Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - cjbruce

Pages: 1 ... 14 15 16 17 18 19 20 [21] 22 23 24 25 26 27 28 ... 49
401
Im pretty sure that he means that the gyro is increased

Ohhh!  Now I understand what you are talking about!

The gyroscopic effects that we are seeing are a direct result of the high angular momentum of the spinners that we are making.  I put in a compensator called a “Spinner Mass Reducer” to reduce this effect and improve drivability for spinners.  If you add a wheel, a spinner mass reducer is not added, so you get the bad gyroscopic effects.

402
Yeah so ever since the 24th build when the weapon falls off, the blur is still there. Even worse, the blur does damage... at this point, I gotta decide between moon gravity and ghost spinners for SvL. I mean... I could always use October 10, but people probably won’t like that. I’m at a dilemma at this point.

Shoot!  Let me look into this.  I definitely want to keep the spinners.  I just need to iron out the bugs like ghost spinners.  It shouldn’t be too hard.  I just need to find and fix all the wired edge cases.

What is the moon gravity issue?  Can you send me an .RR2Bot file to play with?
Moon gravity is nothing much really, its just an thing when you attach a wheel then the weapon on it

I’m not sure I understand.  If you attach a wheel, there shouldn’t be a blur cylinder or any of the associated physics.  Does something weird happen when an object breaks off?

403
Yeah so ever since the 24th build when the weapon falls off, the blur is still there. Even worse, the blur does damage... at this point, I gotta decide between moon gravity and ghost spinners for SvL. I mean... I could always use October 10, but people probably won’t like that. I’m at a dilemma at this point.

Shoot!  Let me look into this.  I definitely want to keep the spinners.  I just need to iron out the bugs like ghost spinners.  It shouldn’t be too hard.  I just need to find and fix all the wired edge cases.

What is the moon gravity issue?  Can you send me an .RR2Bot file to play with?

404
The 28December build is up!

https://robot-rumble.itch.io/builds

This one should allow you to apply decorations without physics colliders. 

Next up is getting chassis armor plates to break off.

Enjoy!  :smile:

405
I'm building the following changes:

[Added] Added a new armor material called "Decoration".  This material has very low density and almost zero strength.

[Changed] Reduced the hit point threshold for disabling colliders to 0.001 HP.

With the two changes above, it is now possible to "paint" a robot using shapes.  If a shape is assigned the "decoration" material its physics collider will be ignored and it will break off with its parent component.

406
the one thing that may be a tad concering, is that if there's no limits to the stats and stuff those components can have, what's preventing someone from having an invincible tinyaf brushless motor with more power than an E-Tek, or having some kind of giant SMEEEEEEE wedge with 3 trillions of HP, you may want some kind of limiters so that some people don't start making an arms race of making the dumbest, most overpowered sh**

oh and other thing, is there any way to fix the bug where the AI just goes in circles like an idiot when it drives over the closed pit ?

I dunno.  How do mods work in any other game?

I suppose people are free to create their own experience with the game as they see fit.  It is kind of self-limiting though.  A crazy powerful motor would be completely undrivable.  A piece with infinite health would quickly make the game trivial and pointless.  You could say the same thing about the millions of Super Mario Maker levels that are completely impossible.  That is part of the fun for some people.

For tournaments it is simple enough to give people a set of acceptable components.

As far as the AI question goes, I’m aware of it.  AI is going to take a complete rewrite as well, now that I (kind of) know what I’m doing.  We can’t begin to tackle AI until all of the inputs are ready.  Specifically the damage system.  AI needs to be smart enough to attack things intelligently, and this is entirely dependent on being able to see which components have what amount of damage.

407
You say that people are gonna be limited by their skills only. Well if thats the case, i might be able to make vrushless motors then!

Exactly. 🙂

408
I thought you might like the idea.  :smile:

409
What if we added a workshop for components? What is made in that workshop could have combined hp, and you could also apply that part to multiple bots without having to do it all again. You could then share stuff like letters and logos with the community.
I don’t know how well this would work, but with my current knowledge, couldn’t you create .partfiles kinda like what you do with the robots? Then you can share your designs that way. Also, all components you make or import could show up under custom weapons and custom extras. This is what I mainly had in mind.

That’s the basic idea behind the RR2 Component Modding Tool.  We are starting work on it and the Arena Builder at the same time. 

