Feature Requests

#81

I’m not certain I understand your point…what is supposed to be rounded that isn’t in that image?

The problem with making a more intuitive interface is that it often tends to be a matter of making it more automatic…which often means that the user ends up backtracking a lot when the function the machine assumes was needed wasn’t. Extruding from a ball and getting a globby little protrusion would take quite a bit of automation to accomplish with splines, as would automatic smoothing; but not everything needs the same smoothness or needs to be connected in the same way…

I think that really the modeling could be streamlined extensively if the five-point patches were automated somehow, though I can’t imagine how.

Still, making patches more like SubD seems impossible–it’s just a totally different dynamic. In some ways, with the total lack of automation and the total control of Hash spines, they’re antitheses.

BTW, the shape is possible in AM, but it took me 10 minutes or so…

#82

Kalimol looks like there are some creases or seems in there.

I agree what people are saying about subd’s and patches and I also agrre hash needs to keep there modeling tools up to date.

I think hash sould make some kind of spline weight feature kind of like subd’s weighting but the spline one would control how sharp or rounded the splines are that pass thought a CP. I know there is the bias setting but it needs to be updates with a slider or something and not freak out when you animate it.

As for 5 point patches I think they should be auto also and as for creases there should be some way like a slider for you to adjust it from smooth to real creasy (if you want creasy) just like the weight controll in Subd’s.

#83

Kalimo,

By rounding, I meant that I would like the shape to be more circular (not peaked). To get technical, I am looking for sine waves rather than porabolas. I wouldn’t want this all the time, just when I have toggled a switch. Another problem that I didn’t mention is that in extreem cases such as the one given, porcilain will make the shading of an object incongruous with its geometry.

I think that 5p patch creation could be automated. After all, 4 and 3 pointers are. Again, a toggle to turn this off would be good. Maybe it would require a lot of hash-power to do it but it would be a BIG time saver.

Also, your last example isn’t rounded at the edges. Proper bevelling means being able to select a load of geometry and slide a slider to vary the amount. In am, things are the polar opposite of that. If you take one of the primitives provided on the CD and stretch it, the bevels get screwed.

#84

Originally posted by dfaris
as for creases there should be some way like a slider for you to adjust it from smooth to real creasy (if you want creasy) just like the weight controll in Subd’s.

In 10.5 there is a slider to vary the amount of porcilain. But having said that, I don’t know why anyone would want creases. I have herd a lot of people say “It is ok that hash patches are creasy because so are peoples faces”. I don’t understand this statement. The fact is that if there are going to be creases in my models then I want to put them there and I want them to be where I put them. Build creases into the geometry or use bump maps but don’t expect flawed geometry to make the right choices for you.

#85

Here is an example of what I want AM to do sometimes:

#86

Kalimol looks like there are some creases or seems in there.

The bias should have been set differently on the lower part. The actual connection was creased a touch too much, and there really should have been another spline ring in there to form the seam…

Also, your last example isn’t rounded at the edges. Proper bevelling means being able to select a load of geometry and slide a slider to vary the amount.

Maya can do that (subjective use of the word “proper” notwithstanding)?

In am, things are the polar opposite of that. If you take one of the primitives provided on the CD and stretch it, the bevels get screwed.

Well, naturally, since it’s just another bit of the model that’s getting scaled/skewed/whatever. I’d think that would give you a good bit of extra control and cause a lot less problems…

I still don’t get the arch thing…isn’t that a simple bias issue?

#87

:blush: Dig a hole, and dig it deep. Ok Let me rephraze a few things. For me to have said that the new porcelain brings A:M’s splines easily on par with sub divs was a poorly made statment. There are advantages and disadvantages to both Spline modeling and subdiv modeling. The new porcelain and bias normals make getting a good looking model much easier then ever before, and so are not a bad alternative to subdivs. There are some modeling situations that subdivs can produce a model quicker; however, there are also situations where hash splines can make a model faster. My attached image has my 3 minute 15 second attempt at the said model, but it also contains a 45 second model that I don’t belive could be modeled as quickly with subdivs either. Yes, I know I’m setting my self up here. ;p Ultimatly, however the point of my message was why I feel that subdivs should be implamented with out trying to stir up the debate that is currently happening. :thumbsdow Bad on me… Oh well…

#88

Your right Bill. . .

