ThreadBoard ArchivesSite FeaturesActiveworlds SupportHistoric Archives |
bugs in the rwx sphere primitive? (General Discussion)
bugs in the rwx sphere primitive? // General DiscussionjitterbugFeb 23, 2002, 3:12am
Sometimes it's just easier to explain things with pictures so
I put this message into HTML form as well: http://memeticdrift.net/jitterbug/rwxsphere/ Its been a while since I have built models, and I have just recently noticed several differences between 3.2 and the older browsers, particularly when it comes to the sphere primitive. I created RWX model shown below to illustrate the problems I'm having. The model contains 4 spheres one on top of the other, each successive sphere has a higher level of detail from the one below. For the RWX code monkeys out there here is the code: modelbegin clumpbegin Color .50 .50 .50 Surface 0.2 0.75 0.2 sphere 0.04 1 translate 0 .1 0 sphere 0.04 2 translate 0 .1 0 sphere 0.04 3 translate 0 .1 0 sphere 0.04 4 clumpend modelend This is what it looks like in RWModeler and how i feel it _should_ look like: http://memeticdrift.net/jitterbug/rwxsphere/Image2.jpg This next image shows what it looks like when i look at it using the AW 3.2 browser using direct 3D. It also looks about the same as this using the software renderer. Here are problems I see with this image: 1. The shading is very flat. 2. The spheres are made using a latitude longitude rather than a rounded octahedron system. 3. At the lowest level of detail the sphere degenerates into a flat diamond. http://memeticdrift.net/jitterbug/rwxsphere/Image1.jpg How it shows up using OpenGL. Notice the strange shading? I'm not sure what has gone wrong here but in addition to the above problems, looks like there are some very screwed up normals on the vertices of the spheres. (I'm using a diamond viper 550 video card if its video card issue.) http://memeticdrift.net/jitterbug/rwxsphere/Image4.jpg Are these problems known bugs, and if so is there a work around for them other than not using the sphere primitive at all? -Jitterbug. silencedFeb 23, 2002, 3:24am
Yeah, they used a new sphere type in 3.2. Try saving it from inside the
modelere then try. It should look better then it does now. --Bowen-- [View Quote] jitterbugFeb 23, 2002, 4:41am
[View Quote]
Actually what I want to do is create the rwx file from my own
program[1] and I'm just trying to figgure out what would be the cleanest way to write out the export file in light of the changes introduced in the 3.2 browser. Have a look at this picture to see an example of what im trying to export to RWX: http://memeticdrift.net/jitterbug/temp/bands.jpg What I have been doing up to now is use the sphere primitive in a prototype and translate it into the correct possitions for the joints as shown with blue spheres in the screen shot. This was why the primitive was very convenient. I do someting similar for the rods only i both translate and rotate them into place. I guess I could create the sphere out of triangles as rwModler does. It would just be more work on my part and would result in slightly bigger rwx files. Yes, I could call the switch of sphere types as a 3.2 quirk and the result of a design choice however the openGL one looks realy broken. (ie: a bug) I think a programmer at AW should have peek at it. I'm still not sure what's up with the lighting between the old and new versions. Something has changed as there doesn't seem to be any hilights in AW anymore. I can't get the shinny plastic look I had before. --Jitterbug -- [1] called SpringDance and its free http://shapeofspace.org/springdance/ silencedFeb 23, 2002, 4:45am
> Actually what I want to do is create the rwx file from my own
> program[1] and I'm just trying to figgure out what would be the > cleanest way to write out the export file in light of the > changes introduced in the 3.2 browser. > > Have a look at this picture to see an example of what im trying > to export to RWX: > http://memeticdrift.net/jitterbug/temp/bands.jpg > > What I have been doing up to now is use the sphere primitive in > a prototype and translate it into the correct possitions for > the joints as shown with blue spheres in the screen shot. This > was why the primitive was very convenient. I do someting similar > for the rods only i both translate and rotate them into place. Wow you lost me on this part :). > I guess I could create the sphere out of triangles as rwModler > does. It would just be more work on my part and would result in > slightly bigger rwx files. > > Yes, I could call the switch of sphere types as a 3.2 quirk > and the result of a design choice however the openGL one looks > realy broken. (ie: a bug) I think a programmer at AW should > have peek at it. Hmm.. I don't think they really care about it right now.. this might be fixed in 3.3? > I'm still not sure what's up with the lighting between the > old and new versions. Something has changed as there doesn't > seem to be any hilights in AW anymore. I can't get the > shinny plastic look I had before. Only one of the lighting modes works now :\. That may be your problem.. using one of the older light settings. -Silenced hal9000Feb 23, 2002, 2:34pm
[View Quote]
If you create 1 sphere in something like modeler, then you can use it
numerous times. at the end of your file add this ProtoBegin Sphere your sphere ProtoEnd then you can put it into your model with "ProtoInstance Sphere", then use your transform and rotate it into position. Yes the files will be slightly bigger, but its better than having 5 of them > > Yes, I could call the switch of sphere types as a 3.2 quirk > and the result of a design choice however the openGL one looks > realy broken. (ie: a bug) I think a programmer at AW should > have peek at it. Roland(roland at activeworlds.com) is the lead programmer for AW, send him an email with whats going on and he might take a look at it, that is if it isn't already fixed in 3.3 > > I'm still not sure what's up with the lighting between the > old and new versions. Something has changed as there doesn't > seem to be any hilights in AW anymore. I can't get the > shinny plastic look I had before. In the 2.x browser lighting had 3 parts to it, Ambient, Diffuse and Specular, Ambient being how bright the model is by default, Diffuse being how much a light will affect it, and Specular being how (for lack of a better word) smoothly the light shines on it. (ie. shiny plastic as opposed to fabric) > > --Jitterbug > > -- > [1] called SpringDance and its free > http://shapeofspace.org/springdance/ jitterbugFeb 23, 2002, 7:12pm
[View Quote]
>
> In the 2.x browser lighting had 3 parts to it, Ambient, Diffuse and > Specular, Ambient being how bright the model is by default, Diffuse being > how much a light will affect it, and Specular being how (for lack of a > better word) smoothly the light shines on it. (ie. shiny plastic as opposed > to fabric) > Just to finnish (no pun intended) the thought, I think you are saying that the 3.x browsers only support ambient and diffuse finnish settings. The specular (shinny hilight) setting knob is not actually connected to anything in the 3.2 browser. I found this in the help pages. http://www.activeworlds.com/help/rwx_specular.html I know, I know. I should read the fine manual! Thanks for your help. just inFeb 23, 2002, 11:25pm
correct, Modeler is showing the specular you have set of .2
any object you make you should change the specular to zero to more accurately reflect what you would see when displayed in the AW Browser. surface a b c where a=ambient b=diffuse and c=specular. In your example I would tend to something like "surface .5 .6 0" to get a nice effect. Note that both modeler and AW interpret shortcut commands into the model before it is rendered using renderware. These shortcuts include all primitives as well as protoinstanting. Thus the interpretation is the result of the interpreter rather than the renderware. As already suggested, the best way around the problem is to save the script in Modeler, and then clean it to reduce the file size. I've not checked if the latest beta has been released, but if it has the latest version does do a big clean-up to reduce the end result file size. Regards, Justin [View Quote] |