I guess I was right on it, thanks Scawen, I'm sure it adds a lot!
Nice! I'm looking forward to see what I could do with it!
That's very nice, now that I'm used to see LFS with linear color space rendering, I can't go back.
Did you converted all the textures to sRGB or do you convert to linear on the pixel shader?
Nice, I was waiting for this one for a long time! Will you still use vertex color to bake indirect lighting/ambient occlusion? I simply hope we don't loose indirect lighting, it would downgrade a lot the sense of depth on tracks.
Anyway, I have so much questions and suggestions for rendering that I could do a complete thread for it, but I think I'll wait to see what you've done first. Keep us updated!
Westhill was primarily a boost for the cruising community and the (big) illegal community. What Scavier wanted to see for racing (full track usage) didn't happen. Yet. Which from a community perspective is logical, there need to be brought more to the table then just some tracks to keep things interesting.
Teasers are a bad idea, community innovation sort of stopped already, this will not return with some teasers as teasers can be pulled just as easily. Remember some Scirocco teaser, Rockingham teaser / track display at the track, tire physics teasers and some unknown never released car teaser? Teasers are pointless on this point in time, real stuff is what counts.
Besides, teasers generate a ton of weekly/monthly "are we there yet?" emotions. Again. Followed by negative emotion because some people run out of patience earlier then others which leads to negativity.
@Scawen, as I noticed, you say that the merging of the graphic code into the version with the new physics of tires is a nightmare for you. Why not in 2006 do not make version code with a branching, this would reduce your time to merge different versions of the game.
I use D3DSAMP_SRGBTEXTURE so we can use standard sRGB textures without needing to convert them in the pixel shader. Then all the processing internally is linear and finally the output is converted to sRGB using D3DRS_SRGBWRITEENABLE.
Yes there is still an offline render for the ambient lighting, but now it is a single offline ambient render shared between all lighting configurations. For anyone who doesn't know what I am talking about, this is the thing mainly concerned with sky lighting, so for example you can have two different places that are not lit by the sun, but one of them might be a lot darker than another because the sky is obscured.
See in the attached images, the first two have all direct (sun) lighting switched off. The first one is simply default ambient lighting without the offline render. The second one has the offline ambient render applied. Finally the realtime direct lighting and shadows are applied in the third image.
NOTE: This is the original South City track with the old textures and no specular lighting added. This is simply a plain conversion of the old South City into the new system and Eric hasn't had a go at it yet.
At the time, I believed the new tyre physics was about to be released and had done a significant amount of work in the new version. It was difficult to avoid clashes in the code because of the number of changes so I deleted the old code out, thinking it was finished other than any emergency patches.
If I had known at the time I would be going for over a decade with two versions, I would have taken a different approach.
@Scawen now that's a nice technique, glad to see that you kept the indirect lightning into the new updated engine, it will save time for Eric, because he doesn't need to back (apply shadow into the final texture) every track objects texture with ambiant shadow .
But it's still a little bit problematic in case of night/day transition, as you cant use the indirect lightning in real time conditon ?
Also is the new dynamic shadow also can be rendred inside the car interior ?
Whatever, I can't wait to see the final result onto the updated tracks.
Nice work, Scawen!
Any plans on non directional sun light bounce?
BTW, you can also use color gradient or spherical harmonics probe based on world space normal for adding variation on ambient lighting.
Yes, because the indirect lighting has "amount of sky visible" and "artificial light colour" stored separately. So the sky colour / brightness can be varied and the ambient lighting can vary in real time.
I have tried to design it to be able to handle changing lighting (day / night transition) although that is complicated and seems lower priority than the physics updates and a lot of other things I need to do and finish, so I don't know when that will appear.
Yes, inside the cars too, cars do cast shadows on themselves.
Yes, there is also a component of ambient lighting contribution from light reflected on nearby objects.
The ambient vertex lighting consists of: sky light, artificial light, light from nearby objects.
Sounds like plans have further aspects as I ( or we as well ) thought. Pretty sure there is thought out of lighting changing visual/physical conditions and vice versa.
Can't wait to see how LFS turns out. I'm leaving to the army the 13th of april. Hopefully I'll come back to something great. Take the time you guys need and good luck.
That's exactly what I had in mind, only sky lighting and constant ambient color is a bit of a downgrade compared to full global illumination, but that's the price to pay for real-time lighting. What I think works well in this case is to have an ambient from a gradient of 2/3 color (sky (/ equator) / ground), as Keiichi said. Works well enough.
What exactly is "light from nearby objects" here? Things like spot lights and stuff, or more light global illumination bounce (which is dependent on weather)?
BTW, on your last screenshot, shadow mapping looks nice
EDIT: Small request bump, if you got time for it, because it can be really frustrating for me sometimes:
@Scawen in fact, i hope you've already started using git to make edits without losing data =)
I also really liked your report of the work with the graphics.
I also see that you have a lot of plans for implementing cool graphics on the DX9. It would be great if you had a well-planned schedule, when and what to implement.
IMHO it would be nice to have some kind of open project on experiments with graphics in which community members could demonstrate some implementation of the code and later integrate them into the mainline.
It's hard to believe that after a long time I still believe that you will make LFS great again
Well, it wasn't what was supposed to be seen, but indeed, a cloudy weather is mostly skylighting with no strong direct lighting (or not at all), so it does looks like a nice british cloudy weather!
Reapiting question scawen.
1. Huge modding system Will be allowed to give community opportunities to make stuff in game? Such as cars/tracks?
2. If we have more options managing races Such as safety car, vsc, etc?
1. As far as I know / remember supporting the community with modding tools has been on "Scaviers" mind for a long time. My best guess is that additional stuff would most likely be limited to cars / tracks. At least that should be the easiest way. Given, that Eric seems to use custom tools (coded by Scawen) it should be a good guess that any official tool might be based on those editors.
However it doesn't seem to be very clever to release those while you are still working on the "core stuff of the game" as you'd have even more stuff and tools to support and people will complain if an update breaks popular mod x or else (see the work that is being done on shaders and track lighting). The approach here is to release stuff only "when it's done properly" instead of a rolling release (which makes sense given the number of coders working on this game which is still one, unless his kids have joined him in his quest by now)
2. I think there is a very long list everyone has of things that need to be implemented (link to the obvious https://www.lfs.net/forum/8-Improvement-Suggestions). Stuff like VSC and SC would certainly be nice, at least SC has been done by several leagues by having an admin drive around slowly in front of everyone else with varying results (<insert joke about admin driving skills here>).
Easier or related stuff usually gets implemented during a round of test patches, if it wasn't already on Scawen's very short "todo-list" (I think he is using continuous paper at this point). My guess here is that if a request seems reasonable, improves the gameplay, fits his shedule and peaks his interest, he'll probably end up at least thinking about it. At least that's how it looked to me on multiple occasions. It also depends on whether the patch is compatible or not (to previous LFS versions) on what get's actually implemented.