Sign in to follow this  

DX11 Getting around non-connected vertex gaps in hardware tessellation displacement mapping

This topic is 1110 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Sorry for the long title, couldn't figure out how to express it shorter without being overly ambigious as to what this post is about.

 

Anyway, I've been poking around with displacement maping using the hardware tessellation features of DX11 for getting some more vertices to actually displace the last few days, for no particular reason other than to try it out so I'm not really looking for other ways to solve some specific problem.

Displacing a sphere or some other surface with completely connected faces work out as intended but issues obviously occur where there are multiple vertices with the same position but different normals (these vertices then get displaced in different directions and thus become disconnected => gaps appear in the geometry). I tried to mock up some simple solution to this by finding out which vertices share positions in my meshes and then setting a flag for these to tell my domain shader to not displace those vertices at all; it wouldn't be overly pretty but at least the mesh should be gapless and it hopefully wouldn't be too noticeable I reasoned. Of course this didn't work out very well (the whole subdivision patches generated from such overlapping vertices had their displacement factors set to 0 creating quite obvious, large frames around right angles and such). What I'm wondering is basically if this is a reasonable approach to try to refine further or if there are other ways to go about it that may be better? The only article on the topic I've managed to find mostly went on about the exquisitness of Bezier curves but didn't really seem to come to any conclusions (although maybe those would've been obvious to anyone having the required math skills).

Thankful for any pointers on this, the more I try to force this, the more it feels like I'm probably missing something.

 

As for my implementation of the tessellation, I've mostly based it around what is described in chapter 18.7 and 18.8 of Introduction to 3D Game Programming With DirectX 11 (http://www.amazon.com/Introduction-3D-Game-Programming-DirectX/dp/1936420228).

Share this post


Link to post
Share on other sites

Thanks MJP.

I don't suppose there's any video / audio recording of the presentation using those slides available somewhere, possibly for a nominal fee?

 

On a more off-topic note, I recognize that avatar of yours but have been unable to remember the name of the show (or was it possibly a book?) it featured in, care to enlighten me? 

Share this post


Link to post
Share on other sites

I think i have implemented index buffer for PN-AEN as described in that pdf document, but once i finished i saw that patch funcion HS_Constant is missing from shader that they have provided in appendix. sad.png

Copy-paste error. laugh.png

 

 


D3D11 ERROR: ID3D11DeviceContext::DrawIndexed: Mismatched topology. Current Hull Shader expects input Control Point count of 3, but Input Assembler topology defines a patch list with 9 Control Points per patch. [ EXECUTION ERROR #2097222: DEVICE_DRAW_HULL_SHADER_INPUT_TOPOLOGY_MISMATCH]

 

Damn. Why did they provide code that does not work.

 

edit 6841:

I get shader to work this time but i am not sure if i have indices right as i see no displacement.

normal.jpg

 

if i add displacement along normal:

... // domain shader
float3 n = mul(f3Normal, (float3x3)g_f4x4WorldView);
    n = normalize(n);
    f3EyePosition += n * 0.02f;

disp.jpg

CRACKS!

void Test::calcPNAENIndices(const std::vector<USHORT>& ind, const std::vector<VERTEX>& verts, std::vector<USHORT>& out)
{
    out.resize(ind.size() * 3);
    for (std::size_t i = 0; i < ind.size(); i += 3)
    {
        out[3 * i + 0] = ind[i + 0];
        out[3 * i + 1] = ind[i + 1];
        out[3 * i + 2] = ind[i + 2];

        out[3 * i + 3] = ind[i + 0];
        out[3 * i + 4] = ind[i + 1];
        out[3 * i + 5] = ind[i + 1];

        out[3 * i + 6] = ind[i + 2];
        out[3 * i + 7] = ind[i + 2];
        out[3 * i + 8] = ind[i + 0];
    }

    struct Edge
    {
        float3 p[2];
        USHORT inx[2];

        bool operator == (const Edge& o) const
        {
            if (inx[0] == o.inx[0] && inx[1] == o.inx[1])
                return true;

            if (Equal(p[0].x, o.p[0].x) && Equal(p[0].y, o.p[0].y) && Equal(p[0].z, o.p[0].z))
            {
                if (Equal(p[1].x, o.p[1].x) && Equal(p[1].y, o.p[1].y) && Equal(p[1].z, o.p[1].z))
                {
                    return true;
                }
            }

            return false;
        }
    };
 
    // reverse edges
    std::vector<Edge> edges;
    edges.resize(ind.size());
    for (std::size_t i = 0; i < ind.size(); i += 3)
    {
        edges[i + 0].p[1]   = verts[ind[i + 0]].pos;
        edges[i + 0].p[0]   = verts[ind[i + 1]].pos;
        edges[i + 0].inx[1] = ind[i + 0];
        edges[i + 0].inx[0] = ind[i + 1];

        edges[i + 1].p[1]   = verts[ind[i + 1]].pos;
        edges[i + 1].p[0]   = verts[ind[i + 2]].pos;
        edges[i + 1].inx[1] = ind[i + 1];
        edges[i + 1].inx[0] = ind[i + 2];

        edges[i + 2].p[1]   = verts[ind[i + 2]].pos;
        edges[i + 2].p[0]   = verts[ind[i + 0]].pos;
        edges[i + 2].inx[1] = ind[i + 2];
        edges[i + 2].inx[0] = ind[i + 0];
    }
 
