Hi Micha,

> WOW, I increase the angle just for fun and the mesher dosn't
> create crap (case (8) - the Rhino "o" with rounds), it looks like
> it should look, also at this extrem raw angle. I'm impressed!

:) It depends a bit on the particular situation though, if you end up with a surface that has only one polygon for the surface part, but an edge gets a refinement in the middle of it, that refined point of the edge will be placed at the actual point on the curved edge's midpoint which is then a far distance away from the single polygon that is approximating the surface. That then makes for a kind of very warped polygon.

It basically causes a single polygon to have a very non-planar shape, which makes for somewhat odd looking polygons.


> ... but I found, that the result is not stable at all angle values -
> maybe it's possible that the mesher can be refined to avoid
> this effect

I guess it would be possible, but it would probably need a special "low polygon" mode that tried to do things in a special way that did not follow the regular method used for normal meshing. Unfortunately that would probably take a fair amount of work so it is not too likely that I will be able to accomplish that anytime too soon. I'd also worry about cluttering up the UI with specialized tools that were only useful to a very special kind of case, that is not so good because it means that the main tools become more complicated to use for everyone and the only benefit is to a very small area of work. That's how adding some things can actually end up with an overall negative effect rather than being an overall benefit to a larger group of users.


Some more notes on the strange-looking result:

Under normal circumstances it is important to be able to have an edge refined with points added to it to an additional amount beyond the surface that it is on, because an edge can easily be curved in shape a lot more than the surface.

Here is a basic example - here I have a plane so the surface is not curved at all, but meanwhile the trim edges of the surface have a very curved shape:




If I did not refine edges to have additional points, then a shape like this would end up with a mesh looking like this:



Which as you can see would not be a good result for the most regular meshing cases such as regular rendering, etc...


Basically for extremely low polygon shapes like you're experimenting with here, I would recommend building those using the tools in a polygon modeler like jonah mentioned, that is a toolset that is specifically designed for that purpose.

The case you are showing here is starting with a very curved shape, that's just not really the best starting point in general if you want an extremely angular result, it is better to use a toolset that is specifically designed for that kind of a thing to get the best results.

- Michael

>> Thanks, the easy examples are examples only. I think, if the easy example works, than complex models should be good meshed too. Most important for complex scenes, where the polycount must be keeped low, but good looking - for example for rendering train / air plane interiors or large ships with many details.

If polygon count is so crucial, I dont think you need fillet on your models, it will increase the poly count a lot.
If fillets are important, you should try to use a smaller meshing angle (12-20), and use the 'avoid smaller than' option to reduce the number of polys. You'll have a much better visual/polycount ratio than just rising the meshing angle so high.

Hi Micha,

> The object I used is a Rhino object (copy&paste to MOI) and I
> attach it now. Could be nice, if the mesh could be same, dosn't
> matter it is a STEP import, a Rhino copy&paste or a MOI3D
> creation. ;)

That would be cool, I agree! But a rounded box can be defined in many different ways, and in order to get one to mesh with all quads the "underlying" surface of the rounded part must be aligned with the other pieces.

The model that you have posted here is not configured in that way at all, if you use Edit/Separate on it to break it into surfaces and then turn on control points for a corner, you can see the surface for it is actually set up like this:



Notice how the surface control points of that corner piece are not aligned with the trim edges?

That corner is defined as a kind of rotated sub-region trimmed out from a larger sphere portion, the sphere is oriented like this:


The problem here is that the UV layout of that sphere (for instance if you were to look at its isoparm grid) does not have any kind of alignment with those edges. Here is dense mesh of just the sphere where you can see how the trim edges collide with the natural surface UV grid at an odd angle:



Because of this lack of alignment, you won't get simple quads from that kind of corner patch.

If you do a similar examination of the rounded box built in MoI, you will see this kind of structure instead:






Notice how the corner pieces there have a natural alignment between the surface and its trim edges? That kind of more simple structure provides the alignment between the surfaces that is necessary to get simple pure quad mesh generation.


So again, don't just assume that every rounded box is constructed in the same way! :) It is possible for some to have a construction method where surfaces do not share natural edges and that will prevent simple mesh output from that kind of stuff.

- Michael

Hi Micha,



You wrote:
> I thought a surface or polysurface has edges

Yes - every surface or polysurface has edges to it - but part of the definition of an edge is a 3D curve that gives the shape of the edge (there is also a UV curve that is in the UV space of the surface). In this case that 3D edge curve is pretty complex, just take a look at the screenshot which I will show here again:




That edge has gone through some kind of "fitting process" to calculate it as an approximation, maybe as an intersection between 2 surfaces or something like that.

