FBA WIP

vbt

Staff member
A new driver is in progress
smile.gif
There is not yet good results. I don't manage to get a correct bitmap rendering.
 

Attachments

  • Image8.jpg
    Image8.jpg
    758.3 KB · Views: 100

Runik

Staff member
vbt said:
A new driver is in progress
smile.gif
There is not yet good results. I don't manage to get a correct bitmap rendering.

Progress is good, even if the display is broken at first
wink.gif


Good work vbt !
 

vbt

Staff member
Runik said:
Progress is good, even if the display is broken at first
wink.gif


Good work vbt !

It's getting a bit better
 

Attachments

  • Image1c.jpg
    Image1c.jpg
    307.9 KB · Views: 64

vbt

Staff member
FacundoARG said:
biggrin.gif
, it's working. When I have more time for this, i will ask about how to reproduce a avi or cpk file.

Sure, you can ask, I think Madroms has biggest knowledge about making vids for Saturn.

Another new pictures :

839f07325c.jpg


91e1a12af1.jpg
 

cafe-alpha

Established Member
vbt said:

What's wrong with this image ? (sorry for the stupid question, but I don't this game)

If there is a problem, don't hesitate to send me a binary, so that I can try to debug the VDP2 stuff in a similar way as I did in last December
smile.gif
 

vbt

Staff member
cafe-alpha said:
What's wrong with this image ? (sorry for the stupid question, but I don't this game)

If there is a problem, don't hesitate to send me a binary, so that I can try to debug the VDP2 stuff in a similar way as I did in last December
smile.gif

not much wrong inside this pic, in fact there is a sprite problem which hapenned on yabause and not on ssf. So, I finally decided to make a PAR release of System 1 (binary with rom inside) and here is the result on yabause :

i62kdz4ptkfrq89g4f3l.jpg


on the left, CD version, there is a little problem with sprites

on the left, PAR version, the sprites look correct

There is something obvious here, the driver source code is the same, just the file loading is replaced. But the result is correct sprites and modified palettes. cafe-alpha, I'm sure you're the right guy to solve this mystery.
 

cafe-alpha

Established Member
vbt said:
not much wrong inside this pic, in fact there is a sprite problem which hapenned on yabause and not on ssf. So, I finally decided to make a PAR release of System 1 (binary with rom inside) and here is the result on yabause :

i62kdz4ptkfrq89g4f3l.jpg


on the left, CD version, there is a little problem with sprites

on the left, PAR version, the sprites look correct

There is something obvious here, the driver source code is the same, just the file loading is replaced. But the result is correct sprites and modified palettes. cafe-alpha, I'm sure you're the right guy to solve this mystery.

Can you send me both versions' binaries ?

Also, can you indicate me the way you use to load the data from CD ? (SBL, SGL, custom library, etc)

First, I will try on real hardware both versions, and see what happens ...

vbt said:
cafe-alpha, I'm sure you're the right guy to solve this mystery.

I will try do my best, but don't expect any guarantee I will actually solve this, nor any fast answer, etc ^^
 

cafe-alpha

Established Member
I took a look at the data you sent me last Tuesday, and I could see that 7 registers values differ between the standalone and CDROM version.

Also, as VRAM contents seem to differ too, I need investigate a little further in order to see why display changes on CDROM version.
 

cafe-alpha

Established Member
vbt said:
not much wrong inside this pic, in fact there is a sprite problem which hapenned on yabause and not on ssf. So, I finally decided to make a PAR release of System 1 (binary with rom inside) and here is the result on yabause :

http://freeebayimagehost.com/images/i62kdz4ptkfrq89g4f3l.jpg

on the left, CD version, there is a little problem with sprites

on the left, PAR version, the sprites look correct

There is something obvious here, the driver source code is the same, just the file loading is replaced. But the result is correct sprites and modified palettes. cafe-alpha, I'm sure you're the right guy to solve this mystery.

cafe-alpha said:
I took a look at the data you sent me last Tuesday, and I could see that 7 registers values differ between the standalone and CDROM version.

Also, as VRAM contents seem to differ too, I need investigate a little further in order to see why display changes on CDROM version.

Just before starting emaulation after selecting a ROM, can you set the registers above to the following values ? (a copy-paste in your source file should work fine)

Code:
(*(volatile unsigned short *)(0x25F80028)) = 0x0C02; // CHCTLA

(*(volatile unsigned short *)(0x25F80030)) = 0x3FFF; // PNCN0

(*(volatile unsigned short *)(0x25F80032)) = 0x3FFF; // PNCN1

(*(volatile unsigned short *)(0x25F80034)) = 0x3FFF; // PNCN2

(*(volatile unsigned short *)(0x25F80042)) = 0x1F1A; // MPCDN0

(*(volatile unsigned short *)(0x25F80046)) = 0x0115; // MPCDN1

(*(volatile unsigned short *)(0x25F8004A)) = 0x0115; // MPCDN2

I am not sure it will work, so please can you tell me if the display is better ?
 

vbt

Staff member
cafe-alpha said:
Just before starting emaulation after selecting a ROM, can you set the registers above to the following values ? (a copy-paste in your source file should work fine)

Code:
(*(volatile unsigned short *)(0x25F80028)) = 0x0C02; // CHCTLA

(*(volatile unsigned short *)(0x25F80030)) = 0x3FFF; // PNCN0

(*(volatile unsigned short *)(0x25F80032)) = 0x3FFF; // PNCN1

(*(volatile unsigned short *)(0x25F80034)) = 0x3FFF; // PNCN2

(*(volatile unsigned short *)(0x25F80042)) = 0x1F1A; // MPCDN0

(*(volatile unsigned short *)(0x25F80046)) = 0x0115; // MPCDN1

(*(volatile unsigned short *)(0x25F8004A)) = 0x0115; // MPCDN2

I am not sure it will work, so please can you tell me if the display is better ?

sorry, i missed your post, i try it today
 

vbt

Staff member
vquqagpphurjxycozx5.jpg


Finally, colors look good, here are the changes :

  • background layer palettes 8x64 => 16x64 (with the help of SatCom)
  • foreground layer palettes 8x64 => 16x64
  • sprite palette 512 colors in CLUT (thanks to Antime)

so 2560 colors used !
 

vbt

Staff member
speed improvements :

- improved color conversion (thanks to Martin) + use of a precalculated lookup table

- foreground layer is rendered using the slave CPU

6lkfseres0zvauh1yp79.jpg


rmvd8f70qc1gnp3riewe.jpg


blkpt764ifag83dh6aia.jpg
 
Top