Chroma-Key, green

Serena wrote on 7/20/2007, 12:31 AM
It seems that I'm not the only one NOT totally satisfied with green screen work (despite Spot's excellent guidance) and I've just come across http://www.bluesky-web.com/greenscreen-6.htmGreen Chroma Key and people[/link]; might be of interest to others. The author is promising another series on the subject with different cameras/formats with examples.

Comments

farss wrote on 7/20/2007, 1:42 AM
My general guidelines to anyone using our green screens:

1) Light the screen with daylight fluros. It must be evenly lit. If your camera has zebras these are the tool to use. You should be able to get the screen to go from no zebra to all zebras in one click of the exposure wheel.

2) Get the talent far enough away from the screen to avoid spill. Light the talent 1 stop over the screen, again with daylight sources. Add some backlight to increase separation if needed.

I guess my formula works, everyone seem to be getting acceptable results. However having said that if you look at production shots of film shoots the green screens are all over the place and yet they extract perfect keys. Or maybe they're doing a lot of manual tweaking. Or they have way better software. Or they have way more data to work with.

I might also mention that Datavideo are selling a green screen kit of the type that uses a green lens ring light and reflective screen for a reasonable price. What I don't know is how durable the screen is. The original reflective material is very expensive and the makers made much of the lack of durability of inferior copies.

Bob.
JackW wrote on 7/20/2007, 3:44 PM
Hi Serena:

Could this be http://www.bluesky-web.com/new-page6.html or http://www.bluesky-wed.com/greenscreen2.htm?

Jack
Serena wrote on 7/20/2007, 8:44 PM
Interestingly the article I was linking to by http://www.bluesky-web.com/index.htmWalter Graff[/link] doesn't load now. It was greenscreen-6. There are useful instruction resources on that bluesky site and, as you point out, others about chroma-key.

The main point in the article (greenscreen-6) was that lighting a greenscreen with tungsten isn't a good idea because that brings green closer to red (which the article demonstrated on the vectorscope) and hence closer to skin colour in the video. So Bob's recommendation (farss) to light with daylight balanced luminaires is spot on. My normal studio setup utilises tungsten-halogen lights, so now I see how things can be improved: big banks of fluorescents!
JackW wrote on 7/20/2007, 9:52 PM
Serena:

As Alice said -- "Curiouser and curiouser." I just tried your original link (at 9:50 p.m., Pacific Daylight Time; July 20) and it works. My earlier experience this afternoon was like yours: no linkage.

It's an excellent article. Glad it's back up and functioning. Thanks for the posting.

Jack
JackW wrote on 7/20/2007, 9:57 PM
Serena:

As Alice said -- "Curiouser and curiouser." I just tried your original link (at 9:50 p.m., Pacific Daylight Time; July 20) and it works. My earlier experience this afternoon was like yours: no linkage.

It's an excellent article. Glad it's back up and functioning. Thanks for the posting.

Jack
farss wrote on 7/21/2007, 1:00 AM
I just discussed this with my boss, an old timer from the days of 2" VTRs and before. To him it was just "duh", they'd always lit green / blue screens with daylight sources, couldn't remember the why of it, it was just one of those things that was always done that way.

I read somewhere that Disney used to use a different approach. They used yellow screens. I think they had a white cyc and the trick was they lit it with sodium vapour lamps. This would achieve what Wlater is saying. Keep the screen in the narrowest band of wavelengths possible, sodium vapour lamps have an extremely narrow spectral emission.

Same thing is achieved using a reflective curtain and green LEDs as the light source.

Bob.