RFactor 2 Graphic Engine

[Max Angelo]

Agree, but your post is a bit out of context.
We are talking about 'why NFS:Shift has nice gfx; why rF can't have it'.

Mate, i think if i am out of context you too was, becouse my post was just a reply to some points you raised in this topic.

In my opinion, comparing EA with ISI is pointless for several reasons, for istance becouse the different market nichle of the two companies.

Where can I find an info that it is not official supported feature. It has been implemented in rf 1.150 and fixed partially in 1.250. There is a note that there are still issues but no one say that is not supported feature

ReadMe.chm, chapter What's new in v1150

<<<Fixed some issues with driver swapping, although feature is NOT fully implemented or tested, and there are several known problems. >>>

Not fully implemented = not officially supported

[MaXyM]

In my opinion, comparing EA with ISI is pointless for several reasons, for istance becouse the different market nichle of the two companies.

Is it the reason that we have not rights to compare gfx engines?

ReadMe.chm, chapter What's new in v1150
<<<Fixed some issues with driver swapping, although feature is NOT fully implemented or tested, and there are several known problems. >>>

Don;t know what version you race, but we have 1255 already
In readme for this version there is no such statement.

Not fully implemented = not officially supported

Max, please... A lot of things are not fully implemented (means has bugs/issues). Is that mean that half of rFactor is not officially supported?

[Max Angelo]

You have the right to speak about what you prefer, but i have the right to say it is a pointless discussion.

Comparing a graphic engine dated 2005 with what?
A 2005 EA game? The last EA game?
Or perhaps you want compare WIP screenshots? Another pointless thing in my opinion.

ReadMe.chm chapter "What's New in v1250" says:

<<<Improved driver swapping functionality. Garage setup is now transferred during multiplayer driver swaps if the server's multiplayer.ini variable "Driver Swap Setups" is enabled (which it is by default). Vehicle state is also transferred, including fuel, tire wear, and damage. Various other minor improvements and fixes have been applied to this feature.>>>

And indeed we all agree rF1255 is improved in the driver swapping, but the key words are:

Various MINOR improvements and fixes ....

So, v1150 notes + v1250 notes describe in the proper manner the feature status.

I dont read anywhere -v1255 (but the last What's New chapter is the V1250 one) has a fully implemented driver swapping feature-.

Also, i can assure you that, between 1150 and 1255, ISI never ran internal tests about the driver swapping.
How can they support a feature tested only by very few trusted leagues, with evident comunication problems, no strict coordination between testers and devs etc?

Again, driver swapping is a feature added "so like it is", just in order to allow to endurance drivers a more realistic experience, even if in a incomplete status.

Last, bugs are a different thing than a not fully implemented feature.
It is no rocket science to understand the difference.

[LesiU]

...and I have the right to say, that you both are right and wrong in the same time

No, really, guys, in fact the truth lies somewhere in between (as always, doesn't it? ). It's not hard to see that community (the modding part of it) is a bit frustrated because rF was provided like "as is" without much real support. I'm a pretty young modder, without much of an experience. But what was a bit of a shock to me at first time, is when asked one question to one of THE top rF modders about one parameter that I was working on about "how it works" and he said "I have no idea, but i do it like that....". Come on, after 4-5 years how there can be something not known to anyone? Ok, might be, because there is no official documentation... that can happen, right? So what should I do then? Ask developers, about how that parameter works. But here's the problem again, because there are lots of simple (yup, about general things) questions that still are not answered at all. So how we should do our modding stuff, without good support? How I should know what such parameteris doing, when I'm not the one that wrote the code? Max, that's what we are talking about lack of support.
You said that rFactor is not supported anymore. So what we should do. Stop making mods and tracks for rF1 and wait unspecified amount of time for rF 2?

The more we have to wait for rF 2, the more we are restricted with rF 1 limitations. We want to organize events, we want to make mods and tracks but what was acceptable a few years back, is not that "OK" anymore and it's not just only my opinion. So yeah, I also would like to see a new revision of the platform, payed one, but with fixed bugs and updates. Making a product valid for a good few years is not good in my opinion, because after that few years it will be very, very outdated. It's like with upgrading your PC once a few years. the older it gets, the more frustrated you are because of lack of performance. It's better to do smaller upgrades but more often. Drop in performance won't be that big. And if you put that on rFactor level, community will be happy because they are getting a product that is not from an ice age ( ;-) ) with its features. Instead, it is more or less up to date. I think that some newer features could have been be easily implemented into newer patches (payed or not) for rF 1. This has one more benefit, I think pretty important for you, the dev team - you have reports from very large community, about how particular features works.
With upgraded rF 1 engine there could be a patch with updated weather system. Then, with reports from the community, it could be probably easier for you to develop more advanced one, that we probably will see in rF 2.

