Thats because the more I investigated what was required the more issues I found. Briefly, these are some.
1. logic mapping of players like in the maze game.
2. visualisation of tunnels/rooms in 3D and walking towards/away from walls in perspective. Turning around by degrees in a room of objects.
3. Nice 3D rooms but what about the 2D players
In short, a lot of work and not very nice results so far.
BTW for a sample of some of the best flash I have ever seen (including a great chess game) see here http://media-division.com/games.htm
Now thats the challenge - to be able to create quality stuff like this with 3dfa.
I did some work on players as child movies, communicating from parent and other child movies using parameters and while that may be usefull in some scenarios its more trouble than its worth for the single player game, Maybe in future XML socket servers the child movie might come up again, controlled by external XML parameters.
So here is a quick look at a tool I created to assist in animating a 3D player. I have decided to have a fully articulated model which will have controllable arms legs torso etc depending on needs.
Such a player will then be required to
1. Walk left/right/towards/away
2. Use tools like weapons
3. Facial expressions etc
For each action you will need selectable images which must still remain connected to the other parts of the body (or else they will float away). The set of movements for say raising a weapon and shooting can then be stored in an array and played back as a subroutine e.g. ShootGun() FallDown() etc.
So for each discrete action walking,shooting,falling there will be a set of array co-ordinates and a subroutine to execute them. The whole shebang is then controlled by you the game player causing actions like walking through a tunnel and encountering the shooter (because the map array for the maze has shooters wandering around like the hunters in my SaveTheGirls game)
Its all too easy huh!
The attached zip file contains a dancer.swf demo of a figure dancing across a stage with articulated legs and feet and a tool to allow you to experiment with moving femur/tibia/foot of a two-legged body along a path e.g. walking or running, with up to 50 steps of motion. There is a default set of data which can be played or single-stepped and you can clear the array and program the limbs individually step-by-step and then copy the results to the clipboard and paste them into a new movie like the dancer.swf Note: The animator and dancer are in 3fda version 3.72
A fully mobile head/torso/arms/legs model will follow this weekend when I have completed it, and will use slider controls to move the limbs rather than the edit boxes used here.
I dont know where this is all going yet but hopefully if the theoretical model works others may be interested in adding to it.
What I am aiming for is a modelable humanoid figure which can execute actions like turning, aiming a weapon, falling down etc and controlled by events during gameplay by you the viewer.
So on entering a room, the figure will be reactive and in say a shootout the player survives and the humanoid falls or otherwise.
The tool for modelling all limbs is coming along and code for navigating a room is too but its all a matter of time (and space!)
Why don't you just make movieclips with motions, for instance:
standing
walking
walking away
shooting
running
falling
dead
etc
and have the different movieclips float around in just 2d. That's pretty much what Doom was about. That's why - if you've ever played Doom in a network - when you died, you would actually turn around (while dead, but not restarted yet) towards any players in the room, because you basically were a 2d drawing of a 3d man, projected in a 3d sphere. I don't know if that makes any sense.
The objective I am aiming at is a better mode of programming the movements of any humanoid/machine object and actionscripts are the way to go. As the racecars here prove, In any movie, the more motion created in transforms, move-events and other built-in methods the more costly in full-length presentations and gaming because the file size of the swf file grows rapidly.
its the same technology motion-programming. Imagine this was a web-advert for a new hispanic market product with advertising text appearing on-screen during the dance. Complex child movie-clips are not easily choreographed and the output files are huge. Thats why implied motion is often created using a few gif sequences dragged across the screen e.g. a 2D-view of a spaceship with engines flaring.
However if, like StarTrek, you want to have a multi-winged space vehicle with pods and antennae, swoop and present theirselves in 'real-time' the array way is the only way -
Shortly after posting my last response here, I have grabbaed a gif off the internet to do another effort using the same tool. Thats how quick it is to modify for different applications.
