Baby Audio Tekno - Drum Synth

[m-ac]

Anyone have any ideas what might be causing/controlling the fluctuation in the sound of the kick? Each midi note is the same velocity and length, and humanisation is off.

https://drive.google.com/file/d/1ypM6iM ... sp=sharing


Screenshot 2025-09-06 at 11.16.24.png

From the picture, nothing. But that's my main issue with this synth. Hard to get stable hats, kicks and so on. The reverb exaggerate the matter (every reverb tail sounds different). I made a feature request to add a phase reset. Because these fluctuations on a drum synths are rudimentary bad side effects of those classic drum machines. Back in the days people had to sample them to get a stable sound.

Well if this is how it's designed, it definitely needs some kind of 'analog' slider which ranges from 0-100 for each engine, because that much variation as a minimum is complete overkill. I thought that's what humanise was for? It's pretty unusable in its current state, at least for me personally.

[kraster]

Anyone have any ideas what might be causing/controlling the fluctuation in the sound of the kick? Each midi note is the same velocity and length, and humanisation is off.

https://drive.google.com/file/d/1ypM6iM ... sp=sharing


Screenshot 2025-09-06 at 11.16.24.png

Kick A is a damped feedback network. ie. An impulse into a resonating filter with nonlinearities.

Length controls the damping factor of the resonance and shape controls the shape of the impulse. A wider shape = more energy into the filter.

I think what's happening here is that the filter is still resonating so that each consecutive hit is carrying some of the energy of the previous hit leading to instability in the sound.

The three drive controls in the calibration section control the distortion and resulting harmonics within the filter and the damping factor controls the shape of the filter.

Those three drive controls can really alter the behaviour. Reducing them can stabilise the sound as can reducing the shape (impulse shape)

Also Kick B is probably better for the kind of sound your looking for.

[Igro]

Anyone have any ideas what might be causing/controlling the fluctuation in the sound of the kick? Each midi note is the same velocity and length, and humanisation is off.

https://drive.google.com/file/d/1ypM6iM ... sp=sharing


Screenshot 2025-09-06 at 11.16.24.png

Kick A is a damped feedback network. ie. An impulse into a resonating filter with nonlinearities.

Length controls the damping factor of the resonance and shape controls the shape of the impulse. A wider shape = more energy into the filter.

I think what's happening here is that the filter is still resonating so that each consecutive hit is carrying some of the energy of the previous hit leading to instability in the sound.

The three drive controls in the calibration section control the distortion and resulting harmonics within the filter and the damping factor controls the shape of the filter.

Those three drive controls can really alter the behaviour. Reducing them can stabilise the sound as can reducing the shape (impulse shape)

Also Kick B is probably better for the kind of sound your looking for.

So what about the short sounds, like hats and so on. Some instabilities are too disturbing. As someone mentioned, what is the point of humanisation slider, if at zero value it is already unstable.

[Gam456]

Anyone have any ideas what might be causing/controlling the fluctuation in the sound of the kick? Each midi note is the same velocity and length, and humanisation is off.

https://drive.google.com/file/d/1ypM6iM ... sp=sharing


Screenshot 2025-09-06 at 11.16.24.png

Kick A is a damped feedback network. ie. An impulse into a resonating filter with nonlinearities.

Length controls the damping factor of the resonance and shape controls the shape of the impulse. A wider shape = more energy into the filter.

I think what's happening here is that the filter is still resonating so that each consecutive hit is carrying some of the energy of the previous hit leading to instability in the sound.
The problem it's a global tweak.

The three drive controls in the calibration section control the distortion and resulting harmonics within the filter and the damping factor controls the shape of the filter.

Those three drive controls can really alter the behaviour. Reducing them can stabilise the sound as can reducing the shape (impulse shape)

Also Kick B is probably better for the kind of sound your looking for.

So what about the short sounds, like hats and so on. Some instabilities are too disturbing. As someone mentioned, what is the point of humanisation slider, if at zero value it is already unstable.

You can tweak how you want to humanize. synth params, time and accent
Unfortunately it's a global setting

[kraster]

Anyone have any ideas what might be causing/controlling the fluctuation in the sound of the kick? Each midi note is the same velocity and length, and humanisation is off.

https://drive.google.com/file/d/1ypM6iM ... sp=sharing


Screenshot 2025-09-06 at 11.16.24.png

Kick A is a damped feedback network. ie. An impulse into a resonating filter with nonlinearities.

Length controls the damping factor of the resonance and shape controls the shape of the impulse. A wider shape = more energy into the filter.

