View Full Version : Blue Screen Versus Green Screen


Peter Moretti
June 2nd, 2008, 09:00 PM
This ? will really show my ignorance, but when does someone use a blue screen as opposed to a green screen (or vice versa)?

Terry Lee
June 2nd, 2008, 09:11 PM
I'd say it would generally be an opinion. The reason for the neon color of the green or blue is so that it will contrast with any other color in the shot. So that everything is easier to key in. In other words, bright neon green or blue is unlikely to be matched by other colors in the shot.

Brian Keith Moody
June 2nd, 2008, 09:14 PM
Back in the day, GREEN was used with VIDEO and BLUE was used for FILM. Now a days you can use either. However, I've noticed movies seem to use GREEN alot more. I guess its a matter of choice.

Jim Andrada
June 2nd, 2008, 10:40 PM
I'd heard, correctly or not, that part of the reason for film using blue and video using green was because of films higher sensitivity to blue and videos's sensitivity to gree.

Could be true, could be nonsense!

Also, the choice may depend on trying to avoid colors present in the subject. I think in some cases red or even yellow are used satisfactorily

Kent Frost
June 2nd, 2008, 11:42 PM
If you can evenly light either color, then at that point it's really going to depend on what the intended background will be. For scenes where the background is well lit or outdoors during the day, use green. For scenes where it will be darker or at night, use blue. The value of each color will assist in blending in with the keyed scene better to make that hard-to-key edge on everything in front of the chroma background less noticeable. If you use a green background for a shot where you're going to key in a nighttime background, you're more likely to the see the bright green edge as opposed to the darker-value blue which would match the darker scene better.

Also, the choice may depend on trying to avoid colors present in the subject. I think in some cases red or even yellow are used satisfactorily

And yes, this is true. Not always will you put a person in front of a chroma key background. If you have a prop you'll be keying, the prop's color will help determine which color you'd use to key with. Refer to the color wheel for the best determination of what the color should be by choosing the opposite of your object/subject. If your prop is orange, use its opposite; blue. This is why green and blue are so efficient for weathermen. The shade of green and blue are very un-fleshtone like, therefore resulting in a more successful key.

Bill Davis
June 3rd, 2008, 01:32 AM
There's another reason to prefer green for DV keying work.

In plain vanilla 25Mbps DV both the blue and red signals are actually calculated by manipulating data contained in the green channel. So there's actually more data in the green channel. Since DV is already hard to key by virtue of it's low data rate and compression, it's somewhat better to pull a key against the channel with the most data.

Most likely Adam Wilt's site has the technical underpinnings of this if you want to look more closely at the how and why.

Harm Millaard
June 3rd, 2008, 02:56 AM
There's another reason to prefer green for DV keying work.

In plain vanilla 25Mbps DV both the blue and red signals are actually calculated by manipulating data contained in the green channel. So there's actually more data in the green channel. Since DV is already hard to key by virtue of it's low data rate and compression, it's somewhat better to pull a key against the channel with the most data.

Most likely Adam Wilt's site has the technical underpinnings of this if you want to look more closely at the how and why.

That makes sense if you talk about single chip cameras that use a Bayer filter. But does this also apply to 3 chip cameras?

Mark Hartopp
June 3rd, 2008, 03:44 AM
I've found the green required less lighting in studios and is perfect if used outside.

I have used limeish green felt fabric works fine. got a roll of it from a fabric shop. Gaffer taped it to some roller doors worked fine, and the client didn't know - it was also cheaper than hire cost of the propper stuff.

As pointed out by Bill & Kent there are other technical issues invovled, so always test first if your new to green/blue screen work, specialy if your going to cut corners.

if anyone is after a cheap option for home/semi pro use pop down to your local fabric shop, see what they have. A 2m sample is ideal for testing.

Shaun Roemich
June 3rd, 2008, 09:14 AM
That makes sense if you talk about single chip cameras that use a Bayer filter. But does this also apply to 3 chip cameras?

Yes. The Y in YUV (the colour methodology used in DV derived Component video) is taken from the green channel, regardless of HOW the colour separation is derived.

Jonathan Levin
June 3rd, 2008, 09:42 AM
My question would be, why not just use a white seamless? Or anything that is well lit? I know in still photography, selecting a background with a selection tool (magic wand) is pretty easy if it is even, and you have the tolerance of the tool set low so it does not select similar values.

Maybe I've answered my own question. The chroma green seems reasonable since no one in their right mind would wear clothing that resembles the chroma green I've seen, so no chance of selecting a similar value. :-}

Jonathan

David H. Castillo
June 3rd, 2008, 03:17 PM
To shoot either green or blue screen is determined sometimes by your subject matter. If you are shooting a person with blonde hair it is best to use a blue screen to get a cleaner key. The set design has a very important factor, if the set for dramatic purpose needs to be cooler in colour than green would be the ideal backdrop. I have used both green and blue and well it really comes down to lighting the screen properly. Todays video cameras can use both with ease.


cheers

Castillo
www.bluebarnpictures.com

Chris Hurd
June 3rd, 2008, 03:26 PM
We've got lots of threads covering this subject... do y'all want a dedicated GS/BS discussion forum? Hopefully more GS than BS. We can do it if you want it.

Pete Bauer
June 3rd, 2008, 04:13 PM
Perhaps these threads fit well in our compositing and effects forum?
http://www.dvinfo.net/conf/forumdisplay.php?f=130

Frank Simpson
June 3rd, 2008, 04:44 PM
Jonathan-

White seamless would be a terrible choice. While there are certainly similarites, a Chroma Key function is very different from Photoshop's magic wand tool. The Photoshop feature selects contiguous pixels whereas Chroma Key selects any pixel based on its color value.

Using white would make peoples' eyes look pretty awful with the whites keyed out. Also any kinds of specular hightlights would key out as well.

John Miller
June 3rd, 2008, 05:26 PM
Yes. The Y in YUV (the colour methodology used in DV derived Component video) is taken from the green channel, regardless of HOW the colour separation is derived.

Y contains contributions from the red, green and blue channels:

Y = 0.299R + 0.587G + 0.114B

Most camcorders sample the green at a higher resolution than the red and blue, with this kind of pixel layout (aka Bayer):

RGBGRGBG
GBGRGBGR
RGBGRGBG

This extra green information is used to artificially increase the luma resolution which makes sense given that green has the largest coefficient in the above equation.

Regarding blue vs. green screen, most analog television broadcasts/recordings (i.e., not film) used blue not green in the color separation overlay process. The rationale is that blue rarely occurs in most human skins tones making it less likely to create issues. This worked well especially when the video signal was processed in the RGB color space.

The advent of digital compression and lower sampling of the chroma components, green is prefered because it provides five times the contribution to Y and some compression schemes deliberately apply more lossy compression to the Cb channel than Cr. Cb is simply blue minus luma (scaled).

Shaun Roemich
June 3rd, 2008, 05:40 PM
What John said.

Next time I won't post anything BEFORE I get some morning coffee in me. I MEANT to say that the Y channel is derived MORE from the green channel and that this was specific to YUV colour space (as compared to RGB colour space) and was indifferent to whether the colour separation was via prism block (ie. 3 CCD/ CMOS) or via filter (ie. 1 CCD/CMOS).

My brain hurts now.