Torque Modeling Project
by Ryan Ackley · in Artist Corner · 02/28/2002 (10:55 am) · 17 replies
Im looking for people willing to help me work on the Torque modeling project, or TMP. If you want to know what i am doing, you can read This (my plan) and Project. Basically, i want to create a model repository, in the long run, but for now, I am aiming to create a fully animated skeleton in milkshape. I posted about this before and a lot of people were interested, and i know the idea of a shape repository has been discussed before. Interst is there, and i am looking for some one, or maybe a few someones, to help me get this off the ground. What i want is a resource section, like code snippits, where people could post models(.dts) and .max and .ms3d shapes for people to download and put in their game. Now, I am a decent modeler (but a kickass programmer :) and less than passable as an animator. Im learning, but im also looking for someone to work with me and help on this project. Ideas? Questions? Comments? Want to help? instant message me at nbCaffeine (thats AIM) or email me at ackleyrj@sbu.edu . Think we could, in the future, set up a modeling section of the resources, GG guys? Let me know what you guys think.
Ryan
Ryan
About the author
#2
I think it is important that some guidelines are established at the forefront to help with quality control, or at least establish a system to make sure the shapes are working in engine correctly.
Having the database clogged with non-working shapes will make using it pretty hit or miss, and might discourage people from using shapes there.
I would also be open to helping out with critique sessions (as Logan suggested).
Joe
02/28/2002 (12:29 pm)
I will volunteer to help out when I can. The time I can contribute is limited, but I will do what I can.I think it is important that some guidelines are established at the forefront to help with quality control, or at least establish a system to make sure the shapes are working in engine correctly.
Having the database clogged with non-working shapes will make using it pretty hit or miss, and might discourage people from using shapes there.
I would also be open to helping out with critique sessions (as Logan suggested).
Joe
#3
A player model must support a certain set of animations (eg. Run, look, ect)
must be under a certain number of polys
sized right, ect.
once it passes the QA, the rankings they use for resources could be the "community QA", good models get 4.0, crappy models get less.
I think that if we made ms3d and max skeletons with all the required anims, "programmer art" could start to be a higher quality, because they could spend time they would have been animating on making the model prettier.
L foster, I'd be interested in looking at your player models... though i am not a (good) modeler (and cant get one, hence this project) and dont quite understand how i would go from max to milkshape... Im not looking so much for the art, as i am looking for something animated that I, as a programmer, can make into a decent model. Any help or files youd be willing to donate would be much appreciated (by me and many other folks, im sure)
ryan
02/28/2002 (7:37 pm)
Well, i have access to 3dsmax (in our cs lab), so exporting and importing into max wouldnt be a problem (assuming the animation stuff is the same, i have no idea). my basic idea for, say a player model is:A player model must support a certain set of animations (eg. Run, look, ect)
must be under a certain number of polys
sized right, ect.
once it passes the QA, the rankings they use for resources could be the "community QA", good models get 4.0, crappy models get less.
I think that if we made ms3d and max skeletons with all the required anims, "programmer art" could start to be a higher quality, because they could spend time they would have been animating on making the model prettier.
L foster, I'd be interested in looking at your player models... though i am not a (good) modeler (and cant get one, hence this project) and dont quite understand how i would go from max to milkshape... Im not looking so much for the art, as i am looking for something animated that I, as a programmer, can make into a decent model. Any help or files youd be willing to donate would be much appreciated (by me and many other folks, im sure)
ryan
#4
There are no limitations as to what a skeleton can be. If you build a working skeleton, it can be made to work in Torque.
The biggest trick here would be to develop a robust skeleton that will allow animators to use. Interesting things can be done, particularly with the shoulder and hip nodes, that will help with certain types of blend animations (hip twist and look animations).
The main skeleton should be a shell of a skeleton that people can add onto if they need to.
The animation files (the DSQs) should be as light as possible, and should have only the necessary nodes in there for the animation sequence to work. If done this way, most sequences done with this setup should load, thus enabling the sharing of animation files.
An ideal situation would be to develop a naming convention that people could choose to adhere to so that all aniamtions made with the skeleton setup can be used.
This can also be scalable. If someone wants to set up a facial aniamtion system and make some canned facial animations, these can be added to the base DTS skeleton and the facial animation DSQs can be exported to use with the DTS shape. The existing animation sequences (walks and such) will not need to be updated.
This skeleton that is built should be a bullet proof animation rig, and the better it is set up for animation, the better the animations will be.
Character animation rigging threory is a subject that could fill a small book, but it would be a good discussion to start here.
For an example of a professional aniamtion rig, you can look here:
Danny's .plan
This is a rig that I would feel very comfortable using, but it might be a little much for the community to use.
