assimp-discussions — Discussions regarding the ASSIMP library.

[Assimp-discussions] visual studio 2022 fail to load obj model

[Alessio M. <ale...@gm...>]

Hello,
I build the assimp library with CMake for visual studio 2022.
It's all ok, but I get a crash when I try to load a simple obj.
I get the correct number faces and vertex but the mFaces data struct is not
contain the face data.

a pieces of code:

Assimp::Importer importer;

    auto scene =
importer.ReadFile("C:\\xampp\\htdocs\\DentalCadWeb\\ogldev\\tutorial31\\Content\\monkey.obj",
aiProcess_Triangulate | aiProcess_FlipUVs);


    if (!scene || scene->mFlags & AI_SCENE_FLAGS_INCOMPLETE ||
!scene->mRootNode)
    {
        return 0;
    }
    auto mesh = scene->mMeshes[0];
    auto faces = mesh->mFaces; // get crash

Please help.

Thanks in advance

[Assimp-discussions] GLTF 2.0

[Yves P. <neo...@gm...>]

Hello everyone,

 

Just a quick question. Based on the repository, it seems like the overall
assimp development has been on hold for almost a year. I'm wondering if GLTF
2.0 has a roadmap of some sort? In the case where there is no current
progress, I might be interested to put some love into it.

 

Cheers,

[Assimp-discussions] PyAssimp ignores all modification before the export

[Yacine K. <yk...@lu...>]

Hi

I loaded a 3d File with PyAssimp then I changed some vertices position and when I exported it all modification were ignored 
This is an example of what I wrote:

>>> pyassimp.export(scene,'D:/ExportedFile.gltf','gltf2')
>>> import pyassimp
>>> scene = pyassimp.load('D:/test.3ds')
>>> v = scene.meshes[2].vertices
>>> v = [ [p[0], p[1], p[2]+200] for p in v ]
>>> scene.meshes[2].vertices = v
>>> print scene.meshes[2].vertices
[[-13.360946655273438, -19.980607986450195, 200.0], [-13.360946655273438, 19.980606079101562, 200.0], ………
>>> pyassimp.export(scene,'D:/ExportedFile.gltf','gltf2')

but when I load the file there are no changes!

Thanks.

Sent from Mail for Windows 10

[Assimp-discussions] Undefined indentifiers when building

[Arno G. <ar...@ag...>]

Hello all,

It's a few months ago I last build the assimp library, so today I updated the sources and tried to build it again. But I get errors about undeclared identifiers, e.g.:

Error C2065 'AI_CONFIG_GLOBAL_SCALE_FACTOR_KEY': undeclared identifier

I already tried to delete my build folder and run cmake completely from the start, but nothing seems to help.

Was anything changed to the building procedures that I missed maybe?

With regards,

Arno

Re: [Assimp-discussions] make mNumUVComponents an importer option

[kim k. <kim...@go...>]

Hi,

I thought about this as well. Does it makes sense to offer 2 channels
instead of three. Any other opinions?

Kim

<c....@by...> schrieb am So., 21. Jan. 2018 um 15:42 Uhr:

>
> Hello all,
>
> I am using assimp in my ongoing C++ port of the Three.js 3D rendering
> framework (https://github.com/bergstr11/three.cpp). Three.js only
> supports 2 UV channels per material, and each UV has two values
> (namely, u and v), whereas assimp maps UV channels to textures and
> uses a 3-value aiVector3t. The latter leads to the fact that when
> reading the model, I cannot directly use the assimp buffer (as I do
> elsewhere) but have to copy the used values over to another buffer.
> This seems quite inefficient to me, in particular with large models.
>
> My question is therefore whether it might make sense to provide an
> option to the Importer which allows the caller to determine what size
> the UV elements have (i.e., mNumUVComponents). I would be willing to
> contribute the relevant changes.
>
> regards,
> Christian
>
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

[Assimp-discussions] make mNumUVComponents an importer option

[<c....@by...>]

Hello all,

I am using assimp in my ongoing C++ port of the Three.js 3D rendering  
framework (https://github.com/bergstr11/three.cpp). Three.js only  
supports 2 UV channels per material, and each UV has two values  
(namely, u and v), whereas assimp maps UV channels to textures and  
uses a 3-value aiVector3t. The latter leads to the fact that when  
reading the model, I cannot directly use the assimp buffer (as I do  
elsewhere) but have to copy the used values over to another buffer.  
This seems quite inefficient to me, in particular with large models.

