What can I say? After looking at this model "from inside" for a while I found it mostly (in term that nothing is absoulutely) perfect and decided to support this model once again. As a result... here it is:
Nice, isn't it? (I admire "the design" more than "the look", actually - when I look at it I can easily believe that it CAN and WILL shoot). That is why my modifications were very light: I enlarged wheels a bit and somewhat changed their disposition (to give the whole machine more stability). Thats it.
Well, of course, we can smooth it a bit to use for icons and this kind of stuff:
So what do you (I mean EVERYONE) think about it? Should we use this model or make another one?
How does the catapult look viewed from the side? Cause that's the actual perspective you'll see in the game. So we have to care that it looks good in this perspective.
As for smooth look: I think it is the best applied to the shooting section (the one that creates hurling force). Wheels look odd, I agree (do not think that clonks with their level of technology are able to create something THAT round anyway). Better leve those octagonal wheels (though I plan on adding small spikes to them - that will give the catapult more "aggressive" look)
I think we stick to the edgy model above. Anyway, the great look can be achieved by a great UV-map (only if the model looks good from the side)
As I said, I like your catapult! But while it looks cool on the title, the side view (ingame view) looks quite odd. One can't see that these strings are supposed to be from the bow of the ballista. Additionally,
After looking at this catapult for a while I came up with a thought that I kinda like its construction. And so I made myself to model it as it is. Here is what I've created:
Well, this is not exactly "as it is" because I could not see some details (to be more precise, it was the core part of a catapult which made the whole thing work) so I imagined them myself. The appearence is rather weird but it definitely looks better from the side (which will be the main ingame angle of view) than Simsi's one (sorry, Simsi, I really liked your design):
I even made a primitive animation to illustrate the shooting process:
The model is a bit high-poly (411 faces, which can be reduced some more - just too tired to go on). Also I didn't have time to make an UV-map for it (I think there will be a lot of comments/remarks/criticism concerning the design, so it is a bit too early for this anyway).
Well, I'm off. Goodnight to everyone. See you tomorrow...
P.S. I think of adding some kind of a rope which will fix the moving part in place (like the one Simsi had) but will definitely not implement it tonight...
P.P.S. Will have to modify it a bit.
- The wheels must have mor polygons. When the wheels are going to be animated, you should not be able to see at 3x that the wheels are not round
- The grey stuff on the other side of the basket is a big phat rock as a counterweight. The arm must be longer so that the heavy rock (could also be bigger and more rock-shaped) is further away from the axis, bearing more power through the Lever-effect. This change is very important, otherwise the catapult looks pretty odd, to me at least.
- Also, the base frame on which the arm lies could be higher
- For a crank-up-animation, a crank like in the previous model would be nice. The shoot-animation would then result in the cutting or untangling of the rope.
> the catapult won't work like this
You forgot another remark:
There is absoulutely NO force which will make the moving part move UPWARDS, thus it will always stay at the lowest point. I have two ideas on how to fix this... I could draw sketches for both of them but I wouldn't (better I fix it right away and show an already complete model).
> The arm must be longer
I tried to make the counterweight part longer but in that case it would collide with with the supporting framework (the X part). It can be avoided by increasing the height of the construction (though some experimentation is needed).
> The wheels must have mor polygons
Luckily it can be fixed in no time :-) The question is how many angles? 16? 24? 32? (currently there are 8 as you can see)
> could also be bigger and more rock-shaped
As for bigger see above (longer arm). As for rock-shape... What about polycount? Already made it as low as possible and it will certainly increase after all modifications which are to be made (for example, crank will give additional (2 + 8) + 8 * (4 +
> cutting or untangling of the rope
At first I thought about a locking cog which is released by pressing the lever but then I realized that we probably do not need such a detailed representation so lets stick to the "rope variant".
Thanks for productive criticism (I really appreciate that)
>You forgot another remark:
Er no? OK here is how I imagine it (and how a catapult works): the counterweight (the rock) is much heavier than the basket, even when a heavy rock is in the basket. So the basket is always on the top, only by pulling down the basket (by rope that is attached to a crank), you get it down but there is a lot of tension. So if you cut that rope on which you pulled down the basket, the rock will immediately crash down and thus catapult the contents of the basket towards the sky.
>As for rock-shape... What about polycount?
The guideline is: it should look good on 3x zoom, not a fixed polycount. But a rock-shape is no that important as long as one can reach the same effect with textures.