Danny and I have worked together on projects before, and we can use each others rigs, but we also have learned that everyone prefers a different setup and that heavy duty animation rigs are not good for less experienced artists.
I suppose the question should be thrown out, what do people want in a skeleton that they would like to work with?
Also, the skeleton and the animations are more important here than character meshes and polycounts. A shooter might need a polycount of 1000-1500, a fighting game (ala Tekken) might have higher resolution characters (3000 polys). If the system were designed correctly though, there is no reason why the skeletons would need to be different.
Joe
02/28/2002 (11:06 pm)
There is a player model example max file included in the tutorial I wrote. Is is a MAX shape and uses a biped skeleton.There are no limitations as to what a skeleton can be. If you build a working skeleton, it can be made to work in Torque.
The biggest trick here would be to develop a robust skeleton that will allow animators to use. Interesting things can be done, particularly with the shoulder and hip nodes, that will help with certain types of blend animations (hip twist and look animations).
The main skeleton should be a shell of a skeleton that people can add onto if they need to.
The animation files (the DSQs) should be as light as possible, and should have only the necessary nodes in there for the animation sequence to work. If done this way, most sequences done with this setup should load, thus enabling the sharing of animation files.
An ideal situation would be to develop a naming convention that people could choose to adhere to so that all aniamtions made with the skeleton setup can be used.
This can also be scalable. If someone wants to set up a facial aniamtion system and make some canned facial animations, these can be added to the base DTS skeleton and the facial animation DSQs can be exported to use with the DTS shape. The existing animation sequences (walks and such) will not need to be updated.
This skeleton that is built should be a bullet proof animation rig, and the better it is set up for animation, the better the animations will be.
Character animation rigging threory is a subject that could fill a small book, but it would be a good discussion to start here.
For an example of a professional aniamtion rig, you can look here:
Danny's .plan
This is a rig that I would feel very comfortable using, but it might be a little much for the community to use.
Danny and I have worked together on projects before, and we can use each others rigs, but we also have learned that everyone prefers a different setup and that heavy duty animation rigs are not good for less experienced artists.
I suppose the question should be thrown out, what do people want in a skeleton that they would like to work with?
Also, the skeleton and the animations are more important here than character meshes and polycounts. A shooter might need a polycount of 1000-1500, a fighting game (ala Tekken) might have higher resolution characters (3000 polys). If the system were designed correctly though, there is no reason why the skeletons would need to be different.
Joe
#5
03/01/2002 (1:27 am)
Oh this sounds like a fun thing, count me in.
#6
(i dont know about 3dsmax, but in milkshape, animating is done with a 'skeleton' of joints and bones. you could, in thoery, make just that skeleton and animate it, and then let other people put the geometry ontop of it.)
support for the basic animations (run, jump, shoot, look, root, foward, backward, fall, die, theres some others i forget)
also, to look into: anims for crouch, prone, positions like that
the geometry of the base "skeleton" isnt important to me right now, as long as it has the base skeleton and has the basic animations down, Id take it and make a dozen or so player models with it. Once we have the base skeleton for player, we should look into doing the same for vehicles (esp. with the new wheeled code) although the vehicles are easy. Once this base work, the "skeleton", is done, this project should grow by leaps and bounds, as im sure alot of people would be willing to contribute some shapes, because like half the work (the animating) is done.
So my question is, would someone be willing to animate a milkshape skeleton if i gave them the base skeleton? well, if you are good at animating, i doubt you need my help with the skeleton, but, hey, i want to help... Id give you a list of animations (prolly short to begin with) and id be able to have it fully tested and such. if so, email me or reply here and ill get ahold of you.
note on max: As i said above, i know almost nothing about max, its out of my price range, and while i do have access (educational edition in the comp sci labs here on campus) i couldnt make money off of the models. Milkshape is fine with me. Maybe a MAX expert would like to take head of say, the Max side of this project, and Id do what i could with the Milkshape side. This project is going to need some work at the begining, but will get easy as it goes on (just making geometry after the skeleton is out) and i dont forsee it taking to long to get everything all set up.
note on the model repository:
now, if this gets off the ground, and we have a place to submit models, aside from a QA standpoint, a submission should list the folowing:
>File format: max, MS, or whatever plugin they have working
>A screenshot, make it easy to decide if its good to DL or not
>skeleton version: Im sure once the basic skeleton is released, someone is going to add on, remove, whatnot. Basically letting you know what work its based on (and what animations it supports)
>poly count, texture sizes (and number of textures), size (how tall the model is) and so forth
ok, theres a lot above, let me know what you guys think... if anyone wants to get to work right away on this, you can contact me by email (ackleyrj@sbu.edu) or AIM s/n: nbCaffeine
ryan
03/01/2002 (10:57 am)
Well, i know nearly nothing of 3dsmax, except that it is daunting the the newcomer. i can model well enough in it, but animating and texturing and such, thats different. Since i started the whole idea of this project, here is what i would want in a skeleton:(i dont know about 3dsmax, but in milkshape, animating is done with a 'skeleton' of joints and bones. you could, in thoery, make just that skeleton and animate it, and then let other people put the geometry ontop of it.)