Of course you have your own way of doing things and I respect that.

As for the gfx engine... what is somewhat killing rF 1 from that point of view is not outdated engine itself... more it's about, that it's not fully utilizing features from SM2.0 model and using some old techniques instead of newer (for that time) and better/more optimized. The same can be with rF 2. It can be new, but if the engine will be using some weird and old techniques then community won't be able to have that much use of it, like they could.

Cheers!

[Max Angelo]

I am not a modder, but i know some of the "old" modders, who to your question "how it works" they say "I have no idea, but i like that" ... in other words, modding and experimentation are strictly linked. To find the solution to a problem is part of the fun, an exercise of creativity.

But, in the years, many great tutorials were released, for istance

"The rFactor HDV file and it's role in defining the car", by Bristow
http://www.racesimcentral.com/forum/...fining-the-car

"The rFactor TBC file and it's role in defining the car", by Bristow
http://www.racesimcentral.com/forum/...fining-the-car

"The rFactor engine.ini file and it's role in defining the car", by Bristow
http://www.racesimcentral.com/forum/...fining-the-car

"The rFactor PM file and it's role in defining the car", by Bristow
http://www.racesimcentral.com/forum/...fining-the-car

and you can find in the net other tutorials, for istance @ The Pits forum.

About the track making, Uzzi posted in this forum his "Track Building Tutorial"
http://isiforums.net/f/showthread.ph...lding-Tutorial

These tutorials are the basis for a good understanding of rF files, and the experimentation does the job about the unanswered questions.

I think Bristow, Uzzi, The pits guys and all the people who wrote guides to the modding, got some info by ISI devs, via e-mail or PM, and i guess private comunication is the best way to get info by the programmers.

ISI released the needed tools for Max, for AIW editing, for mas editing and the plugin exemples (may be i am forgetting some other release?)
In other words ISI supplied the necessary tools.

I think, when modding teams like CTDP, HistorX, Enduracers, VLM etc etc, are able to create masterpieces, i dont understand why talented guys cannot imitate them releasing other great mods!

About the marketing strategy, i dont know if you LesIU are right or if ISI are right, but i realize a small team like ISI is, cant develop updates for rF and, at the same time, develop rF2.
It is a strategic choice, and we all can agree or no, but i think important to realize that ISI is a very small team, and such strategical decisions are needed.

You see how long is the software development process ... at a point, updates for rF could require long long time, with a rebuilding of the most of the engine, so an interrupt point can be a good way to work.
Stop to develop rF, and start a new up to date project.

Time will say if the strategic decision by ISI is right or wrong ... when we all will have rF2 installed in our HDD.

[LesiU]

Thanks for the links but I know them all (I'm not that lazy, okey? ).

As for so called "masterpieces" made by some modding teams... well, that depends on, what you are looking for. If just for fun and about having good time behind the wheel, then OK. You just like some mods more than others... but if you look at them as a physics modder and see what bullsh**t has been put into some of them from authors without even thinking if that has anything in common with reality then, well... maybe you would revise your thinking ;-) Sometimes when looking into physics files of some mods (from teams that advertising themselves as pros with access to complete real world data) I have a feeling that on this planet there are only a few guys that actually creates realistic physics, based on realistic data. Of course with respect to the modding teams, because I assume that everyone put as much effort into creating their mods as they can

but i realize a small team like ISI is cant develop updates for rF and, at the same time, develop rF2.

What I mean is to work on rF 2, but through updates to rF 1. Of course that is not that easy if rF 2 has completly different engine than rF 1 but ISI said it's all about constant improvements, so I think that at some point, having an update (payed or not) called "rF 1.5" with some bug fixed and new features, won't be a bad idea. The community get refreshed product while the devs get feedback about new features.

Time will say if the strategic decision by ISI is right or wrong ... when we all will have rF2 installed in our HDD.