Anyway, as you can see from the number of points in it, the 3D edge curve is not as simple and clean as it could possibly be, to get a better mesh I removed the existing 3D edge curve (by "untrimming") which forced a new one to be calculated from the underlying surface. The new 3D edge curve is more simple in structure and exactly matches the underlying surface, making it a more simple structure overall and helping to build a better quality mesh.


> Could it be possible, that this is internal automatic done for meshing?

Unfortunately it is kind of tough, often times you need a kind of human intelligence to analyze the shape and see what parts can be simplified. It can be difficult to replicate this with a kind of artifical intelligence.

It's also possible that in different situations this kind of process can change the shape of the object slightly. I kind of worry a bit about doing things automatically that change the shape, it is another area that can require judgement.


> Or can I do something at Rhino befor export an object to MOI?

Possibly either "untrim", or "RebuildEdges" in Rhino could be used for a similar process.


> Are the edges needed for meshing?

Edges are definitely used in meshing, without processing edges then you would only ever have meshes that used the "underlying surface" parts of objects and not any trimmed areas which are often formed by booleans, etc...

Edges that are messy or more complex than needed, or wiggly, etc... can introduce complexity into the generated mesh, same as if the surfaces are more complex than needed or wiggly, ...


> Is at MOI at command to select all edges?

You just first select one edge with the "drill-down" click (when you click a second time on an object), then use Ctrl+A to select all.

- Michael

That is that we can name a full answer :)

Hi Micha,

> (5) mesh #3
>
> I attach the original surface, now it should be buggy. ;)

It looks like this one is due to having a surface that has an internal crease (G1 discontinuity) in it.

This is kind of an unusual surface since you have created a degree 1 surface with a lot of points in it - normally when you have a surface with a lot of points like this, it typically is done as a degree 3 surface so that it will be smooth when you pull points around on it, a degree 1 surface is kind of more similar to a polyline, and typically degree 1 surfaces are used only with a smaller number of points to form a simple plane (using just 4 corner points).

At any rate, the bug in this case is in the splitting function that is supposed to split surfaces at G1 discontinuities when they are loaded, it is only splitting it partially in this case.

I've fixed it up for the next v2 beta, so it will get split properly when loaded and mesh well.

Another workaround for the current release is to use Edit/Separate on it, and then Edit/Join to rejoin it, that will cause it to get split up properly and it will mesh well after that.

- Michael

Hi Micha,

> (6) mesh #4
>
> Seems to be the same problem like at (4) and I don't understand it, why edges
> are independent from the surfaces

Well, edges are not restricted to only be along the natural boundary of a surface.

If they were, then you would be unable to trim surfaces in many situations.

Here is an example - say I start with this plane, it is a simple surface where it has 4 edges that are on the natural boundary of the surface:



Now if I take a curve made up of many different segments and trim that plane, it will create a bunch of new edges on that surface:



You can see there that I turned on control points for the surface as well, and the plane surface is still there underneath those 10 new edges that form the new boundary.

Similarly you can have any number of trim curves on a curved surface as well, and it is possible to have trim curves that touch each other smoothly rather than at a sharp corner - that is what was happening with this previous model, if you go in and try to select the outside edges of that cylinder you will see that you can select a small portion of it, which is actually one small edge itself, the boundary is made up of a number of small edges touching one another rather than one single large edge.

Check out this FAQ answer for some more illustrations on how trim curves and an "underlying surface" work:


A larger number of edges makes for a more complex topology and will prevent MoI from doing things like spacing points out evenly across a larger distance like it can do when the edges are larger.

So if you want the most optimal mesh result, it will help to have an optimized topology as well.

A complex topology will also tend to generate a more complex mesh result as well.

The mesh will still be valid on a non-optimized topology (as in no holes, etc...). But since you are very interested in getting the best possible mesh, you will need to also give the mesher clean topology to work with.

In general it is a pretty simple situation, I mean you can't expect the mesher to generate perfect results when it is given lower quality inputs. (sometimes low quality can mean different things, but I've described previously the types of alignment conditions which help for this purpose of all quad generation).

There's a saying: "Garbage in, garbage out", - if you don't start with a high quality result it is not realistic to expect the highest quality result out the other end!


> and how to delete edges.

In MoI this can be done by selecting all the edges of a single surface (not a joined surface, use Edit/Separate) and pushing delete. Check out this object repair tutorial for more info:
http://moi3d.com/forum/index.php?webtag=MOI&msg=446.17


> I would like to
> avoid the effect per Rhino script befor I export the object to MOI. I afraid, at
> complex models the "separate-edge-delete-join" method could be a problem.