The intent for the tools is for them to be separate Unity projects that you can use to create pretty much anything you want, export it as a unity assetbundle, then import it into the game.  The big advantages with this approach is that you will be able to create literally anything you want, limited only by your 3D modeling skills. 

Also, you should be able to use the new tools to create parts that are optimized for the game.  If you were to use our existing in-game tool to create an intricate compound assembly it is going to be extremely inefficient and run very slowly.  I am already having lots of problems trying to run Panic Attack on my laptop.  This isn’t the fault of the robot.  It is an amazing robot.  It is a problem with the tool which can’t handle making really complicated things that run efficiently.

I recommend using Blender to create things.  It’s free and has a massive open source development effort.  It already does everything people have asked for, and is the tool we have been using to create robot parts for the game,

410
Worst case scenario: Make a material called decorative, it would have a 0kg multiplier and it would remove collision, also it wouid be good because then it would fall off when the part that its connected to would fall off.

Im still liking my idea more, as then we could fix the issue where spinners with a lot of small parts would be fragile

Ooo!!!  Great idea! I can make a “decorative” material and get this up and running in very short order.

Lets try it first.  The rest is going to take some work.

411
What if we added a workshop for components? What is made in that workshop could have combined hp, and you could also apply that part to multiple bots without having to do it all again. You could then share stuff like letters and logos with the community.

That’s the idea behind the RR2 Component Modding Tool.  The big advantage to the tool is that you can do things that aren’t possible using the existing tool, like make game engine-optimized parts: a proper high-poly to low poly workflow + PBR texturing.  Building good-looking stuff with the existing in-game tools means that the parts are going to have way too many polygons and way too many physics colliders to be performant in the game.  I want the game to be playable on laptops.

412
Ok so the new update is bad. It breaks robots with thin and small parts, it seems like you tried to do something with group coupling and it didnt really worked that well.
Can i recommend this? Its a longer thing so it is in spoiler below:



Give me a bit to think about it. 

I think it is important to disable as many of the really small colliders as possible in order to minimize the CPU load. 

I think I would like a hybrid approach where you set a threshold and it shows you which parts are too small to add colliders.  With this approach you can always bump up the thickness or switch materials for a particular part to ensure its collider stays enabled.   

This would also allow you to do a bunch of parts really quickly, and reset things instantly. In the case of Panic Attack, this would make it possible to rapidly adjust over 100 small pieces.

413
The driving issues are just as bad as before, if not worse. D:
The skirts now come off way too easily too, just a couple of low speed collisions of the kind they're designed for can snap off the whole thing. And the part that actually runs on the floor doesn't have collisions, so makes the whole skirt pointless.

Shoot! Need to work on things some more.

414
I put the collider disabling feature up as the 26December build.  I am seeing significantly reduced time spent doing physics calculations for Panic Attack.  Hopefully this makes robots work better across the board.

CyarSkirata, please let me know if this fixes the driving issues you were seeing before!

https://robot-rumble.itch.io/builds

EDIT:  With the smaller colliders disabled in this build, we are getting component grouping for free.  Things are breaking off as entire assemblies rather than as tiny little parts.

415
So it looks like the automatic solution is working!

I set an arbitrary limit (5) on hit points.  Most of the robot's colliders are now disabled.  Time spent in physics has dropped significantly.

Moreover, when something big breaks off, it takes all of its tiny little decorations with it.  I will try to get something up for you guys to see today.

Maybe we can do another slider to control the number of hit points before a component becomes breakable in settings?

Please note that this is just a temporary fix.  It will only work for a single weight class.  Beetleweights have way fewer hit points, so the entire robot would be disabled.

416
Ah... Y'know I said less detailed Panic Attacks didn't need the wheelfix?
Only sometimes. It seems like the problem is still with either the double-gearbox or the total ratio of 125:1, because all it takes is a moderately complex opponent to push it over the edge back into passive ultra-gyro.

I'm taking a look at both now.  A few things I've noticed so far:

1. Neither of them have any Blur Cylinders at runtime.  This is good.  It means that any colliders that exist are there because you placed them there.
2. Both robots are running at roughly the same frame rate.  This is good.  It means that they appear to be processing similarly.
3. The regular version has 189 components.  The wheel fixed version has 191 components.  This is consistent with the wheel fixed version having 2 additional foam wheels.
4. Both versions are running at about 15 frames per second when in battle with one other robot in the editor.  This is uncomfortably slow for me.
5. Most of the frame time is spent doing physics calculations.

