Voice capability

#1

Can somebody know-tested how much voices is capable to play at same time from Axoloti wit ARM @ 168Mhz?
I know that depend on complexity of modeled sound-source.
Voice capability
If MI Shruthi wit pretty complex monophonic Synth engine + Arp + Gate + Simple sequencer and lot of modulations can provide one voice wit ATMEGA @ 20Mhz.
Following that logic (One complex sound) @ 20Mhz from ATMEGA
Then ARM 168Mhz can play approx 8 voices poly.
If one voice use 20Mhz from CPU.

#2

I don't think it's really possible to say. It completely depends on the patch.

a|x

#3

they are completely different chips, to start with isn't the atmega used 8 bit? (vs Axoloti 32bit),
so comparing clock speeds is meaningless.
(also shruthi is using a different filter board, so that wont load the cpu)

as you say

you could create lots of voices if the voice was simple (e.g. saw with simple filter), but make a complex enough oscillator with complex filter, and fx and you might be down to one!

so impossible to generalise...
but to give you an idea...
if I'm careful, and not crazy stupid with fx, I can get 16 voices of reasonable oscillator/filters.
if I want more complex fx, and perhaps complex filters this sometimes will drop down to 8.

e.g. some of the mutable instruments oscillators/fx will knock your voice count down...

often now, I put the main voices on one Axoloti, and then feed its output into another Axoloti for the FX.

have a look at my you tube channel and others Axoloti users , and you will get an idea of what is possible.
(or should I say, what we want to use our Axoloti for, rather than is possible :wink:)

its wise to keep expectations 'reasonable', especially when we are all used to very powerful pc, an Axoloti is a microcontroller, which is many times less powerful that your average desktop computer.
with that said, I was surprised how much was possible with Axoloti when I first got mine, particularly for the price... and its standalone nature is awesome.

anyway, as only you know what you have in mind (and your patching/code skills can play a part too), really you can only know what is possible by trying.
I also found once I got patching with Axoloti, my idea of what I was going to use it for changed/evolved, I to play to its strengths.

EDIT: btw, when I refer to voices here, its worth pointing out, often these will have multiple oscillators in them - so even my simple 16 voice patch. will have a few oscillators inside, for detuning, or sub oscillators, etc. ... and of course the number and type of oscillators within one voice, will have a big impact on cpu... (e.g. think about FM, and how many oscillators you might use in one voice for that!)

#4

As you guessed, I would say something between 3 and 8 if you use several factory anti aliased oscillators, some filters and even some fx.

Here is an example of an 8 voice sound (no external fx, all Axoloti):

Arp, gate and basic sequencing have much a lower impact on CPU than audio rate objects such as oscillators and filters.

#5

@thetechnobear
Thank you for all info! :slight_smile:

Wow so many things by you to check on YouTube.
Really nice setup, i must give time to watch them all before ask questions.
Meanwhile, most of all, there will be lot of experimenting-testing-playing to achieve my idea but first i need "green-light" ...
Thank you Mark!

#6

@SmashedTransistors
I never hear so impressive sound. Thumbs UP!
Amazing!!!

Did you sequence that or played live and record?

Sounds like old works of Vangelis synth brass :slight_smile:

#7

The polyphonic aftertouch control of the filter is most responsible of the "Vangelian" color...