Yes, it can definitely be difficult to do a cleanup of messy stuff on a complex model!

But if you don't provide clean inputs, it is just not possible to get the very best quality outputs.

Keep in mind that MoI does not have the same level of human judgment and analysis that you can apply, it is a very difficult problem to do a fully automatic cleanup without judgement to assist the process.

- Michael

Hi Micha,

> (8) I used an angle of 35°. It's the "o" from a Rhino text object with a round.

I think this one was answered by my reply to your additional message, see here for that reply:
http://moi3d.com/forum/index.php?webtag=MOI&msg=2451.21


> This screenshot above was done befor I changed the CentroidTriangulation option
> at the ini. (Sidenote: the Rhino mesher create this overhanging peaks too.)

It probably wasn't so much the change in CentroidTriangulation that avoided this, but rather using a different angle value that didn't cause the edges to get an additional subdivision beyond what the base surface got. Again see the above linked reply for more info on this.


Overhanging peaks are basically when a polygon has a high degree of non-planarity to it. If you have many points that are nearly co-planar and then suddenly one that is a larger distance away from that nearly common plane, it creates that kind of situation.

Rhino's mesher tends to create things like that in many more circumstances even without going to coarse levels of meshing, particularly around closed surface seams.

In MoI you will typically only see that effect if you are torturing the mesher by using very coarse angles like you are doing here! :)

If you see that kind of peak under more normal meshing circumstances in MoI (like say less than 20 degrees angle parameter), then please let me know, but in general I think you won't run across that except for in these very coarse examples.

Again, please refer to that previous message about how it is possible for edges to get refined with additional points than the surfaces and why that is necessary (or else things that are flat but with complex trim boundaries would not work at all, example in that previous message).


> Two other things to this - at Rhino I found, that some triangles can be converted
> to quads (maybe this is fixed in your internal version like you wrote at (9))

This was with the Mesh#6.3dm model? Was this instance where you could convert some triangles to quads happening at the default mesh parameters, or only at a particular parameter?

If you can tell me the parameters you used to run into that, I can test with it over here and see if it is fixed by the other fix or not.


> and could it be possible, that the CentroidTriangulation option is
> part of the general meshing parameters?

Is it something that you expect to be turning on or off very often?

For most people I was figuring that it would be something that they would want to have on all the time or off all the type for more typical usage.

- Michael

"MoI is a one man show" indeed (joke about earlier thread). After the length of what I see here, I'm starting to wonder if there arent a few people named Michael Gibson working on this product! ;)

Hi Nick, I just type fast!

But yeah I do have to try and get some time to write code as well as answer questions! :)

Micha, maybe in the future when you have so many different things to post about you could possibly group them into some more categories and space them out a little bit, like maybe one group of bug reports, one thread a bit later on low poly questions, one thread later on quad generation on particular models, etc....

That would probably be a lot easier for me to manage, either that or I could possibly need to take some items and kind of archive them for future reference rather than trying to deal with them all right now...

- Michael

Oh, that's much to read. :) Thank you for the answers.

First - why I'm fighting with the low poly count? I offer 3D visualisation and get my models from my clients. So, often I can't choose, how detailed the models are. Some production models, most from software like ProE, are so complex, that I must try to get the lowest possible polycount. For example if a designer send me a train seat and like to see a coach full of this seats, best rendered in very high res and without visible tessalation at 100% at the screen - worst case.
So my tests go in the direction low poly count and best quality. I happy to see that MoI doe's a great job for that.

Thank you PaQ for the 'avoid smaller than' hint. That's a good polycount saver. :)

Michael, I see, if I like to get a perfect mesh, than best I model at MoI. ;)

Now I better understand the concept of curves and trimmed surfaces. Thank you for the deatiled infos.

>> Two other things to this - at Rhino I found, that some triangles can be converted
>> to quads (maybe this is fixed in your internal version like you wrote at (9))
>This was with the Mesh#6.3dm model? Was this instance where you could convert some triangles to quads happening at the default mesh >parameters, or only at a particular parameter?
>If you can tell me the parameters you used to run into that, I can test with it over here and see if it is fixed by the other fix or not.

(Rhino "o" model - Mesh#6.3dm)
33° -> "96 triangles were converted into 48 quads."
22° -> "226 triangles were converted into 113 quads."

>> and could it be possible, that the CentroidTriangulation option is
>> part of the general meshing parameters?
>Is it something that you expect to be turning on or off very often?

I don't know - if I mesh an object, than I look at the preview and if I see a problem, than I change the parameter to get a better result. So, best would be good to get all control at one place.