>crank will give additional (2 + 8) + 8 * (4 + 21) = 50 faces
It'd be possible to have the whole crank as a sprite. But for now, just model it. It could also be a lever attached to some cogwheel which has to be pulled down repeatedly for cranking it up
>At first I thought about a locking cog
Yeah, also an option. Would explain that you don't need a new rope for every shot ;-)
> offsetted
Towards which direction?
P.S. A brief illustration would be nice (it is always nice to support every idea with a visual representation)
Not bad from the side either:
And some perspective as well:
For those who have a good Internet connection I also prepared a brief shooting animation.
So this would be more logical:
The details on that "rock" could be also in the texture.
Ok. So now it is time for Concept#3... :-)
Sorry, I didn't have inspiration to draw a rope but you can easily see a bobbin for it as well as a tiny hook to which it should be connected:
Ok, you really did it :-) Now I do not know which one of those two catapults I like most.
> The details on that "rock" could be also in the texture.
I think I will replace it with a mere cube when the concept is finally approved and the time comes to reduce the polycount (right now there are ~750 of them (o_O) )!
Ah, yeah, an animation would also be in order:
> thecrank of course needs to rotate much faster
Yes, of course (I made the simplest possible animation just to illustrate the process)
> Can you post the model?
I wanted to make a lowpoly first. And also a UV-map. Planned on doing that today or tomorow.
P.S. But then again... why not?
- About polygon reduction and -count: rectange polygons count twice because iirc rectangles are broken down to triangles anyway. The polygon count of the catapult is 140% as much as the airship. Look at the wheels for example at 3x zoom and consider if they could have less polys after all.
- Did you read the tips here? http://wiki.openclonk.org/w/Modelling_Workflow#UV_unwrap - identical parts can be mapped to the same position of the map
1) The catapult is not an airship. It is a complex mechanism which should not only work but actually look like it can work. So there are small detailes there which can not be removed that easily (like crank, lever, axis and so on).
2) The reduction of polycount can be achieved by simplifying things. For example, insted of 3-dimensional crank we can have a mere sprite (the same was done for the hook to which the rope should tie) which will greately reduce the polycount. The lever can also be introduced like a single sprite... The question is: which objects are safe to simplify in such way and which are not (for instance, the wheels will definitely look weird if they have no dimension).
3) I thought that catapult will be rendered not only from one side (lets say, left), but from another (right) side as well. So there are no "rear" polygons actually which can be safely removed
4) I have never implemented a 3D-object into the game and do not know which polycount is allowable there. So it is rather difficult to guess how the final model should look. Nevertheless I was hoping to recieve some remarks from more experience members of the community there are a lot of improvements which can be done.
5) ...I've heard briefely about 4->3 transformation but didn't have any information considering this issue.
6) After all, wheels CAN have less polys. I'll see what I can do
My thoughts about texturing:
1) When I thought about identical parts it was too late - I've already posted a message.
2) Anyway I was not hoping that in would be approved right away. I fact I knew that some remerks would certainly be received.
3) I would probably make another UV-map after further polycount reduction.
So thanks for the remarks. I really appreciate that.
EDIT: Right now I am working on polycount reduction. Think that improved model will be right away in a day or two.
All in all: It's not THAT important, 1400 polys is not a big deal, all the tips about the UV map in the wiki are just tips, they are not required. It can also be done later or not at all. Your UV map is very good by the way
>I don't know if or how transparency for textures is implemented.
Yes, it works very well now (by use of PNG alpha). However, one must make sure that backface culling is disabled in the object's Scene.material, otherwise planes will disappear when they are facing away from the camera.
> However, one must make sure that backface culling is disabled in the object's Scene.material, otherwise planes will disappear when they are facing away from the camera.
And if you use a submesh for that then you don't have to disable it for the rest of the mesh.
By the way, is stuff like specular, ambient etc. used in your rendering engine?
> In the Scene.materials, you can have different materials right? So the parts which shouldn't be backface-culled could have that one material, the other parts the "standard" material.
Yes. See the bow for an example.
> By the way, is stuff like specular, ambient etc. used in your rendering engine?
Yes, it is.
The things that were altered:
1) Simplified arm
2) Simplified bobbin
3) Wheels now have 16 faces on edges instead of 32
4) Crank is flat (has no dimension) now
5) Lever is flat as well
6) And the bowl is also flat (completely forgot about that...)
Need some approvement from community members (maybe some detailes - like crank, bowl and lever dimensions - should be returned) before I can make a UV-map for it.
I hereby license the following file(s) under the CC-by license
Powered by mwForum 2.29.7 © 1999-2015 Markus Wichitill