Bringing teletext back from the dead

What's not currently on the box
Post Reply
User avatar
Scary
625 lines
Posts: 271
Joined: Mon Dec 10, 2012 1:30 pm

Bringing teletext back from the dead

Post by Scary »

Interesting article about a new method of retrieving old teletext data from VHS:
The Teletext Salvagers: How VHS is bringing teletext back from the dead

TVT_Dave
405 lines
Posts: 14
Joined: Tue Oct 21, 2014 9:47 am

Re: Bringing teletext back from the dead

Post by TVT_Dave »

What a fascinating item.

As for decoding VHS tapes, I tried this with Sony U-matic recordings as well, but didn't get anywhere. The problem is that the machines of the time weren't that stable. They locked up frame wise, but if you put the signal into a monitor locked to external syncs, the picture would be stable top to bottom, but swinging wildly from side to side. As the teletext signal was inserted into the sync pulses at the top of the picture, putting the VTR through a fieldstore synchcroniser just stripped off the teletext information. Knowing all this came in useful when Engineering Liaison department phoned asking if decoding Umatic tapes would work as they were attending an exhibition where they couldn't get an off air signal, so thought of playing back off air recordings and decoding them from the tape.

As an aside to the above, I worked for a department of the BBC which trained production staff. From time to time we were asked to try out/solve production problems in the small colour training studio we ran. In the early 70s, just as Ceefax was getting started, we obtained a TV with a Ceefax decoder and rigged it in the studio to show the members on one of our courses, what it was all about. The head of the department was showing someone from an overseas TV company around and came in to the studio. At the time there was some problem in the overseas country and by calling up the appropriate page, he was able to read the latest news on the subject.

Dave B

Brian F
D-MAC
Posts: 516
Joined: Mon Dec 10, 2012 4:09 pm

Re: Bringing teletext back from the dead

Post by Brian F »

Another difficulty would be that VHS recorders had noise reduction filters that rolled off the signal above about 3Mhz. I had an Akai that when getting a good off tape signal level would switch the filter off to increase the luma detail. On good tapes bypassing the filter might help, though not with the timing errors.

User avatar
Simon Coward
D-MAC
Posts: 916
Joined: Sat Dec 08, 2012 7:56 pm

Re: Bringing teletext back from the dead

Post by Simon Coward »

From my own experience - and therefore just a limited number of machines - Betamax used to be better than VHS for preserving teletext. I often used to use it for dating programmes when I hadn't kept a note at the time I recorded something.

It would be interesting to know if they've tried it with anything other than VHS.
We all have to eat a peck of dirt before we die.

gmg
405 lines
Posts: 44
Joined: Wed Dec 12, 2012 9:32 pm

Re: Bringing teletext back from the dead

Post by gmg »

Interesting article on a mostly forgotten aspect of archive TV. Quick question: when did manual (i.e. remote controlled) Ceefax/ Teletext become mainstream? My memories are of not having this facility until about 1987/88 and then becoming very reliant on it, particularly for sports coverage/ results. Prior to that I'd generally had to tune into TV coverage to keep up to date but was I late in joining the party?

User avatar
Paul Hayes
625 lines
Posts: 423
Joined: Sat Dec 08, 2012 7:01 pm
Location: Norwich
Contact:

Re: Bringing teletext back from the dead

Post by Paul Hayes »

Interesting stuff, thank you for sharing!

I'd heard of this being done with S-VHS tapes before, but as has been mentioned had always been under the impression that VHS wasn't stable enough for it.

Brian F
D-MAC
Posts: 516
Joined: Mon Dec 10, 2012 4:09 pm

Re: Bringing teletext back from the dead

Post by Brian F »

Often the Page/date time title line will show good but with rubbish for the rest of the screen. I can't remember exactly why but there was a slight difference in the way it was sent/coded if I remember correctly.

ozsat
625 lines
Posts: 209
Joined: Tue Dec 11, 2012 8:55 pm
Location: Oxford

Re: Bringing teletext back from the dead

Post by ozsat »

My S-VHS recordings were often around 90-95% reliable with teletext pages. A static page would usually complete find after a short time if it started with any corruption.

User avatar
Scary
625 lines
Posts: 271
Joined: Mon Dec 10, 2012 1:30 pm

Re: Bringing teletext back from the dead

Post by Scary »

Brian F wrote:Often the Page/date time title line will show good but with rubbish for the rest of the screen. I can't remember exactly why but there was a slight difference in the way it was sent/coded if I remember correctly.
I remember seeing some teletext data on recorded programmes even in LP mode, you could at least work something out. The idea that the data was repeated as the pages were sent in rotation so can be amalgamated is an interesting one, I suppose the status line often appeared good because it was sent more often?

Brian F
D-MAC
Posts: 516
Joined: Mon Dec 10, 2012 4:09 pm

Re: Bringing teletext back from the dead

Post by Brian F »

Scary wrote:
Brian F wrote:Often the Page/date time title line will show good but with rubbish for the rest of the screen. I can't remember exactly why but there was a slight difference in the way it was sent/coded if I remember correctly.
I remember seeing some teletext data on recorded programmes even in LP mode, you could at least work something out. The idea that the data was repeated as the pages were sent in rotation so can be amalgamated is an interesting one, I suppose the status line often appeared good because it was sent more often?
I've checked on Wikipaedia and got this (which I half remembered from what Steve Roberts said on the old Dr Who RT forum)
"For single bit error recovery during transmission, the packet address (page row and magazine numbers) and header bytes (page number, subtitle flag, etc.) use hamming code 8/4 with extended packets (header extensions) using hamming 24/18, which basically doubles the bits used."
So the headers are more secure from errors, but not fully error corrected.

Post Reply