support for the basic animations (run, jump, shoot, look, root, foward, backward, fall, die, theres some others i forget)
also, to look into: anims for crouch, prone, positions like that
the geometry of the base "skeleton" isnt important to me right now, as long as it has the base skeleton and has the basic animations down, Id take it and make a dozen or so player models with it. Once we have the base skeleton for player, we should look into doing the same for vehicles (esp. with the new wheeled code) although the vehicles are easy. Once this base work, the "skeleton", is done, this project should grow by leaps and bounds, as im sure alot of people would be willing to contribute some shapes, because like half the work (the animating) is done.
So my question is, would someone be willing to animate a milkshape skeleton if i gave them the base skeleton? well, if you are good at animating, i doubt you need my help with the skeleton, but, hey, i want to help... Id give you a list of animations (prolly short to begin with) and id be able to have it fully tested and such. if so, email me or reply here and ill get ahold of you.
note on max: As i said above, i know almost nothing about max, its out of my price range, and while i do have access (educational edition in the comp sci labs here on campus) i couldnt make money off of the models. Milkshape is fine with me. Maybe a MAX expert would like to take head of say, the Max side of this project, and Id do what i could with the Milkshape side. This project is going to need some work at the begining, but will get easy as it goes on (just making geometry after the skeleton is out) and i dont forsee it taking to long to get everything all set up.
note on the model repository:
now, if this gets off the ground, and we have a place to submit models, aside from a QA standpoint, a submission should list the folowing:
>File format: max, MS, or whatever plugin they have working
>A screenshot, make it easy to decide if its good to DL or not
>skeleton version: Im sure once the basic skeleton is released, someone is going to add on, remove, whatnot. Basically letting you know what work its based on (and what animations it supports)
>poly count, texture sizes (and number of textures), size (how tall the model is) and so forth
ok, theres a lot above, let me know what you guys think... if anyone wants to get to work right away on this, you can contact me by email (ackleyrj@sbu.edu) or AIM s/n: nbCaffeine
ryan
#7
I am not a milkshape user and I have no plans to start. As far as the animations go (the DSQs) it should not matter which program it was created in.
I don't think too many people are going to be making animations in Milkshape, so it might be better to have those done in MAX.
In order for everyone to wrap their heads around this better, I am going to throw out a few things...
The DTS shape and it's assciated DSQs are assembled at runtime into a new DTS shape.
The DSQ files do not need to have the mesh in it. You can use a DSQ on many DTS shapes as long as the skeletons match.
If this system is set up, people will be sharing animations across shapes and across applications. There should be no reason why one should have to sync animations and meshes (have DTS shapes and DSQs that match perfectly)
Ideally, you want to make a new DTS and plugin any DSQ you can get and have it work.
In MAX, this works fine. I have done this before with many shapes.
I have no idea about Milkshape.
In threory, if the skeletons match, the animations should load. This means the animations and meshes could be created in either program and the DSQs just applied to the DTS.
A test should be done by someone who has access to both (or two people collaborating) to make sure that this actually works and that there are no glitches in the process.
If it does, then the animations can be done in either program, and the mesh can be built and exported in either program, and the animations should still load.
It would be best (and probably quicker to get the animations done) if the rig was made and animated in max first, and then a mirror skeleton was made in milkshape (for exporting DTS meshes, provided that the tests prove the threory)
I think this is what you actually want.
If you want to go this route, I might be able to contribute an animation or two (if you are not too picky about timeframes). My free time is in short supply, but I might be able to help out here.
I would prefer if someone else built the skeletons. I just don't have the time to devote a whole bunch of energy to this at the moment. I can check any MAX skeletons to see if they are good enough to become the 'standard' upon which this system will be based.
Also, in order for this to work and be sustainable for a long period of time, it should be designed so it is not dependent on MAX or Milkshape. Max is too expensive for most, and Milkshape may not evolve with the Torque (read Jeff Tunnel's post regarding 3dtools)
Provided that exporters eventualy become robust enough to get animation working in Lightwave and Maya (with deformable meshes and all the other bells and whistles) then it should not matter where the content originated from.
If a tool for Torque is produced, or if some Mac applications start having exporters, it would be great if they had a base from which to organize their efforts.
