This is another excellent tutorial from BlendedPlanet. I'm sure you all remember the previous tutorials and resources we've posted from BlendedPlanet. This new one covers one of the most difficult things to key, hair.
The thing that I most liked about this tutorial is how enjoyable it is to read. Take these lines for instance:
Good news, friends - I do believe I have stumbled upon the ultimate greenscreen technique, and if you're willing to go with me on a slightly wild ride into the mathematical basis for a perfect matte, I think you will agree. And, believe it or not, we will NOT be starting out with a green minus red color difference matte!
Well, ok, that's not completely true, but humor me for a bit, and let's assume that you HAVE pulled a reasonably good color difference key, but you went ahead and blurred the matte and eroded the edges a couple pixels. So what you have is a great key of your subject, but basically, the edges suck.
From that auspiscious starting we point we ask ourselves - how to restore EXQUISITE edge detail, losing nothing in the process. Well, let's start with a little theory, some assumptions that can guide us on the way. This may not all make sense right now but try to stay with me.
Its very obvious the author is passionate about what he is sharing and it really reaches out to the reader and I know that it made me want to keep reading. The technique really results in some great keying, their appears to be little to no loss in quality and the soft edges of the hair.
Its recommended to download the .blend file as the article doesn't carry itself with a large amount of screenshots, rather the minimum to keep things understandable. Well I'm sure you'd like to check out the tutorial, it can be found here.
21 Comments
impressive stuff- this is about as good a key as I've seen.
Is it possible to group nodes into a single one? With that setup there's not much room for anything else...
markus: Node->Make Group (Ctrl G)
By the way, I wonder how easily this technique can be used in generalised situations. It looks like it's performing very well, I wonder if it could be turned into a node of itself?
To Matt:
After trying out with different green screen footage I can say it doesn't always work.
You need to adjust it and add more nodes considering the situation you're facing. I don't think it could be turned into a node by itself.
You could however create a complex node that would allow you to tweak most of the parameters and get the desired effect. But that wouldn't really map directly to this node structure.
-- Rui --
that's amazing! thank you!
Apologies in advance for cruddy web site performance today!
I noticed this morning my mail was coming in slow, and it occurred to me that BlenderNation must have posted my tutorial, because my server is choking.
It should be better in a day or two when the big rush of views slows down. Meantime if anyone wants to mirror the tute, feel free.
Dave
There is example green screen photos and other (?) matte files :(
Hehe. Sorry Dave. We should warn you when the BN-effect is in full force.
Wouldn't it make more sense to use the "Seperate YUV" node and work in the YUV colorspace? That way you have luminance as a seperate channel, and the "losing contrast" thing isn't an issue.
I made a complex node bundle that does it this way, and also includes a 4 pixel blur on the chroma channels to make keying 4:1:1 DV footage easier. Maybe I should share it.
To knellotron:
Please do share it with us, I'm sure everyone likes to see each other approaches.
Absolutely - this is just one approach, and many times keying is quicker in YUV space. The point here was actually the mindset of understanding what's going on with the pixels so you can apply a little mathematical common sense and get great results without throwing information away unneccesarily, no matter what the keying challenge is.
The site is sooo slow :S
And now, it doesn't work at all :(
Seems to be cool, though :D
Thanks - a brilliant article! I have finally manged to get a decent composited output, and understand the theory in the process.
I've been looking for something like this for YEARS. Hey, do you think he would let me make a video tutorial out of it?
Dave, that is sweet! Great tutorial. Keep up the good work.
In the tutorial he mentions to sample the values of the viewer node in the image editor to get the float point. How in the world do you do that? Wondering if its a keystroke that displays that info. Any help would be awesome. Thanks
This works amazingly well on my crappy DV footage. Thanks Dave.
@Brad: Just hold down the mouse button when the cursor is over the viewer window.
Can you get the same result using the photoshop cs3? I dont know that much about editing pictures and all that. Thanks.
awesome tutorial. I love the example video at the bottom.
Link updated.