I think what's happening here is that the filter is still resonating so that each consecutive hit is carrying some of the energy of the previous hit leading to instability in the sound.

The three drive controls in the calibration section control the distortion and resulting harmonics within the filter and the damping factor controls the shape of the filter.

Those three drive controls can really alter the behaviour. Reducing them can stabilise the sound as can reducing the shape (impulse shape)

Also Kick B is probably better for the kind of sound your looking for.

So what about the short sounds, like hats and so on. Some instabilities are too disturbing. As someone mentioned, what is the point of humanisation slider, if at zero value it is already unstable.

Humanisation affects parameters, timing and accents. It's a completely different thing.

Again, the principle for a lot of the engine is a pulse into a resonating filter.

It really helps to think of Tekno on a circuit level. A lot of the values you've probably noticed use resistance or other circuit based values.

The 808 kick is a good example. It was a pulse into a bridged T filter tuned to 55hz. That caused a "ping" in the filter that was then fed through an amplifier back into the filter to create a feedback loop so that the filter was self-resonating. The amplitude of the signal being fed back into the filter is what determined the decay time.

Hence the decay knob on the 808 kick controlled the amplitude of the feedback and therefore the length of decay.

The characteristic pitch envelope of the 808 kick was a side effect of the initial pulse pushing energy into the filter causing it to start at a higher pitch before settling into its tuned frequency (55hz).

So there was no pitch envelope in the 808 but the pitch slide was a side effect of component overshoot. The accent control on the 808 kick was effectively giving more amplitude to the impulse causing it to start at a higher pitch and excite the filter even more.

So with all that in mind this is the basis for a lot of Tekno's engines but because you can change parameters way beyond something like an 808 you can get all kinds of weird behaviours.

The general rule of thumb is that the less energy you put in the more stable the circuit becomes (less overshoot and less lingering "voltages")

With Hat A I find if I turn down the impulse scale and feedback all the way in its calibration settings it will become more predictable.

Hat B has a lot more going on. It's based on the same principle of the 808 hats which had a collection of square waves that are tuned inharmonically into a filter. You have control over the impulse width and attack. Adjusting the attack can give you less initial energy. You also have access to the filter (resonator) Scale and feedback.

I'm not having much trouble getting relatively consistent sounds from Tekno overall. I say relatively because the point of Tekno is that it behaves like a circuit with all its associated character.

If you want an "every hit is the same" type of approach you can always drag the sample out from an engine and use that.

[m-ac]

If you want an "every hit is the same" type of approach you can always drag the sample out from an engine and use that.

Thanks for the in-depth replies. It's probably just not best suited for my needs. I dig the concept and do like the analog instability, but not that much.

[bite_me]

D16´s Nepheton should be a good sequencer for Tekno. Will try asap. Like the juicy sound of it. Well done Baby Audio! Recommand TAL DAC right after it.

[pekbro]

This is the sequencer I have been liking with this, not perfect, but pretty cool... it generates a whole song of patterns, but the beat generation in something like playbeat is a bit better imo. This still does the job well though, I think.

https://www.axart.net/autobeat/

Anyway, just sayin in case someone finds it interesting and never heard of it.

[Ah_Dziz]

I'd love to see far more flexibility if they wanted to lean on the "analog circuitry" angle. Adjustable supply voltage and signal voltage as well as the feedback resistance and filter capacitance. Being able to trigger the modules with arbitrary audio would also be cool. If they wanted to go so in depth with the models they could have made each voice a similar "circuit" that we could mess with and supplied presets for different sounds. Unfortunately, as it is it's pretty much more useful as an expensive generator of analog style drum samples.

[kraster]

I'd love to see far more flexibility if they wanted to lean on the "analog circuitry" angle. Adjustable supply voltage and signal voltage as well as the feedback resistance and filter capacitance. Being able to trigger the modules with arbitrary audio would also be cool. If they wanted to go so in depth with the models they could have made each voice a similar "circuit" that we could mess with and supplied presets for different sounds. Unfortunately, as it is it's pretty much more useful as an expensive generator of analog style drum samples.

Just like analog drum machines don't use the same circuit topology for different drum sounds the whole idea with Tekno is having various bespoke analogue models tuned to specific tasks.

The 808 had different synthesis types depending on the drum type. Pulse into resonator for the kicks and Toms, twin oscillators plus noise source for the snare, inharmonic square waves for the hats, twin square waves for cowbell, noise burst plus multiple envelopes for handclaps .

