Max-MSP objects based on the phonos library for audio synthesis based on physical modeling. As of now, contains almost exclusively an object related to voice synthesis named phonos.voice~.
The overview patch should appear in the Extras window of Max as phonos-Overview. From there, you can directly open a few demo patches.
The voice object is the main synthesis object of the package. It consists of a physical modeling based object for vocal synthesis. If you are interested in knowing more about the model and algorithm, you'll need to wait until I'm not lazy anymore for me to publish a paper on that... In the meantime, some material is available in the phonos documentation.
At its core, the model can be viewed as a blackbox synthesis block:
The object wraps around this diagram, and, while it gives access to the low level input parameters, it also provide a higher level, reduced-order parameter set consisting of 3 muscle activation parameters instead of the 18 vocal folds parameters, and 2 formant frequencies for tuning the vocal tract geometry. The multichannel output also contains radiated pressure as well as vocal fold masses displacements, glottal flow, and glottal pressure drop.
Here is an example interface built around the object:
A Max for live instruments is built around the voice object. It is not controlled by midi, but rather through low level physical controls.
An audio input can be chosen in the foldable window on the left on the device and is in this case added to the sub-glottal pressure. This can create interesting effects.
The trackpad object only works for MacOS. It outputs a raw access to trackpad measurements (finger positions, pressure, orientation, size...). It supports up to 11 fingers (yes, some guys at Apple apparently thought humans have 11 fingers) in multi-finger mode (the single finger mode allows to "focus" on a single finger). Here is an example of finger touches visualized as ellipses with 4 fingers engaged:
The object can be sphonosd by sending it the message open 0 and stopped by sending it stop 0. Using a higer interger value is possible when external trackpads are connected (multiple trackpads should be able to work together, this is however untested as I don't have access to an external trackpad). The freeze message allows to disable the use of the trackpad for anything other than interactions through the objects (i.e. the cursor disappears, as well as MacOS navigation gestures). Hitting the escape key will release the cursor.
Caution
Due to the lack of better solution, the freeze message internally blocks all communication related to the MacOs Dock process by execution the command pgrep -x Dock | xargs kill -s STOP. As a result, navigation gestures (2 or 3 finger swipes) are deactivated, as well as access to the dock. If you close max normally, everything will come back to normal. However, if Max were to crash, you should recover the use of the mouse cursor but not the Dock-related features, which is REALLY panicking when it happens. The solution in that case is to open a terminal (the cmd+space spotlight research + typing terminal will do so) and execute the command pgrep -x Dock | xargs kill -s CONT. Restarting the computer will also fix the problem in a desperate scenario.
If you wish to build the max package on your computer, you can do so by cloning this repository. After that, you should intialize the needed dependencies included as git submodules by using the following command:
git submodule update --init --recursive
Assuming you have the necessary dependencies for the phonos library (mainly Eigen, see the getting sphonosd for more info), you should be able to build the externals by running
mkdir build
cd build
cmake ../src/phonos-tilde -DCMAKE_BUILD_TYPE=Release
make
cd ..
from the root. After doing so, the built externals must ahve appeared in package/externals such that you can copy the package directory to your max package directory and use the package. If you build for windows, you should remove the phonos.trackpad directory before building as it only works on macos.
The package was initially developed as part of the AVATARS ANR research project. Research was conducted by Thomas Risse, Thomas Hélie, Fabrice Silva and Stefan Bilbao. Code development was made by Thomas Risse and Aloïs Cerbu. Presets for the voice object were kindly given by Tom Mudd.



