Re: [Openexr-devel] OpenEXR geometry...

[Florian Kainz]

Bruno Ripa wrote:
> 
> Hi, another little question about OpenEXR..... :)
> 
> it seems to me that a HDR image can be mapped without any problem on a
> sphere, while i read in OpenEXR docs that pixel coords are relative to a
> rectangular geometry (a plane :) ).
> 
> Let's talk for difference with HDR.
> 
> So, let's suppose i've got a probe image in hdr format, i can use it on a
> sphere and so on....
> 
> And with a openexr image which's the way ?
> Can i map it on a sphere ? I suppose to get a little perturbation of my
> datas because (as sad before) i will use points of a plane and project them
> on a sphere.....
> 
> So what kink of result will i have and how good will be the representation
> of the image (and the data it represents) ?
> 

Hi Bruno,

you can, of course, map an OpenEXR image onto a sphere.
I don't see any reasons why some of the common image
file formats would be better or worse than others for
this kind of operation.

Almost all image file formats that I know of store planar, 
rectangular pixel grids; mapping the pixel grid onto a 
sphere involves warping and resampling the image.  There 
may be image file formats that directly support spherical 
images, but I have never heard of any.

Does this help, or did I completely misunderstand your 
question?

Florian