This example shows how to use an existing source of OSC data.
In this case, we will use Sensors2OSC, an Android app. Refer to the
score documentation (Help > Documentation) to learn how to configure your own
OSC device.
Download and install https://sensors2.org/osc/
In its settings, configure it for your machine's IP address, and have it
send OSC at port 9000, then enable sending data and all the sensors you want
in its main menu.
If you don't know your machine IP, the sysinfo device will show it under:
sysinfo:/network/<index of your network card>/ipv4
Filter for "ipv4" in the device explorer search field to find it quickly.
It will likely be something that starts with 192.168... or 10.0... ; it's not 127.0.0.1.
In the device explorer, we added a few of the control it offers, use the "Learn" feature
to add more. If all is well you should see at least orientation, accelerometer and a few others
move when you move your phone.
Note that it will shut off if your screen turns off, you need to stay on the app.
Sensors2OSC sends its
rotation vector with values
between [0; 1]. We scale them
all to 360
Here we use Object Filter (jq syntax)
to extract each individual touch from Sensors2OSC
multitouch screen:
first object will return the position of the
first finger, second object of the second finger, etc.
This is how we can deconstruct OSC messages
of the shape
/foo <index> <value>
into individual <value> for each index.
Notice the mapping on "hue":
here, we leverage the syntax to
extract an OSC address element directly, .e.g
sensors2OSC:/rotationvector@[0]
If the source OSC address also has a unit / dataspace
set, it is also possible to adapt and convert automatically this way.
For instance, given an OSC RGB input foo:/color, then
the address foo:/color@[color.hsv.s] extracts the current saturation
of the color.