Full Version: V4 Wish List

Show messages:  1-15  …  216-235  236-255  256-275  276-295  296-315  316-335  336-355  …

Hi 3Dfanatic,

re:
> I kept the lines and main objects that I swept before doing the second action which
> is what is causing the removal of the edge.

What is the second action that you're referring to here, is it using Trim ? If so then when you're cutting a solid it is easier to use boolean difference instead of Trim. Trim operates only on surfaces, while the booleans work on solid volumes and are able to make a solid result automatically joining in the pieces of the extruded cutting curve. You can get the same result with Trim but it's more work because you'd need to extrude the curve into a surface and then trim that surface and join the results together. Booleans basically do all that work for you.

So for your case here with your solid you made from your sweep, select the solid and run Construct > Boolean > Difference, and then select your curve as the cutting object:



That will divide it into 2 solid pieces:



Delete the pieces you don't want and you'll be left with a full solid and won't have to worry about closing up holes as you would if you work only at the surface level with Trim:



One thing to note is that although it works ok for your case here, it's generally not good to have your cutting curve overlap right over top of other areas of the model like in this region here:



That can tend to complicate booleans because it's going to extrude out a cutting surface from that curve and the extrusion is going to be skimming right along the same surface area of your main shape and it's more difficult to get a clean intersection curve from stuff like that. In order to do booleans, a well formed closed intersection loop of curves have to be generated to divide the object up into different clean pieces. If the intersection curves are complex and criss-cross over each other that might not happen.

So instead if possible it's better to have a cutting curve like this where it shoots out a ways instead of having spots that hug right along the same shape:



Then for filleting, the problem there is your shape is very thin and so there isn't much room to fit fillets in, about the maximum size that will fit is around 0.01 units, if you try to use a radius much larger than that it would cause the fillets to collide into each other and cause the type of artifacting that you're describing.

So use a fillet radius of 0.01 or smaller for your particular case here, here's an example of what it looks like with 0.01 :


Hope that helps!

- Michael

Image Attachments:
boolean1.jpg  boolean2.jpg  boolean3.jpg  boolean4.jpg  boolean5.jpg  fillet.jpg 

---

Thanks Michael!
That info helped, after all that do you have plans to give more control on skinnier objects so the fillet can be reduced past the .01 setting? Thanks for all your support

---

Hi 3DFanatic, you're welcome!

re:
> do you have plans to give more control on skinnier objects so the fillet can
> be reduced past the .01 setting?

You can do that currently, for your particular case here 0.01 is about the maximum size that will fit, but you can go smaller than that if you want. For example here it is with 0.001 units:



With skinny objects if you get weird results try using smaller and smaller values, don't stop at say 0.1 units, keep going smaller and smaller to get a good idea of what's going to be possible to actually fit there.

With your shape here, a fillet of radius 0.05 takes up this much room just for one side:



It can be a good idea for objects with very small sized features like this to be modeled at a larger scale, like say build this object at 10 times larger than what you've got right now so that you're not working with too tiny values.

- Michael

Image Attachments:
fillet2.jpg  fillet3.jpg 

---

Thank you great advice i will increase the whole thing

---

Not sure if this has been mentioned already but here are two suggestions:

Extend surfaces X amount (like in Catia)

Remove chamfer/fillet by selecting and adjacent surface automatically extend (like in Powershape)

---

Hello,

Extrapolation of curves and surfaces would be a great addition to MoI. Scale that model up x10 and you can fillet the edges at .15 to get a complete round on the edges.

---

Hi Michael.

Have a nice day.

What do you say ? We can except a "hot" October :) ?

V4 is coming ?

Bye!

---

Hi Michael ...
Yes, I am on pins and needles. You have said 'soon' awhile ago. I'm confused about the definition of that word, :-).
cheers,
eric

---

Hi Marco & eric - it's getting awfully close! I've got a good path forward for the viewport display, I've been able to modify the rendering library so it matches up with what I need to do and I've got curves and surfaces drawing on screen now.

The main areas left are shaders, background images and text labels, hit testing, and more exploration of a fast path for high memory video cards. I've been thinking that I might release an early version before all that stuff is completely done.

Here's what it looks like at the moment (model by PaQ):



- Michael

Image Attachments:
v4_screenshot.jpg 

---

Ok Michael!

We're on the right track!

It seems to be a good starting point.

If you create some shader to apply to the base rendering it could become interesting.
I mean, ambient occlusion and some shader with environement illumination/reflection.

I stay tuned.

Thanks.

---

Hi Marco, well the first priority is for shaders that make curve and surface drawing look the same as V3.

For ambient occlusion that's something that I'd like to look at later on, but probably initially in a simple "offline" renderer. That's not something I'm targeting for the first V4 beta release though.

- Michael

---

OK. No Problem.

For me the most important things of the new display engines should be these :

1) Improved antialiasing (both for curve/edges and surfaces)

2) No "display glitches"




3) Speed. Moi must take advantage of the modern GPU, if it is installed on the machine where Moi is executed.

