some news about the 3d background ?

pipbal · 08-28-2002, 02:42 PM

Hi Jim ,

do you have some news about the 3d background 4 us?

Jim Sachs · 08-28-2002, 04:02 PM

First the Freshwater Aquarium, then the 3D background.

flipper · 09-03-2002, 06:05 PM

jim, if you need 3d collision intersection/detection code, i have a number of resources that might prove useful to you. theres some pretty nice royalty-free libraries out there that do all of the drudge work for you and will work both for fish-coral collisions as well as fish-fish collisions. im talking about the OPCODE, RAPID, QuickCD type of libraries.

Jim Sachs · 09-03-2002, 06:44 PM

Those might come in handy when I return to the 3D background. They have to be public-domain, though -- not just royalty-free. Also, some of these libraries are so large that they dwarf the rest of the program.

flipper · 09-04-2002, 08:23 PM

good points i had not considered. here's 2 that are public-domain:

if you pre-determine your bounding boxes for all model geometry, you might want to take a look at http://www.magic-software.com/SourceCode.html . the MgcIntr3DBoxBox.cpp source (in the 3D intersection section) would be ideal for you as the 2 boxes can be any size and with any orientation. you would only have to write your own code to traverse through your box pairings. a pretty simple and very small code footprint solution.

the other one is a complete CD library, OPCODE, and is found at http://www.codercorner.com/Opcode.htm . in dll form its about 1/4 MB (75KB compressed). one thing a library would give you would be faster collision detection since they typically use more intersection tests and also use pre-calculated data structures optimized for quick testing. they also allow for complete flexibility re: model data (same code for different aquariums for instance) and would provide a better set of bounding boxes (better fits) as they calculate the boxes themselves (you just provide the triangles). they also will go down to the triangle/triangle intersection level if you request - which may be used for such things as starting up a fish coral polyp eating animation cycle.

hope these help. i have more if your interested.

Jim Sachs · 09-04-2002, 08:38 PM

I can only use bounding boxes for the rough collision detection. For the fine stuff, like fish pecking at coral (or each other), I need to know if any part on one object bumps into any part of another object. And the fish and coral are constantly morphing.

flipper · 09-05-2002, 03:34 AM

thats why i thought a full library might be best overall. if you want it to it will tell you exactly which triangles from each model are intersecting. so you should be able to determine if its the fishs mouth or the fishs tailfin. there is a big price to pay memory-wise though (20 bytes per triangle just for the CD stuff). and the morphing issue.

08-28-2002, 02:42 PM	#1
pipbal Registered Join Date: Jan 2002 Location: Belgium Posts: 25	some news about the 3d background ? Hi Jim , do you have some news about the 3d background 4 us? Better 2 burn out then 2 fade away

08-28-2002, 04:02 PM	#2
Jim Sachs Developer Join Date: Dec 2000 Location: Southern Oregon Posts: 9,816	First the Freshwater Aquarium, then the 3D background. Jim Sachs Creator of SereneScreen Aquarium

09-03-2002, 06:44 PM	#4
Jim Sachs Developer Join Date: Dec 2000 Location: Southern Oregon Posts: 9,816	Those might come in handy when I return to the 3D background. They have to be public-domain, though -- not just royalty-free. Also, some of these libraries are so large that they dwarf the rest of the program. Jim Sachs Creator of SereneScreen Aquarium

09-04-2002, 08:38 PM	#6
Jim Sachs Developer Join Date: Dec 2000 Location: Southern Oregon Posts: 9,816	I can only use bounding boxes for the rough collision detection. For the fine stuff, like fish pecking at coral (or each other), I need to know if any part on one object bumps into any part of another object. And the fish and coral are constantly morphing. Jim Sachs Creator of SereneScreen Aquarium

09-03-2002, 06:05 PM	#3
flipper Registered Join Date: Mar 2002 Location: Texas Posts: 60	jim, if you need 3d collision intersection/detection code, i have a number of resources that might prove useful to you. theres some pretty nice royalty-free libraries out there that do all of the drudge work for you and will work both for fish-coral collisions as well as fish-fish collisions. im talking about the OPCODE, RAPID, QuickCD type of libraries.

09-04-2002, 08:23 PM	#5
flipper Registered Join Date: Mar 2002 Location: Texas Posts: 60	good points i had not considered. here's 2 that are public-domain: if you pre-determine your bounding boxes for all model geometry, you might want to take a look at http://www.magic-software.com/SourceCode.html . the MgcIntr3DBoxBox.cpp source (in the 3D intersection section) would be ideal for you as the 2 boxes can be any size and with any orientation. you would only have to write your own code to traverse through your box pairings. a pretty simple and very small code footprint solution. the other one is a complete CD library, OPCODE, and is found at http://www.codercorner.com/Opcode.htm . in dll form its about 1/4 MB (75KB compressed). one thing a library would give you would be faster collision detection since they typically use more intersection tests and also use pre-calculated data structures optimized for quick testing. they also allow for complete flexibility re: model data (same code for different aquariums for instance) and would provide a better set of bounding boxes (better fits) as they calculate the boxes themselves (you just provide the triangles). they also will go down to the triangle/triangle intersection level if you request - which may be used for such things as starting up a fish coral polyp eating animation cycle. hope these help. i have more if your interested.

09-05-2002, 03:34 AM	#7
flipper Registered Join Date: Mar 2002 Location: Texas Posts: 60	thats why i thought a full library might be best overall. if you want it to it will tell you exactly which triangles from each model are intersecting. so you should be able to determine if its the fishs mouth or the fishs tailfin. there is a big price to pay memory-wise though (20 bytes per triangle just for the CD stuff). and the morphing issue.