My question is therefore whether it might make sense to provide an  
option to the Importer which allows the caller to determine what size  
the UV elements have (i.e., mNumUVComponents). I would be willing to  
contribute the relevant changes.

regards,
Christian

Re: [Assimp-discussions] 3D Printer Support?

[kim k. <kim...@go...>]

Hi,
I found soe support from the original authors from the 3MF-importer, but as
far as I know no one started to work on an exporrt until now.

It is still on my list, because at SLM-Solutions there are some people
interested as well.

Kim

David Rivkin, PhD <dav...@sc...> schrieb am So., 22. Okt.
2017, 18:30:

> Hi,
>
> I have been away from ASSIMP for a while but getting back to it.
>
> I was just wondering if there was any progress on supporting the export
> of 3D printer formats AMF and 3MF (HP/MS)?
>
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

Re: [Assimp-discussions] 3D Printer Support?

[kim k. <kim...@go...>]

I am wondering why there aren't any slicing metadata provided by 3MF?

kim kulling <kim...@go...> schrieb am Di., 24. Okt. 2017,
13:42:

> Hi,
> I found soe support from the original authors from the 3MF-importer, but
> as far as I know no one started to work on an exporrt until now.
>
> It is still on my list, because at SLM-Solutions there are some people
> interested as well.
>
> Kim
>
> David Rivkin, PhD <dav...@sc...> schrieb am So., 22. Okt.
> 2017, 18:30:
>
>> Hi,
>>
>> I have been away from ASSIMP for a while but getting back to it.
>>
>> I was just wondering if there was any progress on supporting the export
>> of 3D printer formats AMF and 3MF (HP/MS)?
>>
>> --
>> Thank you,
>> Prof. David Rivkin, PhD
>> President & Chief Scientist
>> SciEssence Intl.
>>
>>
>> ------------------------------------------------------------------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>> _______________________________________________
>> Assimp-discussions mailing list
>> Ass...@li...
>> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>>
>

[Assimp-discussions] 3D Printer Support?

[David R. P. <dav...@sc...>]

Hi,

I have been away from ASSIMP for a while but getting back to it.

I was just wondering if there was any progress on supporting the export
of 3D printer formats AMF and 3MF (HP/MS)?

-- 
Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

Re: [Assimp-discussions] Browser Integration

[Tom P. <to...@ma...>]

Hey,

Thanks for getting back to me. I understand - good luck with the
development.

Cheers,

On Fri, Sep 29, 2017 at 3:52 AM, kim kulling <kim...@go...>
wrote:

> I guess we have already this feature request, but not ebnought time to
> start with it.
>
> Kim
>
> Tom Pittlik <to...@ma...> schrieb am Fr., 22. Sep. 2017 um
> 16:01 Uhr:
>
>> Hi,
>>
>> Any chance browser integration could be added to assimp? I have a mesh
>> resource website and it would be great if our users could use assimp to
>> quickly view models out of their browsers. For example, instead of the
>> standard download link, the link to the mesh could be
>> *assimp://[url_to_mesh]*, which would launch assimp out of the browser
>> and download/display the model in assimp.
>>
>> https://stackoverflow.com/questions/389204/how-do-i-
>> create-my-own-url-protocol-e-g-so
>>
>> Thanks,
>> Tom
>> ------------------------------------------------------------
>> ------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot______
>> _________________________________________
>> Assimp-discussions mailing list
>> Ass...@li...
>> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>>
>


-- 
Tom Pittlik

* <https://mail.google.com/>* <http://banana.org/>
*The Gaming Network*

Re: [Assimp-discussions] Browser Integration

[kim k. <kim...@go...>]

I guess we have already this feature request, but not ebnought time to
start with it.

Kim

Tom Pittlik <to...@ma...> schrieb am Fr., 22. Sep. 2017 um
16:01 Uhr:

> Hi,
>
> Any chance browser integration could be added to assimp? I have a mesh
> resource website and it would be great if our users could use assimp to
> quickly view models out of their browsers. For example, instead of the
> standard download link, the link to the mesh could be
> *assimp://[url_to_mesh]*, which would launch assimp out of the browser
> and download/display the model in assimp.
>
>
> https://stackoverflow.com/questions/389204/how-do-i-create-my-own-url-protocol-e-g-so
>
> Thanks,
> Tom
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

[Assimp-discussions] Browser Integration

[Tom P. <to...@ma...>]

