Duration: 0:00

132 views

2

I would like to train Whisper to recognize and timestamp Minmay's voice. But the first one will need to be done manually. This is just a public sample. It's not perfect, and likely cannot be. But it is quite a way to watch Macross.
FFmpeg is used to cut out all the gaps and timestamps, and assembling this video.

Minmay's voice timestamp for this part :
Minmay timestamps

Starting time : 00:17:39.000

00:17:43.562 - 00:18:12.675
00:18:15.594 - 00:18:19.820
00:18:21.767 - 00:18:24.900
00:18:29.358 - 00:18:33.112
00:18:37.960 - 00:18:42.997
00:18:48.419 - 00:18:53.841
00:18:58.304 - 00:19:05.561
00:19:18.224 - 00:19:25.456
00:19:28.084 - 00:19:37.802
00:19:50.756 - 00:19:53.192
00:20:13.300 - 00:20:20.344
00:20:23.013 - 00:20:25.516
00:20:35.359 - 00:20:45.327
00:21:01.300 - 00:21:11.729
00:21:43.277 - 00:21:45.304
00:22:05.900 - 00:22:08.169
00:22:10.900 - 00:22:12.665
00:22:31.125 - 00:22:32.476

Ending time : 00:23:18.000

Gaps for this part :
Part 4 gaps
Start time : 00:17:39.000


1. **00:17:39.000 - 00:17:43.562
2. **00:18:12.675 - 00:18:15.594
3. **00:18:19.820 - 00:18:21.767
4. **00:18:24.900 - 00:18:29.358
5. **00:18:33.112 - 00:18:37.960
6. **00:18:42.997 - 00:18:48.419
7. **00:18:53.841 - 00:18:58.304
8. **00:19:05.561 - 00:19:18.224
9. **00:19:25.456 - 00:19:28.084
10. **00:19:37.802 - 00:19:50.756
11. **00:19:53.192 - 00:20:13.300
12. **00:20:20.344 - 00:20:23.013
13. **00:20:25.516 - 00:20:35.359
14. **00:20:45.327 - 00:21:01.300
15. **00:21:11.729 - 00:21:43.277
16. **00:21:45.304 - 00:22:05.900
17. **00:22:08.169 - 00:22:10.900
18. **00:22:12.665 - 00:22:31.125
19. **00:22:32.476 - 00:23:18.000


End time : 00:23:18.000