Hello All,

6.0 is currently in the "research" phase. I have a few ideas about what I would like. I am considering the following:

1) New renderer

Not everyone agrees with me on this since POVRay produces excellent results and rendering is just a single click on the "POVRay" button.

2) Improved "in place" editing

I would like more of the tools to be "in place" rather than opening a new window. The negative of this is that "custom" tools such as the Lathe tool tend to be easier to use for beginners than "in place" tools.

Nothing is decided yet.

Note that 6.0 is a very long way off. Don't expect it until the Fall/Winter at earliest. That doesn't mean that there won't be updates to 5.5. No doubt there will be new features between now and the release of 6.0. The enhancements to 5.5 will be restricted to scripts/plug-ins/operations.

So, what would you like in 6.0? I can't promise that I will include any "wishes", but I will consider all of them.

Richard

Richard,
As for the new renderer, is it possible to have both a new renderer, and Pov rendering, or would this be an either/or situation?

Tools like unwrapping, uvmap, and uvremap would probably be just as easy to use for "in place" editing, but it seems that the Lathe and Extrude tools would be a bit more difficult to use. I believe the latter two are best as is.

My wishes would be for;

1) An On/Off type option for auto-banking used with aircraft animations.

2) Animation paths, or path names that are visible in the object library.

3) An option for auto-linking(proper term?) of an object to a predesigned path.

Les

I think this would go under the new renderer, but i think that cartoon shading would be really sweet! Hmmm... Nothing else really comes to mind now. Ill tell you if i think of anything else.

Simon

Richard,

about my wish list, I know you already got it, but since this is the thread for 6.0, I would reassume here the main features that I believe would be important to implement or enhance in Canvas.


[size=4]MODELING[/size]

1 ) The solid outline rendering mode. This mode is very useful to use for low-poly modeling, it would be a nice feature to have. Below I reported a picture of the solid outline mode.


the solid outline rendering mode


2 ) Enhanced SDS modeling support. It would be useful to add cycles and subdivide parameters to the Soften tool, so the artist could smooth and unsmooth (using undo) the control cage in a single operation, using the desired degree of smoothing. Below i reported a picture of a smoothed control cage. (by the Bay Raitt tutorial http://www.izware.com/news/indexa446.html).


a SDS model and its control cage


3 ) About inplace operations, they would be really great to improve the workflow ! I totally agree that this would be a very nice and useful feature to add. In my opinion inplace operations would be good for all the modeling tools. While the uvmapping tools need a view on the map and a view on the model, plus a lot of tools to work with, so to have them in a separate window would be probably better. This is also the general solution used in every modeler I know.


[size=4]MAPPING[/size]

[size=1]NOTE: what I reported below about the mapping features is not entirely correct since rotate and weld capabilities was added after this post was published. Apr 9 2002[/size]

4 ) Enhanced mapping support. Though the actual uv-tools are enought for many mapping needs. It is not so easy to use them to get good maps for low-poly models. The picture below shows a very simple model (about 50 polys) and its map, that map has high continuity and low distortion, so it's good enough (you can download the model on my site http://www.theothersideworks.com).


a model with a good map

However to get that map with Canvas is not possible because it misses the weld, align, rotate and mirror operations on the uv-points, that are needed to align and join the map parts. Moreover it misses a resizable checker for fine evaluation of the distortion, and the ability to adjust map parts.

I mean, as you showed in the nice unwrapping tutorial, actually the artist has to redo from start the whole unwrapping process of the map part she/he would wish to adjust, plus the artist has to delete by hand the old unwrap using an external paint application. So to improve the workflow and the features of the uv-tools would be a nice new feature.

A good reference to plan improved uv-tools could be Lithunwrap http://www.geocities.com/lithunwrap.


5 ) An analithical mapper. The next step in the mapping technology. An analithical mapper is a tool which generates the uv-map of a object based on the object shape. It automatically try to achieve low distortion and high continuity. From the point of view of the artist it is simply another mapping type, the same as the classical planar, cylindrical, spherical etc.