Hi,

Any chance browser integration could be added to assimp? I have a mesh
resource website and it would be great if our users could use assimp to
quickly view models out of their browsers. For example, instead of the
standard download link, the link to the mesh could be
*assimp://[url_to_mesh]*, which would launch assimp out of the browser and
download/display the model in assimp.

https://stackoverflow.com/questions/389204/how-do-i-create-my-own-url-protocol-e-g-so

Thanks,
Tom

Re: [Assimp-discussions] Another (crashing related) question

[rdb <me...@rd...>]

I'm not sure if this is what's causing the problem, but the correct syntax
would be "delete[] scene->mMeshes;" assuming it is allocated using
"new[]".  Using "delete" on an array allocated with "new[]" invokes
undefined behaviour.

Kind regards,
rdb

On 12 July 2017 at 13:30, David Rivkin, PhD <dav...@sc...>
wrote:

> Hi all.  Sorry for so many questions but things just keep crashing that
> make no sense to me.
>
> My code (below) is trying to add a new mesh to the end of the list in the
> Scene.  That is all (later I have to add its index to the list of meshes of
> the node but I have not gotten that far yet.
>
> // Add a new Mesh to the scene
>
>     aiMesh ** newMeshArray = new aiMesh*[scene->mNumMeshes + 1];
>
>     for(unsigned int m = 0; m < scene->mNumMeshes; m++) {
>
>         newMeshArray[m] = scene->mMeshes[m];
>
>     }
>
>     aiMesh * newMesh = new aiMesh();
>
>     newMesh->mMaterialIndex = scene->mMeshes[serialFace->meshIndex()]->mMaterialIndex;
>
>     // TODO:  Material selection to Editor and copy to the Mesh here, in future version
>
>     newMesh->mName = aiString(text.toStdString());
>
>     newMeshArray[scene->mNumMeshes] = newMesh;
>
>     scene->mNumMeshes += 1;
>
>     delete scene->mMeshes;
>
>     scene->mMeshes = newMeshArray;
>
> When I go to delete the pointer array, I get HEAP[AddSerial.exe]: Invalid address specified to RtlValidateHeap( 000001931FEF0000, 000001931FF32730 ) on the line:
> delete scene->mMeshes;
>
>
> I dont see why this is happening.  The array is no longer references,
> should have been created on the heap and find to clear since Im done with
> it and have a new one ready for assignment.
>
> --
>
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>
>

[Assimp-discussions] Another (crashing related) question

[David R. P. <dav...@sc...>]

Hi all.  Sorry for so many questions but things just keep crashing that
make no sense to me.

My code (below) is trying to add a new mesh to the end of the list in
the Scene.  That is all (later I have to add its index to the list of
meshes of the node but I have not gotten that far yet.

//AddanewMeshtothescene

aiMesh**newMeshArray=newaiMesh*[scene->mNumMeshes+1];

for(unsignedintm=0;m<scene->mNumMeshes;m++){

newMeshArray[m]=scene->mMeshes[m];

}

aiMesh*newMesh=newaiMesh();

newMesh->mMaterialIndex=scene->mMeshes[serialFace->meshIndex()]->mMaterialIndex;

//TODO:MaterialselectiontoEditorandcopytotheMeshhere,infutureversion

newMesh->mName=aiString(text.toStdString());

newMeshArray[scene->mNumMeshes]=newMesh;

scene->mNumMeshes+=1;

deletescene->mMeshes;

scene->mMeshes=newMeshArray; When I go to delete the pointer array, I get HEAP[AddSerial.exe]: Invalid address specified to RtlValidateHeap( 000001931FEF0000, 000001931FF32730 ) on the line:
delete scene->mMeshes;


I dont see why this is happening.  The array is no longer references,
should have been created on the heap and find to clear since Im done
with it and have a new one ready for assignment.

-- 

Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

Re: [Assimp-discussions] Does anyone have an example of setting new MetaData in a Node?

[kim k. <kim...@go...>]

Have you tried this out:
http://kimkulling.de/2017/05/17/use-the-asset-importer-lib-meta-data-api-right/
.
There is also an example in the unittests.