That is a tricky model to make using subDs.

With AM you get way more precision over the edges. And Lightwave doesn’t even HAVE edge sharpening in its subD implementation. . .

Nice splinin dude. . .

#89

Wanted to take one more stab at it as I wasn’t happy with the original ends I made on the first one. Also I tryed to mimic the angle a little better on this one. took me about another 10 seconds worth of modifications to the lathe shape. This should also show how good the new porcelain is. I think this is very comparable to the subdiv example. My stubs arn’t quite as long, but thats not overly important.

#90

Bill_Young: Would it be possible to get a smooth branching shape like that without porcelain? Porcelain just seems like a bit of a kludge, since the actual geometry still has creases (from what I understand).

#91

I think that an important question needs to be asked though… how practical is the example we are looking at? it seems a bit contrived in my opinion. I just looked at my project and I can’t see one place I have this exact join, how often do you suppose it comes up?

and bill did a much better job of it than I did, which since I equate modeling with painful dental work on the list of things I like to do makes perfect sense… and that’s with patches… god polygons… how can i express my extreme distaste for modeling with polygons… words fail me here. I don’t enjoy modeling in the best of situations but mashing cubes together and trying to get a head? no thanks.

which is kind of what bill was getting at in his original post. I’m a billion times more productive with patches since I can stand to work with them. every attempt I make at polygons or nurbs or what have you just gives me a headache. different tools for different minds.
:slight_smile:

-David

#92

Now I feel really bad…everyone’s a better modeller than me…

Anyway, as for the likelyhood of using a geometry like this…I can imagine some uses, and I built something like it as part of a joint in a robot character recently. But I think the point is that it displays a very machined-looking object that runs counter to Hash’s organic splines–the point all along was that some things are easier with SubD and that they should be added in, right? Not necessarily that they were better? So the shape didn’t have to be practical, just atypical for AM. Not like it matters, with Bill’s success…(You rock Bill!)

But as I’ve said before, I’ve never really had a chance to try out any decent SubD modellers…is edge sharpening like peaking along a specific edge spline?

Bill_Young: Would it be possible to get a smooth branching shape like that without porcelain? Porcelain just seems like a bit of a kludge, since the actual geometry still has creases (from what I understand).

Yeah, the outline doesn’t seem to match the shading in some situations…still, I don’t think it’s really avoidable, and it doesn’t look that fake on lower settings.

I still think photorealism is boring anyway…

#93

Originally posted by Kalimol
I still don’t get the arch thing…isn’t that a simple bias issue?

You could say that it is a simple bias issue but simply fiddling with bias handles all day does my nut.

Look at the first and second examples that I gave. The first is all pointy and the section shape is kind of triangular. The second shape had its biases altered so that it was nice and rounded. If I had a complex model then I realy wouldn’t want to do this by hand.

There are other times when you need to edit bias too, for instance where you have a big patch next to a short one. This is usually fine and can be used for making creases on purpose, but sometimes I am making a background object that I want to be smooth and I don’t want to spend my time making all the patches even in size or editing bias.

Hope this makes my point clear.

#94

Once again I find myself arriving late to this thread, but here’s my penny’s worth.

For me the great advantage of Hash patches is smartskin, especially the new implementation where you can adjust your mesh on any angle of bone rotation. You can achieve really accurate/realistic deformation with smartskin which I believe is more difficult with weight maps and other polygon options.
Now if Hash introduced SubD, wouldn’t features like this type of CP manipulation be difficult to use? If you could tweak the SubD mesh at anytime during your animation like you can with patches, then judging from the examples I’ve seen on the net, SubD would certainly be a welcome addition to AM’s toolset.

