Dimka!!! Написал:
-------------------------------------------------------
> В майе, синеме с этим нет проблем
Смотря что понимать под проблемами. Майя позволит тебе заэкструдить ежа, как тебе нравится, но нормаль-то у фейса всё равно одна. Если задумаешь сглаживать геометрию, где у смежных фейсов нормали в противоположные стороны, получишь в этом месте артефакты. А захочешь сконвертировать в сабдив — Майя расскажет тебе всё, что думает о non-manifold geometry.
В Синеме с HyperNURBS ситуация такая же.
-------------------------------------------------------
> В майе, синеме с этим нет проблем
Смотря что понимать под проблемами. Майя позволит тебе заэкструдить ежа, как тебе нравится, но нормаль-то у фейса всё равно одна. Если задумаешь сглаживать геометрию, где у смежных фейсов нормали в противоположные стороны, получишь в этом месте артефакты. А захочешь сконвертировать в сабдив — Майя расскажет тебе всё, что думает о non-manifold geometry.
В Синеме с HyperNURBS ситуация такая же.
Что неподдерживает?
Это правило построения полигонов вообще.
Почитай лучше тут про, то как устроены полигоны и чем они от меша отличаются:
Historically, 3D Studio DOS and then MAX were based around TriMeshes from the beginning.
The logic is that from a mathematical point of view, 3 points always define a unique plane, it is painfully easy to calculate the normal to that plane (by normalizing the cross product of two edges of the face), and thus it is very simple to render a triangle.
N-gons (where N is 3 or more) usually have to be broken down to triangles at a certain stage of the calculation in order to be able to calculate the normals and the shading of the surface.
The EPoly is not just another way of looking at things, it is based on a completely different internal data structure. Originally, it was developed by the makers of the Nendo and Mirai software (whatever their name is today).
The popular name of the technology is "Winged Edges". The implementation of this new data structure started during Max 3 development (but there was no UI exposure in 3), then the basic UI came in Max 4 and was extended by Laszlo Sebo's Meshtools, and matured in Max 5.
What is the difference?
In the case of TriMeshes, the basic building blocks of the geometry are vertices (stored in a vertex array) and faces (each one expressed as a Point3 value where .x, .y and .z components are indices pointing at one of the vertices in the vertex array).
Edges in this structure do not actually exist. Each face has obviously 3 sides defined by the line connecting vertices 1-2, 2-3 and 1-3, and there is a flag in the structure recording the visibility of that edge. BUT, where two faces share the same two vertices, they don't share a common edge! Each face has its own edge and when changing the visibility of the one, MAX has to make sure internally the opposite one of the neighbour face also gets changed... All in one, there are 3 times as many edges as faces in a TriMesh, since every face has 3 of them. But finding out which edge lies between two vertices or which faces are sharing the same vertices along an edge is rather slow as the whole mesh has to be traversed to find out. As the mesh density increases, this search time gets longer and longer.
In the case of EPoly though, the basic building blocks of the geometry are vertices and EDGES! Polygons are defined by the edges, not the other way round. This means that finding edge loops, splitting edges, cutting through hundreds of thousands of edges with a slicing plane etc. are very fast since the edges are a structural element of the geometry and VERY easy to locate, query and change. Also, two neighbour polygons sharing an edge actually use the SAME edge, so manipulating that edge alters both polygons with a single operation.
A drawback of this structure though is that by definition, an edge can be shared by 1 (open edge, like the outside edge of a 4x4 plane) or 2 (internal edge, like the inside edge of a plane) polygons ONLY! You cannot have 3 polygons sharing the same edge (the case is called a "rat's nest"), which explains why it is not possible to select an internal edge that already has 2 polygons and extrude a 3rd polygon by shift-dragging.
(If you convert to EMesh, shift+drag edge and convert back to EPoly, you will discover that the new poly is a "free-floating element" and not actually attached to the surface it was extruded from - it has no common edge and vertices with the original edge that was extruded).
Another drawback of EPoly is that the graphics pipeline that supplies the realtime renderer with triangles (can be SZB, OpenGL or Direct3D accelerator) has to do some extra work to convert the EPoly data to TriMesh to be displayed in the viewports. So while cutting through an EPoly mesh is much faster than with EMesh, the display is much slower. It also explains why adding a MeshSelect or TurnToMesh to the top of a non-animated EPoly makes it many times faster to redraw in the viewports...