The versioning system you have outlined might work well. I see no reason why the original version should not be robust and complete enough for all character shapes to be more or less interchangable.
Adding to the skeleton is a good thing, people remving stuff from them will be bad. It has the tendency to break things in-game. Removing bones should be avoided as some DSQs may end up having nodes that don't exist in the DTS (and they won't load).
I would also suggest that DTS and DSQ files become the base format for exchange. The source files should be available as well (inclusing a source MAIN skeleton model for all formats that an exporter exists for).
On the max side of things,
I could take on an advisory role, or something to that effect and help to fill in the gaps where they exist, but I can't devote the amount of time upfront to do much of anything and have this happen in a timely manner.
Logan, Pascal? You guys have time to lead up the MAX end?
Here is what I would suggest happens.
Someone build a max skeleton. Someone build a Milkshape skeleton that matches the MAX skeleton.
Export the Milkshape character to DTS. Do a test animation and export a DSQ from MAX and see if the Milkshape DTS loads it without incident.
If it does, then everything looks good to go and this might end up working out quite well for the community.
Also, it might be a good idea for the naming convention for both skeletons to match the Character Studio Biped nmaing convention. I have never tested this, but in theory this should also work. If the node structure and names match, someone could do an animation in Biped in MAX and someone could load the DSQs into a shape created in Milkshape (or whereever)...
This is something that would also be good to test...
Extra care should be taken to document all of this and publish it an get people to try to adopt the process. In the long term, I can forsee twenty different 'flavors' of player models in a respository and it would be better to try to limit this as much as possible or at least plan an organizational system that will allow for the dividing up of different families of shape types.
As the site grows and more content gets out there, it should be as easy as possible for people new to the engine to get content to prototype with without having to do a whole buch of research.
If this goes off, it will be defining some standards (which is a good thing) so it would be a really good idea to have this written down and accesible to al people. It also might be a good idea to hunt down the people working on the Maya and Lightwave exporters (buried in the plans and forums) and get their feedback, as these folks could benefit from conent sharing as well.
Maximum compatiability should be the goal.
Please excuse all the typos, gotta get back to work
Joe
03/01/2002 (12:36 pm)
Just a note,I am not a milkshape user and I have no plans to start. As far as the animations go (the DSQs) it should not matter which program it was created in.
I don't think too many people are going to be making animations in Milkshape, so it might be better to have those done in MAX.
In order for everyone to wrap their heads around this better, I am going to throw out a few things...
The DTS shape and it's assciated DSQs are assembled at runtime into a new DTS shape.
The DSQ files do not need to have the mesh in it. You can use a DSQ on many DTS shapes as long as the skeletons match.
If this system is set up, people will be sharing animations across shapes and across applications. There should be no reason why one should have to sync animations and meshes (have DTS shapes and DSQs that match perfectly)
Ideally, you want to make a new DTS and plugin any DSQ you can get and have it work.
In MAX, this works fine. I have done this before with many shapes.
I have no idea about Milkshape.
In threory, if the skeletons match, the animations should load. This means the animations and meshes could be created in either program and the DSQs just applied to the DTS.
A test should be done by someone who has access to both (or two people collaborating) to make sure that this actually works and that there are no glitches in the process.
If it does, then the animations can be done in either program, and the mesh can be built and exported in either program, and the animations should still load.
It would be best (and probably quicker to get the animations done) if the rig was made and animated in max first, and then a mirror skeleton was made in milkshape (for exporting DTS meshes, provided that the tests prove the threory)
I think this is what you actually want.
If you want to go this route, I might be able to contribute an animation or two (if you are not too picky about timeframes). My free time is in short supply, but I might be able to help out here.
I would prefer if someone else built the skeletons. I just don't have the time to devote a whole bunch of energy to this at the moment. I can check any MAX skeletons to see if they are good enough to become the 'standard' upon which this system will be based.
Also, in order for this to work and be sustainable for a long period of time, it should be designed so it is not dependent on MAX or Milkshape. Max is too expensive for most, and Milkshape may not evolve with the Torque (read Jeff Tunnel's post regarding 3dtools)
Provided that exporters eventualy become robust enough to get animation working in Lightwave and Maya (with deformable meshes and all the other bells and whistles) then it should not matter where the content originated from.
If a tool for Torque is produced, or if some Mac applications start having exporters, it would be great if they had a base from which to organize their efforts.
The versioning system you have outlined might work well. I see no reason why the original version should not be robust and complete enough for all character shapes to be more or less interchangable.
Adding to the skeleton is a good thing, people remving stuff from them will be bad. It has the tendency to break things in-game. Removing bones should be avoided as some DSQs may end up having nodes that don't exist in the DTS (and they won't load).
