Rfactor 2 Beta - A comprehensive constructive review.

Granted allot of the graphics are not using advance methods and new shader techniques , but in the end its the end result that matters. Shadow of the colossus on ps2 is still one of the better looking games purely becuse of what the art team did with it to me RBR also stands out as a fantastic looking game.

Another advantage of using Older methods is that the game will run faster and still be compatable with older hardware.

Putting allot of the weight of artistic output on the artists rather than uniform shaders also allows remote art teems to work with the game and have control over the end result without having to speak to an internal programmer.

You can make valid arguments for both methodologies and work flows . I do agree that for the overall lighting and the shadows DRX10 would probably have run faster for most users on 1-1.5 year old Graphics cards , but allot of RF2 users are still on old pcs.

Personally I prefer realism to come from asset density and I like games to be sharp and crisp , one of the things with lighting used in DRX10 games is that it can be very saturated and quite a strain on the eyes ( i have found this with Cars , Bf3 , F2010 ,Dirt 2 ) and the saturated and over the top lighting is often to hide the pore resolution of consoles but then carries across to pc.

In the end people will like different things , the piont with Rf2 is that its workable and gives enough for mod teams to build onit.

and even if you think it looks awful with no chance it will get better from mods or after the beta the core game play looks like it will be solid and that's what keeps players coming back , CSS is still one of the most played games on pc and that looks like lego land.

 

I cannot agree with you.

1. I'm not talking about "advanced methods". You can achive Fresnel effect from DX9 SM1.0. So since 2003.
2. It really doesn't depends on further DX versions. DX9c (SM2.0) is enough. Newer version gives no additional fps or so.
3. What I'm talking about costs almost nothing (in matter of fps as well as programmer work)
4. What I'm talking about saves gfx creator time, and allow him to stay out of all technicals
5. using universal methods results with consistent gfx appearance, which is may be independed on how artist set it up. means all tracks will look same, in matter of relation between colours, lighting, reflections etc.
6. Those methods doesn't add saturation or other eye candies just by design. Artists do that. Do not mix these.

And finaly, without it you cannot create 3d scene changed its appearance in relation to spectator and light sources position, getting real-looking scene all the time . Depending only on artistic vision, you may get wiped out scene like iRacing, or cartoon-like coloured one like in rF. IRL you may get both.. depending on where you stay, where is the sun, how much blue sky is etc etc.

Most important problem is that a lot of people think that making real-looking gfx is like building space shuttle, resulting NFS-like looking images. Wrong.

Ha.. one more think. Do you think that rendering light glows for 40 cars on track, with use of "fake" 3d objects is "cheaper" than the same (or better) effect resulting from single pass, scene wide post-processing effect? I don't think so. It is not true that new approaches will overload old machines. Often make things more efficient.

time to go to sleep.
best regards

 

Finally...a thread with TRULY CONSTRUCTIVE criticism, and logical arguments (how long will that last, I wonder ). Most people think that "constructive" criticism is just any critical comment that isn't meant to be hurtful. Constructive criticism isn't just saying "it's wrong, fix it", it's saying "it's wrong, here's why, and here's how to fix it."

Regarding the graphics: I have to agree with Maxym. Things like glow effects have been commonplace in games for years now. I'm a bit surprised we haven't gotten it with rF2 yet. However, as we've been told before, we're still in beta, and even once it reaches 'gold', it will still be somewhat of a WIP. Although, I doubt they'd make a change as radical as that without overhauling the entire graphics pipeline, since it would mean that you would have to go back and edit every single car and track to remove the glow objects, and create new 'glow maps' for anything that is supposed to have a glow.

I think the one thing that really stands out to me as being drastically different about rF2 so far apart from everything that's been said already, is how bumpy the roads are. Maybe this is due to the filtering of the FFB--I haven't tested the proposed changes for the G25. But, by and large, every single track feels MUCH bumpier than any other sim I've played. The braking zones and apexes seem to be rife with vehicle-unsettling bumps. I'm not saying that it's 'wrong', and I'm fully aware that we're still in Beta; but I do think that, combined with all of the cars being mid-engine, RWD, is a contributing factor as to why so many people have felt that the cars feel tail-happy when compared with other sims on the market.

 

Actually there are other ways to do this. You could have a separate bass layer, even two that do not pitch shift to the same rate as the main samples. Deep bass wouldn't even need to be pitch shifted at all as it wouldn't be in real life, that's why it's still there.

 

I agree with everything that's happening in this thread.

Especially the part about the shaders. I keep looking at the gloves in rF2 and they just look wrong. Oversaturated cartoonish gloves that are near-constantly glowing. The shadows on the clothing are much too dramatic. I think this may just be a problem with the alpha layers though. Reducing them should help a ton.

 

I don't disagree with the points you have made , I was more making the point that it can go both ways dependent on the resources and choices of the dev team , both ways can work out in the end.

A Talented artist would be able to do well in ether case. I defiantly think You are right that using more modern methods on the code side would allow a less proficient artist to produce better looking work.

