By making it so that reflective objects can show other reflective objects and themselves in reflections. If placing a camera on a mirror will take more than 1 month, I don't know how this would be harder.
"making it so that"... sigh. it seems you have no idea how reflections are actually calculated. the projection that you see in the reflection is taken from a single point in the centre of the car, if you were to make the car visible, all you would see on the outside of the car is a janky projection of the car's interior panelling, seats, etc. if it were as simple as "placing a camera on the mirror" they would have done it way back in 2014 or something, it would be harder because you'd have to completely rethink the way reflections work. the only practical way I can think of is to implement a whole raytracing system, but that would be an absolutely massive undertaking, and with current hardware it'd only be usable for probably 20% of the userbase at most. and that would likely only be used on the cars, not the maps, so there'd be very little general graphical improvement, the only improvement would be seeing your own car in the mirrors.
Not sure why some entirely reasonable suggestions for improvement on this forum devolve into some members arguing so hard against any improvement. It's unnecessary and serves no purpose. Good, functional mirror reflections are nothing revolutionary. Just look at American Truck Simulator. It has multiple high definition mirrors with no degradation in performance. BeamNG does a lot of things great, and even the graphics are really nice, but its mirror system is stone aged in comparison and really in need of improvement, any improvement.
That's pretty much almost every ideas/suggestions thread. 10 billion arguments will be made against requests including Just use mods use the search bar 2 hard 4 devs It'll never be added Bad idea but I won't explain why I don't like it because How old are you And then there's the 1 compliment for a detailed concept Good idea OK The guys that shoot down the "bad" ideas will not complement detailed concept suggestions. People treat scratches like a resource hog, something impossible to implement, and then pretend gta 5/gta 4/any good racing games don't have car scratches. The argument against that? scratches won't work with crash physics have people not seen gta 4 car crashes? welcome to ideas and suggestions, where everything is deemed impossible.
I don't want to derail this thread so much and you definitely won't listen to this and will just instantly pretend that I said "bruh it's impossible but I won't explain why fjeksvirhrvekens". BUT I want to correct your absolutely broken argument about scratches. Which kinda applies to mirrors too. BEAMNG IS FUNDAMENTALLY DIFFERENT THAN GTA, ATS AND WHATEVER OTHER GAME YOU CAN THINK OF. IT IS A BOTTOM-UP SOFTBODY SIMULATOR SO THINGS THAT ARE EASY AND CAN BE DONE IN 5 SECONDS IN DIFFERENT GAME ENGINES ARE AN ABSOLUTE NIGHTMARE TO CODE IN BEAMNG. Part of me just thinks the devs delay the scratches and stuff simply because they are scared of the amount of work they will take which could absolutely freeze all other development. Sure "they could just make it so scratches appear on the surface that deforms what is so hard about it". Well just look at some "mods" from bootleg websites that claim to have working airbags and scratches. It looks extremely cheap on them and doesn't come anywhere near the acceptable level of quality for this game. To properly implement scratches we would need a system that checks for sliding friction forces applied on every surface of the car in real time and then renders scratch textures on them using the 2nd UV map, it would obviously have to calculate the direction and size of the scratches too, it should work different on metal and plastic, also in water or mud, etc. Attempts at this have been made by some modders but it's extremely tricky to say the least, if not even fundamentally impossible without game engine changes. "So why not just make an easier system then???" Because the quality would suffer and it would look cheap, and the devs are all about delivering the highest quality product possible, if a feature doesn't reach their standards they just remove it and restart from scratch later. How do mirrors tie to this? You can't use the argument "If other games have this then it must be super easy!!!!!". Because it might be easy to add new cameras or anything, but it comes out as computionally intensive. Remember that this game suffers a lot from what people call "lack of optimization" which probably for the most part is extreme amount of complexity and calculations it has to do in real time, which just lowers the FPS naturally. If they implemented the mirrors now the way you want, the game would become so laggy it would be unplayable even on the best PCs. So they have to either optimize it a lot, or think of a less laggy solution (ray tracing is probably planned but that can also be extremely laggy). Feel free to add "Bad idea because it would look cheap" and "It would be too laggy" to your list of "invalid arguments", I don't care. Also on another note: do you really think that in GTA scratches work well with crash physics?? That's a joke, watch some videos of GTA races and you will see how bad it is.
