[x3d-public] TextureCoordinate point range…should be reduced?

[Michalis Kamburelis]

Any texture coordinate value is valid, even negative, even -9999, and
you can find 3D models indeed using them. That's true for any 3D
software, format and API (Blender, X3D, glTF, OpenGL..).

The interpretation of values outside of 0..1 range indeed depends on
the "repeatS" / "repeatT" / "repeatR" (for 3D textures) values. The
specification already explains this, see
https://www.web3d.org/specifications/X3Dv4Draft/ISO-IEC19775-1v4-CD/Part01/components/texturing.html#TextureCoordinates
, "18.2.3 Texture coordinates".

Note that TextureProperties node, if present, overrides the
"repeat..." fields, with a bit more capable "boundaryMode...", see
https://www.web3d.org/specifications/X3Dv4Draft/ISO-IEC19775-1v4-CD/Part01/components/texturing.html#TextureProperties
. See also CGE / view3dscene docs,
https://castle-engine.io/x3d_implementation_texturing.php , where we
explain what we support (and why we do not plan to support "clamp to
border", but we support all else).

Regards,
Michalis

sob., 25 lis 2023 o 06:15 John Carlson via x3d-public
<x3d-public at web3d.org> napisał(a):
>
> Can someone explain what a negative point value in a TextCoordinate means when the texture is an ImageTexture?  Does it have something to do with repeatS and repeatT?
>
> A section in the spec is ok.
>
> Thanks,
>
> John
>
> On Fri, Nov 24, 2023 at 10:32 PM John Carlson <yottzumm at gmail.com> wrote:
>>
>> 18.4.7 TextureCoordinate
>>
>> TextureCoordinate : X3DSingleTextureCoordinateNode {
>>   SFString [in,out] mapping  ""
>>   SFNode   [in,out] metadata NULL [X3DMetadataObject]
>>   MFVec2f  [in,out] point    []   (-∞,∞)
>> }
>>
>>
>> I ran into a case where I was expecting texture coordinate point values from [0,1] in a texture mapping file, and I got values like -99999 an other negative values, which cause me to ignore the triangles and totally foul up the texture.  At this point, I don’t know what negative values mean in a texture coordinate point value and what I should accept for an IFS.  It seems like -99999 is dreadfully wrong, but per the spec above, any float value is acceptable.  I will check the IFS spec, but if no one has a use case, I recommend reducing the range of the texture coordinate point values.
>
> _______________________________________________
> x3d-public mailing list
> x3d-public at web3d.org
> http://web3d.org/mailman/listinfo/x3d-public_web3d.org