A good reference to plan an analithical mapper could be Tru-V http://www.texturetools.com.


[size=4]ANIMATION[/size]

6 ) An enhanced bones system. The actual bones system is good enought for many needs, but it has two main limitations. First the bones have a fixed shape for skinning, eg. spherical or cylindrical, so if the surface to skin doesn't fit the bone shape the artist has to use extra bones for skinning. This problem is resolved in some other modelers by allowing vertex assignment to the bones, so a bone can fit any shape on the skin.

The second limitation is the weight distribution system. Since in Canvas the weight of a vertex is equally distributed among the bones influencing it, in the usual joint there are weight 50 and weight 100 vertices only, in some situations this could generate a not optimal bending, depending on the model. Again the artist has to use extra bones if she/he needs a better weighting.

Below I reported the picture of a cylinder bended in Canvas and in another application, no extra bones was used in Canvas to fit the default weighting. I also reported the picture of a bone with vertex assignment, the bone is assigned to the head and fits the whole head shape. Please note that these pictures are meant to show the Canvas limitations, not to denigrate its features in any way.


Canvas bones vs full weighted bones


a bone with vertex assignment fits to any shape

A good reference to plan enhanced weighted bones could be CharacterFX http://www.insanesoftware.de.


7 ) A multitarget morph tool for facial animation. This would be useful for characters animation. A good reference to plan a multitartget morph tool could be the CK gamefactory multimorph plugin http://tokyo.cool.ne.jp/cgf.


8 ) A cloth simulation tool. Useful as well for characters animation. A good reference to plan a cloth simulation tool could be the CK gamefactory clothmotion plugin http://tokyo.cool.ne.jp/cgf.


[size=4]RENDERING[/size]

9 ) About a new rendering engine, personally I would disagree since POV is already excellent in its job. To develop a rendering engine usually takes a lot of time, and as a Canvas user I would prefer you to use that time to implement other features. The only extra feature that POV misses compared to some other renderers is a cartoon shader, I don't know if it's possible to do that with POV.


[size=4]IMPORT EXPORT[/size]

10 ) An enhanced import/export control. It would be useful to add import export options panels to Canvas. This way the artist could fit her/his models to/from any other application, even if those applications don't handle that specific format in a total correct way. This would improve the capability of Canvas to exchange models with other applications.

A good reference to plan import/export options could be 3D Exploration http://www.righthemisphere.com.


11 ) An enhanced md2 exporter. The md2 format is very important for the artist to design game characters ready for the market. So it would be useful to improve the exporter to be friendly to other applications as much as possible.

Though the actual exporter is good for many game engines, a good reference to test the exported format would be the Milkshape MD2 viewer, because it is a "pure" and "official enought" viewer used for this kind of format. So if Canvas md2 models would work with the Milkshape MD2 viewer, then they would have the best chance to work with any application.


[size=4]RELIABILITY[/size]

12 ) GDI resources. Canvas now is very reliable, a great reliability enhancement was done since the first versions. The only thing that still causing some problems is the high requirement of GDI and user resources under Windows 98. When Canvas is running, the user can't run any other major application, even if she/he has a lot of RAM available. It would be useful to find a way to reduce the GDI and user resources requirements.


Well, this is what I believe could be useful to improve in Canvas. Hope this list could be useful for planning precedences for 6.0, or any other future version.


best regards , Alex

Alex,

You know my opinion on Skeletal and Unwrapping. Also, it sounds like you didn't read the updated Unwrapping tutorial.

As for import/export, I have no plans to change this much. If people want 3D Exploration, they should buy 3D Exploration. It is a good product.

There will be no changes to the GDI resources issue because there is little that I can do, and it is not really necessary. I suspect by the end of the year most people who are likely to buy 3D Canvas will have upgraded to XP, where this isn't an issue.