Thanks,
Micha

Hi Micha,

> (Rhino "o" model - Mesh#6.3dm)
> 33° -> "96 triangles were converted into 48 quads."
> 22° -> "226 triangles were converted into 113 quads."

It looks like these are also tuned up to at least some degree for the next v2 release - in the current Jan-19 release with Quads & Triangles, these currently generate these number of polygons overall:

33° -> 635 polygons
22° -> 1211 polygons

In the next v2 release instead these generate a lower number of polygons (some more quads):

33° -> 608 polygons
22° -> 1163 polygons

However, MoI's mesher only creates quads when the quad is a "natural quad" that is aligned with the UV grid of the surface, it doesn't just create any possible quad that may be found by combining any adjacent 2 triangles because that tends to make for somewhat odd looking results in many cases.


I'll see if I can think of a good way to add centroid triangulation as a checkbox option. I kind of hate getting such a huge array of options though, it tends to make the UI much more complex. But I guess in this case since the dialog can collapse to simple version I maybe shouldn't worry about it so much for this.

- Michael

Hi Michael,

here an other issue or example (mesh#8), where could be possible an optimization. I create this object at MoI, it's a first try for a railway sleeper for a visualisation of a train station.
My screenshot shows the unfilleted object that is good meshed and the filleted version with two issues - a large kink, also visible at smaller mesh angles like 16° and missing quads at the big top surface that goes around the sleeper.
It could be nice, if the meshing of the fillet would be determinated by the top surface. But this isn't new, you told me:

"I think it is going to be possible for me to do some work in the future to try and form the initial UV quads of a fillet after the "main" surfaces have been done and in some situations it may be possible for me to make a better initial structure for the fillet that matches an adjacent surface, but probably only when there is a plane to one side of the fillet since matching to pressure from multiple sides of curved surfaces is probably not going to work."

So, maybe this is one more example for a know issue only. But I have the feeling, the big kink at 16° should be avoided without additonal adjusting. ;)

Hi Micha - the big kink at 16 degrees happens because of the curvature of the fillet surface is quite shallow in the longer direction, which I highlight here:




Since the surface is only gently curved in that direction, an angle of 16 degrees does not cause the fillet surface itself to become subdivided.

As I've mentioned before, if you have a situation where a surface does not get subdivided but its boundary edges do, it can lead to these kinds of kinks.

There isn't really any bug here - to solve the big kink you need to do something to cause additional subdivisions of the surface.

The typical method for that is to decrease the angle parameter until it is less than the angle of that shallow curvature for that long direction on the fillet - that will solve the problem. For example try using an angle of 12 instead and your kink problem will be greatly reduced. As you go to a tighter angle yet it will continue to be reduced.

Another way you can solve it is to use the "Divide larger than" setting to force that shallow long fillet piece to get subdivided. For example with Angle 16, try entering a value of 20 for "Divide larger than", and it will force any polygons longer than 20 units in size to become additionally subdivided - that will force additional subdivisions of the surface and again solve the problem.

When you see this problem, it is caused by a surface getting a coarse mesh on it, and the solution is always going to be to force additional subdivisions...

I think that a "Chord height" tolerance value which I am also planning on adding would be a pretty good parameter to help in these situations as well.

But basically if you want to improve the tessellation here, you need to manipulate the settings to produce the result you want - normally the solution is to just make a tighter angle but if you want to avoid that you will need to use some of the other settings instead of angle alone on shallowly curved things.

Keep in mind that 16 degrees is a rather coarse angle, the default that MoI uses for angle is 12, which is already somewhat coarse to start with.

- Michael

Hi Micha, just to give you a better idea on how coarse 16 degrees is (and to better illustrate what is happening), here I have extracted that fillet piece and viewing it from the side where it is curved, I have placed 2 line segments underneath it which have a 16 degree deviation between the lines:




As you can see, the fillet overall does not really bend as much as those line segments bend with respect to one another - that's why an angle of 16 degrees does not cause that fillet to be subdivided in that direction, it is within that angle tolerance.

Again also notice that the angle between the lines is very noticable - that is because 16 degrees is a pretty coarse angle! As you can see lines that are at 16 degrees to one another do not look very much like a smooth curve.

When you generate a coarse mesh you can get that kind of "non-planarity" to n-gons which is what causes those kinds of kinks.

- Michael

Thank you for the hints. Also I'm very courios for the "Chord height" tolerance value.

I play with the options "devide larger .." and run in an other issue. I enabled "curved", but the upper surface is devide in the direction without curvation too. My wish here is a fourth option (planes, curves, all) like "curved UV", so that the upper surface wouldn't devided in the not curved direction of the red arrow. Maybe a fourth option isn't needed and this feature could be implemented for the exist "curved" option.

