Third the questions ...

1 ) I think is there too many reflexion on the body, i've play with the edge blend but without results ... How to ?

2 ) Hot to add metalic effect on this shader ?

thx

If you have my old funcat set from Daz, it came with a bunch of fast rendering metalflake shaders.

Attached Link: And now something completely different - Matmatic car paint by bagginsbill

matmatic shader is not GC,

There is a whole section in this thread on car paint.

http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2722867

You can also get car materials here:

http://www.renderosity.com/mod/freestuff/index.php?user_id=305980

Quote - Third the questions ...

1 ) I think is there too many reflexion on the body, i've play with the edge blend but without results ... How to ?

2 ) Hot to add metalic effect on this shader ?

thx

I see you're looking to make a GC shader. That's good, because the reflections need GC.

However, the reflections look too bright because:

1) The data coming from the Reflect node is from other shaders, and if those shaders are GC shaders, then the data is in GC space, not linear space. It needs to be anti-gamma corrected first. So you want to connect the Reflect to a Pow node. Right now, the GC data from Reflect is then added to Diffuse and gamma corrected again.

2) You did not balance the diffuse amount against the reflect amount. In other words, the shader should add diffuse and reflection in a weighted sum. As you have it, the diffuse effect is the same no matter what. The rule is, when reflect is strong, diffuse is weak, and vice versa. The best way to do this is to combine diffuse and reflect with a blender, and have the ratio controlled by the blending value. The Blender always produces a balanced sum.

3) The math behind the Edge_Blend cannot correctly implement the Fresnel effect by itself. I used to suggest that, but I was wrong. I know better now. You can either go for 100% accurate Fresnel effect (many nodes) or you can approximate it. The approximation I used to suggest is not a very good one. However, as it stands right now, the biggest departure from realism is due to items #1 and #2 above. I wouldn't worry too much about this item until you get those squared away.

The metallic effect we should cover separately. It involves a noise node and a specular node. It's not too hard.

Trying thx !!!

That looks excellent. Well done!

i think there is to much visible reflection infront.

i guess it depends what kind of car material you want. but i think the best way to make realistic car materials is that you have very very little reflection infront and a lot on the edges.

Here's a link to a photo of a real car about the same color.. it looks like it has MORE reflection then what you have now, and IMHO your original image is possibly closer to this.

Here's also the link to the original image, which is much arger and shows more detail:

http://people.mozilla.com/~faaborg/files/20061210-newCar/newCar.jpg_large.jpg

more on the edges not infront IMO.

If he was using an HDR environment, there'd be more reflection like that car photo. It's probably not HDR, so the reflections are limited to k * 1, where k is the reflection value.

However, as I said in point #3, the Edge_Blend alone cannot come close to the reflection value versus angle that is found in the true Fresnel effect. It is only directionaly correct, in that there is less facing the camera and more facing away. That's like saying a boob is roughly the same shape as a bucket.

"Boob and bucket" - name of my next rock band.

he needs to fake HDRI or use a HDR image. infront you only see the brightest part clearly.

Quote - It's probably not HDR, so the reflections are limited to k * 1, where k is the reflection value.

It is an HDRI with your envsphere