Specular

[LsrgcBath]

Another side by side from Abe´s, now with DU2...

https://www.youtube.com/watch?v=bED2Yfdb-JI

[Ian]

Well the mystery of why the light angle is flipped is because there are 2 versions of lemans 24, which I didn't know before

Version 1 (light coming from the right)


version 2 (light coming from the left)


The ROM floating around online for lemans 24 is the first version. Well that solves some potential problems with regards to specular.

[Ian]

Well I think I finally figured out how the real3d is doing specular, and I got the highlights to show in the correct places. Just the scaling or levels are off



Heres the maths

[Bart]

Woohoo! That's some top-class reverse engineering, Ian! Regarding the intensity, as I recall there are two values in the polygon header, which I assumed originally correspond to a multiplicative factor and the shininess coefficient. I suppose the 2-bit value that I took to be shininess could be a lookup into a table of arbitrary values.

[LsrgcBath]

This is so nice, u guys are awesome.

The light angle seems fine now, only the brightness seems a little bit out of order.

Woohoo! That's some top-class reverse engineering, Ian! Regarding the intensity, as I recall there are two values in the polygon header, which I assumed originally correspond to a multiplicative factor and the shininess coefficient. I suppose the 2-bit value that I took to be shininess could be a lookup into a table of arbitrary values.

So, if I correctly understand the tech talk, these values will determine the brightness on the surface of the polygon?

I dont know, maybe im assuming wrong, but it must be difficult for you guys to discover the correct values of brightness and gradient with multiple choices to choose, or by chance is only a matter of: works this way, or only in such other way?

Thank you both for the great work.

[Ian]

Well in the documents for the real3d it says this about specular

The specular component of light can be used to make a surface appear shiny by modelling the surface’s reflective response to a light’s position and angle of incidence.

Which seems to fit with what I found for scud. But it doesn't seem to work for la machine guns, or at least with the limited testing I've done. Perhaps the later hardware actually did proper specular

[Bart]

LA Machineguns should be no different than other Step 2.x games, including Daytona USA 2. There is some specular lighting visible on the vehicles' bumpers (especially on the Phantom car, which is black). I'll try to find some time this coming weekend to finish up the changes I've been working on so I can sync my branch to your current progress.

The best example of specular lighting in LA Machineguns is the Los Angeles boss. Unfortunately, there isn't much (if any) actual arcade footage. The Wii footage might not be 100% accurate (one could easily imagine them using normal specular lighting with adjusted coefficients to make things look 'right').

[Shekel]

I don't know if this will help at all, it's not amazing quality and has rolling black bars but yeah, it's hard to find any clear footage https://youtu.be/UL_SvP3WDv4?t=2m54s

[Joaquim Gonçalves]

Congrats to all.