I would also suggest that DTS and DSQ files become the base format for exchange. The source files should be available as well (inclusing a source MAIN skeleton model for all formats that an exporter exists for).
On the max side of things,
I could take on an advisory role, or something to that effect and help to fill in the gaps where they exist, but I can't devote the amount of time upfront to do much of anything and have this happen in a timely manner.
Logan, Pascal? You guys have time to lead up the MAX end?
Here is what I would suggest happens.
Someone build a max skeleton. Someone build a Milkshape skeleton that matches the MAX skeleton.
Export the Milkshape character to DTS. Do a test animation and export a DSQ from MAX and see if the Milkshape DTS loads it without incident.
If it does, then everything looks good to go and this might end up working out quite well for the community.
Also, it might be a good idea for the naming convention for both skeletons to match the Character Studio Biped nmaing convention. I have never tested this, but in theory this should also work. If the node structure and names match, someone could do an animation in Biped in MAX and someone could load the DSQs into a shape created in Milkshape (or whereever)...
This is something that would also be good to test...
Extra care should be taken to document all of this and publish it an get people to try to adopt the process. In the long term, I can forsee twenty different 'flavors' of player models in a respository and it would be better to try to limit this as much as possible or at least plan an organizational system that will allow for the dividing up of different families of shape types.
As the site grows and more content gets out there, it should be as easy as possible for people new to the engine to get content to prototype with without having to do a whole buch of research.
If this goes off, it will be defining some standards (which is a good thing) so it would be a really good idea to have this written down and accesible to al people. It also might be a good idea to hunt down the people working on the Maya and Lightwave exporters (buried in the plans and forums) and get their feedback, as these folks could benefit from conent sharing as well.
Maximum compatiability should be the goal.
Please excuse all the typos, gotta get back to work
Joe
#8
If what we need is a basic human skeleton i can start working on that although i probably will need some of your help, i don't have much experiance with it.. I'm sure i can do it though so don't worry. There are lots of tuts out there and with a little help i will do just fine. I actually have 2 idle models here textured an all that could use a little animation i will gladly donate them to this project. One is the robot you see in my profile the other is the girl i made that can beseen as a work sample also in my profile. So if your intrested we can use those.
03/01/2002 (1:50 pm)
Yeah the MAX part won't be a problem Joe i will have that covered and i'm sure Logan is willing to do the same.. If what we need is a basic human skeleton i can start working on that although i probably will need some of your help, i don't have much experiance with it.. I'm sure i can do it though so don't worry. There are lots of tuts out there and with a little help i will do just fine. I actually have 2 idle models here textured an all that could use a little animation i will gladly donate them to this project. One is the robot you see in my profile the other is the girl i made that can beseen as a work sample also in my profile. So if your intrested we can use those.
#9
ryan
03/01/2002 (2:48 pm)
heh, whoa... already you've gone over my head... i have no real intrest in max (because, as i said, im a programmer, and i only model because i cant get a modeler) i suppose ill animate a skeleton myself(in milkshape) and post it and see if anyone wants to use it. I suppose ill learn some animating in the process, and make some sort of contribution to the community. If someone wants to do what i was talking about with max, hey great. Im a little (lot) busy, like everyone else, so i dunno how long it will take me to finish.ryan
#10
Wish I could talk more but I have some things I need to finish up here at work ASAP.
Logan
03/01/2002 (3:22 pm)
Ya I would gladly put in what time I can to help out on the Max side of things.Wish I could talk more but I have some things I need to finish up here at work ASAP.
Logan
#11
- Define a standardized trigger list for sequences. This will greatly help in making models usable across multiple Torque projects.
- As Joe already recommended, develop a skeletal system that can be used or created across multiple 3D applications and that will allow us to apply DSQs to nearly any model regardless if they are Biped or Bones skeleton systems.
- Offer honest help, suggestions, and creative critism on art assets so that everyone's indipendant games can be of the highest quality possible.
- Develop (or help edit/critic) various tutorials that group members can create, about creating art assets for Torque projects (my personal preference is to make these intermediate or advanced tutorials, but this is just my personal suggestion).
Well those are my thoughts on what a possible Torque Modeling Project should work towards at this point and time.
Logan
03/04/2002 (7:39 am)
Ok now that I have more time to think about this (and had a chance to talked to Pascal and Joe about it) here are my thoughts on what a Torque Modeling Project should try to do at this point in time...- Define a standardized trigger list for sequences. This will greatly help in making models usable across multiple Torque projects.
- As Joe already recommended, develop a skeletal system that can be used or created across multiple 3D applications and that will allow us to apply DSQs to nearly any model regardless if they are Biped or Bones skeleton systems.