General I miss quads at the upper surface like the object without fillets show it. I marked two critical areas. So, it's not only a question for the polycount, also for the look.



Also nice would be, if the polygon edges of the polygons could go straigt upward. But maybe that could be done by a inteligent mesher with a sense for beauteousness only. ;)
I know, you like a simple interface, but I could imagine, that it could useful to set independent parameters for "divide larger than .." of curved and plane surfaces.

Hi Micha,

> I play with the options "devide larger .." and run in an other issue.
> I enabled "curved", but the upper surface is devide in the direction
> without curvation too.

Yeah, the "curved" option will apply the division to any surface that is not planar, and will divide in both directions.

I think that the chord height option will work better for this particular thing rather than having another variant of "Divide larger than".

That is basically what will be useful about a chord height metric - directions that are flat will already satisfy the chord height metric. By the nature of that metric it only applies to curved areas.


> General I miss quads at the upper surface like the object
> without fillets show it. I marked two critical areas. So, it's not only
> a question for the polycount, also for the look.

Well, you've used an angle of 100 degrees in your screenshot there, you should generally expect to get a jagged look to your object with such a coarse angle, no surprise about that!

Re: quads - as far as I can tell the same thing that I described earlier in this thread applies here - you're not getting quads because the initial UV quad layouts of those surfaces do not align, here is a link back to that explanation and illustration again:
http://moi3d.com/forum/index.php?webtag=MOI&msg=2451.6


> I know, you like a simple interface, but I could imagine, that
> it could useful to set independent parameters for "divide larger
> than .." of curved and plane surfaces.

The separate chord height metric will basically allow for something pretty similar to that since it essentially ignores planes. So that would let you use "Divide larger than" and set it to only apply to planes, and meanwhile use Chord height separately which will only apply to curved areas.


> Also nice would be, if the polygon edges of the polygons could go
> straigt upward. But maybe that could be done by a inteligent
> mesher with a sense for beauteousness only. ;)

I'm not quite sure what you mean by striaght upward? But yes unfortunately many things that would be nice to have tend to require a kind of "artifical intelligence" which is rather difficult to code. Maybe some day! :)

- Michael

Hi Michael and Micha,

I've been following this thread with interest, I am not by all means a meshing expert, my question is, does model geometry/surfaces play a part in good meshes, the reason I ask is
that I noticed Micha has modelled the top of railway sleeper with one surface, typical of a poly modeller, as I've learnt from the forum.
Would it be better in the meshing sense to model all the fillets as fillets instead of a single spline?

I've attached Micha's railway sleeper as I would have modelled it, it is the third model along, hope you don't mind.
Does this make a difference ? Just curious and educating my self.

Thanks

The chord height options sounds very interesting. I'm curious for it.

Straight upward means something like this ...


... 100 degree - I set this high angle because I tried to get control only by the devide option. I'm quite happy with the result, only the missing quads. The initial uv layouts could be alligned, if fillet and top surface would be created per sweep allong a side curve. I understand.

I tested how good could be a mesh (#9) created only by "devided ..". I wonder me, that the right surface show some triangles. The initial uv layout should be quite good to create a mesh of quads only or?



Or here - additional polygons at the lower edge ...

... and from down view

... also "avoid smaller than" 2 dosn't help.

Hi Danny, that can actually help somewhat to have it structured the way you show there, with more of a partitioning between different regions of that long surface rather than just one long single piece.

With a long single piece that goes through a lot of bends, what will happen is the mesher will attempt to fit a regular uniform mesh on it but will have to give up on that since it would take too many polygons to do a completely uniform evenly-spaced mesh across the whole thing.

When you have it separated out into more individual regions like you have done there, that can help each section to get a uniform base mesh on it, which can help to avoid adaptive subdivision of the mesh.

Adaptive subdivision is the kind of thing shown here:
http://moi3d.com/forum/index.php?webtag=MOI&msg=2379.3 which will divide only certain areas of the surface into smaller pieces instead of having a totally uniform division throughout the entire thing. You won't get simple quads anymore in areas where adaptive subdivision happens, where there is a transition from one level of subdivision to the next.

But at any rate, some of these issues are things that can come into effect with any kind of model when you want to make a coarse, low-polygon version of it that doesn't have odd things sticking out from it. Coarse meshing and jagged shapes kind of tend to go along with one another, so it can take some targeted manipulation of the meshing parameters to get low polygons without too many artifacts. That can be kind of a separate issue from the modeling technique used since it applies to a wide variety of models.

- Michael