I’m running Shotcut 25.10.31 on a Mac running Sequoia 15.6.1. I’m following the instructions given here: Subtitles > Speech To Text but either no text is generated, or it’s something wildly inaccurate. I’ve tried several different language models but nothing has worked. I’ve also tried reinstalling Shotcut.
Here’s the end of the log from my latest attempt:
main: processing ‘/private/var/folders/w2/jtg42mwx3c7bs80dm1vnvdrr0000gn/T/shotcut-hJohme.wav’ (255360 samples, 16.0 sec), 15 threads, 1 processors, 5 beams + best of 5, lang = en, task = transcribe, timestamps = 1 …
[00:00:00.000 → 00:00:15.950] this is a. this is a. this is a. to. to.
[00:00:15.950 → 00:00:30.000] a.
whisper_print_progress_callback: progress = 188%
output_srt: saving output to ‘/private/var/folders/w2/jtg42mwx3c7bs80dm1vnvdrr0000gn/T/shotcut-bQoVFY.srt’
whisper_print_timings: load time = 412.48 ms
whisper_print_timings: fallbacks = 1 p / 1 h
whisper_print_timings: mel time = 8.28 ms
whisper_print_timings: sample time = 97.07 ms / 199 runs ( 0.49 ms per run)
whisper_print_timings: encode time = 118950.10 ms / 1 runs ( 118950.10 ms per run)
whisper_print_timings: decode time = 8073.94 ms / 52 runs ( 155.27 ms per run)
whisper_print_timings: batchd time = 14528.77 ms / 139 runs ( 104.52 ms per run)
whisper_print_timings: prompt time = 0.00 ms / 1 runs ( 0.00 ms per run)
whisper_print_timings: total time = 142146.61 ms
ggml_metal_free: deallocating
Completed successfully in 00:02:22