- Offer honest help, suggestions, and creative critism on art assets so that everyone's indipendant games can be of the highest quality possible.
- Develop (or help edit/critic) various tutorials that group members can create, about creating art assets for Torque projects (my personal preference is to make these intermediate or advanced tutorials, but this is just my personal suggestion).
Well those are my thoughts on what a possible Torque Modeling Project should work towards at this point and time.
Logan
#12
03/04/2002 (7:56 am)
I've got a bunch of dts and difs, with their corresponding ms3d and map files, you gonna have an approval process for this? Where you guys gonna put this?
#13
A TMP page, and on that page have the requirements for a model to get posted, the basic skeleton files for all supported modeling platforms, an explanation of TMP, and tutorials and such on modeling.
A resources section ( think code snippits for models ) that would have the following:
A (decent sized) screenshot of the model, the model itself (.max or .ms3d and the dts), what sequences it supports (dsq or in the dts), poly count, a short description, and a rating. What should happen is, before a model gets approved, it gets a score from a reviewer(say 1-5) and models under 3 or some other benchmark get denied. Then we could slowly build a collection of quality models that all support the same animations. Ideally, we could take one of these, drop it into a pristine copy of the demo, and it would work flawlessly.
Ed: i could set up a preliminary site for the TMP and post these for download. Also, if they are done well, would you mind if i used them as a starting point on working on a milkshape skeleton? like getting rid of the geometry and setting animation names to whatever the standard ends up being.
My question is, for those who know, what are in the player model right now? my knowledge goes as far as:
root, foward, backward, left, right, jump, fall, fire, die... i think i missed one or two, but thats pretty much it. What we should look into is:
crouch, prone (at least have the ability) lean left and right, maybe a few different death animations.
i think it was lfoster who was all for the standardised animation across platforms, so also we should decide on what joints (i dont know what they are called in MAX) should be included. Standardised naming for all of the above is more important in max than in MS (if you download a skeleton, its already going to be animated for you... right therei n the ms3d file) so, ill let those who know max more than i do decide on that.
whew. long post. if i forgot anything, lemme know. And let me know what you guys think on the resource section stuff, i want to get that submitted to GG as soon as possible.
ryan
03/05/2002 (12:01 pm)
OK guys, im workin on a proposal to send the GG guys to host the models for us. Basically i prepose the following:A TMP page, and on that page have the requirements for a model to get posted, the basic skeleton files for all supported modeling platforms, an explanation of TMP, and tutorials and such on modeling.
A resources section ( think code snippits for models ) that would have the following:
A (decent sized) screenshot of the model, the model itself (.max or .ms3d and the dts), what sequences it supports (dsq or in the dts), poly count, a short description, and a rating. What should happen is, before a model gets approved, it gets a score from a reviewer(say 1-5) and models under 3 or some other benchmark get denied. Then we could slowly build a collection of quality models that all support the same animations. Ideally, we could take one of these, drop it into a pristine copy of the demo, and it would work flawlessly.
Ed: i could set up a preliminary site for the TMP and post these for download. Also, if they are done well, would you mind if i used them as a starting point on working on a milkshape skeleton? like getting rid of the geometry and setting animation names to whatever the standard ends up being.
My question is, for those who know, what are in the player model right now? my knowledge goes as far as:
root, foward, backward, left, right, jump, fall, fire, die... i think i missed one or two, but thats pretty much it. What we should look into is:
crouch, prone (at least have the ability) lean left and right, maybe a few different death animations.
i think it was lfoster who was all for the standardised animation across platforms, so also we should decide on what joints (i dont know what they are called in MAX) should be included. Standardised naming for all of the above is more important in max than in MS (if you download a skeleton, its already going to be animated for you... right therei n the ms3d file) so, ill let those who know max more than i do decide on that.
whew. long post. if i forgot anything, lemme know. And let me know what you guys think on the resource section stuff, i want to get that submitted to GG as soon as possible.
ryan
#14
From what I can see of your outline, you wish to create a site where people from various project's can download models, thus creating a library site that is the likes of Polycount or Turbosquid. Although this would help out a lot of projects, I fail to clearly see how artist's can get a real benefit from it? I also believe that a site full of model assets is already possible through the e-commerce setup that GG has going to sell games.
Don't get me wrong, I agree that the creation of a model library would be a valuable asset, but at this time most of the artists in GG are too busy creating content for thier own projects to contribute towards the creation of a library (unless it suited their interests).
What artist's really need at this time is a set of standards for Sequence Triggers and a cross-application skeletal setup that can use a wide range of DSQ files. These two tasks alone are very daunting and will involve much debate and hard work, they should not be taken lightly.
