I spent some time this month to try and get [list] to work for the heavy compiler.

There are some limitations, but things are mostly working: https://github.com/Wasted-Audio/hvcc/pull/381/changes

We could really use some nice examples to try with this!

How to create a Polyphonic Synthesizer using the Befaco Lich:

@EEight Last year Link was updated to include sending and receiving audio.

The original Link was only used for synchronizing tempo/transport.

Sorry, can you elaborate?

@dreamer This is resolved now, although it doesn't seem to be fully functional on my machines (mainly not getting audio).

[edit: ok the instructions in the help files were not clear (or missing). The peer and channel names have to be adjusted so they match up. Then it works!]

@dialer_vox I recommend that you use plugdata and enable "compiled mode" in the menu.

This will give you early warnings about incompatible objects, so you don't create a patch that ends up not working with the compiler.

I don't have any immediate suggestions on how to replace these two objects. Sorry.
(since we can't do FFT in Heavy, due to its single sample processing, I don't think you can)

@julienbayle The repo is missing pdlibbuilder third party module and a Makefile in the pd directory

So right now it's not possible to build this.

Please also submit to deken!

Ok it's tagged with version 0.1.0 on deken and the package is called PDNAM and should contain both externals, help files and testing models.

How are these different from the Heavylib polybleb oscillators? (I see you are still using some Heavy based subpatches)

I see only minor modifications, like to the square there is a phase control and the width has direct signal input.

The triangle is new I suppose. Maybe we can clean it up and port it

So I decided to split this into two separate externals.

There is now pdnam~ which uses the same MicroNAM implementation as in Heavy. It can load only WaveNet models Nano, Lite, Feather, and Standard

Then there is pdnam+~ which uses the full NeuralAudio implementation with LSTM and various model weight sizes.

This ensures that the basic implementation is the same for PD vs Heavy and we can now allow feature-creep to happen on the NeuralAudio version if users request functionality in the future

Test builds are here: https://github.com/Wasted-Audio/PDNAM/actions/runs/25001888719

Yeah deken search is very picky, the exact external name is pdnam~ and you can't use any fuzzy search like just nam or pdnam without the ~. (plugdata search is better at this)

Unless you us a wildcard like *nam*

v0.0.1 is now available on Deken!

Please test so I can move the version up

Ps; for macOS the minimum required version is 10.15 die to std::filesystem requirements.

Last weekend I made a very basic NAM loader external: https://github.com/Wasted-Audio/PDNAM

This is still very beta, but it seems to work alright on my machine! So please test and let me know if it works

Latest builds are here: https://github.com/Wasted-Audio/PDNAM/actions/runs/24035689961

What I'd also like is some help with proper packaging for deken. I'm using pdlibbuilder right now, but I'm not sure how to get multiple architectures into a single package. Or if there are any other requirements that I'm overlooking.

Are people publishing to deken directly from CI, or is this done by hand? Right now there are separate CI builds, but I assume for deken everything has to be combined?

On Thursday April 2nd, 18:00 there will be a pure-data workshop with Mike Moreno @TimothySchoen and I, hosted at HKU IBB-Laan, Utrecht!

The topics will be livecoding, the heavy compiler, plugdata and more.
We know it's a bit last minute, but we hope to see some of y'all there! 🇳🇱 🙂

https://www.hku.nl/en/contact-and-locations/location-ibb-laan

@asher-vaynne said:

Max has a more polished interface

We have https://plugdata.org for that

@Balwyn Nice! DSP settings don't do anything in Heavy btw, it's always running (and it's not needed for metro).

Changing to a boolean value for the toggle probably fixed it

@Balwyn Oh you need to setup special receive objects to create plugin parameters:

https://wasted-audio.github.io/hvcc/getting-started/patching/
https://wasted-audio.github.io/hvcc/generators/dpf/

@Balwyn Eh?

Probably because we don't sign binaries on Windows.

Also: who still uses McAfee? o.O

I could also export it with the heavy compiler, but the effect is really silly

@Orangery You have to install plugdata and open the pd patch from there.