Everyone will be happy when rF 2 will be released, no doubt about that. It's more about, what happens in the mean time ;-) So, we just have to wait

[MaXyM]

You have the right to speak about what you prefer, but i have the right to say it is a pointless discussion.

Comparing a graphic engine dated 2005 with what?
A 2005 EA game? The last EA game?

I have a feeling that you thinking about comparison like some kind of war.
People compare things to get to know about best possible results they can get/see in nowadays products.
That's why this topic has been created (check first post). I believe he didn't expected notes about companies business model or dev teams abilities - it is irrelevant to what the thread author wanted to say.

I also want to see rF2 with models/track physics like in iR, playability of GT5 and gfx engine comparable to CryEngine2 (or at least 1). Even if rF2 will not have such features, even if I will buy it and have fun with it, even if I make rF2 main sim supported on my portal, I WILL compare to some references. Comparing things is the only way to make progress.

Or perhaps you want compare WIP screenshots? Another pointless thing in my opinion.

So.. you may compare WIP screenshots of other products to rF2. Then you will see the point and be able to compare. OK?

Note, that some skilled persons are able to justify even from WIP stuff. For example we can predict that rF2 gfx engine will consist the same limitations as rF1 one (in some areas shown on screenshots). Comparing those predicted features/issues to other products is not pointless. Especially if we give feedback to devs. Of course if they can filter out those messages from tens of 'hypes'

PS. I don't get args about implemented but not supported features, open modding platform without docs and support, modding as reverse engineering and so on. but it is not place to discuss it.

[Max Angelo]

I am reading here a lot of blah blah.

I posted here just to reply to some points raised.

Drivers swapping is not an officially supported feature. Period.

ISI supported for free rFactor for around 3 years. You forgot how rF v1020 was, related to v1255? Period.

ISI support works pretty well, the game issues are replied and often are replied also the issues related to OS/Hardware. Period.

Last, in my opinion, WIP shots are useless, at this stage, for comparisons with other software. You do think my opinion means comparisons are a sort of war?
Not a my problem.

[dragon]

Is it so difficult to understand that as players we want to know when the desired simulator will be released?

PS. Have you ever happened to some developer to write something I could not have been done better? I sincerely doubt it: / and the process of "correction" could go on forever what worries me.

[beatnik]

Why are people getting so upset that there is no definitive release date? I mean here you have the option to interact with developers and people associated with ISI, yet people get so riled up over the unknown.

We can't expect these guys/gals to update us every time they add a new line of code or pixel into the sim.

Also to anyone expecting Need For Speed graphics...don't get your hopes up. EA sacrifices A LOT realism for those graphics. You can't have the world's best graphics and the world's most accurate physics overnight or at this current computing power. Let these developers work instead of going back and forth with them.

I'm not trying to be mean or rude to anyone because everyone is entitled to their opinion, but at least step back and think about what you just wrote before clicking send. We have a great thing going here with the Official ISI Forums, let's keep it that way.

[Gonzo]

Everyone should think about how programming works and that there are reasons for that there is no release date yet.
Even internal exists not more than goals and wishes.
Its simply impossible to calculate time for programming projects.

For example:
Lets decide to develop a new tiremodel, and think about 12-20 weeks of work.
While that 2 million things may happen that will increase the development time.

After that, you have to test this out and find bugs, but you cannot calculate how much bugs you build in before.
So we have a development time and a testing time, both with +X of time and while that the others of the development team developing other things with development time and time X to fix bugs.

And then the new tiremodell is finished and the graphics are finished and the content is available and the customers want you to release.
But what is a tiremodell, new graphics and new content worth if the netcode isn't working or the user interface is not finished?

You simply cannot release or predict while in development.
If you give a releasedate which is too early and you cannot fit it-> people will be angry, because they plan their leagues with the new product as their base.
If you give a releasedate which is too early and you have to fit it with a buggy / rushed product-> people will be angry, because they cannot play in their leagues cause of the bugs
If you give a releasedate which is propably too late-> people will be angry and lose their interest, and you got forgotten maybe.
If you give no releasedate -> people will be interested and are angry even if you give them screenshots, informations, a forum and directly-developer contacts.

Whats your choice?

[Siggs]

I've been reading this thread, and I never thought I'd see one, but Gonzo is what I might call an rFactor fanboy. The phrase 'open-minded' is apparently not in his train of thought.

