 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> clipka <ano### [at] anonymous org> wrote:
>> Am 11.10.2017 um 15:19 schrieb And:
>>> clipka <ano### [at] anonymousorg> wrote:
>>>> Am 11.10.2017 um 09:53 schrieb omniverse:
>>>>
>>>>> I got the STL file too, and converted to OBJ then exported to POV but that won't
>>>>> show the interior via CSG difference.
>>>>
>>>> You'll have to add an `inside_vector` statement (unless your converter
>>>> software already does that for you).
>>>
>>>
>>> So I see that it states an "inside_vector <0,0,1> "(from poseray) in the last
>>> item in each mesh2 object. It does not contain sufficient information.
>>
>> With that statement, the mesh should work fine in CSG difference (or
>> intersection or union, for that matter).
>>
>> Not sure what you mean by "it does not contain sufficient information".
>
> Because there is just one vector <0,0,1> at the end, I suppose that it should
> be a lot of vectors. Maybe each face have one vector, and should vary from face
> to face, I don't know。。。 I suppose it should look like
> that.
>
There is never more than a single inside_vector par mesh.
I think that, if you have several vectors, then, only the last will be
used, each new one replacing the previous one.
You may need to slightly change it if there are triangles that are
parallel to it. But nothing else.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
clipka <ano### [at] anonymousorg> wrote:
> Am 11.10.2017 um 16:35 schrieb And:
>
> >>> So I see that it states an "inside_vector <0,0,1> "(from poseray) in the last
> >>> item in each mesh2 object. It does not contain sufficient information.
> >>
> >> With that statement, the mesh should work fine in CSG difference (or
> >> intersection or union, for that matter).
> >>
> >> Not sure what you mean by "it does not contain sufficient information".
> >
> > Because there is just one vector <0,0,1> at the end, I suppose that it should
> > be a lot of vectors. Maybe each face have one vector, and should vary from face
> > to face, I don't know。。。 I suppose it should look like
> > that.
>
> No; there is only one `inside_vector` per mesh; it tells POV-Ray that in
> order to test whether a given point is inside the mesh or not, it should
> shoot a ray in the direction specified by that parameter, and count the
> number of surfaces encountered in that direction. If the count is odd,
> the point must be inside the mesh; if the count is even, the point must
> be outside.
Well, that it seems poseray adds a symbolic vector, because <0,0,1> does not
inside my model in fact. And it shouldn't get such an information because the
mesh was output from sketchup it does not contain that, moreover most of the
mesh 3d model does not have a well construct that contain a inside/ outside
distinction.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 11-10-2017 7:46, And wrote:
> I downloaded a sketchup 3d model on the 3d warehouse. Last week I processed it
> to a section view, and output to povray, then rendered this picture.
>
Nice!! Didn't know the site, but I am sure to visit in future
--
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 12/10/2017 07:36, And wrote:
>> No; there is only one `inside_vector` per mesh; it tells POV-Ray that in
>> order to test whether a given point is inside the mesh or not, it should
>> shoot a ray in the direction specified by that parameter, and count the
>> number of surfaces encountered in that direction. If the count is odd,
>> the point must be inside the mesh; if the count is even, the point must
>> be outside.
> Well, that it seems poseray adds a symbolic vector, because <0,0,1> does not
> inside my model in fact. And it shouldn't get such an information because the
> mesh was output from sketchup it does not contain that,
That is probably because PoseRay was designed to create mesh2's for
PovRay. So when things change in PovRay FlyerX updates PoseRay to
reflect that.
> moreover most of the
> mesh 3d model does not have a well construct that contain a inside/ outside
> distinction.
>
Which is probably because Sketchup was designed to upload content to
Google Earth. Not to a quality renderer.
Which is not a criticism of Sketchup. It produces good results as your
images show.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Stephen <mca### [at] aolcom> wrote:
>
> Which is probably because Sketchup was designed to upload content to
> Google Earth. Not to a quality renderer.
> Which is not a criticism of Sketchup. It produces good results as your
> images show.
>
> --
>
> Regards
> Stephen
Thomas de Groot <tho### [at] degrootorg> wrote:
> On 11-10-2017 7:46, And wrote:
> > I downloaded a sketchup 3d model on the 3d warehouse. Last week I processed it
> > to a section view, and output to povray, then rendered this picture.
> >
>
> Nice!! Didn't know the site, but I am sure to visit in future
>
> --
> Thomas
By the way, many models of 3d warehouse were made for google earth in the past.
So its models lack details. But there are still many authors upload detailed
models.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 12.10.2017 um 08:36 schrieb And:
> clipka <ano### [at] anonymousorg> wrote:
>> Am 11.10.2017 um 16:35 schrieb And:
>>
>>>>> So I see that it states an "inside_vector <0,0,1> "(from poseray) in the last
>>>>> item in each mesh2 object. It does not contain sufficient information.
>>>>
>>>> With that statement, the mesh should work fine in CSG difference (or
>>>> intersection or union, for that matter).
>>>>
>>>> Not sure what you mean by "it does not contain sufficient information".
>>>
>>> Because there is just one vector <0,0,1> at the end, I suppose that it should
>>> be a lot of vectors. Maybe each face have one vector, and should vary from face
>>> to face, I don't know。。。 I suppose it should look like
>>> that.
>>
>> No; there is only one `inside_vector` per mesh; it tells POV-Ray that in
>> order to test whether a given point is inside the mesh or not, it should
>> shoot a ray in the direction specified by that parameter, and count the
>> number of surfaces encountered in that direction. If the count is odd,
>> the point must be inside the mesh; if the count is even, the point must
>> be outside.
