motiondescriptors / modosc Goto Github PK

A set of Max abstractions designed for computing motion descriptors from raw motion capture data in real time.

License: GNU General Public License v3.0

Max 97.16% HTML 0.15% JavaScript 2.68%

motion-capture motion-tracking motion interaction-design descriptors real-time movement

modosc's Introduction

modosc

modosc is a set of Max abstractions designed for computing motion descriptors from raw motion capture data in real time. The library contains methods for extracting descriptors useful for expressive movement analysis and sonic interaction design.

VIDEO TUTORIALS: https://www.youtube.com/playlist?list=PLMrDazzs9wCQET95Mel3v_Ujmq0uP7XCT

Installation

Download the .zip file, open it, then move the 'modosc' folder in your Max 'Packages' folder.

Dependencies

Using modosc requires the o.dot externals for Max. The official o.dot releases can be found here: http://cnmat.berkeley.edu/downloads. However if you are running Windows in 64-bit mode you will need a more recent beta release from the o.dot github page here: https://github.com/CNMAT/CNMAT-odot/releases

mo.myo requires the myo for max external: https://github.com/JulesFrancoise/myo-for-max

Documentation and further reading

Details can be found in the wiki: https://github.com/motiondescriptors/modosc/wiki

For more information see the following two papers on the initial release of Modosc:

F. Visi and L. Dahl, “Real-Time Motion Capture Analysis and Music Interaction with the Modosc Descriptor Library,” in NIME’18 – International Conference on New Interfaces for Musical Expression, 2018. (This is included in the repository as modosc_NIME_2018.pdf.)
L. Dahl and F. Visi, "Modosc: A Library of Real-Time Movement Descriptors for Marker-Based Motion Capture", in MOCO'18, Proceedings of the 5th International Conference on Movement and Computing, 2018. (This can be found here: https://dl.acm.org/citation.cfm?id=3212842)

GIMLeT is a set of Max patches for interactive machine learning that uses modosc for motion features extraction: https://github.com/federicoVisi/GIMLeT

modosc's People

Contributors

Stargazers

Watchers

Forkers

maccallum yrammos

modosc's Issues

mo.vel/acc/jerk arguments reordering

mo.velocity et al. should have the point as the first argument and the filter coefficient as an optional second argument with a default value if no second argument is entered. This will improve usability and consistency with how other objects work, but is a breaking change.

Coding style review

switch to camelCase for all OSC addresses.
Discuss other style changes in order to be consistent throughout.

getting started

hi Federico,
I wanted to get started with mo.myo and tried to make the simplest patch to verify it's working. I receive the OSC packets but they are all zeros. Any ideas?
AT

myomodoscmaxpat.maxpat.zip

New input: posenet

New mo.posenet abstraction to bind posenet data to the modosc namespace.

Multiplier for mo.contractionIndex

mo.contractionIndex often returns a very small value. Currently, I added a *100 multiplier, which is quite arbitrary and perhaps not the best solution. Perhaps we could add an additional argument which acts as a scaling factor?

C3D player

c3d file player streaming c3d to modosc OSC namespace

qualisys/miqus cost, alternatives?

Has modosc been tested w/different camera environments? What was the pricing for your setups? I'm hoping for some practical information for replicating your y/t setup. FWIW, this strikes me as the most significant contribution to kinect skeletal data, especially for max users. Isadora software will soon support other devices like orbbec and intel realsense. Have these been tested, and do you foresee this supporting for other systems?

angular velocity in mo.velocity

mo.velocity could also be used too calculate angular velocity by having an additional @Attribute that when set to 1 will make the abstraction compute the velocity using /rot_rpy instead of /pos. This way the abstraction can also be used with IMUs. I have made a temporary solution to achieve the same result with mo.imu.velocity.

Add buffer-based descriptors

Buffer-based descriptors allow the storage (inside the o.dot code) of the last N frames of some value. This will allow computing statistics features, e.g. mean, median, max, min, etc.

Implementation will use a circular buffer to minimize moving memory around.

Possible downside: the contents of the buffer will appear in any display of the o.dot frame, adding visual clutter.