Messages

Yes. GPBikes' core concept is about real-life simulation. In no way the content should be adapted to purely subjective perspectives, often (if not always) inefficient in accuracy driven projects. I am pretty sure a big part of us bought PiBoSo's products with a certain idea in mind: experiencing unbiased unrestricted developmental accuracy driven race simulators. In other words, 'dem best racing sim of all tiem!'. Don't change that.

Edit: It makes even less sense to follow option #2 if some specific mods are meant to incarnate real motorcycles. I didn't sign for this!

Tuning audio is about altering a sample. Procedural audio using CPU generated waveforms (as shown earlier) is another thing. If this method you mentioned is effective though, then it's entirely fine. First target is to get clean, loop-able and well transitioning samples. Try to locate where the sound is different (not just the pitch) on the rev range and it should be good for the current system. On a more abstract scope, I think there should be a sample set per type of engine; using cylinders (and valves?) numbers, crankshaft configuration (flat-plane/cross-plane) and exhaust system (all in one, divisions, one per cylinder etc...) as difference factors.

It's all just idealistic ideas, though. If we can achieve one of them, the quality of the simulation would be greater, in my opinion. If we can get decent and pretty accurate sounding samples, then it's perfectly fine once again. Hopefully nobody got me wrong in my approach.

Technically, this Discord server's scope encapsulates all of PiBoSo's simulators. Don't hesitate to leave me a message either by pm here or on Discord so I can give you member rights.

It is very likely the clip won't be fitted for sample creation. Since sample frequency modulation is a very capricious method, as anything unwanted in the sample will also get the changes, a recording made without certain preparations is expected to give a flawed and pretty uncomfortable playback in plus of giving a harder job to the one that configures the simulated audio.

Preferably, you would need to run the vehicle on a dyno, in a silent open space, and vary the revs on specific levels. You would also need a certain setup to efficiently record the sound on the desired location (engine, exhaust, pilot seat etc..). A technique used by professionals, with another kind of means, unfortunately.

Ideally, granular synthesis or procedural engine audio simulation is what we, or likely PiBoSo since he is the only one with control over the engine, should be going for. It demands a consequent amount of knowledge in different domains, but can do wonders when mastered.

This video's simulation is procedurally generated/processed by the CPU. No sample were used for the simulation.
https://youtu.be/vKnfO3LFPls

This is what PiBoSo claims.

1. System Requirements
----------------------

Operating System: Windows 98/ME/2000/XP/Vista/7/8, 32 or 64 bits
Processor: 1.5 GHz
RAM: 512 MB
Video Card: 64 MB OpenGL 1.2.1-compliant
DirectX Version: DirectX 8.1 or newer libraries

Surprising. You shouldn't have any performance issue with such component. Are we talking about a clean/vanilla build or one using mods ?

With this camera i see a little problem with the windshield, need a double texture/shader because it seems to disappear 

If the windshield mesh is not double layered, it could be due to a probable specific face culling mode; which you could theoretically be able to set on any shader language if I am not mistaken ? Also, isn't it possible to re-use any material or shader function for different objects or sub-meshes ?

On a side note, is GPBikes using FFP ?

Video Card: 64 MB OpenGL 1.2.1-compliant

Not commenting on any system in particular but i expect most modders would be targeting decent performance for cards in the £100-150 bracket, these typically represent the value end of gaming systems assuming the rest of the system is of similar supporting spec. Today that gets you from GTX750ti to  GTX960 with 2GB VRAM (or equivalent AMD) quite a powerful card relative to GPB base specification and should provide around 150 to 200fps single player at 8*AA and everything else maxxed on most tracks.

Well, I do not have the money to spend on that at the moment. And it would be pretty hard to install one of those on a laptop. Well, not on mine at least, hah!

As the graphics are all hosted locally and only the 'game' positional data is transmitted via the server it should not really impact on line play if the local machine is sensibly configured.

Oh it doesn't really have to do with the network algorithm, as -- theoretically and for example -- only physical (positions, rotations, velocity, angles or axises, player state etc...) and informational (client state, chosen bike, chosen paint etc...) data would be transited; which is a very fast and normally lightweight process. The rest is indeed handled locally (including bike data loads).

It is important to optimize the temporary model in line with PB's specs (low poly/no texture) to ensure rider joining delay/stutter is minimised.

In my opinion, the use of a general 'default' (and geometric adaptive if in need of accurate physics?) low poly and poorly rendered model loaded on connection would be more adapted. It would permit to have a clear visual on the user in plus of preventing connection freezes; thus not harming the experience of those who are already playing. This relates more to the simulator's development concepts and solutions, though.

Modern graphics cards will not be very troubled by anything in GPB. If you are using very old hardware you should turn down your graphic options or force low performance modes in the game launcher.

Modern doesn't especially mean decent. Also, it depends on what you call 'modern'. I don't think my computer is that old, but it sure has quite cheap components. Only certain maps are a problem, though.