    // compare
    for (std::size_t i = 0, j = 0; i < out.size(); i += 9, j += 3)
    {
        Edge e;

        // edge 0
        e.p[0]   = verts[out[i + 0]].pos;
        e.p[1]   = verts[out[i + 1]].pos;
        e.inx[0] = out[i + 0];
        e.inx[1] = out[i + 1];

        if (e == edges[j + 0])
        {
            out[i + 0] = edges[j + 0].inx[0];
            out[i + 1] = edges[j + 0].inx[1];
        }

        // edge 1
        e.p[0]   = verts[out[i + 1]].pos;
        e.p[1]   = verts[out[i + 2]].pos;
        e.inx[0] = out[i + 1];
        e.inx[1] = out[i + 2];

        if (e == edges[j + 0])
        {
            out[i + 1] = edges[j + 0].inx[0];
            out[i + 2] = edges[j + 0].inx[1];
        }

        // edge 2
        e.p[0]   = verts[out[i + 2]].pos;
        e.p[1]   = verts[out[i + 3]].pos;
        e.inx[0] = out[i + 2];
        e.inx[1] = out[i + 3];

        if (e == edges[j + 0])
        {
            out[i + 2] = edges[j + 0].inx[0];
            out[i + 3] = edges[j + 0].inx[1];
        }

        // edge 3
        e.p[0] = verts[out[i + 3]].pos;
        e.p[1] = verts[out[i + 4]].pos;
        e.inx[0] = out[i + 3];
        e.inx[1] = out[i + 4];

        if (e == edges[j + 1])
        {
            out[i + 3] = edges[j + 1].inx[0];
            out[i + 4] = edges[j + 1].inx[1];
        }

        // edge 4
        e.p[0] = verts[out[i + 4]].pos;
        e.p[1] = verts[out[i + 5]].pos;
        e.inx[0] = out[i + 4];
        e.inx[1] = out[i + 5];

        if (e == edges[j + 1])
        {
            out[i + 4] = edges[j + 1].inx[0];
            out[i + 5] = edges[j + 1].inx[1];
        }

        // edge 5
        e.p[0] = verts[out[i + 5]].pos;
        e.p[1] = verts[out[i + 6]].pos;
        e.inx[0] = out[i + 5];
        e.inx[1] = out[i + 6];

        if (e == edges[j + 1])
        {
            out[i + 5] = edges[j + 1].inx[0];
            out[i + 6] = edges[j + 1].inx[1];
        }

        // edge 6
        e.p[0] = verts[out[i + 6]].pos;
        e.p[1] = verts[out[i + 7]].pos;
        e.inx[0] = out[i + 6];
        e.inx[1] = out[i + 7];

        if (e == edges[j + 2])
        {
            out[i + 6] = edges[j + 2].inx[0];
            out[i + 7] = edges[j + 2].inx[1];
        }

        // edge 7
        e.p[0]   = verts[out[i + 7]].pos;
        e.p[1]   = verts[out[i + 8]].pos;
        e.inx[0] = out[i + 7];
        e.inx[1] = out[i + 8];

        if (e == edges[j + 2])
        {
            out[i + 7] = edges[j + 2].inx[0];
            out[i + 8] = edges[j + 2].inx[1];
        }

        // edge 8
        e.p[0] = verts[out[i + 8]].pos;
        e.p[1] = verts[out[i + 0]].pos;
        e.inx[0] = out[i + 8];
        e.inx[1] = out[i + 0];

        if (e == edges[j + 2])
        {
            out[i + 8] = edges[j + 2].inx[0];
            out[i + 0] = edges[j + 2].inx[1];
        }
    }
}
Edited by belfegor

Share this post


Link to post
Share on other sites


I don't suppose there's any video / audio recording of the presentation using those slides available somewhere, possibly for a nominal fee?

 

Not that I know of, sorry.

 


On a more off-topic note, I recognize that avatar of yours but have been unable to remember the name of the show (or was it possibly a book?) it featured in, care to enlighten me?

 

It's Rocko!

Share this post


Link to post
Share on other sites

I think i had some mistakes in code, having hard time to understand what exactly i need to do.

Here is corrected version but still wrong results. sad.png

void Test::calcPNAENIndices(const std::vector<USHORT>& ind, const std::vector<VERTEX>& verts, std::vector<USHORT>& out)
{
    struct Edge
    {
        float3 p[2];
        USHORT inx[2];

        bool operator == (const Edge& o) const
        {
            if (inx[0] == o.inx[0] && inx[1] == o.inx[1])
                return true;

            if (Equal(p[0].x, o.p[0].x) && Equal(p[0].y, o.p[0].y) && Equal(p[0].z, o.p[0].z))
            {
                if (Equal(p[1].x, o.p[1].x) && Equal(p[1].y, o.p[1].y) && Equal(p[1].z, o.p[1].z))
                {
                    return true;
                }
            }

            return false;
        }
    };

    std::vector<Edge> edges(ind.size());
    out.resize(ind.size() * 3);

    for (std::size_t i = 0; i < ind.size(); i += 3)
    {
        // initial values
        out[3 * i + 0] = ind[i + 0];
        out[3 * i + 1] = ind[i + 1];
        out[3 * i + 2] = ind[i + 2];

        out[3 * i + 3] = ind[i + 0];
        out[3 * i + 4] = ind[i + 1];
        out[3 * i + 5] = ind[i + 1];

        out[3 * i + 6] = ind[i + 2];
        out[3 * i + 7] = ind[i + 2];
        out[3 * i + 8] = ind[i + 0];