Also, you can't have #1 without a new rendering engine.

But, thanks for your other suggestions,

Richard

Richard

You might guess what I would like...

GROUP ALL OBJECTS in workspace (rather than one at a time)and then change the coordinates XYZ of ALL the Objects...

Move ALL OBJECTS to upside down

Move ALL OBJECTS to stand up and lie down

Move ALL OBJECTS to lie on side

Move ALL OBJECTS to face forwards or backwards

This would be usefull for Importing different file formats Objects and then Moving them to the correct XYZ positions for Exporting to different Games...

Also I like the idea of freehand modelling...

1st point...zeroXYZ click mouse ! then for 2nd point use 4 arrow keys left/right/back/forward and 2 keys up/down click  mouse for next point...(the single arrow clicks could be set in units of 0.1 to 10)

You would be able to see the object appear in the workspace !

...its ok I think this would require a completly new 3D Modeller

Here is something I would like to see. Sometimes when I open up a model that I have downloaded or a model that I get from one of my games and open it up in 3D Canvas, the model is so huge that I have to keep moving back and back and back until the model will fit into the entire view in 3D Canvas. I would suggest that whenever you open up a model that the view will automatically adjust to fit the entire model in the screen. This would be nice, or is there already a way to do this and I just don't know?

Support for TrainZ? Why stop at MSTS?

I know this is probablly too hard but...

I would love to be able to 'control' what i create

sort of like a game..

but im just dreaming lol.

I think that would be too impossible probablly!

other than that i cant think of anything new (at least right now)

maybe something will come to me lol

later...

Richard

Dropdowns for Topbar...

Edit                      View

select All               Zoom 1:1

Group (All)             Zoom + 2
Triangulate (All)      Zoom + 10

Flip Vertical 180      Zoom -2
Flip Vertical   90      Zoom -10

Flip Horizontal 180
Flip Horizontal   90

He he he  ... Just for me !

For the MSTS export, more distance models are needed so that we can match the M$ supplied models. I think they use 5. It could be fairly simple to control, any objects in the top level of the heirachy (main) would appear in all models, objects in the next level down would appear in all but the farthest, and so on, then anything in the fifth level, or lower, would only appear in the closest distance model.

Also, it would be nice to be able to view the different distance models as a check that they look OK, or if objects need to be moved up or down a level.

 hi, i thought of something i would LOVE to see in 3dc...


animated textures!!

this is probablly very diificult but would be extremely coool!!

i want to simulate water moving thats how it came to me!!

bye..

;D Hi again, just one more lol :

more advanced lighting system where you can make it even darker or brighter-(for individual lights)

thanks.

Hi all,

Thanks for all of the suggestions. It appears that most of the suggestions can already be done with 3D Canvas. Alex had quite a few suggestions that were new. Also, Cartoon rendering is something I think would be cool.

A couple of points:

Ian: 3D Canvas' technique for creating optmized distance levels (I think it create 5 or 6) is quite sophisticated and given the right model produces what I would consider to be an "optimal" progression with the final distance level being just those objects in the "main" group. But I definitely can see your point.

As for flipping/lighting/mirroring/animated textures etc. These can all be done in 3D Canvas right now. Probably the first step is to ask "How do I do this in 3D Canvas?". And when you are sure that it can't be done add it to this list.

Richard

Richard,
Since my only use of Canvas Pro so far is with train simulator I have to say that it does nearly everything that I could wish.
I would agree with Ian that it would be useful to be able to see what the distance models look like.
The only other thing I would like to see is more control over the background. ie :-
The ability to scale and move the background individually in each view.

cheers
Kevin

Richard


The Properties/Group Orientatation Tool is Excellent

If I GROUP ALL the objects to one another and use
The Right Click Grouped Objects and Properties/Group Orientatation Tool, I can Rotate to the required 90 or 180 Degrees...this is Excellent

But, I'm not sure if I understand the flip tool, it will not flip GROUPED OBJECTS, like above...only merged or single objects !