Hope this helps!
Cheers,
Bobo (Frantic Films)
Взято с cgtalk.com
Это правило построения полигонов вообще.
Почитай лучше тут про, то как устроены полигоны и чем они от меша отличаются:
Historically, 3D Studio DOS and then MAX were based around TriMeshes from the beginning.
The logic is that from a mathematical point of view, 3 points always define a unique plane, it is painfully easy to calculate the normal to that plane (by normalizing the cross product of two edges of the face), and thus it is very simple to render a triangle.
N-gons (where N is 3 or more) usually have to be broken down to triangles at a certain stage of the calculation in order to be able to calculate the normals and the shading of the surface.
The EPoly is not just another way of looking at things, it is based on a completely different internal data structure. Originally, it was developed by the makers of the Nendo and Mirai software (whatever their name is today).
The popular name of the technology is "Winged Edges". The implementation of this new data structure started during Max 3 development (but there was no UI exposure in 3), then the basic UI came in Max 4 and was extended by Laszlo Sebo's Meshtools, and matured in Max 5.
What is the difference?
In the case of TriMeshes, the basic building blocks of the geometry are vertices (stored in a vertex array) and faces (each one expressed as a Point3 value where .x, .y and .z components are indices pointing at one of the vertices in the vertex array).
Edges in this structure do not actually exist. Each face has obviously 3 sides defined by the line connecting vertices 1-2, 2-3 and 1-3, and there is a flag in the structure recording the visibility of that edge. BUT, where two faces share the same two vertices, they don't share a common edge! Each face has its own edge and when changing the visibility of the one, MAX has to make sure internally the opposite one of the neighbour face also gets changed... All in one, there are 3 times as many edges as faces in a TriMesh, since every face has 3 of them. But finding out which edge lies between two vertices or which faces are sharing the same vertices along an edge is rather slow as the whole mesh has to be traversed to find out. As the mesh density increases, this search time gets longer and longer.
In the case of EPoly though, the basic building blocks of the geometry are vertices and EDGES! Polygons are defined by the edges, not the other way round. This means that finding edge loops, splitting edges, cutting through hundreds of thousands of edges with a slicing plane etc. are very fast since the edges are a structural element of the geometry and VERY easy to locate, query and change. Also, two neighbour polygons sharing an edge actually use the SAME edge, so manipulating that edge alters both polygons with a single operation.
A drawback of this structure though is that by definition, an edge can be shared by 1 (open edge, like the outside edge of a 4x4 plane) or 2 (internal edge, like the inside edge of a plane) polygons ONLY! You cannot have 3 polygons sharing the same edge (the case is called a "rat's nest"), which explains why it is not possible to select an internal edge that already has 2 polygons and extrude a 3rd polygon by shift-dragging.
(If you convert to EMesh, shift+drag edge and convert back to EPoly, you will discover that the new poly is a "free-floating element" and not actually attached to the surface it was extruded from - it has no common edge and vertices with the original edge that was extruded).
Another drawback of EPoly is that the graphics pipeline that supplies the realtime renderer with triangles (can be SZB, OpenGL or Direct3D accelerator) has to do some extra work to convert the EPoly data to TriMesh to be displayed in the viewports. So while cutting through an EPoly mesh is much faster than with EMesh, the display is much slower. It also explains why adding a MeshSelect or TurnToMesh to the top of a non-animated EPoly makes it many times faster to redraw in the viewports...
Hope this helps!
Cheers,
Bobo (Frantic Films)
Взято с cgtalk.com
n-gons нет
сейчас опять пробую в максе моделить - так себе впечатление. Хотя раньше нравилось..
Совсем другой принцип.. в максе почти все делается екструдом эджей и фейсов - от частного к общему.
А в синеме наоборот - накидал болванку лупами, кистями подправил и пошел деталей добалять - очень быстро получается.
буду разбираться.
сейчас опять пробую в максе моделить - так себе впечатление. Хотя раньше нравилось..
Совсем другой принцип.. в максе почти все делается екструдом эджей и фейсов - от частного к общему.
А в синеме наоборот - накидал болванку лупами, кистями подправил и пошел деталей добалять - очень быстро получается.
буду разбираться.