        // store reversed
        edges[i + 0].p[1]   = verts[ind[i + 0]].pos;
        edges[i + 0].p[0]   = verts[ind[i + 1]].pos;
        edges[i + 0].inx[1] = ind[i + 0];
        edges[i + 0].inx[0] = ind[i + 1];

        edges[i + 1].p[1]   = verts[ind[i + 1]].pos;
        edges[i + 1].p[0]   = verts[ind[i + 2]].pos;
        edges[i + 1].inx[1] = ind[i + 1];
        edges[i + 1].inx[0] = ind[i + 2];

        edges[i + 2].p[1]   = verts[ind[i + 2]].pos;
        edges[i + 2].p[0]   = verts[ind[i + 0]].pos;
        edges[i + 2].inx[1] = ind[i + 2];
        edges[i + 2].inx[0] = ind[i + 0];
    }

    for (std::size_t i = 0; i < out.size(); i += 9)
    {
        // i think i should skip first 3 indices as they point to triangle and i need to check edges
        for (std::size_t j = 3; j < 9; ++j)
        {
            std::size_t first  = j;
            std::size_t second = j + 1;
            if (second == 9)
                second = 3;

            Edge e;
            e.p[0]   = verts[out[i + first]].pos;
            e.p[1]   = verts[out[i + second]].pos;
            e.inx[0] = out[i + first];
            e.inx[1] = out[i + second];

            for (std::size_t k = 0; k < edges.size(); ++k)
            {
                if (e == edges[k])
                {
                    out[i + first]  = edges[k].inx[0];
                    out[i + second] = edges[k].inx[1];
                }
            }
        }
    }
}

Can someone take a look please and decipher instructions given for this to work:

 

 

1.  Create an output IB that is 3 times the size of input IB.

2.  For each input Triangle in IB, with indices i0, i1 and i2:
    a.  Write out an initial output entry of:  i0, i1, i2, i0, i1, i1, i2, i2, i0, which sets edges to

        initially be neighbors of themselves. This would produce identical results to PN
        Triangles.
    b.  Lookup the positions p0, p1, and p2, using i0, i1 and i2 to perform a lookup for
        position of the associated vertex in VB.
    c.  Define 3 Edges, which consist of the two indices and two positions that make up
        the corresponding Edge. An Edge should consist of the origin index, the
        destination index, the origin position and the destination position.
    d.  For each edge, store the reverse of that edge in an easily searchable data structure
        for the next step. The reference implementation uses an stdext::hash_map<Edge,
        Edge> for this purpose. Reverse simply flips the sense of the edge (originating at
        the destination position and index and heading to the origin position and index).

 

3.  Walk the output index buffer (OB) constructed in step 2. For each patch of 9 indices:
    a.  For each Edge in the current Patch, perform a lookup into Edge->Edge mapping
        created in step 2d.
    b.  If found, replace the current indices with the indices found in the map. Note that
        two edges should be considered matching if their "from" and "to" indices match,
        OR if their "from" and "to" positions match.
    c.  If not, continue to use the existing indices.

Upon completion of this algorithm, a buffer suitable for usage with PN-AEN will be
available.

 

Edited by belfegor

Share this post


Link to post
Share on other sites

That is that. Thank you very much. smile.png

 

With "initial" IB (notice cracks):

crack.jpg

 

PNAEN

good.jpg

 

Now i would like to displace position along normal using heightmap but i dont know how to get average normal so i can pass it to domain shader. Any pointers?

Share this post


Link to post
Share on other sites
Now the real challenge starts. Here the sample goes the easy way and just passes the normals from the inner triangle and interpolates linearly. "Correct" PN triangles uses a quadratic bezier patch for the normals, examples of which you can find in the (June 2010 SDK) samples or in the Hieroglyph 3 engine (there's also full chapter in the Practical Rendering book).
There's also a displacement tesselation sample using decals in the SDK which might be worth a peek.

Either way it sounds complex as the presentation MJP linked to shows. Can't help you any further now, I haven't done displacement mapping (not counting terrain). For a simple start maybe you can get away averaging normals at corners (not using a bezier quad patch and taking averaged normals from all adjacent triangles). But that's really just an idea.

Anyway, looks like it also needs special care on the content creation side (citing above presentation).

Speaking of which (and out of curiosity): Why does that pig/boar generate cracks ? I wouldn't expect them on organic surfaces (the initial index buffer results in usual PN triangles). You got a link to that mesh ?

Share this post


Link to post
Share on other sites


It's Rocko!

Ah yes, haha, remember watching that in the mid-late 90's. Maybe worth a rewatch now that I'm old enough to actually understand it better than back then...