This would be Good:

The Flip tool (or Rotate 180 Degrees for SINGLE or MERGED OBJECTS)

EXTRA :
The Rotate tool (90 and 180 Degrees for ALL GROUPED OBJECTS) would be good

So, for quick modelling...A GROUP ALL TOOL and a Rotation Tool which could do the same as the Orientation Tool to Rotate ALL the GROUPED OBJECTS quickly either 90 or 180 Degrees XYZ !

This is only to explain my thinking before and only a thought...

Alex,

I have been looking at bones in more detail. I did a lot of thinking yesterday about the requirements for a more sophisticated weighted bones system for 3D Canvas. After thinking about it for a while I realized that there were real problems with bone systems that exclusively use "weighting". I've decided that 3D Canvas' bone system is actually superior to the one you have demonstrated. But perhaps you convince me otherwise.

I've attached a picture from a program you recommend. The unweighted skin on the left is obviously skewed and look wrong. However, the skin on the right that is supposed to show how a sophisticated weighted bones system is superior, to me, shows the exact opposite. This is not a proper look for a boned object. It is more like a deform. There should be a crease not a continous curve. And based on what I see, it could be quite difficult to produce a proper look using this style of weighting.

If one were deforming a spine this might be a good look. For example a dolphin deformed using this technique would look good. If you were deforming an object like this with 3DC you would use the Bend operation.

Richard

Here is the same thing in 3D Canvas. In my opinion this better shows how a "joint" should look. This is especially true when you view the shaded version on the right.

And finally...

The first image (from the left) is a replication of your demonstration of how "weak" the 3D Canvas' bone system is. The second image is the same object with properly set up bones. Not perfect obviously. There is too much creasing on the inner part of the joint. The third image shows the effect of adding a "floating" bone. This "floating" bone can be moved based on the orientation of the joint as needed to improve the appearance. The fourth picture is a smooth shaded version of the third image.

So, I would say that 3D Canvas' bone system is not "great", but pretty versatile. And I have no immediate plans to change it.

Richard

Richard,

since you said "but perhaps you convince me otherwise", I guess you expect a reply. So here it is, directly from the bones and mapping grumbler ;D. Don't worry, I'll be quick to the point .

1) All the bended Canvas skins that you showed have the 50/50 weight applied only to one section of the skin, all the other sections have weight 100. This technique is the same used to bend skins with unweighted skeletons, eg. Milkshape's :P. Unfortunately this doesn't show that the bone system is good, on the contrary this shows that it is comparable to an unweighted one :-/.

2) The picture showed by CFX is only an example, with the CFX bones system you can achieve any type of bending that you wish, because you have full control on the weight. And, as you can see, any section of the skin has a different weight distribution, I guess this is why they used that picture as example . Anyway this effect is much better visible in the picture from my previous post, where you can see a more bended cylinder with all the sections having different weights.

3) However, the major problem with Canvas is not the limited weight distribution. For low poly models even unweighted bones are good enought. The major problem with Canvas is the fixed shape of the bones :'(.

Suppose you have a part of a model with two flat shapes one near the other, and you have to bone them with two bones. You can't do that with cylindrical bones because the two cylindrical shapes would intersect each other. A common way to resolve this problem is to use vertices assignment to fit the bones influence to any skin geometry.

This feature is much more important than weight control, because without it, depending on the shape of the model, it could be practically impossible to bone the skin. In the example above you would have to use a lot of thin cylinders to bone the two flat shapes. From this point of view even a unweighted bone system with vertex assignment would be better than the current bone system.

Apart all of this, the Canvas bones system is good enought if the skin is designed to work with it, it just can't work with any skin. So from this point of view you are right when you say that it doesn't need to be improved. While of course I have to disagree when you say that CFX is not superior, since differently from Canvas, it can fit to any skin and any type of bending.


best regards , Alex