher 1000Hz that filters out some harmonics of the sound. Or choose 200Hz to filter out the whole sound to get an extreme effect where you only hear the pops.


Now you can hear that low frequency pop sounds at the transition points. Especially when the filter closes (from 20000Hz to 0Hz for example). This is because the filter needs some time to make the frequency jump. What you hear is some type of fast filter sweep from 20000Hz to 0Hz. This is totally natural.

The solution is easy. Just switch the filters and filter the high frequncy with the first one and do the filter modulations (evelope, rect LFO) with the second filter.

Not sure I understand your explanation. But it results from the dual filter architecture, right?
In serial mode there are three cutoff knobs involved since one can't just bypass either filter. Maybe adding a Bypass option in the filter mode drop-down menu would help. Oops, forget it, that's what the parallel mode is for...

Just loaded that patch again and tested stuff.
I still think it has to do with the envelopes or voice stealing as I had originally suspected. From my patch, when I reduce the filter env attack to zero, oddly the clicks go away. With other synths it is the other way round
When the filter env attack is slower than the amp env attack, many synths make the typical synth brass bump, but it seems to be something different.
And as I said earlier, it goes away when I add more voices. With two voices it is still there, but more rarely. With 4 voices I no longer hear the clicks.

Thanks for the additional information

The synth does some fades when he steals a voice, but this only affects the amp.

You can only hear it when the synth re-uses a voice, because it has to make that frequency jump (go down to the attack start point of the filter ADSR). A fresh note (especially when it has filter TRIG enabled) starts with fresh initialized values and don't have to make that jump. It just starts at the start frequency and does not have to jump to the attack start frequency. It also does not have older values of the higher frequency in his history. Maybe all values are zero.

IIR filters remember old values for some time and the filter needs it's time until he forgets the older values and is completely closed.

Sure there are other ways to do voice stealing. You could fade out a sound and play the next one with a fresh initialized filter. We did this in TAL-Sampler, but this also has it's drawbacks. The filter always has to build up for example resonance oscillation and it also introduces a short delay. The most analoge way to keep the filter running

Just switch the order of the filters to avoid this "issue".

[fluffy_little_something]

But what I mean doesn't have to do with filter order. When I load my demonstration patch, switch from serial to parallel filter mode and disable filter 2, the clicks remain the same. Disabling the Trig button doesn't change anything about it, either.

With exactly the same filter settings ancient Sylenth1 doesn't make such low-frequency clicks.

I think the issue is closely related to the other issue I described a few posts ago, i.e. the little bump I hear when pressing the same key repeatedly, cutting off the previous notes release. Sylenth1 also does the complete retrigger thing, but in a smooth way without any audible bumps or clicks, which is important with slow pads.

[fluffy_little_something]

Just made a little comparison where I switched back and forth between two chords, where only one note changes, while the other two notes of the chord always remain the same, first Mod, then Sylenth1.
I tried to match the envelopes as closely as possible, which is not easy as they are rather different. But the last chord for each synth is open-ended, so you can hear the two patches sound very similar...
https://app.box.com/s/caf3dyznbqntd9s7snbnj3zgod4oqprb
(11 MB)

[dreamvoid]

Whatever little bugs there still may be - this software synthesizer sounds just wonderful. While it has not enough modulation possibilities for me, I made some sounds with the demo and listen to the presets and was totally convinced by the smooth, yet powerful sound and bought a license.

[Lotuzia]

I suggested to Mr TAL a few mods and changes :

- ReCalibration of the ADSR Decay Knob in the middle range so that's its easier to find sweets spot for middle range values.
- KB tracking available as modulation source for all mod inputs (Very important parameter to balance patches across keyboard, and make more sophisticated/musical RM patches for example, or rescaling oscs frequency for somewhat 'more subtle FM')
- Enhanced handling of the EG Release parameter 'tail'
- A 3rd EG, and discussing bipolar/unipolar behaviour of the Mod EG

And, afaik, these mods should be available in the next update.

So a big to Mr TAL. I think/hope it will help make Tal-Mod easier to program, and more versatile and pleasurable to hear in certain situations.