Some thoughts:

- 15 frames per second is an unpleasant experience even when things are working properly.  At this speed the AI control signals cause a ton of overshoot and the robots drive like they are intoxicated.
- I suspect that the problems you are seeing with the weird gyro might be due to the fact that the physics system frame rate is dropping so far that colliders find themselves teleported into physically impossible positions.  The system then tries to recover by applying ridiculously large impulses.
- I think the key to increasing the frame rate will be to reduce the number of colliders.  I suspect Panic Attack might have 300+ colliders.  I would like to cut the total number down to around 30.
- For a first cut, I'm going to try disabling all of the colliders for components with extremely small numbers of hit points.  If it works, then I might pursue @kix's suggestion to group components.  I would like to try the automatic fix first before we commit to building more UI and reworking the .RR2Bot file.

417

Yeah, the wheelfixed versions have normal handling everywhere.
These are the two versions of the current tournament model, wheel-fixed and not. I was gonna remove the paintwork from them for the sake of your CPU load, but this time - which surprised me given it never worked for the previous versions of the problem - removing the decals did actually fix the in-battle handling problem. So this time the decals do undoubtedly relate to the issue.
  [ Quoting of attachment images from other messages is not allowed ]    [ Quoting of attachment images from other messages is not allowed ]  

At this point, I've been wondering if the problem might relate to the super-low gearing too.
So at this point, I imagine the issue would probably be mostly if not entirely fixed by the whole part grouping thing.

Taking a look now...

418
I'm hoping to ramp up development in 2020 and bring on a few more people to the official dev team.  It would be great to get things into a beta state by the end of 2020.

Community Manager
Component Modding Tool Developer

Is anyone interested in the above?  Any other ways that you have been dying to contribute?

Please send me a PM if you are interested!

Would it be possible for us to contribute by, for example, submitting ideas or designs for house robots? Idk how many pre-built robots you want to have in the finished game, but more is always better, and across a wide range of weapon types and weight classes.

Mind you, maybe it's too early to be thinking about stuff like that, considering the number of breaking changes that keep getting introduced...

House robot and arena design suggestions are always welcome. 

The problem with house robots is that they are currently built in the Unity editor.  They don't use the in-game Robot Workshop at all.  Every time we add a house robot, I have to make sure it works more-or-less the same as an equivalent robot in the Robot Workshop.  This is a MAJOR drain on my development time.  I always feel like I should be spending that time on making sure you can do things in the workshop and have them work in a battle scene.

That's why we need somebody working on the RR2 Component Modding Tool.  My vision for the tool is that you can build any components you want to any level of fidelity that you desire.  Because it is a workshop component, it will slot right into the workshop.  You can then use the parts to create any robot you want.

Example #1 - Lets say you want to build a tube with a hole.  Build it in the RR2 Component Modding Tool.  Then you could make robots with holes in them.
Example #2 - Imagine that you needed brushless motors.  They are incredibly powerful for their size and you really need them for the replica of your real-life beetlweight.  Build the motors you need in the RR2 Component Modding Tool.  Then you could make robots with brushless motors.
Example #3 - Imaging that there was this incredible real-life heavyweight with awesome paintwork and really cool actuating mechanisms.  Build its subassemblies in Blender, texture them in Substance Painter, then import all of the pieces into the game via the RR2 Component Modding Tool.  You could then assemble all the parts and use that amazing new robot in the game.

The bottom line is that a good RR2 Component Modding Tool could massively speed up the development of the game, and allow the community almost unlimited freedom to create new things.  We could really use a dedicated person working on it.

Any volunteers?  :bigsmile:

419
Tbh i want to help, but im fine giving you feedback like we all are doing on gtm.




Now how about we talk about Chassis damage, part grouping, tube components and brushless motors?

Love it!  :claping

I like the order that you put them.  Chassis damage is not too intimidating.  I just need to get the plates breaking off.  I like the idea of part grouping, but it will take some UI work.  Tubes and brushless motors are a little lower on my priority list, but totally doable.

420
I'm hoping to ramp up development in 2020 and bring on a few more people to the official dev team.  It would be great to get things into a beta state by the end of 2020.

Community Manager
Component Modding Tool Developer

Is anyone interested in the above?  Any other ways that you have been dying to contribute?

Please send me a PM if you are interested!

Pages: 1 ... 14 15 16 17 18 19 20 [21] 22 23 24 25 26 27 28 ... 49