Kim

David Rivkin, PhD <dav...@sc...> schrieb am Fr., 7. Juli
2017 um 14:59 Uhr:

> Hi everyone,
>
> I see that Node MetaData is a problem for me and perhaps I am doing
> something wrong.  If anyone has an example where this works properly,
> please let me know.  I only get an error when I try to delete the scene,
> not even when I write the new file to disk.
>
> Here is my code below.  If you see something I am doing wrong, please let
> me know!
>
> if(nodePointer()->mMetaData == NULL) {
>
>             meta = new aiMetadata();
>
>         }
>
>         Q_ASSERT(meta);
>
>         if(meta->mNumProperties < 1) {
>
>             meta->mKeys = new aiString(ADD_SERIAL_STRING);
>
>             meta->mValues = new aiMetadataEntry();
>
>             meta->mNumProperties = 1;
>
>         }
>
>         // For the string as JSON.
>
>         aiString metaKey;
>
>         QByteArray json = face->toString();
>
>         metaKey.Set(json);
>
>         meta->Set<aiString>(0, ADD_SERIAL_STRING, metaKey);
>
>         nodePointer()->mMetaData = meta;
>
>
>
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

[Assimp-discussions] Does anyone have an example of setting new MetaData in a Node?

[David R. P. <dav...@sc...>]

Hi everyone,

I see that Node MetaData is a problem for me and perhaps I am doing
something wrong.  If anyone has an example where this works properly,
please let me know.  I only get an error when I try to delete the scene,
not even when I write the new file to disk.

Here is my code below.  If you see something I am doing wrong, please
let me know!

if(nodePointer()->mMetaData==NULL){

meta=newaiMetadata();

}

Q_ASSERT(meta);

if(meta->mNumProperties<1){

meta->mKeys=newaiString(ADD_SERIAL_STRING);

meta->mValues=newaiMetadataEntry();

meta->mNumProperties=1;

}

//ForthestringasJSON.

aiStringmetaKey;

QByteArrayjson=face->toString();

metaKey.Set(json);

meta->Set<aiString>(0,ADD_SERIAL_STRING,metaKey);

nodePointer()->mMetaData=meta;

-- 
Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

Re: [Assimp-discussions] Exporting Scene question

[kim k. <kim...@go...>]

Callstack?

Thanks,
Kim

David Rivkin, PhD <dav...@sc...> schrieb am Fr., 30. Juni
2017 um 14:42 Uhr:

> Hi again all,
>
> Does Export(...)  clear the aiScene * memory?
>
>
> In my app, as I need to edit the scene object, I first copy it:
> aiCopyScene(importScene, &mScene);
>
> After I export mScene to a file and then go to close my app (window) when
> the window destructor calls aiFreeScene(mScene); and this crashes.
>
> Any thoughts?
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

Re: [Assimp-discussions] Question about the "best format"

[kim k. <kim...@go...>]

Hi all,

at the moment obj is a stable one. To store more meta information the
meta-API can be used. Collada is stable as well, but the format is more
complex.

Best regards,
KIm

David Rivkin, PhD <dav...@sc...> schrieb am Fr., 30. Juni
2017 um 11:41 Uhr:

> Hi all,
>
> I would like to get your opinion as to the "best", ie most complete file
> format that I can use as an intermediary  file format.  It should
> support as much of the ASSIMP datastructure as possible, including Node
> and Mesh names, materials, colors.  ASSIMP needs to be able to read and
> save to this format.
>
> OBJ seems pretty good.  3MF and AMF would be ideal but are not yet
> supported for saving it appears.
>
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

[Assimp-discussions] Exporting Scene question

[David R. P. <dav...@sc...>]

Hi again all,

Does Export(...)  clear the aiScene * memory?


In my app, as I need to edit the scene object, I first copy it:  
aiCopyScene(importScene,&mScene);

After I export mScene to a file and then go to close my app (window)
when the window destructor calls aiFreeScene(mScene);and this crashes.

Any thoughts?

-- 
Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

[Assimp-discussions] Question about the "best format"

[David R. P. <dav...@sc...>]

Hi all,

I would like to get your opinion as to the "best", ie most complete file
format that I can use as an intermediary  file format.  It should
support as much of the ASSIMP datastructure as possible, including Node
and Mesh names, materials, colors.  ASSIMP needs to be able to read and
save to this format.

OBJ seems pretty good.  3MF and AMF would be ideal but are not yet
supported for saving it appears.

-- 
Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

Re: [Assimp-discussions] Where can I find the definition of TReal?

[kim k. <kim...@go...>]

Per default TReal is set to float, but you can enable it to double. There
is a switch for cmake -DASSIMP_DOUBLE_PRECISION=ON to enable this feature.
With double precision all numerical operations will run with doubles ...