The only other feature that I can see artist's requiring at this time is the ability to maybe get a few tips or tutorials to help them out a bit, get support for problems that they encounter, and get some honest critiquing of their work.
As I stated, my biggest concern right now is that I don't feel that your idea of the Torque Modeling Project is really geared towards artist's, but towards gathering a free library of art assets for people to pick and choose from. This likely would have the negative result of keeping artists away from the project :(
I think your best bet would be to get in touch with Pascal, Joe, myself, and anyone else who has the time and ability to contribute and let us pick one another's brains, so that we can get this idea flushed out so that it can benefit the artists first (and then in turn benefit their project's).
Logan
03/05/2002 (5:40 pm)
Ryan, perhaps I have a different idea of what the Torque Modeling project must do than you do, this is why I must disagree with your proposed outline.From what I can see of your outline, you wish to create a site where people from various project's can download models, thus creating a library site that is the likes of Polycount or Turbosquid. Although this would help out a lot of projects, I fail to clearly see how artist's can get a real benefit from it? I also believe that a site full of model assets is already possible through the e-commerce setup that GG has going to sell games.
Don't get me wrong, I agree that the creation of a model library would be a valuable asset, but at this time most of the artists in GG are too busy creating content for thier own projects to contribute towards the creation of a library (unless it suited their interests).
What artist's really need at this time is a set of standards for Sequence Triggers and a cross-application skeletal setup that can use a wide range of DSQ files. These two tasks alone are very daunting and will involve much debate and hard work, they should not be taken lightly.
The only other feature that I can see artist's requiring at this time is the ability to maybe get a few tips or tutorials to help them out a bit, get support for problems that they encounter, and get some honest critiquing of their work.
As I stated, my biggest concern right now is that I don't feel that your idea of the Torque Modeling Project is really geared towards artist's, but towards gathering a free library of art assets for people to pick and choose from. This likely would have the negative result of keeping artists away from the project :(
I think your best bet would be to get in touch with Pascal, Joe, myself, and anyone else who has the time and ability to contribute and let us pick one another's brains, so that we can get this idea flushed out so that it can benefit the artists first (and then in turn benefit their project's).
Logan
#15
We work on setting up set of standards for Sequence Triggers for the community to use. (by sequence trigger, im sort of confused, are you talking about the actual sequence names [eg: run, whatever] or are you talking about the code to play it [eg: playcel("celdie"); or whatever]) now if im correct in my thinking, you are referring to the first, and if thats the case, once the standard is set, we work up an example shape (nothing too pretty, look at the example player shape for milkshape that tim (i think) worked up... heh ) and let the artists pretty up the model as each team needs. Then, if people want to submit their work to a central model database, thats fine, if not, then they have the example shape, at the minimum.
on a side note, a modeling question:
can a milkshape exported dsq (that doesnt have any animations) use a max exported dsq? Its my thinking that its possible, if the joints are named the same as in max and are the same size and whatnot.
Ryan
03/05/2002 (6:06 pm)
Heh, you are about right, my idea was to help out indie teams that couldnt get a modeler. My original idea was to have a fully animated milkshape skeleton that any idiot could throw some geometry on and have a functional model (even if its ugly :) But i dont fail to see your side, i agree that artists (and would-be artists) do need some Tutorials and walk throughs and standards for all the reasons you stated. Maybe we could do both in one swoop:We work on setting up set of standards for Sequence Triggers for the community to use. (by sequence trigger, im sort of confused, are you talking about the actual sequence names [eg: run, whatever] or are you talking about the code to play it [eg: playcel("celdie"); or whatever]) now if im correct in my thinking, you are referring to the first, and if thats the case, once the standard is set, we work up an example shape (nothing too pretty, look at the example player shape for milkshape that tim (i think) worked up... heh ) and let the artists pretty up the model as each team needs. Then, if people want to submit their work to a central model database, thats fine, if not, then they have the example shape, at the minimum.
on a side note, a modeling question:
can a milkshape exported dsq (that doesnt have any animations) use a max exported dsq? Its my thinking that its possible, if the joints are named the same as in max and are the same size and whatnot.
Ryan
#16
The reason for developing a standard skeletal setup is not just to give users a high-end skeleton system to drop in and begin using, but to also give them a skeleton system that will work with any existing DSQ file. An added benefit of having a standard skeletal system is that should models be shared or liscensed out they will be incredible simple to integrate into your Torque project. You can read Joe's previous comments for additional reasons and specifications that need to be aimed for in the creation of the skeleton.
