"Gilles Tran" <tra### [at] inapginrafr> wrote in message
news:3c714ce8$1@news.povray.org...


> 3c6fae4a@news.povray.org...
>
> > I'm intrigued/puzzled/fascinated by your use of the image_map. How did
you
> > know where to place it in the first instance with a 'complicated' image,
> > trial and error? Or is there a set way to determine this? I can guess
that
> > the large area of free-space around your logo is to cover the rest of
the
> > area on the mug, ie, the inside, outside of the mug, but why does it
have
> to
> > be that way? Couldn't a smaller image be applied to one specific area of
> the
> > mug instead of having a large area of free-space? I don't understand
this.
>
> I guess I can answer this one since I modelled the mug in Rhino.

    Gilles, my sincere apologies to you. I'm looking at the code now and see
your name....  :(

     My sincere apologies to Chris too....


 >The answer
> is simple : it's a uv-mapped mesh. The bitmap template is obtained
directly
> using uvmapper (the original obj file is needed for that). It's large
> because it covers the whole surface of the mug, interior and exterior (the
> handle has its own map, not used here). Note that it could be very easy to
> change the template and rewrite the mesh uv coordinates to give more space
> to the logotype (again, using the original obj). One problem though is
that
> the uv grid lines are not parallel so that distortion occurs (tricky to
> fix).


     Yes, as I found out. As I mentioned, I had great fun trying to line up
my version without some kind of distortion, but due to 'natural' distortion,
ie, the curve of the mug, I found it difficult to find the right
perspective, (which is why my version looks distorted towards the base even
after trying to distort the image to compensate in an outside paint
program).


See the attached screen shot of the mug in uvmapper pro (demo) showing
> the repartition of the map surface.
>
> Note that the logo could be done with a regular plane or cylinder mapping
> after getting rid of the uv info, but then the image would get replicated
on
> the other side or in the inside and the normals would be messed up on the
> edges. For simple objects like this, it's possible to find workarounds for
> regular povray mapping, but uv mapping is much more flexible.


    I see, thanks for that.


>
>
> >    Also, can you explain why it's a reversed/mirror image before
applying?


>
> For some reason, there are problems of inversion with meshes and their uv
> maps. Basically original (obj) mesh geometry tend to be inverted after
> conversion to povray. This may be fixed during the conversion process, but
I
> tend to follow the easiest path...

   Thanks. I understand a touch more now Gilles: I need to experiment with
UV mapping myself, (I haven't yet).

   I've learnt a bit from playing with this image though, which takes me a
bit further with my enjoyment in using PoV-Ray.

   Thanks for your informative reply.

      ~Steve~


>
> G.
>
>
> --
>
> **********************
> http://www.oyonale.com
> **********************
> - Graphic experiments
> - POV-Ray and Poser computer images
> - Posters
>
>
>
>

Post a reply to this message