Kim

Kim

David Rivkin, PhD <dav...@sc...> schrieb am Di., 6. Juni
2017 um 13:18 Uhr:

> I have tried greping the code but have strangely been unable to find a
> header file that defines TReal.
>
> Clearly everyone would want to include the header that defines TReal in
> their codes so that type actual types (float vs double) are used correctly.
>
> At this point I am assuming that TReal is only in the Makefile as a
> #define but this is not really useful to applications using Assimp that
> need to do the math with matrixes etc.
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

Re: [Assimp-discussions] Where can I find the definition of TReal?

[rdb <me...@rd...>]

TReal is not a type, but a template parameter.  The meaning of this becomes
defined where the template is instantiated.  For example, in vector3.h,
aiVector3t is a class template that defines a single typename parameter
called TReal:

template <typename TReal>
class aiVector3t

This template is instantiated as aiVector3D with TReal being substituted
for ai_real (which is defined in defs.h):

typedef aiVector3t<ai_real> aiVector3D;

Kind regards,
rdb



On 6 June 2017 at 13:18, David Rivkin, PhD <dav...@sc...>
wrote:

> I have tried greping the code but have strangely been unable to find a
> header file that defines TReal.
>
> Clearly everyone would want to include the header that defines TReal in
> their codes so that type actual types (float vs double) are used correctly.
>
> At this point I am assuming that TReal is only in the Makefile as a
> #define but this is not really useful to applications using Assimp that
> need to do the math with matrixes etc.
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>
>

[Assimp-discussions] Where can I find the definition of TReal?

[David R. P. <dav...@sc...>]

I have tried greping the code but have strangely been unable to find a
header file that defines TReal.

Clearly everyone would want to include the header that defines TReal in
their codes so that type actual types (float vs double) are used correctly.

At this point I am assuming that TReal is only in the Makefile as a
#define but this is not really useful to applications using Assimp that
need to do the math with matrixes etc.

-- 
Thank you,
Prof. David Rivkin, PhD
President & Chief Scientist
SciEssence Intl.

Re: [Assimp-discussions] AMF and 3MF Exporting?

[kim k. <kim...@go...>]

Hi,

right, another topic we need to adress. Do you need it?

Kim

David Rivkin, PhD <dav...@sc...> schrieb am Sa., 27. Mai
2017 um 02:25 Uhr:

> I see that the current github MASTER of Assimp is supporting reading AMF
> and 3MF 3D Manufacturing Format file.
>
> Is anyone working on the exporting?  This should be easier than importing.
>
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>

Re: [Assimp-discussions] aiProcess_GenNormals question

[Giuseppe B. <el...@gm...>]

As already David Nadlinger mentioned, it's like that because of the
real-time 3d graphics, where unfortunately you can't have normal and color
per triangle/face.

More specifically, the vertex shader in the graphic pipeline expects always
data for each vertex attribute.

For example, here, you need position, normal and color:

https://github.com/g-truc/ogl-samples/blob/b856b8ca236e1b46d40cf514784a25d3eaec5f4f/data/gl-430/buffer-uniform.vert#L17-L19

then each vertex has to come with a position, a normal and a color data
(vec*) inside your gpu buffer objects.

This is the norm unless very specific cases.


Giuseppe

2017-05-30 12:49 GMT+02:00 David Rivkin, PhD <dav...@sc...>:

> Why does aiProcess_GenNormals (option to Importer::ReadFile()) generating
> a list of Normals for each vertex when there should be just 1 per Face, and
> hence the array of mNormals should be equal to the aiMesh.mNumFaces, not
> aiMesh.mNumVertices?  The normal should be both a aiVector3Dand a Real
> scaler (offset typically referred to as 'w' in code for Omega) as well.
>
> In fact, the documentation is very confusing stating "aiVector3D
> <http://www.assimp.org/lib_html/vector3_8h.html#a47f446977120a9ca7d868353aad70b09>*
> aiMesh::mNormals
> Vertex normals
>
> The array contains normalized vectors, NULL if not present. The array is
> mNumVertices in size...."  This should be Face Normals.
>
> I would consider this a bug.
> --
> Thank you,
> Prof. David Rivkin, PhD
> President & Chief Scientist
> SciEssence Intl.
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Assimp-discussions mailing list
> Ass...@li...
> https://lists.sourceforge.net/lists/listinfo/assimp-discussions
>
>