# RECCoon

A minimal Android app that records uncompressed WAV audio and transcribes it
on device.

- **Format:** 44.1 kHz, 16-bit, mono, PCM (no compression)
- **Controls:** one big round button — tap to start recording (timer starts), tap again to stop and save
- **Saving:** the WAV is written to the public **Downloads** folder as
  `reccoon-YYYY-MM-DD-HHMMSS-XXXXX.wav`, where `XXXXX` is a persistent
  5-digit progressive counter
- **Library:** a list of saved recordings, with a player and a scrubber to jump
  around long recordings
- **Transcription:** on-device, offline speech recognition with
  [sherpa-onnx](https://github.com/k2-fsa/sherpa-onnx) using multilingual
  Whisper — works with **English and Italian**
- **Name:** RECCoon · **by** wuhei · **version** 1.1.0-alpha
- **Icon:** raccoon with a microphone

## Requirements

- Android 8.0 (API 26) or newer, arm64-v8a or x86_64
- `RECORD_AUDIO` permission
- On Android 9 and older, `WRITE_EXTERNAL_STORAGE` is also requested
- Internet access the first time a transcription model is downloaded

## Build

```sh
./gradlew :app:assembleDebug      # debug APK
./gradlew :app:assembleRelease    # release APK (signed with the debug key)
```

Outputs:

- `app/build/outputs/apk/debug/app-debug.apk`
- `app/build/outputs/apk/release/app-release.apk`

A prebuilt copy is available at `dist/RECCoon-1.1.0-alpha.apk`.

## Recording

- `WavRecorder` uses `AudioRecord` (MIC source) to capture 16-bit PCM mono at
  44100 Hz into a temporary file, then rewrites a correct 44-byte RIFF/WAVE
  header when recording stops.
- `MainActivity` handles the runtime permissions, the timer, and copying the
  finished file into Downloads. On Android 10+ this uses `MediaStore`; on
  older versions it writes directly to the public Downloads directory and
  notifies the media scanner.

## Library and player

- `RecordingsActivity` lists the `reccoon-*.wav` files in Downloads.
- `PlayerActivity` plays a file with the platform `MediaPlayer`, a scrubber
  (`SeekBar`), ±10 s jumps and a playback speed control (0.5×–2×). Tapping a
  transcript line seeks the player to that moment.

## On-device transcription

- `TranscriptionEngine` streams the WAV, resamples it to 16 kHz and segments it
  with Silero VAD, then decodes each speech segment with a multilingual Whisper
  model through sherpa-onnx. Streaming keeps memory bounded for long
  recordings (offline Whisper only handles 30 s at a time).
- `ModelRepository` downloads the models on first use into app-private storage:

  | Model | Size | Notes |
  |---|---|---|
  | Whisper tiny (int8, multilingual) | ~99 MB | faster, lower accuracy |
  | Whisper base (int8, multilingual) | ~154 MB | default, better Italian |

  Both include the small Silero VAD model (~0.6 MB). Files come from the
  sherpa-onnx Whisper model repos on Hugging Face and the sherpa-onnx GitHub
  release.
- Language can be forced to English (`en`) or Italian (`it`), or left on
  **Auto detect**.
- Transcripts are stored as JSON sidecars in app-private storage
  (`filesDir/transcripts/<recording>.json`) and are shown in the player. A
  `TXT` badge marks recordings that already have a transcript.

## Native libraries and APK size

The APK bundles the sherpa-onnx/onnxruntime native libraries for `arm64-v8a`
and `x86_64`, so it is around 70 MB. Removing `x86_64` from the `abiFilters`
in `app/build.gradle.kts` roughly halves the APK size for phone-only builds.