Fix diarization: use soundfile instead of torchaudio for audio loading

torchaudio 2.10 unconditionally delegates load() to torchcodec, ignoring
the backend parameter. Since torchcodec is excluded from PyInstaller,
this broke our pyannote Audio monkey-patch.

Fix: replace torchaudio.load() with soundfile.read() + torch.from_numpy().
soundfile handles WAV natively (audio is pre-converted to WAV), has no
torchcodec dependency, and is already a transitive dependency.

Also added soundfile to PyInstaller hiddenimports.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

python/voice_to_notes.spec

@@ -21,6 +21,7 @@ a = Analysis(
     hiddenimports=[
         "torch",
         "torchaudio",
+        "soundfile",
         "huggingface_hub",
         "pysubs2",
         "openai",