This thread has devolved into arguing because nobody will listen to a real technical explanation on anything and people keep bumping a thread where the topic has already been concluded previously just so we can bicker back and forth But reflections yo? It's not complicated. Just put a camera bruh... But taxing performance. It's not that simple, current systems aren't implemented to do that the way you think. Picture in picture (multiple camera perspectives at once displayed to one end user) isn't possible. It's why split screen is also not a thing... (in current Torque3D) Then everyone repeats the argument back and forth because nobody will listen to actually people who have a clue how difficult of a question people are trivializing constantly. Also funny thing about GTA... Correct me if I'm wrong but the mirrors in that game are just cube maps that look like stretched vaseline. Completely unusable and show no valid information what so ever to the driver. Hmm even GTA couldn't pull that off so easily... They didn't even bother with their insane budget and team.
1. I read messages(in detail) 2. GAME ENGINES ARE A BASE FOR ACTUAL WORK, THEY ARE NOT THE FINAL PRODUCT. AND ALSO CALM DOWN ON THE CAPS LOCK 3. They can make different departments for each project, jbeam department, texturing department, the misc department, the programming department, and the features department. features department, programming department, and texturing department would work on scratches. There, development won't freeze in time for 1 day. 4. The AI for this game started as an artificial idiot, didn't know a single thing about pathfinding, and crashed its car 10 times. Now, the AI uses tactics, slows down, utilizes lanes on top of paths, has multiple modes, and also catches up quickly. scratches could start as burnout marks on the fenders/doors and Improve over versions. 5. mud on tires is a different idea, and another heated, rude, debate. 6. the only guy that I could find was some banned user who had a scratch dot png mod in the works, but no download link. 7. removing and starting from scratch is only applicable if that code does not even resemble the original idea, or is far from it. They probably don't start with a new skin texture after finding a small coloring error. 8. again, people with the highest settings in grand theft auto 5 have had no problem increasing vehicle density to 100%. 9. The less optimized games are optimized as time goes on. If it's about/has been 10 years, why is optimization lacking? for example, gta 4 was resource intensive, gta 5 has been certified steam deck playable. If there's a lack of optimization, then the game should be optimized(duh) 10. Alright, so putting in ray tracing calculated for everything loaded is less laggy than being able to see your car in the rearview mirror? great. 11. invalid arguments are arguments that aren't related to the subject, or are strawman arguments. I can't count the amount of times I've seen messages linking to search queries, they tell you to not bump, then they tell you to not spam. plus the things like 'way too hard to implement' 'why?' no answer given. 'just use mods' linking to mod that went outdated 1 year ago. side response: here's a clip from destruction nation This game had good crash physics for an open world game and then this Those are dirt marks, and scratches seem impossible huh?
Why don't you calm down on the bold and why doesn't anyone use enter to create paragraphs that are legible? If you have multiple statements break it up so people can understand what you are saying. It really helps deliver information in a digestible manner to the reader you intend to view your words. That way you can stop using bold so much too.
