I was wondering is there a way to set your texture map for frontal projection mapping. Any suggestions?

If you mean camera mapping, i don't think there is. Maybe if you explain a bit more about what you're trying to do, there could be alternative ways ?

 Basically What I'm looking for is a change the way a texture map is laid out on a model. In Cinema 4D this is done by the "Texture tag's projection properties. i.e. frontal, spherical, uvw mapping, shrink wrapping, flat, etc. I believe Poser's set to UVW mapping and I wandering if there is a way to change that.

Not inside Poser.  You can change the mapping itself before you take the
model into Poser, but Poser just uses the existing UV.  

You can change the scale and starting point of U and V, so you could
e.g. rotate the image around a cylinder, or stretch and shrink linearly.  
In some cases those would give the same result as changing the
projection, but it's really not the same thing.

Sigh. Sigh, again.

I will never cease to be amazed that people can make such statements about something being impossible in Poser, just because they do not know how it is done.

Here is how you set up frontal mapping. See that Texture_coords parameter in the Image_Map node? That's where you get to pick the coordinate system. The one you want for frontal mapping is XY. It can be global or local to the model. If global, then rotating the model will have no influence on the pattern - it will always face forward.

The scale needs to be much bigger than 1 - I haven't tested what the underlying units are in this mode, but if you use the default scale of 1, the entire image is very tiny.

Depending on what you want, you can also try "Create Perspective UVs" in the grouping tool. 

Stewer,

What does the setting "Mapper" do? I've never figured that one out.

Quote - Sigh. Sigh, again.

I will never cease to be amazed that people can make such statements about something being impossible in Poser, just because they do not know how it is done.

'frontal' mapping isn't planar mapping in XY, global or otherwise. Its what's C4D calls camera mapping, the main difference being that planar mapping is orthographic, while camera mapping applies perspective distortion that matches the settings of your camera. This way it can be used to apply textures to standins in your 3d scene and do 'fake' 3d effects.

Like stewer suggested, the create persp. UV thing is a kind of hackish way of doing it and it can work for some cases. But poser will not do proper camera mapping, no matter how much you insult fellow forum members.

ghonma,

I didn't insult anybody. I simply made an observation.

Further, I agree camera mapping is not XY mapping, but the OP didn't use that phrase. The phrase was "frontal projection" and orthographic projection is a kind of projection. I acknowledge that it isn't perspective projection. Had the OP asked specifically for perspective projection, I would have suggested "Create perspective UVs".

I'm no expert on the Poser shader tree, but what you've shown looks like flat mapping not frontal mapping. True frontal mapping projects the texture from the camera position. Flat mapping projects the texture on a given plane - XY in the example you quoted. The OP mentioned C4D - the image above shows flat and frontal mapping in Cinema. Maybe if you use the global option you mentioned you would get a frontal mapping effect, but that isn't what you've shown.

And the clarified question, was "I believe Poser's set to UVW mapping and I wandering if there is a way to change that."

ockham specifically and directly responded to that question, which is what I objected to as there is a dedicated parameter for that purpose.

geez, what is going on here?

The second question was "Poser's set to UVW mapping and I was wondering if there is a way to change that."

The question was not "change that to frontal mapping" or "change that to camera mapping". I acknowledge that, not being a C4D user, there were certain connotations of "frontal" that might need to be qualified as perspective projection that I was unaware of.

However, ghonma started out with "If you mean camera mapping". Now why would he say such a thing unless "camera mapping" is different from "frontal mapping". I make no claim to know the C4D terminology for mapping modes. I do make the claim that there is a way to change the texture mapping modes in Poser.

Further, in general, I can't see how "frontal" could be equated with "camera perspective" unless you can guarantee that the camera is in front.

Quote - geez, what is going on here?

As a multi-lingual person (I'm fluent in Engineer, Gamer, Norm and l33t, with a minor in ChatText) I can tell you that it's just the normal noise in the signal.

