win-dictate/src

Win Dictation - Source Notes

Architecture (current)

Push-to-talk, batch mode. Press record, speak, press again. The full audio clip is passed to whisper_full once on stop. No streaming, no VAD, no ring buffer.

• Audio: SDL2 capture at 16kHz mono
• Inference: whisper_full with physical-core threads
• UI: Single-surface GDI+ immediate-mode painting (no child-window chrome)
• Progress: Self-calibrating linear estimator fused with whisper's chunk-boundary callbacks

Key files

File	Purpose
main.cpp	Window, painting, interaction, settings, clipboard
transcriber.h / transcriber.cpp	Audio capture, whisper preload/inference, callbacks
timing.h	Per-model online least-squares timing model + live progress estimator
settings.h	INI file read/write for persistent settings
text_util.h	Transcript concatenation
logging.h	Timestamped log to win-dictation.log

Building

cmake -S . -B build -G "Visual Studio 18 2026" -DSDL2_DIR="deps/SDL2-2.28.5/cmake"
cmake --build build --config Release

Output: build\bin\Release\win-dictation.exe

Target machine: 2-core / 4-thread Intel i5-7th-gen, GPU CUDA disabled.

Model placement

Drop .bin files in models/ next to the executable. The app scans for:

• ggml-tiny.en.bin
• ggml-tiny.en-q8_0.bin
• ggml-base.en-q5_1.bin
• ggml-base.en.bin

First available is used. Toggle in the model popup.

Settings

Stored in win-dictation.ini next to the executable. Sections:

• [app]: window position, hotkey, model, capture device, pinned/autopaste/autohide flags
• [timing-ggml-*.bin]: per-model timing accumulators (learned transcription speed)