 

As for the displacement mapping I eventually settled on a rather simple approach for use with pre-existing triangle list meshes (it is by no means perfect but it seems that the "real" solution is really just having the 3D / texturing artist(s) being aware of your intents and have them author appropriately mapped bump maps and ensure corners actually are rounded, albeit with extremely short edges between corner vertices, as seemed to be the main point that the chapter in Zink, Pettineo & Hoxley (2011) as referred by unbird in the last post). Basically I average all disjoint vertex normals and create a 6-control patch index buffer that holds the initial triangle in the first 3 indices and then any "dominant" vertices sharing the position of these original vertices in the last 3 indices. The dominant vertices are arbitrarily chosen as the first found in the vertex data for a given position and are set to the same vertex as in the first three indices if there are no shared vertices at a given position. The dominant vertex is used to ensure that all overlapping vertices will sample the same value from the displacement map, while still allowing them to sample other textures by their UV coords.

Not perfect but good enough for some general-purpose examples.

As for actually employing this kind of displacement mapping in a more professional game / visual demo / what-have-you, the artist should ensure that there are *no* disjoint vertices (as said, such vertices can be moved slightly apart and be connected by a short edge, allowing the corners to appear mostly sharp) and a secondary set of texture coordinates should be provided for the displacement map / alternatively the displacement map should be handcrafted such that when displaced (with a reasonable strength / height factor), faces won't extrude through each others.

Of course there exists more elaborate ways as described in the previous posts, I just thought I would share this if anyone found it interesting, After tracking down a copy of that practical rendering book just for that tessellation chapter, it didn't say that much more about this particular problem, so might as well save others the trouble.

Share this post


Link to post
Share on other sites

Hi,

i'm struggling with PN AEN now for two weeks and I'm not sure where my mistake could be... I'm pretty sure, the indices are correct, since I tried out unbirds exmaple and get the same indices. However my result looks like this:screenshot.png

 

I get the same result as with PN Triangles, no difference.

 

My TCS is:

#version 420
layout (vertices = 9) out;
in vec3 vp[];
in vec2 UV[];
in vec3 vn[];
out vec3 vpos[];
out vec2 outuv[];
out vec3 vnor[];
uniform vec3  tessLevelOuter;
uniform float tessLevelInner;

patch out Patch
{
    vec3 b210;
    vec3 b120;
    vec3 b021;
    vec3 b012;
    vec3 b102;
    vec3 b201;
    vec3 b111;
    vec3 n110;
    vec3 n011;
    vec3 n101;
	vec2 t110;
    vec2 t011;
    vec2 t101;
} OutPatch;

#define b300 vp[0]
#define b030 vp[1]
#define b003 vp[2]

#define n200 vn[0]
#define n020 vn[1]
#define n002 vn[2]

#define t200 UV[0]
#define t020 UV[1]
#define t002 UV[2]

void main()
{

	if( gl_InvocationID == 0 )
		{
			gl_TessLevelOuter[0] = tessLevelOuter[0];
			gl_TessLevelOuter[1] = tessLevelOuter[1];
			gl_TessLevelOuter[2] = tessLevelOuter[2];
			gl_TessLevelInner[0] = tessLevelInner;
		}
		vpos[gl_InvocationID]=vp[gl_InvocationID];
		vnor[gl_InvocationID]=vn[gl_InvocationID];
		outuv[gl_InvocationID]=UV[gl_InvocationID];
		
		if( gl_InvocationID == 0 )
    {

						 
						 
		OutPatch.b210 = (2.0 * b300 + b030 - dot( b030 - b300, n200 ) * n200 + 2.0 * vp[3] + vp[4] - dot( vp[4] - vp[3], vn[3] ) * vn[3] ) / 6.0;
        OutPatch.b120 = (2.0 * b030 + b300 - dot( b300 - b030, n020 ) * n020 +2.0 * vp[4] + vp[3] -dot( vp[3] - vp[4], vn[4] ) * vn[4] ) / 6.0;
        OutPatch.b021 = (2.0 * b030 + b003 - dot( b003 - b030, n020 ) * n020 +2.0 * vp[5] + vp[6] - dot( vp[6] - vp[5], vn[5] ) * vn[5] ) / 6.0;
        OutPatch.b012 = (2.0 * b003 + b030 - dot( b030 - b003, n002 ) * n002 +2.0 * vp[6] + vp[5] - dot( vp[5] - vp[6], vn[6] ) * vn[6] ) / 6.0;
        OutPatch.b102 = (2.0 * b003 + b300 - dot( b300 - b003, n002 ) * n002 +2.0 * vp[7] + vp[8] -dot( vp[8] - vp[7], vn[7] ) * vn[7] ) / 6.0;
        OutPatch.b201 = (2.0 * b300 + b003 - dot( b003 - b300, n200 ) * n200 + 2.0 * vp[8] + vp[7] -dot( vp[7] - vp[8], vn[8] ) * vn[8] ) / 6.0;
				 
		OutPatch.b111 = (OutPatch.b210 + OutPatch.b120 + OutPatch.b021 +
                         OutPatch.b012 + OutPatch.b102 + OutPatch.b201) / 4.0 - (b300 + b030 + b003) / 6.0;				 
						 
						 
						 

        const vec3 d0 = b030 - b300;
        const vec3 d1 = b003 - b030;
        const vec3 d2 = b300 - b003;
        const vec3 n0 = n020 + n200;
        const vec3 n1 = n002 + n020;
        const vec3 n2 = n200 + n002;
        const vec3 v0 = (2.0 * dot( d0, n0 ) / dot( d0, d0 )) * d0;
        const vec3 v1 = (2.0 * dot( d1, n1 ) / dot( d1, d1 )) * d1;
        const vec3 v2 = (2.0 * dot( d2, n2 ) / dot( d2, d2 )) * d2;
        OutPatch.n110 = normalize( n0 - v0 );
        OutPatch.n011 = normalize( n1 - v1 );
        OutPatch.n101 = normalize( n2 - v2 );
     }
}

Any ideas what I'm doing wrong?

