moblog uk

view from my pocket

by mat

user profile | dashboard | imagewall | contact user

« older newer »

Buy my stuff! - modernist wooden jewellery, award-winning computer accessories and more.

"Do you realise that every blade of grass is a different shade of green?" - from My Voice Will Go With You

Creative Commons Attribution-NonCommercial-ShareAlike License

Search this moblog

Recent visitors

subtitling challenge

(viewed 1840 times)
Bookmark and Share
Jay-Z, live. They're trying, but they're not doing very well...
28th Jun 2008, 23:42   | tags:

nige says:

haha. that was funny.

winehouse' performance was another story. my favourite part: when she elbowed that guy from the front row in the face.

29th Jun 2008, 10:23

rhys says:

It's funny watching subtitles for live performances/programmes, they usually make a few mistakes, but someone at BBC had to pick the short straw for Jay-Z.

29th Jun 2008, 10:57

goode says:

Jay-Zed as he's known in the UK

29th Jun 2008, 12:54

spongevid says:

surely on the continent he'd me called "yay2 instead of jay because of the way the pronounce the J's....

29th Jun 2008, 12:59

mat says:


Like Hay-sus.

29th Jun 2008, 13:01

JokerXL says:

"yay-zet" indeed Spongevid

29th Jun 2008, 13:18

afternoon says:

I know a few people in the Beeb's subtitling dept. The thought of them tackling Jay-Z is quite funny.

They don't type it, they re-speak the words clearly into trained voice recognition software.

30th Jun 2008, 16:53

mat says:

Really? It seemed a bit slow for voice recognition (well, when there were slower/clearer bits that could have been subbed). More like non-hacker/non-secretary's typing speed.

Dragon, on a fast machine with lots of ram, is scary good - even untrained. You can talk at slightly-fast-normal speed into it, no problem.

Although now I have a fantastic image of some genteel BBC types sitting around with commentator's microphones, repeating "I have got 99 problems but a bitch is not one" in lovely plummy received pronunciation..

30th Jun 2008, 17:09

afternoon says:

@mat Now you have the exact same image I had :-).

The ones I know are relatively restrained white girls. It really is quite humourous.

1st Jul 2008, 12:40

Richcolour says:

I saw a cow called Day-Z.

1st Jul 2008, 13:18

kel says:

@mat and noon: I don't think they would have used the live voice recognition for this - probably would have attempted to get a set list as quick as possible and used the lyrics they had.

I could be wrong though... were the words appearing one by one as opposed to set phrases? And were there lots of misspelled words and homophones? That's usually a clue as to whether they're doing it live or not. It happens loads on the news.

Once had an first stage interview for a subtitling position where they gave me numerous news reports to try and 'respeak'. When Rory Cellan-Jones turned up on screen I wasn't happy. Didn't get the job!

4th Jul 2008, 12:03

mat says:

one by one, and with some serious lag - like I said, kinda typing speed lag, not VR lag. Plenty mispelled words.

LOTS of very long gaps. They tended to do better at the bits they'd have had lyrics for, they didn't even try when he was freestyling (which was quite a lot)

4th Jul 2008, 12:08

kel says:

Sounds like it was then! If I'd had been on shift that day I would have wandered up to the third floor and eavesdropped for a bit of a giggle.

4th Jul 2008, 12:11