Looks awesome.
Was just trying it out with some wavetables I already had, which are 128 waves of 2048 samples each. They don't look right. Such big waves, I guess will not work?
nbWaves is how many waves? And length is the size of each individual wave?
Thanks!
The preparation for antialiasing implies 8 bytes per sample... and you've got 8MB of RAM...
not 100% sure what that means?
What size should I make wavetables in to be able to use them with this object? 
Sorry for the short answer, i was in the metro.
Now i'm in front of my computer and i can answer more accurately.
The length of the waves are limited to 512 (that's because of arithmetic limitations of the anti aliasing algorithm).
A wave should represent a cycle of a waveform.
The number of waves can go up to 1024.
Many instances of the load object can be used but the total SDRAM used cannot go beyond 1 mega samples.
Hi @SmashedTransistors, the wavetable objets are great. The interpolation and the phase modulation sounds amazing! Thank you!!!
Just a few questions:
What disto exactly does? It seem to no work properly to me. When active I get a very low frequency signal...
What loadAtinit does?
Also I'm starting to experiment with the sss/SuperTable object that can play a detuned stack of waves (like a supersaw but with wavetables). But it works with 32 bit wavetables, so I have to convert mine to 32 bit. Also I think your phase modulation sounds better (I have to do furter comparitions). Can you add somenthing like the supersaw feature?
Thank you.
The XT/load is designed to work with the XT/osc it won't be compatible with other table oscillators / readers.
XT/load loads 16 bit littleendian wavetables (as provided by @lokki) and performs a pre integration so that the XT/osc can perform integro-differential antialiasing (including when phase modulated).
I did not experienced low frequency buzz. Can you post the patch or a snapshot.
In distortion mode, the module behaves as a distortion, in this mode, the audio input is the pm input.
when loadAtinit is 'yes' the file indicated by the string filename will be loaded on init. It will be reloaded (or another one if you use a string multiplexer) when you will use the trig input.
When loadAtinit is 'no' the file will only be loaded when triggered.
Thanks for your response!
I know that XT/load and XT/osc only work with 16 bit tables. I tried with my own tables without problem, I'm using this format for other projects.
May be I wasn't clear (or messed up what I tried to say). I used other wavetable modules to make a polysynth, recently I was trying to use sss/SuperTable that uses 32 bit x 1024 samples waves, so I'm converting my wavetables to that format, but considering the limited resources of Axoloti, using tables in that format is a bit overkill for what I want to do.
Your module is pretty close to what I want (wavetable morphing with FM input and very good antialiasing), but I'm missing that stack detuning feature (like a supersaw). I tried using a chorus module in the past (a 4 voices one) and works well, but if I want to insert the chorus before the filter (to mimic the stack detune) in a poly patch the DSP consumption becomes too high.
So I was asking if it is possible for you to add that feature to the XT/osc. Sorry if I keep asking too much...
About the low frequency buzz, I didn't save the patch as I was just trying your module, but basically I added a toggle button in the disto input and a bipolar knob in the disto offset input in your example patch.
I was expecting that the output will be the wavetable distorted in some way... Now that you explained what it does it totally makes sense (it works as a waveshaper as I understand), and it is a great feature.
Thanks for this module!!!
With the PM and antialiasing process there is almost no way to optimize detuned osc.
You should use a single chorus common to all voices. Is there a specific reason why you need it to be before the filter ?
@SmashedTransistors, yes, I want it before the filter mostly because it sounds different if I filter the detuned oscilators (or chorused) that if I add chorus after filtering a single oscillator, specially in a polyphonic patch.
I think I described what I'm trying to accomplish before.
I have this digital module (Neutron Sound Orgone Accumulator) based on Teensy 3.1 (Open Source if you want to look at it) that have wavetables (you can morph between 3 waves), you can do FM and AM with another wavetable oscillator (and also have another kind of modulation called Pulsar) and you can have a detune FX (5 copies of the oscillator detuned), plus some other features. I use to send the output of this module to a VCF module modulated by an ADSR. I love the sounds I get out of it, but is monophonic... So I'm trying to do something similar in Axoloti but polyphonic. Until now I did some good patches that I like but the detune feature is a bit elusive... with the chorus a the output I get a nice sound but not the same as the detuned oscillators before the filter stage... and if I add the chorus before the filter I can't make it polyphonic.
sounds great! i just gave it a try.
pm modulation introduces aliasing again, i guess that is to be expected, right?
really great sounds, thanks!!
PM should not introduce much aliasing, if the modulating signal is not aliased itself.
Aliasing can also be induced by post oscillator distortions and saturations.
hmm. i tried modulating with a sine oscillator (not much aliasing) and with another of your wavetable objects (quite heavy aliasing).
Can you post the patch and wavetable ?
I always check for saturation in the audio path (most standard axoloti objects have a __SSAT instructions in them) they cause aliasing when the audio signal is too strong.
I have made some 64x256x16bits raw wavetables, do you know how i can share them ?
I will have a look at those, i never used these before.
At first, I can share them in the forum but i don't know if it is OK to post 32kbytes files on the forum. (@thetechnobear ?)
Wetransfer is really easy to use. You just drop the file into wetransfer window and they will give you a link or they can send a link to you, that you can pass on to others. But one issue is, that it will deleted after 7 days.
I don't think you even can post those kind of files here. I think you can only post .axo files, pictures and text. Everything else needs to be hosted externally. It probably has to do with the amount of space it would take up, if we were allowed to post all kinds of files, like audio files. Audio files alone could fill up the server pretty fast, I imagine
i would discourage from using WeTransfer on forum links, since the download becomes unavailable after 7 days as @jaffasplaffa pointed out. a dropbox link works much better.
that is probably it! i will watch my levels in a simple patch, if i get aliasing again, i will post the patch here, thanks
Yes it's not the best solution, I agree, for forum sharing.
One could also use some of the free cloud services around, that gives you some free space. I used to use MEGA a lot but over time they added some limitations for non payers, like limited amount of downloads pr day and stuff like that. Dunno if other companies that offers free services does the same, though. Some probably does.