1
00:00:00,000 --> 00:00:02,720
Thanks for making the time.

2
00:00:02,800 --> 00:00:06,240
I want to start with the part you said was
messy.

3
00:00:06,320 --> 00:00:09,690
Right, so the messy part is that the file

4
00:00:09,770 --> 00:00:12,720
never arrives the way you expect it.

5
00:00:12,800 --> 00:00:14,817
Someone records on a phone,

6
00:00:14,897 --> 00:00:19,400
someone else on a laptop, and the levels
are nothing alike.

7
00:00:19,480 --> 00:00:22,800
We stopped sending the audio anywhere.

8
00:00:22,880 --> 00:00:25,890
Once the whole thing ran on the machine
that

9
00:00:25,970 --> 00:00:29,120
recorded it, half the problems were just
gone.

10
00:00:29,200 --> 00:00:32,538
What surprised me was how much of the time
went

11
00:00:32,618 --> 00:00:35,520
into the download, not the transcription.

12
00:00:35,600 --> 00:00:41,400
The model is about 200 megabytes, and
after the first run it is cached.

13
00:00:41,480 --> 00:00:45,851
On the laptop an hour of audio takes about
three minutes.

14
00:00:45,931 --> 00:00:50,458
On the desktop with a proper graphics
card, well under one.

15
00:00:50,538 --> 00:00:54,440
That was the moment it stopped feeling
like a demo.

16
00:00:54,520 --> 00:00:56,177
Let me add a pause here.

17
00:00:56,257 --> 00:01:00,960
And then continue with a second paragraph
about subtitle formats.

18
00:01:01,163 --> 00:01:05,323
SRT works on every video platform.

19
00:01:05,403 --> 00:01:08,523
VTT is what browsers use.

20
00:01:08,603 --> 00:01:10,683
Plain text is for notes.