So pretty much People arguing for eternity The game engine can't handle simple tasks vaseline mirrors Okay then
Hmm perhaps someone working on a system that could set back many other teams wouldn't work so perfectly smoothly as you presume. But do go on, I'm sure your the most well-rounded person on the entire forum who understands game development better then anyone on this 8 billion person rock. --- Post updated --- *Vaseline mirrors in GTA V* I said the game engine doesn't have that feature implemented and they'd have to build it into the engine. I did not say the game engine can't handle simple tasks. (Even if I hate the way T3D handles UV/Material/Textures)
> Comes to a random bumped thread > Drops his uneducated idea > Ignores all counterarguments > Fails to recognize how game development work > Argues everyone else is wrong > Expects a soft body sim to have every feature other games have Absolute madman
Also thanks for rehashing the history of BeamNG AI. As a long time player since day 1 I've clearly forgotten that history. It's also irrelevant because it's not a graphics engine task and doesn't conflict with artists and other teams you talked about so much. It's a computational task that you implement into the engine separately which can be handled by a team that really should have far less conflicts and issues. All they'd have to do is try to write the best AI algorithms that are efficient and functional. Develop it further and further with those same goals as time goes on... In theory that doesn't conflict the dev team elsewhere. Invalid argument really.
the main issue I have with what you're saying, is that you claim something to be easy when you have no idea how the current system, or a theoretical system, is practically implemented. there may be some peculiar graphical cheat that the devs are aware of and intend to implement, but from what I know, and from what I've heard from them in the past, there is currently no feasible way to get more accurate reflections, the current system, bodgy as it is, is as good as it gets. theoretically, if they were to implement in the game engine a way to have more than 2 camera angles rendered, you could have more accurate (but still slightly bodgy) mirror reflections (as in ETS), and have dynamic reflections on multiple vehicles at the same time, and have split-screen multiplayer, but despite all these possibilities they haven't gone after it, so it must be more convoluted than it seems.
They said "Comes to a random bumped thread", not bumps a random thread. Hmm I do believe previously you said you read messages.
There is support for many angles at the same time, but it must be way more complex than that since they aren't used on mirrors
When you bump a thread, you have to visit that thread, so you've visited a thread you bumped. I'm peacing out of this 'arguing for eternity' thing. also, I have to read your message to reply to it. You're from day 1, not the user I was replying to. there are people from 2020 that exist. Different teams for each thing means development won't freeze until a project is finished. Alright, where did I say I'm the most well-rounded person, or that I'm smarter than 8 billion people? And jeez, calm down with the attitude! Times I used bold: 3 what exactly is wrong with bold? it's just a darker color agent_y committed vocabulary against humanity by having the whole message in 1 paragraph(not line breaks), and I read the whole thing to respond. then me not pressing enter 2 times made mine unintelligible. I disagree.
Peace guys Some insight from behind the curtain for you all: The current mirrors work by chance and use the reflection of the cubemap we generate for the dynamic reflections. The cubemap (see also https://en.wikipedia.org/wiki/Cube_mapping ) is generated at low framerates and in the center of the player vehicle, hiding it in the process. Proper mirrors will have to render the scene from another viewpoint again. To have them working really well, we will need to hide certain parts of the car and have a special frustum for the mirror as well, depending on the view angle to it. There is a lot of complexity in the topic, not to mention non-flat mirrors even. All this complexity makes it hard to implement and the additional frame to render is a hard hit on performance (think as in every additional render halves the framerate) So yeah, its on our TODO list, but its a lot of work and research. However, with the new renderviews implementation we are going a step into the correct direction. Its a matter of time and resources if and when we get to it really
Proposed solution : Wait 5 years for the devs to do it, along with scratches, dirt, etc. I'm sure it's on their to-do list, but they again it's not the only thing on there and they first have to deliver on promised content before they pursue bigger projects like PIP mirrors and other stuff like that. Also remember that BeamNG is constituted of many employees, but that maybe only 2 or 3 of them are assigned the necessary role for making working mirrors and scratches. For now my guess is that they're working on something else that's more important and actually in their near-future scope. We're not inside the BeamNG headquarters so any attempt at defending or turning down those ideas makes very little sense. EDIT : Welp I guess tdev replied before I had the chance to press enter lol, at least you're getting official insight now
I used bold only because you did to prove how unneeded it is. If you'd pay attention I have this thread watched because I posted in it some 18months ago... I did not bump anything I wasn't already previously a part of...
