Cineversity

I posted this on the C4DCafe.com forum and am wondering if anyone has an answer here. I have a PMan file that I’m working with. I need to cross over some of these textures to the V-Ray engine. To do this, I believe I need to bake my PMan projections. When I do this, I somehow lose the alpha channel. Now, this is a bit of a pain to have to go back in and repaint my alphas. Has anyone experienced this? I can’t figure out how to resolve this.

Hi Koboldstudios,

I focus here on C4D only: You can bake Alpha as well Illumination (and other) channel at the same time. You get for each channel a texture written out.

========

Note that camera mapping is nothing else than a special way of organizing your UV data. If you have set up everything and you need no change anymore (otherwise make copy perhaps), select all polygons and Assign them to the UV(W) Tag. You might test this on a small scene to get your workflow to your target renderer in a savvy way.

In the moment the UV has the Camera data as “camera-indepentend” data, it should work as any UV textured object. (For this UV workflow you stick with the given textures! No “Texture Baking”. Baking always lowers the quality, it never gets better.)

Again, I do not touch here V-Ray at all.

If you have a small (reduced) scene, I like to convert it for you. This might help to see if there is anything else that I can’t know without the teh specific scene of yours.

All the best

Sassi

Hi Sassi,
Thanks! Well, the point here is to bake everything, then I can proceed with the VRay stuff. First things first: when I am baking out my PNG’s with Illumination, I am only getting a file that says “####_Color.png”. This is odd to me, since there is an Alpha Channel being used by Projection Man. So where is the Alpha? The Photoshop file is about 100 MB since it’s pretty layered. Reducing it might be a problem. However, if you are feeling up to it, you can email me through my profile and I’ll send you a link to my DropBox where the C4D file is.

Thanks!

P.S. I should add that it is only the color file that is exported/baked no matter what format I choose and in the files in Photoshop, there isn’t even an alpha channel. Something strange is occurring. I do have “Illumination” checked.

I will not do 3rd party project work here. A reduced file all the time, what ever fits the forum, and what helps to clear the question, nothing more. :o) Thanks for understanding.

(I have done it over a longer time, with an average of 3-4 hours in each case, as each thing creates follow up questions. Only to note that none of these people helped later when a similar problem was here in the forum. Which was my hope though, to establish here something. So, what fits in the forum and is accessible for everyone, no e-mail or pm’s with such requests please, thanks. I will just not answer anymore. The forum is the place for information exchange. :o)

Baking Textures: You might merge the luminance and the alpha if the alpha is not in your png file. (Sorry I haven’t used png files since a decade, so no practice anymore with it, no experience…)

UV: I would go if needed with the UV data, as that will not lower your quality. From projection to bake I assume a loss in quality of at least 20% -40%. A reason why I suggest to stick with the files you have. Perhaps you have large files for the set up, then it might be no problem.

Best

Sassi

P.S. “Baking Objects” for Camera-Projection and leave the Keep UV checked on, might result not in the quality you might like. It can, but mustn’t, change the density in some areas heavily. To be save, test both (un-checked and checked)

Sassi,
No worries!! I completely understand! I was able to reduce the file quite nicely (took some hacking through it, but this should serve the point). So, the problem still remains, PNG, PSD, TIFF, JPEG, none of these files produce an embedded or separate alpha when the texture is baked. As you can see, the PSD has alpha in it. So what is happening makes absolutely no sense to me.

Thanks a million for your help!

Mike

p.s. I am unable to attach a 2.1MB zipped file…for some reason. I have reduced the PSD to 2 objects and my Cinema file has been reduced to 5. It’s the only way I can share this file at the moment because the Cineversity file management system doesn’t work.
http://dl.dropbox.com/u/75499435/PMan_Failure.zip

Hey Mike,

Thanks for understanding. I had many night shifts with such requests and those cases haven’t supported anyone else, not as file here, not active. So I stopped it all together. Nothing personal, great you take it that way.

I found that sometimes two or more periods in a zip file disable the option to attach a file. Sometimes it is (in my case) just Safari as a browser, and I use another one.

All the best

Sassi

Sassi,
I am unable to attach any files to the Cineversity system. I have tried Safari and Firefox. I’m really not sure what the problem is. It is a ZIP file, which the system should be able to handle. It is also a total of 2.5MB, which is well below the allowed size. I’m not sure how else to share the file other than my public Dropbox folder, since the methods provided in the forum don’t work.

Best,
Mike

Hey Mike,

The first thing that cleared the file for me was that you didn’t used the “Bake Texture” tag, you had the Bake Object in use, so far I can tell from the entries in the “Bake Object”.

So first of all, thanks for sharing the file, even while obstacles where given to do so.

The “Bake Object” function is designed to merge all the information into a single channel, the Luminance channel. I’m not certain if there are any options to expand on that.

I use - if needed the “Bake Texture” tag, which helps to create way more information, but yes, it might be not the one single click approach.

