im thinking of implementing this over the next couple of days, and have wondered if ppl are already. if so are they using it for all objects eg camera/materials/geometry or just for a limited set. im thinking of just implementing the materials (textures/shaders/blending etc) stuff to start off with,
does anyone have any collada files that use glsl i could have a scuzzy at, curious how theyre handling uniforms etc
ta zed
not many users of collada here, apparently. i can say with confidence that i have never used it. but then again, i have no idea what it is, beyond what i could glean from a cursory overview.
i see it’s xml based, for better or worse. i’m partial to my own script formats so i might just bow out early on this one.
it seems interesting though.
regards,
bonehead
Will Maya, 3D Max or lightwave import & export your own formats effortlessly with a freely downloadable plugin? Will other software be ready to interoperate with your own stuff? Will some cool renderer, tool or mesher you have be able to play with other people’s assets & art path or will it remain an irrelevant footnote?
The art asset path is a mess of proprietary formats, tools & scripts in proprietary languages. Getting something standard is a way to start to leverage each others tool chains through interoperable formats. Hopefully it won’t just add to the polygon babel. The history of 3D graphics is littered with the wreckage of misguided attempts to design 3D interchange formats, maybe this one will have some legs :-).
Personally I don’t have a high opinion of XML or it’s advocates, but that’s a secondary issue here.
P.S. You should realize that the guys posting the specs to this are at the heart of the Sony PS-3 software team with years of experience in real-time rendering & file format conversion, loading & saving for real-time graphics systems.
Will Maya & 3D Max or lightwave export your own formats effortlessly with a freely downloadable plugin? Will other software be ready to interoperate with your own stuff? Will some cool renderer, tool or mesher you have be able to play with other people’s assets & art path or will it remain an irrelevant footnote?
this is the interesting part. as it is now, my resource world (what there is of it) is its own microcosm. but something like this would be very cool for sharing, were that a requirement. i’ve been leaning towards something more or less text based, sorta like the stuff doom3 is doing. it’s friendly…all you need is a text editor…it’s dirt simple. and simple is good, for me anyway :-).
regards,
bonehead
The history of 3D graphics is littered with the wreckage of misguided attempts to design 3D interchange formats, maybe this one will have some legs :-).
the collada forums have never been busy (checked a few months ago + now) why is this? is nobody using it? there seem to be a lot of big names behind it, so why the lack of apparent support?
most of us use standard texture formats dds/jpg/tga same with sound. but with meshes theres less standardization, and materials hardly at all.
Personally I don’t have a high opinion of XML or it’s advocates, but that’s a secondary issue here
some implementations of it are a joke eg for a vert x=“12121” y=“343” z=“343”
I’m not here to advocate Collada, but some forum’s traffic is no indication of support. AFAIK this is a key open piece of Sony’s PS-3 content path and that should produce all kinds of support, it may already be getting used by developers under NDA with internal real support rather than noise on public forums.
It’s impossible to really gauge how significant it is.
Anyone with a clue understands that as an intrinsic format it is uninteresting, the only interesting aspect is that as a standard of sorts that might be used to share/transmute reasonably rich data with a reasonable level of support and interoperability. That depends 100% on the unknown factor you point to.
most of us use standard texture formats dds/jpg/tga same with sound. but with meshes theres less standardization, and materials hardly at all.
dds is a great thing…all in one texture format that has good compression. all in a simple, tidy binary format.
if you buy that you might be willing to say the same about microsoft’s x format. probably not as many takers here, because here you get into all sorts of issues with game related additions to meshes. ever used unreal ed? then you know about all the nifty things you can do with models, including adding attach points, building and adding collision meshes, and a host of other cool things. it’s too complex to stuff into a readily exchangable format, at least not an xml based one (tee hee). sure, you could make it work, but why would you want to if what you have already works…
as for materials i’m guessing you mean shaders and such? same thing goes here. if you serialize your content into a binary packgage with scads of dependencies in your object graph, why would you change this if you didn’t have to? using unreal as an example agin, their binary packages are notorious for having loads of dependencies. you can export stuff from the editor into text formats, but by itself its not going to amount to much. i think this is the main detraction from such a thing. forcing a break with a complex system mired in dependencies. and this is to say nothing of the extended load time incurred by loading and parsing text files. although personally i find this obstacle the least objectionable of all.
You don’t wake up one day and decide to create something like Collada because it’s an interesting idea. Inherently it implies some kind of content tools and rendering path that Sony is working on at a minimum. You can bet it will express what that system requires in useful detail.
Clearly Sony is going to do more than throw driverless hardware over the wall to developers this time around. Collada suggests they have something bigger than just OpenGL support planned.
I don’t think the aim of collada is to provide a format that you read into your runtime application. (thinking games/realtime here, not art packages)
I think the aim is to provide a means for different tools and art packages to interoperate.
Ie.
a) Artists use any tool they want to create art
b) Process art with common tools (ie generate normals, remove degenerates etc) and re-save as collada.
c) Convert collada files into game/app specific format (that is optimal for your game).
d) Load app specific files at runtime.
This makes artists happy:
- they can use whatever art package they want to create art
- art can be re-used between projects. (if origionals are lost)
It makes programmers happy:
- as they can re-use processing tool chains between applications.