Here is the result, which would better if I added further moveable elements and also a mirror-image which is required if we wanted to produce the correct perspective as the turn continues,
This example also highlights one of the deficiencies, that the model is still really two-dimensional. An example with a cube rotating would work poorly here, and notice that Annabella is always one-face perspective, so for turning face-on to the X-wing and then turning further, requires about eight perspective views for 8 x 45 degrees that the model turns and more still if the model has a roof and a floor.
At this point scale and x,y,z co-ordinates start to lurk in the animation, however for a lot of flash games this will still work better than ray-tracing, morphing or motion-events, primarily when the game has the smarts to control the anime.
It sounds a bit like overkill if you ask me. I think just the idea of making a Doom like game using flash was outrageous enough, but to give such rich detail to some reasonably simple movements sounds like overkill. I thinking primarally on file size and framerate, is it going to flow nicely when you have all these moving elements? Wouldn't it be better if you focused on getting a 3d engine up and running before starting on character animation? Once you have a 3d motor, then starting with simple characters and building them up from there I think perhaps would be a more suitable way of building up this game.
I dunno, what do you think?
And that's how I got here.
Years ago I was using 3dfa v3.42 sans actionscripting and Swish 2.0 sans actionscripting. Last year I investigated DarkBasic which truly is an animation tool for the gamer world but the results are .exe which many folks wont download over internet and cant integrate with a browser html world.
Now with imminent Swish v3.0 and 3dfa v4.0 we have user friendly internet savvy tools. Just last month I upgraded from v3.42 to 3.72 of 3dfa and re-visited swf gaming. I made the SaveTheGirls maze game based on a sketch by Blanius and from there a series of 3D walls which led to a 3D room holding 2D characters from the maze game.
So, allowing for some future code refinement the ability is here to have
1. Complex 3D figures performing canned activity like walking, falling controlled by subroutines
2. Simple 3D room representations with floor plan and object positioning
3. Overall control by Actionscript maze /adventure logic
Result DarkBasic style adventure in an swf at reasonable file size.
Easy, no, but worth pursuing. Incidentally, none of the above precludes using child movie clips. Just as an artist uses different brushes and styles to fool the eye, think of any of the good flash swf work you have seen with innovative 'clever' use of whatever the medium let the author do.
I can see a tool for manipulating a 14-jointed human figure with a 16-jointed close-up tool for a human hand. Now create a slightly different tool based on the same array logic which can allow 3D visual control of an object like a domestic automobile with opening doors. Once built these tools allow easy positioning and articulation of the object(car,human,hand) and ease of replay until the action is smooth and visually correct (instant replay/single-step is one of the arrays best features), then saving to a swf automobile sales advesrtisement where a car arrives on-screen, the door opens and Annabella steps out, then close up to hand and face shot.
The current tools for gorgeous 3d image creation like Swift 3D will still be used by pro-shops along with Adobe and MX but I want to develop low-end techniques for 3dfa/swish
Heres the challenge for items such as vehicles, starships, cubes etc - to animate in 3 axes.
The solution is to still use the array to hold the choreography but using multiple views of the item broken down into running-left, running-right top,bottom etc and then assign a datapoint plus a hide/show visibility for each element.
In this intermediate sample View_1 shows left-to-right while Views_2 and 3 will handle right-to left and top/bottom views
Also a single slider controls rotation and a single brass ball for dragging, selectable by using the checkboxes. This is not a properly functional tool yet but I'd like your inputs
Still thinking along the lines of 3D and everything...
Here's a quickie solution to visualizing the 3D world into which games etc could be placed. I was looking at a chess game over the weekend and wanted to see the pieces on a board which could be rotated or moved around. Lots more work to do and most important how to make the black-white squares. Should be capable of Doom use too. I animated the position of one piece only, leaving the other two static.
Finally for all its visual impact look how small the swf is. Its in v3.72 of 3dfa
Reply With Quote