Sequence Triggers are values that are sent to the engine when they are hit, when the engine sees one of these triggers in the sequence animation it can then associate the trigger to something. For example, you could trigger the exact frame the left or right foot hits the ground. There are a few important reasons why a standarized sequence trigger list is beneficial 1) it allows artists to assign the same triggers to actions in their animations, 2) programmers will have an easy time associating triggers to something in their code, and 3) it allows shared or liscensed models to be brought in with minimal fuss.
As for your question regarding DSQs, as long as the names of the skeletal pieces are the same, there should be no problem applying one DSQ to another DSQ (hence the whole reason why we create a standard skeletal setup).
Logan
03/05/2002 (9:13 pm)
Ryan, although giving users a setup skeleton rig would be quite useful, it is unfortunately not as easy as dropping a mesh on top of it unfortunately :(The reason for developing a standard skeletal setup is not just to give users a high-end skeleton system to drop in and begin using, but to also give them a skeleton system that will work with any existing DSQ file. An added benefit of having a standard skeletal system is that should models be shared or liscensed out they will be incredible simple to integrate into your Torque project. You can read Joe's previous comments for additional reasons and specifications that need to be aimed for in the creation of the skeleton.
Sequence Triggers are values that are sent to the engine when they are hit, when the engine sees one of these triggers in the sequence animation it can then associate the trigger to something. For example, you could trigger the exact frame the left or right foot hits the ground. There are a few important reasons why a standarized sequence trigger list is beneficial 1) it allows artists to assign the same triggers to actions in their animations, 2) programmers will have an easy time associating triggers to something in their code, and 3) it allows shared or liscensed models to be brought in with minimal fuss.
As for your question regarding DSQs, as long as the names of the skeletal pieces are the same, there should be no problem applying one DSQ to another DSQ (hence the whole reason why we create a standard skeletal setup).
Logan
#17
Also, it should not matter what type of skeleton it is.
A stanbard bones skeleton named like a Charactert Studio biped (exported to DTS) should be able to load a DSQ animated with Biped and vice-versa
An animation or shape made in Milkshape should be able to load or be loaded in any other shape.
Again, this is in theory. It has not been tested fully, and it should be.
If it works, and a standard convention can be set, it should work not only for Max and Milkshape, but for Lightwave and Maya and whatever other 3D app gets an exporter.
It should not take too long to work though it and test it, but it should be tested before a standard gets that doesn't work properly.
Ideally, a model and it's animations are done in different files, so you could put the animations directly in the Milkshape file, but I don't see the value of doing that if the loading of DSQs can be made to load in any character DTS shape file.
Total cross compatability is the ideal situation, as it allows everyone access to the greatest amount of content.
If done correctly, the programmer should be able to get working shapes and animations without ever touching a 3d tool, and mixing and matching different animations can be done in script, wihtout having to re-export.
Joe
03/05/2002 (10:12 pm)
Ryan, the answer to your question is 'maybe'. In theory, it should work, but it has not been tested.Also, it should not matter what type of skeleton it is.
A stanbard bones skeleton named like a Charactert Studio biped (exported to DTS) should be able to load a DSQ animated with Biped and vice-versa
An animation or shape made in Milkshape should be able to load or be loaded in any other shape.
Again, this is in theory. It has not been tested fully, and it should be.
If it works, and a standard convention can be set, it should work not only for Max and Milkshape, but for Lightwave and Maya and whatever other 3D app gets an exporter.
It should not take too long to work though it and test it, but it should be tested before a standard gets that doesn't work properly.
Ideally, a model and it's animations are done in different files, so you could put the animations directly in the Milkshape file, but I don't see the value of doing that if the loading of DSQs can be made to load in any character DTS shape file.
Total cross compatability is the ideal situation, as it allows everyone access to the greatest amount of content.
If done correctly, the programmer should be able to get working shapes and animations without ever touching a 3d tool, and mixing and matching different animations can be done in script, wihtout having to re-export.
Joe
Associate Logan Foster
perPixel Studios
I was actually planning on releasing my Max bones skeleton that I am using for Captain Zap, a list of the bone names (for those of you with Milkshape who want to duplicate what I have done and use my DSQs) and all of the basic DSQ actions. This way people will have some generic files that they can load up and quickly (as long as they want something prebuild and basic).
Another feature I would like to bring up is/are the triggers for sequence objects. I definately think that we need to create a basic list of which triggers are assosiated to which actions. I feel this is important for two big reasons 1) if someone loads a model from the free model library, they will not need to worry about finding out which triggers are associated to which actions, 2) this is something that is very important for each and every Torque project as triggers can be used with each and every project and its something that artists and programmers should know.
A final feature that I would recommend is a professional community where artists can get honest and constructive criticsm or suggestions on their work.
Logan