Aligning an existing transcript

[samuelbradshaw]

Does this library provide a way to align an existing transcript to an audio recording?

[LauraGPT]

Yes. For an existing Chinese transcript, use FunASR's fa-zh forced-alignment model:

from funasr import AutoModel

model = AutoModel(model="fa-zh")
result = model.generate(
    input=("audio.wav", "transcript.txt"),
    data_type=("sound", "text"),
)
print(result)

It returns token/character timestamps in milliseconds. See the timestamp/forced-alignment documentation.

Scope note: fa-zh is intended for Chinese (including Chinese/English text), not generic multilingual forced alignment. Fun-ASR-Nano's timestamp stage aligns its own decoded result; it does not align an arbitrary supplied transcript.