Each of these had specifically tuned filters, envelopes, impulse responses, oscillators etc.

That's what Tekno is leaning into in terms of analog modelling. It's not an open ended system because analog drum machines themselves are very specifically tuned to produce their results.

A lot of the engines DO have things like feedback resistance, capacitance and voltage amounts. Pulse Scale is effectively a voltage setting. A lot of the feedback controls have resistance values. But there's only so much you can expose to the user before the use case falls apart.

The problem with having too open ended a design within this approach is that a lot of the component value ranges are sweet spot ranges just like in the real drum machines and anything outside of that becomes useless for the purpose.

You'll note in Tekno that the ranges of the parameters you can alter are very narrow for that reason.

In other words it's as flexible as is practicable.

I think Microtonic is great as a one size fits all style drum synthesiser but there's definitely space for a different approach that's more rooted in classic techniques that models the per sound particulars.

[bite_me]

Simplicity is their best feature, me thinks, while it delivers a good range of sounds quickly. Like BA-1, which has no this and that. Not everything has to be "The Grid". Cymbals and rides are not my cup of novichok though.

[SebAV]

..
I think Microtonic is great as a one size fits all style drum synthesiser but there's definitely space for a different approach that's more rooted in classic techniques that models the per sound particulars.

Exactely. Just a better UI and it could be the Microtonic friend a lot of people are wanting.

[DrGonzo]

I think Microtonic is great as a one size fits all style drum synthesiser but there's definitely space for a different approach that's more rooted in classic techniques that models the per sound particulars.

What continue to impress me about Microtonic year after year is that the only engine it's using is so damn capable. In some regards it is way more capable than my Elektron Analog Rytm, although that one have so many specialised engines.

/C

[morphex]

Now that I have spent some time working with Tekno, I must say that you can dial in some really cool tones. But it isn’t immediate, and I agree with others regarding some of the points made about the UI.

Personally, this thing is absolutely screaming for the addition of an onboard sequencer. I understand that it requires another layer of complexity, but it just feels incomplete without it.

I think any drum machine style plug-in in 2025 should be capable of generating sequences and patterns, and have it sync to the DAW’s tempo. It’s part of what makes groove boxes and drum machines so fun to use.

Does anyone know if Baby Audio has commented on whether or not they would be willing to include a sequencer in a future update? Having that along with the randomization features would elevate this product, and I would be much more likely to use it on a more consistent basis.

[Ah_Dziz]

I'd love to see far more flexibility if they wanted to lean on the "analog circuitry" angle. Adjustable supply voltage and signal voltage as well as the feedback resistance and filter capacitance. Being able to trigger the modules with arbitrary audio would also be cool. If they wanted to go so in depth with the models they could have made each voice a similar "circuit" that we could mess with and supplied presets for different sounds. Unfortunately, as it is it's pretty much more useful as an expensive generator of analog style drum samples.

Just like analog drum machines don't use the same circuit topology for different drum sounds the whole idea with Tekno is having various bespoke analogue models tuned to specific tasks.

The 808 had different synthesis types depending on the drum type. Pulse into resonator for the kicks and Toms, twin oscillators plus noise source for the snare, inharmonic square waves for the hats, twin square waves for cowbell, noise burst plus multiple envelopes for handclaps .

Each of these had specifically tuned filters, envelopes, impulse responses, oscillators etc.

That's what Tekno is leaning into in terms of analog modelling. It's not an open ended system because analog drum machines themselves are very specifically tuned to produce their results.

A lot of the engines DO have things like feedback resistance, capacitance and voltage amounts. Pulse Scale is effectively a voltage setting. A lot of the feedback controls have resistance values. But there's only so much you can expose to the user before the use case falls apart.

The problem with having too open ended a design within this approach is that a lot of the component value ranges are sweet spot ranges just like in the real drum machines and anything outside of that becomes useless for the purpose.

You'll note in Tekno that the ranges of the parameters you can alter are very narrow for that reason.

In other words it's as flexible as is practicable.

I think Microtonic is great as a one size fits all style drum synthesiser but there's definitely space for a different approach that's more rooted in classic techniques that models the per sound particulars.

These are all good points about analog hardware. Not so much about software models of analog hardware. There's no constraints here about parameter ranges or available ICs or cost of adding knobs (they've already made a clunky second parameter area). These are pretty good sounding models but they have (for me) badly tuned parameter (or needlessly limited) ranges and other parameters that aren't adjustable at all. Adjusting the available headroom in each circuit would be a simple to implement tweak that would yield massively different sounds.