>
> Well, that it seems poseray adds a symbolic vector, because <0,0,1> does not
> inside my model in fact. And it shouldn't get such an information because the
> mesh was output from sketchup it does not contain that, moreover most of the
> mesh 3d model does not have a well construct that contain a inside/ outside
> distinction.
As described above, `inside_vector` does not specify a location, but
rather a direction (in this case parallel to the z axis).
The "given point" mentioned above will arise from the rendering
algorithm, more specifically the algorithm used to render CSG
differences, intersections or merges. As part of this algorithm, POV-Ray
frequently needs to answer the question, "this point I've found on the
surface of object A, is it inside or outside object B?"
The procedure I described earlier, which makes use of `inside_vector` as
a direction, is how POV-Ray tries to answers this question for cases
where object B is a mesh (or mesh2).
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> Well, that it seems poseray adds a symbolic vector, because <0,0,1> does not
> inside my model in fact. And it shouldn't get such an information because the
> mesh was output from sketchup it does not contain that, moreover most of the
> mesh 3d model does not have a well construct that contain a inside/ outside
> distinction.
>
It's not a sympolic vector, it's a DIRECTION vector.
To see if ANY point is inside the mesh, POV-Ray shoot a ray from that
point in the direction of the inside_vector.
With the default vector rovided, the test ray is parallel to the Z axis
toward +Z.
It's similar to the normal vector of a plane.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
clipka <ano### [at] anonymousorg> wrote:
> Am 12.10.2017 um 08:36 schrieb And:
> > clipka <ano### [at] anonymousorg> wrote:
> >> Am 11.10.2017 um 16:35 schrieb And:
> >>
> >>>>> So I see that it states an "inside_vector <0,0,1> "(from poseray) in the last
> >>>>> item in each mesh2 object. It does not contain sufficient information.
> >>>>
> >>>> With that statement, the mesh should work fine in CSG difference (or
> >>>> intersection or union, for that matter).
> >>>>
> >>>> Not sure what you mean by "it does not contain sufficient information".
> >>>
> >>> Because there is just one vector <0,0,1> at the end, I suppose that it should
> >>> be a lot of vectors. Maybe each face have one vector, and should vary from face
> >>> to face, I don't know。。。 I suppose it should look like
> >>> that.
> >>
> >> No; there is only one `inside_vector` per mesh; it tells POV-Ray that in
> >> order to test whether a given point is inside the mesh or not, it should
> >> shoot a ray in the direction specified by that parameter, and count the
> >> number of surfaces encountered in that direction. If the count is odd,
> >> the point must be inside the mesh; if the count is even, the point must
> >> be outside.
> >
> > Well, that it seems poseray adds a symbolic vector, because <0,0,1> does not
> > inside my model in fact. And it shouldn't get such an information because the
> > mesh was output from sketchup it does not contain that, moreover most of the
> > mesh 3d model does not have a well construct that contain a inside/ outside
> > distinction.
>
> As described above, `inside_vector` does not specify a location, but
> rather a direction (in this case parallel to the z axis).
>
> The "given point" mentioned above will arise from the rendering
> algorithm, more specifically the algorithm used to render CSG
> differences, intersections or merges. As part of this algorithm, POV-Ray
> frequently needs to answer the question, "this point I've found on the
> surface of object A, is it inside or outside object B?"
>
> The procedure I described earlier, which makes use of `inside_vector` as
> a direction, is how POV-Ray tries to answers this question for cases
> where object B is a mesh (or mesh2).
Hmm...it seems ok, if mesh shape is closed. But just like Alain said, need to
slightly change it.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 12/10/2017 10:00, And wrote:
> Stephen <mca### [at] aolcom> wrote:
>>
>> Which is probably because Sketchup was designed to upload content to
>> Google Earth. Not to a quality renderer.
>> Which is not a criticism of Sketchup. It produces good results as your
>> images show.
>>
>> --
>>
>> Regards
>> Stephen
>
>
> Thomas de Groot <tho### [at] degrootorg> wrote:
>> On 11-10-2017 7:46, And wrote:
>>> I downloaded a sketchup 3d model on the 3d warehouse. Last week I processed it
>>> to a section view, and output to povray, then rendered this picture.
>>>
>>
>> Nice!! Didn't know the site, but I am sure to visit in future
>>
>> --
>> Thomas
>
>
>
> By the way, many models of 3d warehouse were made for google earth in the past.
> So its models lack details. But there are still many authors upload detailed
> models.
>
So I noticed. A big improvement since I last visited the site.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 12/10/2017 à 18:41, Alain a écrit :
> Le 17-10-12 à 02:36, And a écrit :
>
>>
>> Well, that it seems poseray adds a symbolic vector, because <0,0,1>
>> does not
>> inside my model in fact. And it shouldn't get such an information
>> because the
>> mesh was output from sketchup it does not contain that, moreover most
>> of the
>> mesh 3d model does not have a well construct that contain a inside/
>> outside
>> distinction.
>>
>
> It's not a sympolic vector, it's a DIRECTION vector.
> To see if ANY point is inside the mesh, POV-Ray shoot a ray from that
> point in the direction of the inside_vector.
> With the default vector rovided, the test ray is parallel to the Z axis
> toward +Z.
> It's similar to the normal vector of a plane.
And to finish the explanation: POV-Ray counts the number of
intersections from that point along that vector, to infinity.
When odd, the point is inside,
when even, the point is outside.
(Unless inside & outside were inverted)
Hence it's better for the vector to never be perpendicular to any normal
of the triangles of the mesh, should bad luck occurs for a point in the
exact plane of that triangle, the inside/outside test might be "fluffy".
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