Share this post


Link to post
Share on other sites

Off the top of my head each individual face of your cube is tessellated and then displaced.

You need to ensure that the edge vertices are shared in each (subdivided) side-face or else these seams will occur since all vertices on the top face are displaced only along the up axis and all vertices of the front face are displaced only along the depth axis.

A simple solution is to displace along the vertex normals and ensure that whereever you have overlapping vertices (such as at the corners of a cube) you set the normal of all such vertices to the average of all "actual" vertex normals at that position. This will make the edges a bit more bulky but keep the faces connected.

 

My previous post in this thread (just above yours) describes how I solved this in a relatively simple way in more detail.

Share this post


Link to post
Share on other sites

I'm not doing any Displayment mapping so far.

My TES looks like this:

	vec3 vp=vpos[0] * w * w * w +
                          b030 * u * u * u +
                          b003 * v * v * v + 
						  InPatch.b210 * 3.0 * w * w * u + 
                          InPatch.b120 * 3.0 * w * u * u +
                          InPatch.b201 * 3.0 * w * w * v +
                          InPatch.b021 * 3.0 * u * u * v +
                          InPatch.b102 * 3.0 * w * v * v +
                          InPatch.b012 * 3.0 * u * v * v +
                          InPatch.b111 * 6.0 * w * u * v; 

	vec3 vn= normalize( n200 * w * w + n020 * u * u + n002 * v * v +
                             InPatch.n110 * w * u + InPatch.n011 * u * v + InPatch.n101 * w * v );

About the screenshot above: I'm not applying any displacement yet. The cube simply cracks by smoothing the surface. As far as I understood this is a known issue with PN, but PNAEN shouldn't have this problem.

I know that you meant dominant UV to solve displacement map cracking, but I still have PN cracking with PNAEN.

Share this post


Link to post
Share on other sites

I just fed a standard cube to my shader and it stays a cube with flat sides, no matter the tesselation factors (both PN and PNAEN). Makes sense if you use the axis-aligned normals and not averaged, like suggested several times (the bezier surfaces are indeed flat if all normals are equal). Maybe you stumbled across the same bug I did (vertex shader, not normalized normals therefore wrong control point calculation).

 

Edit: How does your PN behave for a nice model like a sphere ?

Edited by unbird

Share this post


Link to post
Share on other sites

I just fed a standard cube to my shader and it stays a cube with flat sides, no matter the tesselation factors (both PN and PNAEN). Makes sense if you use the axis-aligned normals and not averaged, like suggested several times (the bezier surfaces are indeed flat if all normals are equal).

At first I used a standard cube generated by Maya and the cube stayed flat exactly as you said. Then I changed the normals of the cube like this:

cubes.png

PN got the cracks as expected, but PN doesn't look smooth like in the graphic above and looks 100% like PN.

 Maybe you stumbled across the same bug I did (vertex shader, not normalized normals therefore wrong control point calculation).

I tried that, adding normalize in the Vertexshader, but no difference, still cracks.

 

Edit: How does your PN behave for a nice model like a sphere ?

Looks smooth, no cracks. No difference to flat tessealtion or PNAEN.

Share this post


Link to post
Share on other sites
OK, now I made you move in circles, sorry about that. I can only guess: Either your re-indexing goes wrong or the shader has a bug (well, not very helpful either, I know).
 
You might show your re-indexing code, maybe someone spots something. Also, my test case is probably too small. Can you put that cube up for download (preferably in a common exchange format like Wavefront) ? I'll run it through my reindexer and give you the results.
 
As for the shader: Cannot help you there, I'm really just a D3D guy wink.png. You might have more luck in the OpenGL forum. Though if you carefully translated it from the paper it should work (apart from the vertex shader bug).

Share this post


Link to post
Share on other sites

You might show your re-indexing code, maybe someone spots something. Also, my test case is probably too small.