I hope these things will be done within the V4 beta period.

Have a nice day.

---

Hi Michael
Thanks for the response. I can hardly wait to start using the beta 4.
cheers,
eric

---

Hi Marco, for #1 curve antialiasing in MoI is pretty much already as good as it's possible to do so I don't really expect to be able to improve that in particular. Do you have an example where it isn't working well?

For #2 I'm afraid that display glitches in the viewport display are difficult to avoid without taking a hit in speed at the same time. I do have some ideas on possibly improving that particular type of artifact but it remains to be seen how much of a speed hit will be involved.

For #3, that is something that I do intend to focus on, however your other requests would generally come at the expense of speed so you're kind of asking for mutually opposing things on your list here... ;)

- Michael

---

Hi Michael.

@You : "..for #1 curve antialiasing in MoI is pretty much already as good as it's possible to do so I don't really expect to be able to improve that in particular. Do you have an example where it isn't working well?.."

Please look at these screen shots and this video : http://take.ms/YULPS



As you can see even if I force the antialiasing with my Gtx 1080 Ti (a beast of Gfx card), the results doesn't change.



And this is the 3dm file : http://take.ms/f2aR5

I made the model intentionally in that way, obtained by curve projected on curved surfaces and then lofted and blendcap.

@You : "..For #3, that is something that I do intend to focus on, however your other requests would generally come at the expense of speed so you're kind of asking for mutually opposing things on your list here... ;).."

Well...I can't believe that with a card like the Gtx 1080 Ti, or even a 1080, 1070 or a modern AMD card, MOI should be slow.

I can understand that you want that Moi runs on every system, also slow computers, but in 2017 it's a shame not to take advantage of modern GPUs.

In this way all the owners of fast and powerful GPU are always penalised.

I hope you understand my point of view.

Thanks a lot for your support.

Bye.

Marco (mkdm)

---

"""""Here's what i looks like now, with PaQ's model"""""""

Dont know how much i appreciate the beautiful antialiased curves of MoI until their gone!!! Lol

---

Hi Marco, none of the areas that you are indicating have any curves being drawn there, so that has nothing to do with curve anti-aliasing. It's shaded surface anti-aliasing that you want there. I think that could be possible to add as an option for v4, it just hasn't been a priority since most models have many of those areas covered by edge curves which are anti-aliased and in most cases there isn't a lot of contrast with shaded surfaces silhouettes against the background so aliasing in those spots usually isn't very distracting. You're using a customized background color which makes it more apparent. Additionally there is a speed penalty that comes along with it as well.

I have done some initial experiments with taking advantage of the plentiful amount of VRAM that is available on newer cards. I'm not sure yet if that will be done for the first beta release or not. I think I've got it worked out how to speed up shaded surface rendering when there's a lot of VRAM available but I haven't yet got it worked out for curves yet.

> As you can see even if I force the antialiasing with my Gtx 1080 Ti (a beast of Gfx card), the results doesn't change.

That's good, because if the driver changed that without MoI asking for it that would break selection.


> Well...I can't believe that with a card like the Gtx 1080 Ti, or even a 1080, 1070 or a modern AMD card, MOI should be slow.

Not for a limited size model like you've got there but if you increase the model complexity a lot more then it's a different case.


- Michael

---

@Burr,

> Dont know how much i appreciate the beautiful antialiased curves of MoI until their gone!!! Lol

Not to worry, I expect it to return! :)

Just getting pixels to be drawn is an important step first. Getting them to look nice again will probably be just about the last thing.

- Michael

---

Hi Michael.

Good morning.

@You : "...It's shaded surface anti-aliasing that you want there. I think that could be possible to add as an option for v4..."

Yes! It would be very desirable to have that feature. Thanks.

@You : "...I think I've got it worked out how to speed up shaded surface rendering when there's a lot of VRAM available but I haven't yet got it worked out for curves yet..."

Ok. This is good.

@You : "...Not for a limited size model like you've got there but if you increase the model complexity a lot more then it's a different case..."

Yes, you're right. But for very, very complex assembly or model you can always play with the Mesh Angle option.
Please consider that actually, on my laptop with a almost outdated Quadro K3100M with 4GB Ram, I always use a Mesh Angle of 5 without any speed problem also for a (not simple) model like this one (the model it's not mine).
And with my Gtx 1080 Ti I can easily copy and paste the same model many times to have no noticeable speed issue.
Nothing that really impact the modelling process.



I stay tuned.

Have a nice day.

---

Hi Marco, well although that's not a simple model you show there complexity can go up quite a lot more than that, particularly for large assemblies that someone is importing from another CAD system.

I hope at some point to maybe get rid of the mesh angle setting and try to make it more automatic, like start out with a rough angle and then if things are going well with that do higher density ones on a background thread.

- Michael

---

Show messages:  1-15  …  216-235  236-255  256-275  276-295  296-315  316-335  336-355  …