Fastest way to make whole-scene prebaked textures?
Hi,
I bought Cheetah 3d just this week. I'm pretty sure I'm up to speed on the things I'm about to talk about, but don't spare my feelings if I'm not...
I'm experimenting with Cheetah as part of a game-oriented workflow. I want to model in Cheetah 3d, have it prebake light maps and get the various bits of geometry (with object names preserved) into my code where an ambient map will be multiplied into the light map on the GPU as memory constraints mean that the light map will likely be of a quite low resolution.
Quick youtube video of efforts to date (the target device is an iPhone, hence the relatively low quality video that features quite a lot of my finger): http://www.youtube.com/watch?v=gfzazLI46kM.
So far it seems like this is going to be a labour intensive way to proceed. Can anyone tell me if I'm correct about the following assertions?
(1) you may convert only exactly one object at a time to polygons (eg, multiple selecting objects via the object browser and selecting 'make editable' will convert only the first selected);
(2) there is no Javascript interface for converting objects to polygons;
(3) texture baking is exclusively a per-object function (eg, you cannot say "I want to pre-bake a single texture to cover objects X, Y and Z"); and
(4) the 'Import Children' tool is not available in Javascript.
Assuming all those assertions are true, is it correct to say that probably the most optimal way to automate the light map generation task would be to write a script that saves the document out to an obj (which seems to convert all objects to polygons), issue a system tool to execute some command or another that can parse the OBJ and kill all the 'g <object name>' lines (assuming for argument's sake that I want a single light map for the entire scene though smarter methods could be applied), have the result OBJ opened so that all objects are collapsed to a single object, then manually set up the baking parameters I want and bake?
I'm actually a really awful artist and the basic castle model in the video is probably the single most artistic thing I've ever done. I therefore claim no real experience of modelling or modelling packages, and may still be missing obvious solution or trying to fight the user interface rather than just figuring out how to work with it...
Thanks in advance for any responses!
Hi,
I bought Cheetah 3d just this week. I'm pretty sure I'm up to speed on the things I'm about to talk about, but don't spare my feelings if I'm not...
I'm experimenting with Cheetah as part of a game-oriented workflow. I want to model in Cheetah 3d, have it prebake light maps and get the various bits of geometry (with object names preserved) into my code where an ambient map will be multiplied into the light map on the GPU as memory constraints mean that the light map will likely be of a quite low resolution.
Quick youtube video of efforts to date (the target device is an iPhone, hence the relatively low quality video that features quite a lot of my finger): http://www.youtube.com/watch?v=gfzazLI46kM.
So far it seems like this is going to be a labour intensive way to proceed. Can anyone tell me if I'm correct about the following assertions?
(1) you may convert only exactly one object at a time to polygons (eg, multiple selecting objects via the object browser and selecting 'make editable' will convert only the first selected);
(2) there is no Javascript interface for converting objects to polygons;
(3) texture baking is exclusively a per-object function (eg, you cannot say "I want to pre-bake a single texture to cover objects X, Y and Z"); and
(4) the 'Import Children' tool is not available in Javascript.
Assuming all those assertions are true, is it correct to say that probably the most optimal way to automate the light map generation task would be to write a script that saves the document out to an obj (which seems to convert all objects to polygons), issue a system tool to execute some command or another that can parse the OBJ and kill all the 'g <object name>' lines (assuming for argument's sake that I want a single light map for the entire scene though smarter methods could be applied), have the result OBJ opened so that all objects are collapsed to a single object, then manually set up the baking parameters I want and bake?
I'm actually a really awful artist and the basic castle model in the video is probably the single most artistic thing I've ever done. I therefore claim no real experience of modelling or modelling packages, and may still be missing obvious solution or trying to fight the user interface rather than just figuring out how to work with it...
Thanks in advance for any responses!