How do PSX games manage a Z buffer?
Jason Main
but some use what seems like a z-buffer for water effects?
How do people mannage this on the machine (wich after all is
limited for anything not suported with hardwear....)?
Thanks in advance!!!
Aaron Arndt
Which games are you referring to? Some PSX games with a constant view
point use a static z buffer but in a very loose sense of the term.
Justin Heyes-Jones
wrote:
Sounds interesting, could you expand a bit?
Justin Heyes-Jones, just...@hotmail.com.
EuroDog
Aaron Arndt wrote in message <35C933FC...@radical.ca>...
>Jason Main wrote:
>
>> I know that most PSX games use painters for all polygons,
>> but some use what seems like a z-buffer for water effects?
>>
>> How do people mannage this on the machine (wich after all is
>> limited for anything not suported with hardwear....)?
>>
>
>Which games are you referring to? Some PSX games with a constant view
>point use a static z buffer but in a very loose sense of the term.
>
memory (2 megs), most people opt for a z-sorted ordering table.
Marc Coupar
>
> On Thu, 6 Aug 1998 05:41:32 +0100 , Aaron Arndt <a...@radical.ca>
> wrote:
>
> >
> >> I know that most PSX games use painters for all polygons,
> >> but some use what seems like a z-buffer for water effects?
> >>
> >> How do people mannage this on the machine (wich after all is
> >> limited for anything not suported with hardwear....)?
> >>
> >
> >Which games are you referring to? Some PSX games with a constant view
> >point use a static z buffer but in a very loose sense of the term.
>
> Justin Heyes-Jones, just...@hotmail.com.
Resident Evil 2 uses Z-Buffering which has static views but these are
precalculated for each scene. You can actually use the psx GPU to
render a Z-Buffer (of sorts) then dump it to main RAM but as Aaron says,
it's only really suitable for static cameras. I don't know of any
currently available titles doing this but there will be before Xmas :)
Marc Coupar
Psx programmer
Intelligent Games Ltd.
mco...@igl.co.uk
ma...@taste.demon.co.uk
Jason Main
between where Loria's in water and out. Shes dyed blue because
of hardware restrictions but think this is happening. Now
that you mention it I have to have a closer look....
EuroDog wrote:
>
> Aaron Arndt wrote in message <35C933FC...@radical.ca>...
> >Jason Main wrote:
> >
> >> I know that most PSX games use painters for all polygons,
> >> but some use what seems like a z-buffer for water effects?
> >>
> >> How do people mannage this on the machine (wich after all is
> >> limited for anything not suported with hardwear....)?
> >>
> >
> >Which games are you referring to? Some PSX games with a constant view
> >point use a static z buffer but in a very loose sense of the term.
> >
Christer Ericson
<jas...@hpbs1079.fooultrabar.boi.hp.com> wrote:
>I know that most PSX games use painters for all polygons,
>but some use what seems like a z-buffer for water effects?
>
>How do people mannage this on the machine (wich after all is
>limited for anything not suported with hardwear....)?
They 3D clip polygons to the plane of the water (or whatever),
rendering twice to get the lower part of the body to appear
underwater.
However, as Marc and Aaron pointed out, it is definitely possible
to implement a (somewhat limited) z-buffer (as outlined in
Marc's post). For nothing more than a body-submerged-in-water-
effect that would be a lot of overkill though.
Christer
Neversoft
Jason Main
that I might get the particle update working but would have difuculty
rendureing them. I cant interspearce them with the polygons because
I need to have well defined boundries. I can z fill a couple relevent
polygons, but mergeing them would probally overdue the PSX bus. I
was wondering if I missed something.....
Thanks for the reply!!!!
-Jason Main
Aaron Arndt
> The idea I am working on now uses mucho-particles. I was gueesing
> that I might get the particle update working but would have difuculty
> rendureing them. I cant interspearce them with the polygons because
> I need to have well defined boundries. I can z fill a couple relevent
> polygons, but mergeing them would probally overdue the PSX bus. I
> was wondering if I missed something.....
>
To do a limited z-buffer is pretty damn slow on the PSX. You can get the
GPU do the z-buffer but even that is slow. If you don't care about speed
just a cool looking effect then the GPU z-buffer can work. However if you
need mega speed you'll have to detect which particles are colliding and 3d
clip them against each other. Hmm.. that can get rather tricky too. :) Do
you have access to the Sony Developer newsgroups or website? They have an
article + source on how to do the GPU Z-buffer.
Russ Williams
[...]
>>Which games are you referring to? Some PSX games
>>with a constant view point use a static z buffer but in a
>>very loose sense of the term.
>
There's some sample code on the web site and
check the toptips newsgroup for another method.
---
Russ
Christer Ericson
<jas...@hpbs1079.fooultrabar.boi.hp.com> wrote:
>The idea I am working on now uses mucho-particles. I was gueesing
>that I might get the particle update working but would have difuculty
>rendureing them. I cant interspearce them with the polygons because
>I need to have well defined boundries. I can z fill a couple relevent
>polygons, but mergeing them would probally overdue the PSX bus. I
>was wondering if I missed something.....
Particles, as implemented on a Playstation, are nothing but ordinary
GPU primitives (like, say, tiles) so there is no distinction "particles"
versus "polygons". Basically they're all polygons (lines excluded),
so it all comes down to how to sort your polygons and unless you give
more detail, it's hard to suggest any other approach than the standard
ordering table/bucket sort one.
Generally people avoid sorting problems by carefully organizing their
levels, and by tagging polygons with different sorting criteria (eg.
nearest, average, or furthest z). BSP-trees aren't too hot on the
PSX due to the amount of computation and RAM access. An idea I've
toyed with would be doing a (limited) Newell-Newell-Sancha step to
reduce sorting problems with the basic OT approach, but I have no
clue as to how well it would perform.
Christer
Neversoft