Hair revisited

[sibarrick 18]

From the Rythm and Hues notes:

The importance of hair-to-hair shadowing to the realism of rendered hair cannot be overstated. Using traditional depth maps for this purpose would fail to capture the gradual shad-owing arising from multiple layers of partially transparent hair. Deep shadow maps are a better option, but they require considerable computation and storage. Our solution is to pre-compute and store self-shadowing information in each hair primitive

[Marc 274]

aye, I read that.

Definitely a job for point clouds. :thumbsup:

BTW, I remember the bunny demo... but I don't remember seeing the fur. Are there any images on the net? SESI?

[Marc 274]

A furry mandrilball!

[Jason 134]

Thats pretty cool!

[Marc 274]

Here's a small update. Perhaps not as cool as the Mandriball™, but cool nontheless.

edit: I'm cheating shadows at the moment by darkening the roots. There is a zdepth shadow as well, but I don't think it contributes too much to it.

Turntable coming soon.

[thekenny 3]

very cool.

would be interesting to see the skin shading without the hair...

could you post the same frames without fur?

the darkening at the roots is a good trick to use.

very good stuff.

-k

[Marc 274]

There is actually no skin shading to speak of at the moment.

I have a skin surface underneath that's basically a very dark clay shader. Once I have the texture painted then I'll probably use that.

The good thing about a horse is that the hair is pretty short and very close to the skin, so you hardly ever see the skin unless you go up close and move the hairs about. Cleverly, I won't be doing that .

Although I'm going to have to come up with something for the eyes and mouth.

Ciao

M

[sibarrick 18]

Hmmmm very soft and velvety... nice. There are a few artifacts but I guess that is the jpeg compression.

[Marc 274]

Here's a turntable for you to take a look at.

There is still a slight amount of noise in the fur, but I'm sure it will go away if I turn up the shading quality a little more.

This turntable has shown some issues in the geometry, so I'll be fixing those today.

Enjoy :

Hturntable.avi

M

[Marc 274]

Just another little sum'in sum'in for the weekend.

Gotta blend the mane in with the neck a little better.

[SvenP 2]

The hair doesnt look bad but still is not really looking like the hair of the horse ( just out of the head i would say the specular is wrong ) . My wife has 2 horses so i will take you some close up photos for reference this weekend .

I hope they will help you to make that horse look really " kick ass ".

Sven

[Marc 274]

Hey Sven

Thanks for the comments. I was thinking yesterday that he's a little fluffy too.

I did kinda decide that I don't want him looking too much like a real horse. He does have a human body, so its definitely not a real thing I'm trying to replicate here

Having said that though, some pictures would ber great for reference.

Thanks

M

[SvenP 2]

You have a new PM

Sven

[MatrixNAN 0]

Hey Marc,

I was going to say Point Clouds with Subsurface Scattering which is what all the major studios seem to be using when I was at the 2004 siggraph Hair/Fur Course. They basically showed what really made their hair look realistic was the subsurface scattering. Man what a difference between that and normal shading procedures I saw sold after that on SSS.

Cheers,

Nate Nesler

P.S. By the way your fur is looking really good.

[bfx 0]

Hi there,

I work on fur sometimes. I used the displacement from the "line as tube" you talked previously. Thats really great.

I have not read yet all the references you talk about (stuart little,...) but i'll do it.

I have looked at the R&H cat fur explanations, using DeepShadows map. Is it a way to generate that kind of map with mantra ? I guess this a renderman feature ?

I play around the "fake fur" rendering from B Goldman (1997)

http://www.cs.washington.edu/homes/dgoldman/fakefur/

I'de like to know what you think about it ?

The opacity method is what i hack in my own shader. I very quickly see how it improved the fur but i don't understand how to have an optimal setup for the probability setting "number of hair in a sample region" & "local density of hair". Any clue ?

the result i've for now:

http://b.fx.free.fr/image/cat.sit.jpg

thanks

[Jason 134]

Did anyone point out this?

http://www.cs.cornell.edu/~srm/publication...r-abstract.html

It talks about the tinted secondary specular that occurs in (mostly blonde) hair and the transmited light. The paper is a little, er, dense, but I wonder if anyone has tried out this paper yet?

[edward 350]

FYI: Link doesn't work.

[Jason 134]

Thanks Ed - try this link:

http://graphics.stanford.edu/papers/hair/

[edward 350]

Yeah, saw this at Siggraph when it came out but don't know if anyone has done it yet (outside of perhaps Joe Alter). Did anyone notice the hair shading in the Incredibles? It was beautiful.

[stevenong 3]

Yep, it is beautiful, especially when it is wet!

Cheers!

steven