The Veremin control panel is powered by dat.GUI.
The control panel allows you to configure various settings and tweak your experience.
Below options are available for configuration in the control panel:
Option |
Default |
Description |
algorithm |
multi-pose |
Version of algorithm (i.e., single-pose, multi-pose) to perform the pose estimation. |
outputDevice |
first detected MIDI output device (or browser if none) |
List of available output device (i.e., browser plus detect MIDI output devices) |
chordIntervals |
minor0 |
Chords/notes to select from when computing the note values as defined in chord-interval.js. |
noteDuration |
300 |
Duration (in milliseconds) for how long a note is ON. |
notesRangeScale |
1 |
Scale for the range of notes in respect to the vertical zone height (i.e., 1 equals entire zone height, 0.5 equals half the zone height, etc.). |
notesRangeOffset |
0 |
Offset for the range of notes in response the the vertical zone bottom edge (i.e., 0 bottom of zone equals start of notes range, 0.5 bottom of zone equal half notes range) |
Option |
Default |
Description |
mobileNetArchitecture |
0.75 |
Corresponds to a MobileNet architecture and checkpoint. The larger the value, the larger the size of the layers, and more accurate the model at the cost of speed. |
outputStride |
16 |
Desired stride for the outputs when feeding the image through the model. The higher the number, the faster the performance but slower the accuracy. |
imageScaleFactor |
0.5 |
What to scale the image by before feeding it through the network. The lower the number, the faster the performance but slower the accuracy. |
Option |
Default |
Description |
minPoseConfidence |
0.1 |
Minimum confidence score before a pose is registered. |
minPartConfidence |
0.5 |
Minimum confidence score before a body part is registered. |
Option |
Default |
Description |
maxPoseDetections |
5 |
Maximum number of poses to detect. |
minPoseConfidence |
0.15 |
Minimum confidence score before a pose is registered. |
minPartConfidence |
0.1 |
Minimum confidence score before a body part is registered. |
nmsRadius |
30.0 |
Non-maximum suppression part distance. Two parts suppress each other if they are less than nmsRadius pixels away. |
Option |
Default |
Description |
showVideo |
true |
Whether or not to show the web camera video stream. |
showSkeleton |
true |
Whether or not to show the skeletal information detected by PoseNet. |
showPoints |
true |
Whether or not to show the joint information detected by PoseNet. |
showZones |
true |
Whether or not to show the left and right activation zones. |
showWaveform |
true |
Whether or not to show the sound waveform in the background. |
Option |
Default |
Description |
on |
false |
Whether or not to publish position data to configured MQTT broker. |
secureWebsocket |
true |
Whether or not to try to connect to MQTT broker using a secure WebSocket connection. |
brokerUrl |
test.mosquitto.org |
The URL to a MQTT broker that supports WebSockets. Some publically-accessible MQTT brokers can be found here |
brokerPort |
8081 |
The WebSockets port for the MQTT broker. |
eventTopic |
veremin/{event} |
The topic to use to publish MQTT messages. If the topic consists of the string {event} , it will be replaced with the appropriate message event type (i.e., nose , angle , etc.). |
clientId |
blank |
Unique ID to use to identify the client when sending messages. If left blank a unique ID is generated and used. |
username |
blank |
The username to use when connecting the MQTT broker (if the broker requires credentials). |
password |
blank |
The password to use when connecting the MQTT broker (if the broker requires credentials). |
cameraFOV |
120 |
Estimated field of view of the web camera being used. |
distanceMult |
1 |
What scale to use when trying to determine distances. |
log |
false |
Whether or not to also subscribe to the eventTopic and log on incoming messages. |