[fluffy_little_something]

More factory presets (v 1.0.2)
Bug fixes (v 1.0.2)
Different and cleaner note stealing (v 1.0.2)
Keyboard tracking added (v 1.0.2)
ADSR slider behavior optimized (v 1.0.2)
Another ADSR added (v 1.0.2)
Smoother ADSR release (v 1.0.2)

[ToguAudioLine]

But what I mean doesn't have to do with filter order. When I load my demonstration patch, switch from serial to parallel filter mode and disable filter 2, the clicks remain the same. Disabling the Trig button doesn't change anything about it, either.

With exactly the same filter settings ancient Sylenth1 doesn't make such low-frequency clicks.

I think the issue is closely related to the other issue I described a few posts ago, i.e. the little bump I hear when pressing the same key repeatedly, cutting off the previous notes release. Sylenth1 also does the complete retrigger thing, but in a smooth way without any audible bumps or clicks, which is important with slow pads.

Ok, i can't live with that. I have changed the key stealing behaviour to the common one with the very short fade out before triggering the next note (around 4ms) like other synths do. This way no cutoff sweep is hearable. After some test i also like it more and there is no reason why doing something the analog way when we have such possiblities in the digital world. This also has the advantage that every note (with triggered oscs and filters) has exactly the same attack sound. Thanks for this one, i always like to learn

There are also other improvements like a 3rd ADSR, a more natural release in the ADSR, keyboard tracking as modulation source and more presets(trance and tech house) in the latest beta release. I think the beta is almost finished and feature complete. Thanks everyone for the feedback:

https://tal-software.com/products/tal-mod

[Echoes in the Attic]

Just curious - Did any elements of this synth take inspiration from any classic analog synths? Filters loosely modeled on anything?

[layzer]

Thanks for this one, i always like to learn

Fluffy and his/her close attention to detail will do that to you.
Fluffy was a major contributor to the design of the OBXD gui when i was
arranging knob positions, graphics and such.

Getting input from this forum can be invaluable, having many eyes and ears
at your disposal to hear /see things you may have missed otherwise.

thanks for taking time out to interact with us, Patrick!

[fluffy_little_something]

Ok, i can't live with that. I have changed the key stealing behaviour to the common one with the very short fade out before triggering the next note (around 4ms) like other synths do. This way no cutoff sweep is hearable. After some test i also like it more and there is no reason why doing something the analog way when we have such possiblities in the digital world. This also has the advantage that every note (with triggered oscs and filters) has exactly the same attack sound. Thanks for this one, i always like to learn

There are also other improvements like a 3rd ADSR, a more natural release in the ADSR, keyboard tracking as modulation source and more presets(trance and tech house) in the latest beta release. I think the beta is almost finished and feature complete. Thanks everyone for the feedback:

https://tal-software.com/products/tal-mod

Maybe I am missing something, but the problem is there just like before.
And for some reason it seems that the CPU load is much higher than before. My DAW actually overloads now when playing two 3-finger chords in a row Can anyone confirm the increased CPU load?

[fluffy_little_something]

Fluffy and his/her close attention to detail will do that to you.
Fluffy was a major contributor to the design of the OBXD gui when i was
arranging knob positions, graphics and such.

Well, at times I am a pain where the sun never shines

[fluffy_little_something]

Maybe I am missing something, but the problem is there just like before.
And for some reason it seems that the CPU load is much higher than before. My DAW actually overloads now when playing two 3-finger chords in a row Can anyone confirm the increased CPU load?

Seems the CPU load varies a lot, not sure it has something to do with Mod. Now I did the same thing as hours ago and there is no more overload. Will keep observing...

Regarding the other issue, I noticed that Mod does not use multiple voices for the same note when played repeatedly, while in Sylenth1 playing the same long-release note twice in a row doubles the number of voices used. So the first voice is fading out while the second is already fading in, which creates the smooth transition.

[PatchAdamz]

Some more fun with Tal-Mod:

https://soundcloud.com/vintage-synth-pa ... -tal-mod-1

[AnX]

@TAL: Can we lose the cables a have a MM, as thats badically all it is. Much more user friendly.

Thanks