:biggrin:

One more point. It is also possible in Poser to do any projection of a texture, including camera perspective, spherical, box, equirectangular, slope, height, or anything else that can be expressed mathematically. You just have to know how to connect the right nodes to the Image_Map to change its coordinate system to anything you want. Of course those are not as easy as in C4D or other tools. Nevertheless, when the question is "is it possible" and it is possible the correct answer is yes, not no. If the question is "is it easy", then ok, the answer is no.

The OP did not state the reason for asking about this projection mode. So perhaps I'm off-base in assuming the actual goal has nothing to do with projection modes, per se, but rather the goal is to integrate a 3D object so that some elements of the photo appear to be in front of the object and others appear to be behind. If that's not the goal, I'm sorry to waste your time.

However, if that is your goal, then perspective mapping is one way to do it, but not necessarily the only way to do it.

Here is a demonstration. There is a background photo (actually an equirectangular image mapped onto my environment sphere, so it completely surrounds the 3D subjects.)

The figure is 3D and the goal is to render him behind the fence rails.

But since he's entirely in front of the photo, he occludes all of it.

I set those boxes to use refraction with an IOR of 1. This means they are essentially 100% invisible.

However, I made the figure invisible to raytracing. So when we see through the boxes, we can't see the figure, only the photo behind him.

There's a small amount of error in my placement of the boxes. More careful alignment would solve that.

bill, ISTR one of the developers explained what the mapper function was (can't find the forum post, of course), and it pertained to some script or menu function that was not successfully integrated into poser IIRC.  something like cage's poser uvmapping script in their python forum here.  the camera-mapping (perspective UVs) thing is the main mapping tool now.  this may imply they won't be adding the mapper GUI to poser 8, either, hence stefan may not be able to say much.

Quote - The OP did not state the reason for asking about this projection mode. So perhaps I'm off-base in assuming the actual goal has nothing to do with projection modes, per se, but rather the goal is to integrate a 3D object so that some elements of the photo appear to be in front of the object and others appear to be behind.

I don't know what the OP wants either but that's not the main use of camera mapping. Camera mapping is a way to take a 2D texture/image and by projecting it on dummies, make it appear to respond properly to camera motion. This way you dont have to make a fully textured 3D object for everything in a scene but can just do a nice detailed matte image and use cam mapping to fake the rest. It sees extensive use in film work where most of the backgrounds you see (like the cities in the star wars prequels or the environments in LOTR) are done this way.

To show a simple example, start with a photo of a box:

Import to your app and match it with a dummy cube:

Apply a camera projection on the dummy and try a render. Right now it matches the original photo:

But since its actually projected onto a 3d object, we can now move the camera around like this:

Now of course you could have created a 3D box, textured it to match the photo, then setup lighting to match, but it would have taken far longer then what i just did in 5 minutes. There are also artifacts because the source image is low rez and you're limited in the kind of cam moves you can do, but it saves so much time for complex backgrounds that its more then worth it. It's a pity we don't have proper support for it in poser, as it's one of the most powerful CG techniques out there.

Interesting - thanks for the demo ghonma.

As I've never used it, I won't assume it's the same, so what does Poser's "perspective UVs" do? I'm told it hoists a photo onto a prop at a given starting orientation, as you've shown here. But I won't assume that's the case. Plus there may be subtle but important distinctions that lead to failure.

blender can do this better and its free.

Quote - As I've never used it, I won't assume it's the same, so what does Poser's "perspective UVs" do? I'm told it hoists a photo onto a prop at a given starting orientation, as you've shown here. But I won't assume that's the case. Plus there may be subtle but important distinctions that lead to failure.

It does the same thing but it's fairly limited in practice. Ideally it should be a projection choice in the image node with an input for a camera name. That way it would be easy to tweak and you could setup multiple projections. The way it is now is mainly useful for simple scenes, but in simple scenes you may as well just model and texture everything rather then bothering with cam projections.