For reference the rider helmet alone has a 2048*1024 texture and the rider suit is 2048*2048

Seems pretty detailed for such small objects; the most surprising being the helmet (even lesser surface/more compact).

I am suspicious as to how GPB handles models/textures, it seems big highly textured models can be made to load faster than simpler models and it seems to be dependent on how the model and textures are arranged rather than their size - you can observe it when switching in the showroom and when loading models in bikeED - not sure why.

That is kind of unusual, and odd (at least to me). If you split your textures in a set of smaller images, it should theoretically load faster and handle better in plus of keeping the same quality. Is there any information about the renderer ? It's version ? It's language ? What data does the EDF format store exactly ?

The biggest hit/game processing load is definitely the current track view, the lack of track LODs means peak loads vary wildly depending on view (even without other bikes on track), LODs would help to average the graphical workload.

Could explain certain maps not suited for my current configuration.

Totally agree with H

I use just a texture 4096x2048 being lazy because I never grouped the individual elements in a 2048x2048 texture. It is not recommended to do this, but it does not create problems

The only thing I would have to say about this is don't oversight critical details and think of the worst possible situation when it comes to optimizations. It's just about being careful and serious, not any reproach or something offensive (at least not meant). I am just a little concerned about certain of your decisions/how you look at certain things. The rest is good to me, I really think it.

Glad this is still on going! And it seems to be progressing well!

There is no connection between cores and textures ... the default livery is 4096x2048 and the bike has already been tested in a race with many users, (10/15?) I do not remember any core.exe 

I meant that making a new skin for this bike, I noticed too low quality with half res. So also the template need a upscaling.

I wasn't referring to the common, fatal, unhandled exceptions but the system 'core' stability and performances. I am personally used to see 1024 (mostly square) images with a very detailed/on point render; 2048 in extreme cases. I am afraid this template needlessly uses a tremendous amount of memory; which might be too much for lower-end configurations (RAM/VRAM). Just saying.

Aren't images that big memory costly ? Won't it compromise the core stability ? I am surprised it isn't enough.

Revision 6 is up. Finally fixed the default rear sprocket (I hope...).

More realistic. I approve.

So, there is nothing to do for the engine brake ?

Sorry SAS but can not agree with that about modders. If it was just Piboso we would still have only one track and that for the last how many years???

I have learned to be patient from other mod expendable projects. Also that having content is great, but without core functionalities and stability, it isn't much useful. The pair is progress inducing.

And how many bikes are only Pibosos and how do they handle???

PiBoSo made four bikes with stunning overall quality, using his systems fully. This is a GP racing simulator if I am not mistaken; he doesn't need much variants of the same bike, paints, other maps and categories having no relation with the GP world. What is present is actually enough as a testing and playable/enjoyable base.

No disrespect to Pib as he has a lot to do and knows that the modders are here to help him and us all so saying modders are not as depended on as Piboso is soooooooooooo wrong.

Modders sure are of a certain help, at the same title as those who play. They establish thorough tests of the known systems and can dig up issues not especially noticeable at first sight. Also, people can give more or less objective feedback and suggestions for the development. An extern point of view should always be welcomed of a content creator; not especially adopted, but at least thoroughly thought. Still, mods are just a plus to the simulator, in my opinion.

Not just that but look at Blacky, Manu and H's bike models...........FUCKING AMAZING!!!! Piboso does not have the time to dedicate to such quality in his models right now.

Never pretended the contrary. Also, I think I have somehow -- probably pretty implicitly -- already mentioned the later.

They do what PiBoSo couldn't achieve in the current situation (well, he could, but that would be silly and the result quite different).

I desperately want the 900 from Blacky as from what I have seen it would make GPB so personal for me and a few others I know ride that bike IRL with a passion.
I am sure Blacky will at one point be back full power but until then he needs his time and space.

I second that with full hopes.

If people need to take a break(which I can understand) then take a break, but I'm sure we can do without the declared "GPB is dying" dramatics as their parting gesture.

My core point, exactly!

This thread is a shit now.

It isn't. Sure the subject might have spread a bit, but a single relevant message at any time could patch it up. I will try to stay on topic from now on, sorry.

there is no interest for gp bikes

What causes this lack of interest ? Constraints ? Lack of communitarianism/help ? Working alone ? Repetitive tasks without much victory earned/in sight ? It is very likely that nothing will happen if you don't speak your mind.

without modders gp bikes is dying ... 

That is untrue. GPBikes only depends on PiBoSo. Mods are just a more or less legal plus brought to an unofficial, commercial, motorcycle-centered simulator in expansion. Don't forget where things stand, Blackheart. Of course, I don't mean any offense to modders. They do what PiBoSo couldn't achieve in the current situation (well, he could, but that would be silly and the result quite different).

I don't force you to do anything, but whatever is blocking you -- unless private -- shouldn't be kept to yourself, in my opinion. Who knows ? You could regain motivation/more/else from anyone here or outside ? So many possibilities.

Heh. I am not really the one that can and should give advises, but I am just trying to help; in my manner. Choice is yours.