 

Here is my indexing code:

void indexPNAEN(std::vector<unsigned short> indices, std::vector<glm::vec3> & in_vertices, std::vector<unsigned short> & out_indicesaen)
{
	out_indicesaen.resize(indices.size()*3); //step 1 Create an output IB that is 3 times the size of input IB.
	//step 2 c define edge
	struct Edge
	{
		glm::vec3 p[2];
		short ind[2];

		bool operator == (const Edge& o) const
		{
			if (ind[0] == o.ind[0] && ind[1] == o.ind[1])
				return true;

			if (p[0] == o.p[0] && p[1] == o.p[1])
			{

					return true;
			}

			return false;
		}

	/*	bool operator != (const Edge& o) const{
			return !(*this == o);
		}*/


		Edge reverse()
		{
			Edge returnedge;
			returnedge.p[0] = p[1];
			returnedge.p[1] = p[0];
			returnedge.ind[0] = ind[1];
			returnedge.ind[1] = ind[0];

			return returnedge;	
		}
	};

	struct KeyHasher
	{
		std::size_t operator()(const Edge& k) const
		{
			using boost::hash_value;
			using boost::hash_combine;

			// Start with a hash value of 0    .
			std::size_t seed = 0;

			// Modify 'seed' by XORing and bit-shifting in
			// one member of 'Key' after the other:
			for (int hashi = 0; hashi < 2; hashi++)
			{
				hash_combine(seed, hash_value(k.p[hashi].x));
				hash_combine(seed, hash_value(k.p[hashi].y));
				hash_combine(seed, hash_value(k.p[hashi].z));
				hash_combine(seed, hash_value(k.ind[hashi]));
			}

			// Return the result.
			return seed;
		}
	};



	std::unordered_map<Edge, Edge,KeyHasher> edges(indices.size()); 

	//step 2
	for (int i = 0; i < indices.size(); i += 3) //For each input Triangle in IB,
	{
		out_indicesaen[3 * i] = indices[i]; //i0
		out_indicesaen[3 * i + 1] = indices[i+1]; //i1
		out_indicesaen[3 * i + 2] = indices[i + 2]; //i2

		out_indicesaen[3 * i + 3] = indices[i]; //i0
		out_indicesaen[3 * i + 4] = indices[i + 1]; //i1
		out_indicesaen[3 * i + 5] = indices[i + 1]; //i1

		out_indicesaen[3 * i + 6] = indices[i+2]; //i2
		out_indicesaen[3 * i + 7] = indices[i + 2]; //i2
		out_indicesaen[3 * i + 8] = indices[i]; //i0

		//2b and 2d

		Edge edge0;
		edge0.p[0] = in_vertices[indices[i + 1]];
		edge0.p[1] = in_vertices[indices[i]];
		edge0.ind[1] = indices[indices[i]];
		edge0.ind[0] = indices[indices[i + 1]];
		edges.emplace(edge0.reverse(), edge0);

		Edge edge1;
		edge1.p[0] = in_vertices[indices[i + 2]];
		edge1.p[1] = in_vertices[indices[i + 1]];
		edge1.ind[1] = indices[indices[i + 1]];
		edge1.ind[0] = indices[indices[i + 2]];
		edges.emplace(edge1.reverse(), edge1);

		Edge edge2;
		edge2.p[0] = in_vertices[indices[i]];
		edge2.p[1] = in_vertices[indices[i + 2]];
		edge2.ind[1] = indices[indices[i + 2]];
		edge2.ind[0] = indices[indices[i]];
		edges.emplace(edge2.reverse(), edge2);
	}


	//step 3 Walk the output index buffer (OB) constructed in step 2. For each patch of 9 indices:
	for (int i = 3; i < out_indicesaen.size(); i += 9)
	{

		//3a For each Edge in the current Patch, perform a lookup into Edge->Edge mapping created in step 2d.
		for (int k = 0; k < 6; k += 2)
		{
			int i0 = out_indicesaen[i + k];
			int i1 = out_indicesaen[i + k + 1];
			Edge temp;
			temp.ind[0] = i1;
			temp.ind[1] = i0;
			temp.p[0] = in_vertices[i1];
			temp.p[1] = in_vertices[i0];

			auto foundIt = edges.find(temp);
			if (foundIt!=edges.end()) //look up in edge vector
			{
				const Edge& second = foundIt->second;
				out_indicesaen[i + k] = second.ind[1];
				out_indicesaen[i + k + 1] = second.ind[0];
				
			}

		}

	}

}

Can you put that cube up for download (preferably in a common exchange format like Wavefront) ? I'll run it through my reindexer and give you the results.

Oh, thank you! IThe cube can be downloaded here. (I'm not permitted to attach an .obj file blink.png ). I'm starting to get frustrated with PNAEN...I can't find the bug for two weeks now. It's part of my master thesis.

Share this post


Link to post
Share on other sites

Ok. Had to adjust the import (Assimp, join identical vertices), so only 24 instead of 36 distinct vertices. In any case, I included a dump of the vertices, so you could reconstruct the mesh programmatically. It's a bloated html-log (and the indices are dumped as tables and "raw" for easier copying).

 

From a quick glance, I think your reindexer is fine. 

 

Edit: Wait, try something first. Hash only the positions (x,y,z), the map will fail otherwise (at least it did for my C# dictionary).

Share this post


Link to post
Share on other sites

My reindexer is indeed wrong, I get totally different indices. I copied your indices manually into the vector and the cracks are gone! The shader works!

I'll try to hash only the positions.  I hope that's the bug! Thank you very much for your help, the html log is awesome!

 

Edit: I found out that my edge Map seems to be already wrong. i have only 32 entried while your map has 36 entries.

Edited by windschuetze

Share this post


Link to post
Share on other sites

I solved my problem! I had serveal bugs:

  • filling the edge map: indices[indices[i]] instead of indices[i]...

  • I removed the hashing of the index and it worked!

Thank you very much, I couldn't have done it without your help!

