Indigo 0.9 test 2

[rgigante]

Thanks Ono!

Any news or performance comparision in the case of rollback to the old transformation matrix form?

Thx in advance, Riccardo.

[jurasek]

Ono thank you for sse1

greetz,
jur

PS. Works fine via Wine 0.9.36.

[OnoSendai]

Thanks Ono!

Any news or performance comparision in the case of rollback to the old transformation matrix form?

Thx in advance, Riccardo.

Not yet, will look at that soon

[ortega]

crash again



its my computer i think

[ortega]

mmmm works now ... but with simple scene

[ortega]

just i dumb ... works fine!


thx

[Whaat]

Thanks Ono!

I am still having problems with textured materials on spheres. UV set cannot be bound. (Did you change the XML?) Also, did you fix the exit portal bug that causes the direction of the light to change?

Thanks again! (especially for fixing all the startup crashes!)

Whaat

[OnoSendai]

try 'default' for the sphere UV set.

As far as the light changing direction, it's not really a exit portal bug, but more a consequence of the way environmental lighting is done with bidirectional path tracing.

[Whaat]

try 'default' for the sphere UV set.

OK, that worked but it creates a problem if you create a material with multiple uv sets and apply it to a sphere primitive. (eg. doing a material preview on a sphere).

As far as the light changing direction, it's not really a exit portal bug, but more a consequence of the way environmental lighting is done with bidirectional path tracing.

So...using exit portals should give you the same environmental lighting as if not using exit portals as long as you don't change the setting for bi-directional....I will test this.

I am getting crashes with this simple scene. It seems to work if there is no texture applied to the cube. Any ideas?

Using Pentium 4 with Windows XP Pro

Whaat

[CoolColJ]

wow way faster, man it's gonna fly with quad core

some included test scenes don't work with the new gamma element

any examples on how to syntax it? and the new texture syntax?
becuase the speed icnrease makes it worth using over v0.8 stable....

[OnoSendai]

rgb spectrum looks like:

Code: Select all

				<rgb>
					<rgb>10 0 0</rgb>
					<gamma>1</gamma>
				</rgb>

textured emitter looks like

Code: Select all

<meshlight>

    <rotation>
      <matrix>
        1 0 0 0 1 0 0 0 1
      </matrix>
    </rotation> 
    <pos>0 0 0</pos>
    <scale>0.5</scale>
    <mesh_name>lightmesh</mesh_name>

    <!-- define the emission spectrum-->
    <spectrum>
      <uniform>
        <value>100000000</value>
      </uniform>
    </spectrum>
    
    <!-- texture the light -->
    <texture>
      <uv_set>albedo</uv_set>
      <path>../177875_1168716264_large.jpg</path>
      <exponent>2.2</exponent>
    </texture>
    
  </meshlight>

[rgigante]

rgigante wrote:
Thanks Ono!

Any news or performance comparision in the case of rollback to the old transformation matrix form?

Thx in advance, Riccardo.

Not yet, will look at that soon

Thanks Ono... waiting for any news.

But... in the meanwhile... any chance to see the emitting property and "exit portal" property becoming integral part of material definition removing the meshlight items?

Thanks really again.

Riccardo.

[ZanQdo]

yay it works on wine, does wine take advantage of SSE?

[Kram1032]

If it wouldn't none of the Indigoversions'd work, there

[OnoSendai]

Thanks Ono!

Any news or performance comparision in the case of rollback to the old transformation matrix form?

Thx in advance, Riccardo.

Looks like this works alright.
So I'll probably relax the orthogonality restriction next release.

The matrix will still have to be invertible tho