Jump to content
Sign in to follow this  
itriix

UV Texture/ Cd and Particle color

Recommended Posts

Well hello everyone :)

I've succesfully gotten an animated model of a girl exported from maya to houdini using exportSeries.mel... *it was a little clunky though but that's not the point of this post*

I feel like i'm doing something wrong in my network and have a few questions.

I'm taking a SINGLE bgeo of my animated girl, and applying a UVTEXTURE node with camera set to PERSPECTIVE... I'm then using a POINT SOP, and using the PIC function and MAPU and MAPV to point to a single frame of "footage" in the IMG network. At this point, i've colored my model and applied Cd info to it.

I then bring in the animated bgeo sequence of the girl running model. I do an attribute copy from the "colored single bgeo girl model" to the "animated bgeo sequence"... i've attached an image that shows the settings....

After that i'm using a material sop, and a vex layered surface shader, and pointing to the "single" frame of footage that i've used over and over to texture the girl... Anyways, when I look at this girl now, and render it. It's IDENTICAL to the color of the "footage" it looks GREAT... but here is where my problem is... the Cd information gets lost after the Attribute Copy... And so, after the material sop, i've layed down ANOTHER POINT SOP and did the PIC FUNCTION AGAIN... but then at this point the color looks SUPER saturated... and not like the original "before" adding the new Cd again...

THE REASON THOUGH I need the Cd, is because i'm going to plug this into a POP NET to emit particles from the girl.

I need the Cd...

And so, i'm not sure what i'm doing wrong here. But before the Point Sop, it looks EXACTLY how i want!!!!!

Thoughts?

Thanks

uvtext.tiff

Share this post


Link to post
Share on other sites

Hey Jon,

In your AttribCopy SOP, you don't have to set the Group Type parameters because you're copying wholesale. Next, instead of setting Texture UV as the Attribute to copy, you should set the parameter to "Other Attribute" and specify the attributes you want, for example, "uv Cd".

Hope that helps!

Cheers!

steven

Share this post


Link to post
Share on other sites

Thanks Steven,

That cleans up my network a bit however, i'm still getting a somewhat weird issue with the "coloring" when i do that. Everything turns very "oversaturated" looking...

Whereas, when I had the settings of the attributeCopy set to "vertices", "vertices", "UV Texture", the render with the "material" node, before the "Cd" Point Sop, gave me the image that has the "perfectly" textured girl. However, back to the same problem again, I need the Cd attribute in order to get the particle color... The thing is as soon as I add the "Cd" the color changes to the oversaturated texture.

Any thoughts on that one?

Thanks so much for taking the time to help me

Jonathan

I've uploaded 4 pics of what the image looks like with the AttribCopy Sop set as I had before which gave me the correct color, BEFORE the Cd was added using the Point Sop *my initial issue*

And then the oversaturated pic is from changing the AttribCopy to your suggested method. This results in this oversaturated look as well...

So it seems like as soon as Cd is added, the color becomes oversaturated

pics.zip

Edited by itriix

Share this post


Link to post
Share on other sites

Check out your material.

You might actually be using the point colors to tint your object. This might effectively be coloring the object twice, once from the Cd from the points and again from the texture.

In the material you are using for the object, turn off "Use point color".

The Cd attribute will still be there on the mesh so you can use it for your particles but will now be ignored in the material. When Cd is not there on the mesh it defaults to not being used in the material either.

Share this post


Link to post
Share on other sites

Thanks for the suggestion Peter,

That definately was causing the oversaturation!... Interesting thing was when I first turned off the "Tint with Point Color" attribute, to test it, I was still getting oversaturation. So then I noticed that in the POINT SOP, where I did ADD COLOR with a PIC FUNCTION --> In order to obtatin Cd, at the bottom of the POINT SOP, there was KEEP TEXTURE, I thought maybe that was part of it, so i turned that off, and that got rid of it. So I guess with KEEP TEXTURE and ADD COLOR, the colors were being multiplied.

Thanks again for your suggestion

Jonathan

Edited by itriix

Share this post


Link to post
Share on other sites

Can my particles inherit the color from the geometry that is strictly using the "uvtexture sop" and a "material sop with a constant shader and a color map referencing the footage" piped into a popnet. Instead of using the SUPER SLOW POINT SOP and PIC function to get a Cd attribute for the particle color?

Just doing that doesn't seem to work, so maybe i'm missing something or it just can't work like that?

Having to calculate the point color frame by frame on the geometry is quite time consuming. Also it just seems like since i can render the correct color of the geometry and texture color with only UVTEXTURE SOP set to my Matchmoved Camera, and then a MATERIAL SOP with a CONSTANT shader pointing to a COLOR MAP of the FOOTAGE... Then it seems like that information should be able to be used to give my particles color without having to use the POINT SOP?

there was a post:

http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=13620&highlight=particle+color+colour+texture

and it talked about using a uvtexture node and then a vopsop that turned the uv color into cd attribute, and then plugged that into the pop net... the problem here is that the colormap node in vops, cannot use a color map that is based on TIME... so is there a workaround? i need to be able to use an image sequence.

Edited by itriix

Share this post


Link to post
Share on other sites

I have another question:

say I get the texture projected color on the 3d girl model, looking how i want it. The model, in fact, isn't EXACTLY perfect duplicate of the real actress. So it's matched up the best it can be, but for example, the 3d model has "pants on" whereas the actress has capris on, so her legs are much thinner at the bottom. Therefore, during the texture projection, there are parts of the CG model where you can see the "bg image", such as the sidewalk, or a building or something, being projected on. Well here is the thing, that i'm curious about. I'd like to try and clean that up a bit, and make it have alpha 0 in those areas. Now I know this might be tricky but i'm looking for some hacks.

Here is what stuff I have to work with. I have rotoscoped the girl in after effects. And therefore I have a MATTE to work with. I'd like to remove the "unwanted color" BEFORE sending this into my POPNET to emit the particles with the color. Because I don't want to be emitting sidewalk color or cars or whatever off of the girl :)

So, I was clicking around on the image and looking at the HSL of the colors, and it seems that most of the unwanted color has a fairly "standard" H value of around 45 - 60... And that most of the other color values don't... *this might not hold up through the whole animation however, but it's a start"... So i'm wondering, are there any tricks that I can do to isolate that H value in the color of the cg model, tell it to turn to alpha 0? Or is there a way that I can use my rotoscoped footage from after effects to somehow, EAT out the unwanted areas? BEFORE the POPNET?

Thanks,

Jonathan

SOLVED:

I just brought in my rotoscoped footage into a copnet, used it as an alpha matte for the footage, turned on ADD ALPHA in my POINT SOP and used a PIC function to get the alpha, and then in my constant shader, i turned on the opacity map... seems like whenever i just use the "point sop and pic" the color or opacity information doesn't look as clean, as when i ALSO turn on the color map and opacity map in the constant shader as well....

Well anyways, still if there is a more efficient way of doing this or it sounds like i'm doing this wrong, please let me know my other ways and alternatives...

and as of now, i'm still trying to get away from using this point sop for my particles inherited color

Edited by itriix

Share this post


Link to post
Share on other sites

Hey Jon,

I know eetu helped you over at the SESI forum so if you can post the link to the thread, that will help others who run into the same problem later.

Thanks!

Cheers!

steven

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×