 

Now my cube looks smooth and crackfree:

screenshot16111.png

Edited by windschuetze

Share this post


Link to post
Share on other sites

 


It's Rocko!

Ah yes, haha, remember watching that in the mid-late 90's. Maybe worth a rewatch now that I'm old enough to actually understand it better than back then...

 

As for the displacement mapping I eventually settled on a rather simple approach for use with pre-existing triangle list meshes (it is by no means perfect but it seems that the "real" solution is really just having the 3D / texturing artist(s) being aware of your intents and have them author appropriately mapped bump maps and ensure corners actually are rounded, albeit with extremely short edges between corner vertices, as seemed to be the main point that the chapter in Zink, Pettineo & Hoxley (2011) as referred by unbird in the last post). Basically I average all disjoint vertex normals and create a 6-control patch index buffer that holds the initial triangle in the first 3 indices and then any "dominant" vertices sharing the position of these original vertices in the last 3 indices. The dominant vertices are arbitrarily chosen as the first found in the vertex data for a given position and are set to the same vertex as in the first three indices if there are no shared vertices at a given position. The dominant vertex is used to ensure that all overlapping vertices will sample the same value from the displacement map, while still allowing them to sample other textures by their UV coords.

Not perfect but good enough for some general-purpose examples.

As for actually employing this kind of displacement mapping in a more professional game / visual demo / what-have-you, the artist should ensure that there are *no* disjoint vertices (as said, such vertices can be moved slightly apart and be connected by a short edge, allowing the corners to appear mostly sharp) and a secondary set of texture coordinates should be provided for the displacement map / alternatively the displacement map should be handcrafted such that when displaced (with a reasonable strength / height factor), faces won't extrude through each others.

Of course there exists more elaborate ways as described in the previous posts, I just thought I would share this if anyone found it interesting, After tracking down a copy of that practical rendering book just for that tessellation chapter, it didn't say that much more about this particular problem, so might as well save others the trouble.

 

Hi,

im trying to implement the dominandt data, but I'm struggling again...I'm not sure how to understand the  GDC 2012 slides, when i'm generating the index buffer:

? Can use AEN edge data as well

vs:

? All shared vertices must have the same dominant data
? Both edge vertices must be from the same primitive

 

I tried with the PN AEN Indexbuffer adding 3 arbitrarly chosen vertices, but it didn't worked.

Then I found on the Maya Homepage a despription of a PNAEN 18 implementation using dominant data with a 18 size index buffer and I tried to adapt it. But the result isn't good either, the UVs doesn't look smoother:

 

screenshot0312.png

 

My index buffer now looks like this for unbirds example:

0,1,2,0,1,4,3,2,0,0,1,4,3,2,0,0,4,2,3,4,5,2,1,4,5,5,3,4,3,4,5,5,3,4,4,5

(See also xls Attachment for details).

I'm pretty sure that the shader interpolates right:

			float 
		uCorner =  (u == 1 ? 1:0),
		vCorner =  (v == 1 ? 1:0),
		wCorner =  (w == 1 ? 1:0),
		uEdge =    (u == 0 && (v * w)!=0 ? 1:0),
		vEdge =    (v == 0 && (u * w)!=0 ? 1:0),
		wEdge =    (w == 0 && (u * v)!=0 ? 1:0),
		interior = (u * v * w)!=0 ? 1:0;

		vec2 displaceCoord= 	uCorner*InPatch.domVert[0]
							+vCorner*InPatch.domVert[1]
							+wCorner*InPatch.domVert[2]
							+uEdge*mix(InPatch.domEdge0[1],InPatch.domEdge1[1],v)
							+vEdge*mix(InPatch.domEdge0[2],InPatch.domEdge1[2],w)
							+wEdge*mix(InPatch.domEdge0[0],InPatch.domEdge1[0],u)
							+interior*UV;

Any clue?

Share this post


Link to post
Share on other sites

This topic is 1110 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

  • Forum Statistics

    • Total Topics
      628740
    • Total Posts
      2984472
  • Similar Content

    • By GreenGodDiary
      Having some issues with a geometry shader in a very basic DX app.
      We have an assignment where we are supposed to render a rotating textured quad, and in the geometry shader duplicate this quad and offset it by its normal. Very basic stuff essentially.
      My issue is that the duplicated quad, when rendered in front of the original quad, seems to fail the Z test and thus the original quad is rendered on top of it.
      Whats even weirder is that this only happens for one of the triangles in the duplicated quad, against one of the original quads triangles.