[Gonzo]

Fanboy? LOL
no, not really, no.
I am just a softwaredeveloper myself and know how the business works.
But maybe i should register as rffanboy here, too?
Hmm....

[MaXyM]

Everyone should think about how programming works and that there are reasons for that there is no release date yet.
Even internal exists not more than goals and wishes.
Its simply impossible to calculate time for programming projects.

Lol... I should print it with big letters in my company
Did you never heard about software projects and schedules?
I'm software dev too, that's why I know that serious software business cannot live without it. but you are saying it is impossible...

BTW: Note, that ISI has worked on other projects for the last few years. So, don't think that developing rF2 is almost 'missing impossible' specially 'cause they have basis (rf, rfPro). They simply have no time for it. Worse is that we are getting inpatient/angry because we are waiting for promised rF2 as upgrade to rF1 (fixed things, unsupported features turned into supported ) more than 2 years now (2 last Christmas was suggested as release date)

[K Szczech]

I think this topic was started because someone wanted to know what are ISI's goals in terms of graphics engine.

So let's just stay on the topic, shall we?

[MaXyM]

I posted here just to reply to some points raised.

[K Szczech]

I have one suggestion (actually I have a lot of them ). It's about cubemaps.

In rF1 we had cubemaps defined in vehicles. That's bad because when I go to Monaco I'm surrounded by buildings, when I go to Belgium I'm surrounded by trees and in Bahrain I'm surrounded by sand.

It has been said, rF2 will have per-track cubemap. That's a good thing and now we see that there will also be dynamic cubemaps.
So, there's an improvement I would like to suggest here.

1. Static per track cubemap should not be provided by track author. Instead, track author would just point out one place on the map, where this cubemap is taken from. Since you have realtime cubemap implemented, it would be trival to implement.
2. Do not render sun on this cubemap (but light scattering around sun should be here).
3. This static per-track cubemap should not be that static - once every few seconds it should be updated.
4. This cubemap should be in floating-point format (you can't do proper reflections if you can't represent incoming light intensity properly).

Now this gives us a lot and it's actually trival to implement considering what you have implemented so far. With this we would have car reflections corresponding to actual scene.

Next suggestion:
5. Use fully dynamic cubemaps on nearby cars, and static ones for cars further away.

And finally:
6. Make a blurred version of this cubemap:
-downscale base image 2x and additionally blur it - create half a size cubemap (base level mipmap)
-downscale this blurred cubemap again and blur again - this is next mimpap level
-keep going until all mipmaps are generated

This way we have a cubemap with blur increasing quickly with mipmap level.

What's it for?
- reflections on rough surfaces (like plastic or skin) - you would simply define LOD bias for cubemap in material properties to get desired blur level for reflection
- ambient lighting (yup, constant color is not good enough ) - ambient lighting should be an integration of incoming light (surroundings) over a 180 degree area of cubemap with NdotL function which of course would be computationally expensive, but small, heavily blurred cubemap should do the trick just fine.

[Gonzo]

Lol... I should print it with big letters in my company
Did you never heard about software projects and schedules?
I'm software dev too, that's why I know that serious software business cannot live without it. but you are saying it is impossible...

What kind of software does your company develop?

[MaXyM]

Does it really matter?
If company wants to make money, they must create business plan balancing planned incomes against outcomes. Without ability to plan time of developing software, it is impossible to predict outcomes. I think it is clear.

I worked on software for pharmacy, industry (db and visualization systems), made-to-measure crm/cms systems for various companies. Believe me or not, you wouldn't be able contracting a job without knowing how much time developing will take. Currently I'm working in company developing internal system. And even if we don't sell this software, I couldn't do that not being able to predict a cost of developing changes. In that case cost is equal to time spent on it (programming + documenting)

I know that developing in ISI is also scheduled. But I'm, as long as rF2 is not a space-shuttle-like project, the title is not delayed due to developing issues. It's rather because rF2 has quite low priority comparing to their other projects which make real money.

[Gonzo]

Yeah, but you have an idea in how the final-product has to be.
Game-development is different, you have to experiment with it, test out things and in 2007 you do not know how it should look and be in 2011.
Especially if you are a small company.

If you follow the EA-strategy, they have 1 year time to add minor things, polish it and release it, but thats another story.

There is a reason for that there are NO release dates for every game out there.
Once you reach point X in development you can feel and calculate a release within 3 months or 6 months, etc.
But before this you cant.