 

Yes, talented artist will be able to provide glow-like effect, using some workarounds. But doesn't matter how talented he is, he will not be able to get correct reflections (Fresnel) or correct-looking lighting. Simply because these efects depends on angle spectator watching them. These are fully dynamic. If not done correctly on engine level, cannot be changed on customize one (by artist). Unfortunately these things are mainly responsible for how result will appear: as real-like picture or as artificial one.

If you still don't believe me, I will provide you pictures of what artist never achive withous proper support of rendering engine.

best regards

 

Great Review, thx for that.
May i'am be allowed to quote that review to an other forum?
there ar people, who are not rgistered in any forum.

 

Very constructive review with some great points... hope ISI guys are combing through all these threads.

 

Nice review and I agree with almost everything except for the understeer part. I think that the lack of communication from the FFB is because of your wheel of choice. I own both a g25 and t500rs. With new games I always test on both wheels just out of curiosity. I noticed that with the g25 I wasn´t experiencing the distinctive scrubing of the tires during understeer as I felt with the t500rs. In fact when I first driven a couple of laps with the t500rs the first thing I noticed was the scrubing because it was so different from any other sim I had driven. Tweeking the FFB settings made it come alive even more.

 

Yep, low speed grip seems to be difficult for just about every simulation out there and it's one I'd like to see being improved upon. A remnant from Pacejka thinking, even if his formulas aren't used anymore?

 

I know what you mean not having some things force artists to do stuff that's maybe more stylized and ultimately wont be as realistic as if you had shaders using better underlying math allowing for a proper more dynamic output.

I think part of it for me comes down to the fact that we are still at a point that even the best looking DRX 10/11 games still look stylized and real time graphics are quite far from photo real. So in my mind so long as the art is internally consistent that's what matters to me for immersion and suspension of disbaleaf.

The biggest thing I found in Rf2 for immersion and practical driving is the asset density allowing me to set braking points without thinking and also getting a far grater sense of speed as objects fly past.

There is probably some good reason why they haven't implemented shaders and some effects as you have said , I have worked with some game engines in the past that were a nightmare for lighting and with the Gmoter back ground in pure simulation the engine might use really non standard ways to do certain things making some specific FX and shader implementation harder or more time consuming.

Some engines I have worked with were really modular and a coder could come in and upgrade core parts right away with no issues then others it would take a coder months to do even basic things.

I would love to Visit ISI and talk about the core workings of the engin and how they got to where they did what factors influenced them and how the business side interacts with the creative productive side.

Internally building game engines scares me to death now from a business point of view. With the availability and low price of engines like unity that seems to be where most people are going now. Unless they are some huge company or a crazy indi dev

 

Or just from Eastern Europe/Russia. =)

 

The assets density are very well done indeed. For me it is one of the most important things.

The glare may could be done by one one post process pass... I am curious about it too, why rF2 faked it?

 

I think there's a not covered aspect in car physics in all simulators available which is the transition from the mechanical grip to aerodynamic grip. As the cars are at low speeds, there are no wing effects to add grip, then the cars just feel loose and without control.

Some bad physics RF1 mods exaggerates this wrong effect. Some prototypes mods are just laughable at this aspect: the same car is capable to contour a 350 km/h curve but it spins immediately if you try to do a 40 km/h corner at 41 km/h. No mechanical grip at all.

However, this is just wrong. At low speed the car don’t have the aerodynamics help to pull the chassis down, but there’s a lot of mechanical grip owing the friction (you’re below the slip angle of the tires). This problem was in GTR1, RF1 and looks like present in the RF2 new tire model. GTR2 seems to deal with this aspect in a better way than the other softwares.

 

I think GTR2 just fudged it by having overly gripy physics in general not that the math behind the phics was any better it was just a case of having more overall grip meant that the cars were then more stable at low speed, it also made the cars a bit stale and joyless to drive as it was harder to get them to nose in on the brake or fling them around like you can in Rf2 / Net Car pro.

maybe ISI can just fudge it by having it so that as car goes from higher to lower speed an extra artificial grip value is added but only takes effect from 20mph down to 5mph lol.

 

I'd rather they not fudge it, but do a realistic solution. Low speed handling is always going to be a tricky part when doing physics.

 

well thats interesting.
If thats the case, it would be nice if ISI could confirm whether understeer is supposed to be directly conveyed in the FFB, and if so, what form is it supposed to take.
maybe theres something us g25 users need to adjust to magnify whatever youre experiencing.

 

Definitely , but it seems noone has worked out how to do that yet , and in the end if the "fudge" is more realistic and works then it works.

Presumably to get physics working well and running in real time its going to use lots of work arounds and massive oversimplifications , I would love to have a go on a professional grade simulator using one of there insane tire models see if evan then it has the same issues.

 

no problem.
my only intention is for ISI to take note, really, and respond: whether the things ive said can actually be done by the game engine, either by them, or by the modders.