Stubborn in terms of quality as I’m, I suggest to go with the UV Assign method! Why? Based on my experience with night time photography, to gain a decent image, if not shot in HDRI, or at least MDRI, you have to clip the highlights, as otherwise anything else will be dark. If you go the route with at least three exposures (MDRI) you have for any light source values above 100% to expect. With a fast pan (as in your scene) this will “paint” a large an long line for each light in the motion blur (MB). On the other hand, those lights tend to bloom (Bokeh) a lot if a shallow depth of field (SDOF) is used and the lights are not in focus. Perhaps MB and SDOF are both in the final image, and then it becomes even more interesting. If only 8 or 16 bit/c integer is used, the lights are clipped and SDOF and MB can produce such effects only with questionable tricks, like setting the threshold for highlights lower (if that is given in software) It is questionable, as any value that is originally not as “hot” as the light sources, might be treaded as well as “hot”

So all that integer stuff disables it, baking textures mixes pixel values and perhaps, if done with an integer format as intermediate solution, lowers the quality even more.

The texture for camera mapping need to be always 1.5 time higher than what you can see in the camera view of the render camera. To bake just in project dimensions, if, is certainly on the way for a very bad result. Everything is perhaps changing in a project, and to start over, just to save some texture storage, not my idea these days anymore.

I would as well use the “Convert To Object” command to get ride of the Bend Deformer, but that is just me. ;o)

All the best

Sassi

Sassi,
Thanks for the reply! I’ll give the Bake Texture tag a try.

I like what you say about the quality. In this instance I had to use a licensed stock image, so I was limited. My image size is almost twice the size of my output, which gave me some flexibility in zooming.

Cheers!!
Mike

Thanks Mike,

I write here of course always to a forum, and address this not only to you. (I assume that we have always a wide variety of user skill level here, so, thanks for taking it that way.)

If you ever need L.A. night shoots, perhaps I have something in my library, but more than that, just ask here, we have people from all over the world, networking is everything in our business.

As I started with to convert my analog film knowledge two decades ago into a digital one, I went through a lot of mistakes, which were at one point just the top options to have. Well, time changes everything.

Good luck with the project

Sassi

Yes! I should have thought of that!! Perhaps one small issue is the “licensing” of images. I am freelancing for a cable TV network (based out of LA) with an office in Albuquerque, NM. I wonder what the legality is regarding using someone’s images, other than retaining the email with a sentence that says, “Here, you can use this in broadcast.”

So, I have investigated your suggestion further. The problem is that when using VRay, the Projection Camera doesn’t seem to like using VRay shaders. So if I convert the shader to a VRay shader, the link in Projection Man breaks. This means that the “Bake Texture” method will not work for VRay.

The work around I have developed is this:
1] Duplicate my object, which Projection Man immediately sees.
2] Duplicate my texture.
3] Assign my duplicate texture to my duplicate object.
4] Delete the image from the Luminance field in my duplicate object. This provides me with a Luma Matte.
5] Bake out both objects. One is a luma, the other is my luminance.
6] Make a VRay Advanced Shader.
7] Load in my luminance and my luma matte.
8] Replace the material on the object.

I will discuss this method with the VRay for C4D developer. This seems slow and arduous, but it works.

Hey Mike,

When the time comes to check images, we find a way. Yes, everything needs to be documented, to avoid stress and upset. I’m very careful with such things as well.

=============

I may sound like an old record by now, but if you would try this one time, and all the baking effort fades away to zero:

1) Have your object in polygons and leave the textures and tags as they are, but with nothing else attached (like Deformers).

3) Object manager>Tags>Assign UVW Coordinates.

Done.

Perhaps check the Texture tag if it is now set to UVW Mapping.

That is all you need to do. The projection texture is nicely mapped as if the camera would affect the scene, but the camera is NOT connected anymore to it.
NO Baking needed! You might optimize your texture resolution, but well…

It is that simple.

Good luck

Sassi

P.S. I just did my own points 1-3.

The file has not your texture inside, you have it anyway, why sending it again.

Sassi,
Thanks! I’m trying this on my large file (taking a bit of time but thus far seems to work. I must have misunderstood above. I’ve found that I have to “Generate UVW” instead of Assign. Seems to be going well thus far!
Cheers,
Mike

Sassi,
This is the way to go (at least in this instance). I went through, deleted all my UVW tags, made sure all objects were polygons and deformations baked in, “Generate UVW,” which then broke my textures from the projection camera, and then I could convert to VRay materials.

Thanks a million!!
Mike

Hi Mike,

Thanks for the post/feedback! :o)

“Generate” works as well. “Assign” works for all or partially, with an existing UVW Tag. Either way, I like this, and use it often. (In all the years I had only twice trouble with it, but couldn’t really tell why. Until I know, I always suggest a copy to use, but the camera projection is easy to get back. So, not really a “brainer”.

The Assign option is great if you have several cameras working on one mesh (not overlapping). Then you can select only the polygons of the projected area and “store” them newly in the UVW tag.

Again: good luck with the project.

Sassi