- They do not have to write exporters for different art packages.
- They can change the binary format of the app whenever they want to and batch convert all the collada files. (instead of telling the artists to re-export everything) <== This is a big advantage.
i’m sure you’re right dorbie. i’ll bet sony has their beans polished with a fine shammy. it’ll be interesting to see what happens with this 
Only a quick note. We are using Collada in our engine and the Shader Designer will be released soon with Collada importer/exporter. At least for us, Collada is agreat help. Now we are independent of the 3d modelling tool used (except for the bugs in the exporters… ) and we dont have to spend a lot of time writting complex exporters for different programs. Only one parser and only one mesh processor. That’s good . We didn’t reach the shader part yet, but is our next thing in our TODO list (think in a conversion between 3DSMax or Maya to SD and viceversa)
Originally posted by sqrt[-1]:
I don’t think the aim of collada is to provide a format that you read into your runtime application. (thinking games/realtime here, not art packages)
I’m sure that’s correct, and never posted anything to the contrary, however a key element is to massage art assets into optimized targeted formats using the tool chain. With this it makes sense to have a final target and the foundations of a path to get there.
Just creating Collada in isolation as a “nice idea” would be a distraction unless there’s something underlying this probably multiple things, i.e. tools and runtime help. Collada is (in one sense)an open door invitation to interoperate with that.
Think, Sony’s answer to Microsoft’s XNA, but with an emphasis on open interoperability.
I’m giving COLLADA a try at the moment.
The 3ds exporter seems robust enough to handle any mesh/material/shader.
The format itself is very very easy to learn/read/parse due to its XML structure.
At the moment I’m using .NET to parse the files with makes my life very easy.
The only worry I have at the moment is the COLLADA forum itself, I always visit it once in a while and it makes me think noone its using collada due to the forum innactivity.
So to conclude >> lovely format, but no one knows yet if its a succesful format.
Originally posted by dorbie:
(snip)
P.S. You should realize that the guys posting the specs to this are at the heart of the Sony PS-3 software team with years of experience in real-time rendering & file format conversion, loading & saving for real-time graphics systems.
Those who’ve worked with Sony graphics APIs and hardware in the past might see their involvement as a cause for concern
well i see from the spec you can reference external binary images so this is encouraging. the more i read of this thing the more i like it…except of course for the whole xml business.
it seems to be nothing more than a file schema, albiet a highly extensible one. which immediately raises the question concerning who’s responsible for dealing with this extensibility. how much does my parser need to know to parse xyz’s export correctly, robustly? seems like you’re in the same boat all over again.
you could argue however that at least you would only have a single parser to worry about, and that is good…very good
yes, good for shuttling raw assets around, if everyone embraces it wholeheartedly.
Originally posted by mikef:
[quote]Originally posted by dorbie:
(snip)
P.S. You should realize that the guys posting the specs to this are at the heart of the Sony PS-3 software team with years of experience in real-time rendering & file format conversion, loading & saving for real-time graphics systems.
Those who’ve worked with Sony graphics APIs and hardware in the past might see their involvement as a cause for concern[/QUOTE]Since you’re casting vague aspersions on people, they didn’t work on earlier Sony APIs AFAIK.
I’m tempted to ask “What APIs?” since a packet spec and DMA engine isn’t an API by any stretch of the imagination, it’s not even a driver, but maybe it depends on what end of the pool you jumped in the water. I don’t have a dog in the hunt. The most interesting part of this is the reaction of all those VU programmers IMHO. At least they have OpenGL ES with vertex & fragment programs :-). It’s going to be tough to find fault with that.
What about the new CgFx 1.4 Format ? How does it compare to collada ?
Only a quick note. We are using Collada in our engine and the Shader Designer will be released soon with Collada importer/exporter. At least for us, Collada is agreat help.
excellant news, im sure once a certain number of good apps use it, critical mass will be reached resulting in even more ppl using it.
pity it wasnt used in doom3 than every man + his dog would be using it (based on the number of md2/md3/md5 viewers etc ppl have written)
I always visit it once in a while and it makes me think noone its using collada due to the forum innactivity
thats cause theyre all posting elsewhere ahem (www.opengl.org)
seriously though i dont wanna waste time on something thats gonna not be used, eg 3-4 years ago i based my engine on the intel (quasi)replacement to opengl32.dll but that died.
Originally posted by dorbie:
[quote]Originally posted by mikef:
Those who’ve worked with Sony graphics APIs and hardware in the past might see their involvement as a cause for concern
Since you’re casting vague aspersions on people, they didn’t work on earlier Sony APIs AFAIK.
I’m tempted to ask “What APIs?” since a packet spec and DMA engine isn’t an API by any stretch of the imagination, it’s not even a driver, but maybe it depends on what end of the pool you jumped in the water. I don’t have a dog in the hunt. The most interesting part of this is the reaction of all those VU programmers IMHO. At least they have OpenGL ES with vertex & fragment programs :-). It’s going to be tough to find fault with that.[/QUOTE]I didn’t mean any offence to the Collada folks; I’m very glad to hear that none of them bear any responsibility for the PS2. And I too look forward to a return to sanity with the PS3.