To further elaborate with an example, if you look at the pix above, note that i'm only moving the camera so that the same 3 sides of the box are visible at all times. This is because the other 3 sides are a total mess :) To do a camera move in which all 6 sides may need to be seen, you would need 2 separate projections on each of the 3 sides and then blend between 2 different photos using them. For a more complex object, you may even need a whole lot more. But this sort of thing is not easily doable with the tool poser currently has.

Quote - blender can do this better and its free.

True enough. Blender punches way above what it's lightweight price would suggest. Some of it's features, like fluids, even put high end apps costing thousands of $$ to shame.

Thanks, ghonma. I'll have to give this a try in Poser.

So if I understand this correctly, I can quickly build a library of a few simple background props, mostly boxes or cylinders - simple stuff,  based on photos that can be tossed in easily without a lot of manual texture building/straightening etc. Things like windows, doors, paintings, that don't need to be viewed from a lot of different angles, or lit particularly accurately either. Once built, I can position these with certain limitations, but still have some freedom to move them around. I can quickly assemble little "stage props" for demo renders. These should liven up my typical dull demo renders quite a bit.

Am I understanding this correctly?

"camera mapping' is more properly known as Photogrammetry ( http://en.wikipedia.org/wiki/Photogrammetry )

Align with the reference image loaded as background.

Quote - Thanks, ghonma. I'll have to give this a try in Poser.

So if I understand this correctly, I can quickly build a library of a few simple background props, mostly boxes or cylinders - simple stuff,  based on photos that can be tossed in easily without a lot of manual texture building/straightening etc. Things like windows, doors, paintings, that don't need to be viewed from a lot of different angles, or lit particularly accurately either. Once built, I can position these with certain limitations, but still have some freedom to move them around. I can quickly assemble little "stage props" for demo renders. These should liven up my typical dull demo renders quite a bit.

Am I understanding this correctly?

this is what camera mapping is.
http://www.3dfluff.com/cameramapping/cameramappingtut.htm
http://www.digitalartform.com/archives/2004/11/camera_projecti_1.html

Quote - So if I understand this correctly, I can quickly build a library of a few simple background props, mostly boxes or cylinders - simple stuff,  based on photos that can be tossed in easily without a lot of manual texture building/straightening etc. Things like windows, doors, paintings, that don't need to be viewed from a lot of different angles, or lit particularly accurately either. Once built, I can position these with certain limitations, but still have some freedom to move them around. I can quickly assemble little "stage props" for demo renders. These should liven up my typical dull demo renders quite a bit.

That's a pretty cool idea actually. It's easy to find photos of things like furniture, doors, windows, accessories in online catalogs and google images. Shouldn't be too hard to get a dummy to match one of those and do a quick texture with persp UVs like stewer demonstrated. After which you would use it like any other prop, within reasonable angles of view of course.

Quote - Now of course you could have created a 3D box, textured it to match the photo, then setup lighting to match, but it would have taken far longer then what i just did in 5 minutes.

Well, this certainly depends on ones skill with perspective matching.

I just spent fully 20 minutes trying to project a photo of a china cabinet onto a dummy box. I agree it only takes a couple minutes, provided you are willing to accept error in the projection, i.e. you cut off some of the object or you catch some of the part that is not the cabinet. However, if you're trying to capture the edges correctly, particularly with furniture or doors, I have to say this is NOT faster. It is nearly impossible. It is hours of work. For me at least.

I can create a very believable 3D china cabinet from scratch in 10 minutes, using Poser primitives. Then, given that I have a very big library of photo-realistic procedural textures I've made, I could texture it instantly. My wood textures, for example, do not even need UV coordinates, so I don't care if a model is UV mapped or not.

So for me this technique is not such a time-saver. Given that I can photo realisticly texture a prop instantly, and I'm an amateur, I'm guessing the pros can do it too. If they can't, they should be paying me to consult.

I could have built a whole kitchen by now. Sigh.