      Here's a video to show you what happens: Video (ignore the stretched textures)

      Here's my GS: (VS is simple passthrough shader and PS is just as basic)
      struct VS_OUT { float4 Pos : SV_POSITION; float2 UV : TEXCOORD; }; struct VS_IN { float4 Pos : POSITION; float2 UV : TEXCOORD; }; cbuffer cbPerObject : register(b0) { float4x4 WVP; }; [maxvertexcount(6)] void main( triangle VS_IN input[3], inout TriangleStream< VS_OUT > output ) { //Calculate normal float4 faceEdgeA = input[1].Pos - input[0].Pos; float4 faceEdgeB = input[2].Pos - input[0].Pos; float3 faceNormal = normalize(cross(faceEdgeA.xyz, faceEdgeB.xyz)); //Input triangle, transformed for (uint i = 0; i < 3; i++) { VS_OUT element; VS_IN vert = input[i]; element.Pos = mul(vert.Pos, WVP); element.UV = vert.UV; output.Append(element); } output.RestartStrip(); for (uint j = 0; j < 3; j++) { VS_OUT element; VS_IN vert = input[j]; element.Pos = mul(vert.Pos + float4(faceNormal, 0.0f), WVP); element.Pos.xyz; element.UV = vert.UV; output.Append(element); } }  
      I havent used geometry shaders much so im not 100% on what happens behind the scenes.
      Any tips appreciated! 
    • By mister345
      Hi, I'm building a game engine using DirectX11 in c++.
      I need a basic physics engine to handle collisions and motion, and no time to write my own.
      What is the easiest solution for this? Bullet and PhysX both seem too complicated and would still require writing my own wrapper classes, it seems. 
      I found this thing called PAL - physics abstraction layer that can support bullet, physx, etc, but it's so old and no info on how to download or install it.
      The simpler the better. Please let me know, thanks!
    • By Hexaa
      I try to draw lines with different thicknesses using the geometry shader approach from here:
      https://forum.libcinder.org/topic/smooth-thick-lines-using-geometry-shader
      It seems to work great on my development machine (some Intel HD). However, if I try it on my target (Nvidia NVS 300, yes it's old) I get different results. See the attached images. There
      seem to be gaps in my sine signal that the NVS 300 device creates, the intel does what I want and expect in the other picture.
      It's a shame, because I just can't figure out why. I expect it to be the same. I get no Error in the debug output, with enabled native debugging. I disabled culling with CullMode.None. Could it be some z-fighting? I have little clue about it but I tested to play around with the RasterizerStateDescription and DepthBias properties with no success, no change at all. Maybe I miss something there?
      I develop the application with SharpDX btw.
      Any clues or help is very welcome
       


    • By Beny Benz
      Hi,
      I'm currently trying to write a shader which shoud compute a fast fourier transform of some data, manipulating the transformed data, do an inverse FFT an then displaying the result as vertex offset and color. I use Unity3d and HLSL as shader language. One of the main problems is that the data should not be passed from CPU to GPU for every frame if possible. My original plan was to use a vertex shader and do the fft there, but I fail to find out how to store changing data betwen shader calls/passes. I found a technique called ping-ponging which seems to be based on writing and exchangeing render targets, but I couldn't find an example for HLSL as a vertex shader yet.
      I found https://social.msdn.microsoft.com/Forums/en-US/c79a3701-d028-41d9-ad74-a2b3b3958383/how-to-render-to-multiple-render-targets-in-hlsl?forum=xnaframework
      which seem to use COLOR0 and COLOR1 as such render targets.
      Is it even possible to do such calculations on the gpu only? (/in this shader stage?, because I need the result of the calculation to modify the vertex offsets there)
      I also saw the use of compute shaders in simmilar projects (ocean wave simulation), do they realy copy data between CPU / GPU for every frame?
      How does this ping-ponging / rendertarget switching technique work in HLSL?
      Have you seen an example of usage?
      Any answer would be helpfull.
      Thank you
      appswert
    • By ADDMX
      Hi
      Just a simple question about compute shaders (CS5, DX11).
      Do the atomic operations (InterlockedAdd in my case) should work without any issues on RWByteAddressBuffer and be globaly coherent ?
      I'v come back from CUDA world and commited fairly simple kernel that does some job, the pseudo-code is as follows:
      (both kernels use that same RWByteAddressBuffer)
      first kernel does some job and sets Result[0] = 0;
      (using Result.Store(0, 0))
      I'v checked with debugger, and indeed the value stored at dword 0 is 0
      now my second kernel
      RWByteAddressBuffer Result;  [numthreads(8, 8, 8)] void main() {     for (int i = 0; i < 5; i++)     {         uint4 v0 = DoSomeCalculations1();         uint4 v1 = DoSomeCalculations2();         uint4 v2 = DoSomeCalculations3();                  if (v0.w == 0 && v1.w == 0 && v2.w)             continue;         //    increment counter by 3, and get it previous value         // this should basically allocate space for 3 uint4 values in buffer         uint prev;         Result.InterlockedAdd(0, 3, prev);                  // this fills the buffer with 3 uint4 values (+1 is here as the first 16 bytes is occupied by DrawInstancedIndirect data)         Result.Store4((prev+0+1)*16, v0);         Result.Store4((prev+1+1)*16, v1);         Result.Store4((prev+2+1)*16, v2);     } } Now I invoke it with Dispatch(4,4,4)
      Now I use DrawInstancedIndirect to draw the buffer, but ocassionaly there is missed triangle here and there for a frame, as if the atomic counter does not work as expected
      do I need any additional synchronization there ?
      I'v tried 'AllMemoryBarrierWithGroupSync' at the end of kernel, but without effect.
      If I do not use atomic counter, and istead just output empty vertices (that will transform into degenerated triangles) the all is OK - as if I'm missing some form of synchronization, but I do not see such a thing in DX11.
      I'v tested on both old and new nvidia hardware (680M and 1080, the behaviour is that same).
       
  • Popular Now