tapwithus/tap-web-sdk/tap-raw-sensors

Raw sensors

For research, datasets, and custom gesture models. For pointer movement or orientation, prefer tap-imu-motion (less data, already processed).

Packet format (both firmware classes)

The callback gets a list of packets: cb(identifier, packets).

Units: raw LSB counts, or mdps (gyro) and mg (accelerometer) when scaling is on. Rate is about 200 Hz per sensor, delivered in batches.

TapSDKWeb2

import { DeviceFeatures, ImuAcclSensitivity, ImuGyroSensitivity } from '@tapwithus/tapsdk';
sdk.registerRawImuDataEvents(onRaw); // registerRawDataEvents also works
await sdk.setFeature(DeviceFeatures.RAW_IMU_DATA, true);
await sdk.setImuSensitivity(
    ImuAcclSensitivity.G2,       // G2 / G4 / G8 / G16
    ImuGyroSensitivity.DPS125,   // DPS125 ... DPS2000
    { scaled: true },            // payload in mg / mdps
);
// later
await sdk.setFeature(DeviceFeatures.RAW_IMU_DATA, false);

connect() already started notifications. There is no start(). await sdk.getImuSensitivity() returns [ImuGyroSensitivity, ImuAcclSensitivity].

TapSDKWeb

import { InputModeController, InputModeRaw } from '@tapwithus/tapsdk';
import { FingerAcclSensitivity, ImuAcclSensitivity, ImuGyroSensitivity } from '@tapwithus/tapsdk';
sdk.registerRawDataEvents(onRaw);
await sdk.setInputMode(new InputModeRaw({
    scaled: true,
    fingerAcclSens: FingerAcclSensitivity.G2,  // default G2
    imuGyroSens: ImuGyroSensitivity.DPS125,    // default DPS125
    imuAcclSens: ImuAcclSensitivity.G2,        // default G2
}));
// leave raw mode
await sdk.setInputMode(new InputModeController());

• To change sensitivity, leave raw mode first, then enter it again with new values. The SDK ignores a raw-to-raw change.
• Raw streaming needs Developer mode turned on in Tap Manager.
• Tap Strap streams finger accelerometers. Tap Strap 2 and TapXR also stream the thumb IMU.

Log to CSV

Open scripts/log-csv.html from this repository with a local static server after npm run build. It works on TapSDKWeb and TapSDKWeb2. It records for the chosen number of seconds, then downloads a CSV file.

Columns: host_time, device_ts_ms, type (imu or accl), then the payload values.

In the user's app, push packets onto an array in the callback and build the CSV outside the callback. Download it with a Blob.

Plot

Parse the CSV in the page and draw it on a canvas, or open the file in a spreadsheet. imu accelerometer axes are payload columns 3, 4, and 5 (after gyro x/y/z).

Tips

• Callbacks get batches at a high rate. Append to a list. Write the file or draw outside the callback.
• Use ts for timing between samples, not Date.now().
• Lower sensitivity (G2, DPS125) gives more resolution. Raise it for fast, strong motion so values do not clip.

Docs

• Demo: https://tapwithus.github.io/tap-web-sdk/
• Packet type: RawDataMessage in src/types.ts