Here’s an example of some truly impressive SubD modelling (it’s a bit risque)

http://www.cgchannel.com/forum/viewthread?thread=1629&offset=29

The girl’s nose for example, shows exquisite modelling detail that I would find quite difficult to achieve with patches.

#95

John Keates–You could say that it is a simple bias issue but simply fiddling with bias handles all day does my nut.

Look at the first and second examples that I gave. The first is all pointy and the section shape is kind of triangular. The second shape had its biases altered so that it was nice and rounded. If I had a complex model then I realy wouldn’t want to do this by hand.

Yeah, but you can just select them and set the bias mag in the properties box…

Or are you saying that you want a way to set the automatic bias to 167ish% instead of 100% so that it’s round on a 90* arc?

In any event, I can think of just as many uses for the first form as the second…

Edit–Saw the girl model–Daaang. I’d sell my right arm to be able to model like that.

#96

Originally posted by Hookflash
Bill_Young: Would it be possible to get a smooth branching shape like that without porcelain? Porcelain just seems like a bit of a kludge, since the actual geometry still has creases (from what I understand).

Ok. Porcelain is sort of a “Kludge” but then again almost all shading on models is some form of “Kludge” and this isn’t that bad of a one. Porcelain from my understanding is samples the patches more like polygon shaders typicaly do with a series of sampled normals blended togeter with the normals around them. The sample rate per patch was increased in version 10.5 to produce a more accurate reproduction of the original surface. This means, if I’m not mistaken, that the patches are more aware of how the surounding patches are blending to produce the surface. With the normal renderer the patches shade them selves independently of one another, and look smooth because they are butted up against one another at a sub pixel level. In theory this is a far more precice representation of each patch surface, but because the surfaces are not “aware” of one another if they don’t line up creases may occure. Again this is speculation based off the information I’ve herd about how hash patches work. I could be way off base; however, To answer you question in theory you could get a normal patch surface to render like this, but it would take alot more work and mesh detail than it would be worth. IMO Porcelain is a much easier way to get this result. It’s not overly harmfull you a model so long as you are aware of how it’s going to effect youre model. I think of porcelain as hitting the tab key in lightwave to turn on subdivs. Technicaly they are just as much of a “kludge”. Poly’s normaly shade facited unless you make them normal aware. After all they are flat.

#97

I’d really like to be able to tag an item as alpha object only. Or exclude from alpha map. that way one could reduce the number of layers you’d have to render if you were thinking of using compositing to tie the video together.
Anyone think this would be useful?
I am aware of the other methods one can take to get all the channels one needs but with this option you could really cut down on the number of different renders you need to do.
Mike Fitz

#98

Originally posted by Kalimol
[B]
Or are you saying that you want a way to set the automatic bias to 167ish% instead of 100% so that it’s round on a 90* arc?

In any event, I can think of just as many uses for the first form as the second…

Edit–Saw the girl model–Daaang. I’d sell my right arm to be able to model like that. [/B]

Well, if I edit the biases as you suggest, thhen start moving cps about, the geometry is no longer rounded. The bias that you want is dependant on the relative cp distances. If all the patches were the same size then setting the bias to 167 would do but this is not always the case.

If want to make a model fast then I don’t want to worry about getting all the patches even in size. I just want to put the cps here and there and have a nice smooth model. Is this selfish?

#99

so what was the final outcome to all who joined the subdiv discussion?
Is it just a question of supporting polymodelers or is it about keeping hashspline or AM technology up to date?

Bill- we seem to have gotten of on a bad footing (by my reading of your reply to me) I have no desire for a battle. May I apologise if you saw my comments as “personal” in any way!

#100

Well it took me a while to get around to it but here is my subd model that Bill wanted us to make and compair to the AM model.

It took me about 3 mins to make using C4D but I did have to think about the best way to do it because I just got C4D and dont know the best way to do some things. It was pretty easy once I knew what to do.