Discussion in 'General Discussion' started by bakayoyo, Feb 18, 2017.
Let's go vulkan
I used to be all excited about RR changing to UE4, but after ACC I really hope Sector 3 comes with a different solution. Good lighting and weather effects, as good as they can be using UE4, don't mean much if the sim is otherwise a huge mess of blurry pixels and popup shadows/LODs.
Driving Paul Ricard in ACC and looking at the moire effect going on with the track lines in the distance and the tree LODs changing right in front of me, just makes me feel gratefull that RR is "still" running this old but very clean and sharp looking engine.
Same here. UE4 can look really good, but it's not "free", given folks tend to tweak a lot for custom titles. The biggest benefit is the tools for artists and tons of features for animation, effects etc. that go along with it. But while originally I thought a move to ue4 was great choice, after personal experience coding with the engine, my opinion has changed.
The main issue imo is that if you deviate from the way ue4 "works" (physics, tripple screen etc.) you need experts in the engine's low-level design. However it's success is mostly from people focusing on the game, not messing with its tech (which you s3s would have to do, to keep its old physics etc.).
Granted if you do it yourself, you also need experts, but ideally you "grow" complexity yourself as you need to (in theory).
What are the benefits moving to the ue? And which of those could also handeled by keeping the current one and expand to dx12?
I don't like the graphics of ACC too.
Yes, disiking the graphics in VR of ACC, too.
But what we need is a thing, which enables all the cores of our processor.
It's a real problem with a processor, where one core is on more then 85 percent. R3R is really struggling then.
Do you think, DX12 can solve those issues?
I really like the ISI engine. It is build for racing games and really smooth, not like UE4 with its troubles.
Is the engine using all of the cores? As far as I know is that not usual in games industry up to now but I can be wrong. I just read somewhere sometime...
To be honest I don't know the capabilities of dx12, but I see all the issues with ACC in a full grid (and this is not 100 cars, but 20+). All graphic settings on fire it struggles a lot....really a lot.
Because of that it would be interesting for me to understand where the bottlenecks are right now and how die UE can solve that. And if there would be also an option (what perhaps doesn't exist cause Sector 3 propably have their route map regarding the engine) to cover this by going up to dx12.
Going vk would give you benefits within Win7 still over dx12. going dx11 is "cheaper" for a developer, however, as dx12/vk require more work on the developers side, but offer the latest and greatest tech etc.
UE4 offers gazillion features, cross platform... so it makes "generally" a lot of sense to use a ready-to-use solution. Likewise there is tons of artists and freelance developers that are trained with it. Content production is typically the highest cost in gaming, so there are your benefits why to choose ue4, unity over something custom. It's impossible to do as much features compared to those companies having large teams building the engine and the tools.
Now that said, if you lower your ambition a bit, and because you develop something "purpose-built" you can still make up for that but you must be aware that anything that adds more complexity can become harder and harder to pull off. "wet surfaces", what about wet pit crews, what about entering dry pits for repairs etc.
ACC is still early access, hardcore performance tuning is mostly likely done towards the end, so I would not judge things now. It just shows that investment needs to be done. And we live in a wold of post-release performance update patches
It was mentioned in older interviews that R3E doesnt use ISI for graphics, but has its own render engine. They mostly used it for physics but also improved that themselves.
The key difference from the "old" engines (rf2, r3e, ams, iracing, ac) vs the new ones (pcars2, acc) is lighting and shading systems. Which is why the older engines will always look a bit "cheaper" than the new ones when it comes to rain & dynamic lighting. So without major investments in rendering/shading/effect tools for artists
As for multi core usage, that is really up to the developers.
In the end all this is a matter of costs, and in all cases (ue4/unity, or improve your own engine) you need experts to do the job, or you need to invest time to build up expertise yourself over years.
Thanks for the explanation. In the end my hope is it will not lose its own style and the picture will remain as crispy as it today is.
I am not excited at all. Every UE4 engine games I have these two dramatic options to me as an end user: it looks really good on a high end pc (which I don't have) or it looks really (I mean REALLY) ugly and still performs bad. I'm yet to see a UE4 game which convinces me that it is the next best thing since sliced bread.
if we can have day night and weather and keep the physics we have then id go for it , i just dont think UE4 can offer that , id prefer to stay as we are until we can keep the physics side than go all sparkly and moist
got exited when i hear ACC got UE4 but after see the results i am more afraid when R3E does this step...
As has just been mentioned its maybe ok on high end pc, mine is an average pc and i just cant run high graphics and as a result game looks bad, if i use epic settings it runs really bad with loads of flickering and stuttering. Hoping that final optimization will help enough for it to be playable.
even on epic it is not epic...
AA settings are to limited in UE4,mirrors looks horrible etc etc...
then better not to have night/rain and a "clean" graphic where you can see something on the track (and got a working triple screen).
It will take two more years for all the sand to fall.
Meanwhile let's enjoy R3E.
In my devious head, I think S3 delivered some fake news a while back, how GT3 and RR would be making the leap to UE4. With AC having the backing of 505, they jumped the gun and announced ACC using the Unreal 4 engine and are in the process of delivering said Sim. S3 and Simbin get a good look at what can be done with it, then make an educated decision on whether UE4 is the right way to go or not. Saved S3 a bunch of dollar on R&D.
On a personal note, there is not a lot wrong with what we have visual wise, that of course is not to say things can't be improved but I have faith S3 will make the right decision when they finally make the leap.
Some time in the future is a fairly safe, if not ambiguous statement. Having seen 'prettier' sims that deliver a significantly less driving experience than R3E, I am content with what it delivers now.
Out of curiosity does graphic representation have any effect on physics, or vice versa, when it comes to allocating PC resources?
I'm totally satisfied with R3E graphics. It's not only sharp and clear but also pretty. But on the other hand, weather changes, day and night cycles, track grip changes... these features would add really much to general realism of R3E
Strictly speaking not. However the larger engines often come with their own built-in physics representation (they are complete engines, not just rendering) and high level representation of objects is wired with that, therefore replacing or changing those to something you had before, can be tricky. Which is why the lack of moving pictures about GTR3 or their slightly less hardcore sim statements could mean there is some obstacles.
Stumbled upon this good article
I'm pretty sure @J-F Chardon mentioned somewhere that they would prefer to use the existing physics in any leap to DX11/12, so if I was a gambling man I'd bet on RR adopting a hybrid engine, S3 physics, enveloped in a DX11/12 framework. I swear when Reiza & S397 made the leap sod all changed for the end user, it just opens up the developers environment and allows for a broader selection of tools to enrich that environment.
Separate names with a comma.