Last week I was following a train of thought (as I often do on the commute to and from work) and I wondered - would it be possible to recover the colour from a B&W 16mm film if the source was in colour?

For those of you who don't know, a "kinescope" or "telerecording" was the process of recording TV from a CRT onto 16mm film. It was the only way it could be done prior to videotape. Even after, it was still done because videotapes were expensive and were wiped and re-used. A backup (in B&W) was taken on film.

I wondered - some "kines" were quite good quality, although many were dreadful! Has anybody tried to recover the colour from the chroma patterning in the image?

Well yes they have. Here is one of the results:

https://www.youtube.com/watch?v=KE-VXlOeaIA

And here is a link to how it was done:

http://www.techmind.org/colrec/

Hi Ian, that is amazing but logical when I realise how a PAL colour signal is composed.
Sort of like backward engineering the signal.

Actually in a rock clip, weird colours would just look like normal Molly Meldrum "special effects"!

Fascinating and wonderful when you see how much more "alive" a colour picture looks compared to B/W.

Fred.

The legacy B&W Kine recorders would have lacked an appropriate chroma notch filter which they should have implemented, because dot artefacts would shift gamma/greyscale linearity.

Actually it seems that they did in fact have notch filters, which were often not used because they'd degrade the sharpness of the image.

There have been some disappointments when thought-to-be-lost kines have been found, only to discover they were made with the notch inserted.

Seems it was the practice to switch in a notch on playback of the film as required.

That link I put up shows quite surprisingly good image quality, considering it is a kine and late 60s cameras.

What impresses me is the way the colour reference is obtained from the colour information. You can do that with PAL because of the phase inversion on alternate lines.

It can track the unavoidable image non-linearity and actually correct it.

Perhaps they had to recover/re-sync the 8-field PAL-colour-framing. Good luck with that.

At first I thought "of course you can't get a colour image from a B/W telerecord" but all of the chroma information is essentially there if the 4.43 notch is not in use.
A fascinating read.