[QuadList] OT?-BBC recovers color from B&W kinescopes--old tapes next?

[Ted Langdell]

And now for something totally different: (with a likely followup by  
Tony Quinn)

BBC Viewers in the UK will get to see a 40-year old episode of "Dad's  
Army" in color this weekend, long after the original 1969 Quad PAL  
color videotape was wiped after being transferred to a 16mm film copy.

http://www.guardian.co.uk/technology/2008/dec/11/digital-video-restoration-dad-s-army

The process uses software to decode the "chroma dot" pattern left in  
the image by technicians who chose not to use a filter to remove the  
color information at the time the telerecording" (kinescope in the US)  
was made.

It isn't clear whether the monitor used in the BBC tape to film  
transfer was a color or monochrome monitor. Or whether that makes a  
difference. The major dependency is whether the chroma dots are in the  
filmed image, which is the result of telerecording technicians not  
using a filter to kill the chroma dots during the tape to film transfer.

The article reports:

"Says Insell "I was watching a black and white [originally colour] Jon  
Pertwee [Doctor Who] episode on UK Gold. On the end titles I could see  
some red breaking through. 'Where is this coming from?' I thought.  
'How is there colour coming out of this black and white film  
recording?'"
Ghost of colours past

The colour Insell saw, he would later go on to discover, came from a  
series of electronic artifacts burnt into the film from when the black  
and white telerecording was first made - a ghosting of the programme's  
original colour.

Manifesting as a pattern of faint dots across the picture, these  
'chroma-dots' are the key to colour recovery, a process which has  
caught the eye of the BBC (Putting the colour back into the Doctor's  
cheeks, March 6).

http://www.guardian.co.uk/technology/2008/mar/06/research.bbc

The breakthrough for the project, which has led to the rescue of the  
Dad's Army episode, came with the development of useable software  
created by award-winning computer programmer Richard Russell. A  
retired former BBC employee, Russell had previously helped develop the  
BBC microcomputer in the 1980s."

The March 6 article describes the process more fully:

"Telerecordings (as they are known) were created by shooting, on 16mm  
film, a required programme that was showing on a high clarity  
television monitor. However, there were several problems with the  
process. Principal among these was the change of the frame rate - from  
the 50 frames per second of video to the 25 frames per second of film.  
This can now be corrected by a technique called VIDFire.

However, there is a more relevant problem. Any black and white  
telerecording of a colour programme is prone to pick up interference  
from the colour encoded video signal. This manifests itself as a  
pattern of small grey dots, called chroma-dots, across the picture.  
There was a way to stop this from happening, by using a special filter  
to cut out the electronic artefacts. However, the interference was  
often deemed so minor that the technicians doing the transfers used no  
filter and so the resultant film prints often contain a burnt in  
pattern of these chromadots.

Insell suggested that it might be possible to decode the original  
colour signal of the show from these chromadots, since they contain an  
electronic remnant of the original video signal. Since then, Insell  
has set up an independent group - outside the BBC - to put together a  
technology to extract this coded pattern within the black and white  
film and decode it.

"There are various provisos," Insell says. "The quality of the film  
has got to be good enough to have captured this pattern. We're really  
talking about working from the original negatives and having an HD  
scan made to get as much information as possible from the film."  
Insell also stresses that the process would only work on programmes  
that were originally recorded in colour."

This has me wondering whether the "chroma dots" this process uses are  
the ones often referred to by US engineers as "chroma crawl," visible  
as a series of dots "crawling" up the screen if you look at an NTSC  
picture, closely.

--(Tony Quinn says "Yes" in a post elsewhere)

In any case, let's hope that the concept or software now in use can be  
adapted to NTSC images and used to recover color from monochrome Kines.

I'd opine that actually implementing such a concept will depend on how  
archives and rights holders can develop a way to make back the cost of  
doing so—"monetize" the content in today's lingo.

The technology also might presage a way to recover color from  
videotapes that were monochrome recordings of color programs...  
depending on a lot of "ifs."

Ted

Ted Langdell
Secretary

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://quadvideotapegroup.com/pipermail/quadlist_quadvideotapegroup.com/attachments/20090102/a36b4365/attachment-0003.html>