Archived

This topic is now archived and is closed to further replies.

morphy

Anti-aliasing on PS2

Recommended Posts

Sorry if I bother you all for this over-heard question, but I really would like to now WHY do the PS2 game (DOA2, TTT, RRV)suffer from jagged graphics. It shouldn''t be! On the paper the ps2 is so powerfull. Sony should be crazy pretending this is the revolution! Do not mind getting technical. I will understand. Just tell me why and how can the future games cope with this strange problem. thank you all m

Share this post


Link to post
Share on other sites
That''s because the PS2 lacks anti-aliasing capabilities. Or at least according to developers. Developers say it doesn''t ahve anti-aliasing features and Sony says that it does deep down inside somewhere but developers just haven''t found it yet. Not sure who''s right.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
Actually PS2 hardware does anitalising (or least it does according to the specs i have read), but the hardware does not support true "hi-res" at 640x480 instead it uses an interlaced mode running at 640x240. Thus those nasty jaggies are created since only every other line is updated each frame. Makes no sense why sony would go all out on the machine and not support true 640x480 like the dreamcast hardware does.

Share this post


Link to post
Share on other sites
Fine. So it seems PS2 can throw millions of poligons without any anti-aliasing and at a 640*240 interlaced resolution. I''m wondering what kutaragi was drinking when he designed the EE and the GS...

Do you really think this could be the Achilles'' heel of the PlayStation2?

I just hope the future games will come up with a 800*600 resolution. That could solve the aliasing probs.
m.

Share this post


Link to post
Share on other sites
That isn''t the only problem. The PS2 has two vector processors that have to be kept in sync. That''s what basically killed the Saturn because developers couldn''t create any decent games cuz they had to spend 90% of the time trying to make sure the processors were synced.

Share this post


Link to post
Share on other sites
oh that''s interesting...but I''m just wondering why all the developers I have been talking with seem to be in love with the ps2. They all seem considering programming the ps2 just a "exiting challenge".

I really cannot believe sony did the same mistake sega did with the saturn. They should have the knowledge to avoid it.
m.

Share this post


Link to post
Share on other sites
The biggest difference is that Sony (and others) provide decent middleware for PS2, so that developers don''t have to get down to the nuts and bolts of it all, unless they really disagree with the middleware.

And look at what happened with the original PS - games that bypass officially licensed middleware and do their own thing with the hardware often don''t run on the PS2...

On the other hand, Saturn was pretty nasty, and Sega didn''t support developers all that well. Er, allegedly .

TheTwistedOne
http://www.angrycake.com

Share this post


Link to post
Share on other sites
Actually, pretty much all you guys are wrong :p

I''m currently working on a PSX2 RPG, and here''s my take on it :

- the PSX2 does have explicit hardware edge antialiasing.
- to use it effectively you have to draw polygons from back to front so the edges can be blended together properly.
- the overall architecture of the PSX2 is not very conducive to sorting polygons. Its geared towards throwing tons and tons of polygons into the graphics synthesizer with as little processing as possible. Sorting all the polygons in a scene typically is a _huge_ amount of processing.
- the vector units on the PSX2 can run completely independantly of the main CPU, but individually they''re kind of like one way streets. You send data _into_ them, they do their thing, and then spit out raw polygons to the graphics synthesizer. Their individual memories are so teeny tiny, and the assembly language they run in is so weird, that sorting operations are almost prohibitively expensive. Its not the "Sony Philosophy".

As next generation games come out I have no doubt people will come up with tricks to get around these issues, but its tough enough so that I''m not surprised first generation games are having problems. Don''t worry though, it''ll happen

Share this post


Link to post
Share on other sites
DC have full screen antialiasing but with a ''crappy'' output like a TV it don''t need to be used, a good res might be enough. (640*480 is nice ,800*600 may not be required)

Daveb, I wonder how you manage the textures and what screen res you use.
Cause with only 4Mb of VRAM, you might have a problem.
I think that you use far less textures, or textures or far less quality than PC games, OR that you use the main memory to store the textures...

Maybe do you use streaming to load textures in memory too...

PS2 is far harder to program than a DC, and the DC have many interesting features not yet used in games.

Both console''s games will improve quality over time.

The PS2 will probably outcome the DC, but it''s not sure, and the time it''ll take will maybe give the PS2 edge @ the time of the DC2 or PS3 release...


-* Sounds, music and story makes the difference between good and great games *-

Share this post


Link to post
Share on other sites
If you want to see what the DC can really do, buy a VGA box and plug it into a computer moniter

the reason DC is better is because of the following:

o the SDK saga provides is bloody brilliant
o newer developers can just the the winCE architecture
o true 640x480 mode
o "simple" architechture (compared with the PS2)
o brilliant hardware manufactors
o networking support INCLUDED, while on PS2 it''s only a fantasy
o quake3 is comming out on it this summer

sorted

MENTAL

PS: i know it''s a bit offtopic, but i love bashing the PS2 and a love my DC.

Share this post


Link to post
Share on other sites
Vram isn't an issue on the PS2 simply because the DMA and bus bandwidth is so insanely high. When we were getting our renderer up in its early stages we actually had a one entry texture cache. That is to say, _whenever_ we needed a new texture we had to upload it. On a level with lightmaps and maybe 100+ textures, this can result in thousands (or 10's of thousands) of uploads per frame. The performance hit was negligible. Any experienced 3d engine guy on the PC will tell you that on _any_ PC with _any_ video card, this would bring the whole machine to a halt. We're talking multiple seconds-per-frame. The PC's bus is like one of those mini stirrer-straws compared to the PSX2's NYC Aqueduct. Literally.

This is not to say that you should lean on this feature heavily. On the PSX2, lots of developers literally talk about _cycle_ counts per vertex. That's a level of precision and speed PC programmers generally don't even think about. And as such, having to do crazy amounts of texture uploading can seriously impede on an ultra-high performance engine. But, a smart, simple texture cache can alleviate pretty much all of this.

Yes, it would be nice to have more Vram, but it really hasn't been an issue at all for us.

Edited by - daveb on 4/13/00 9:17:35 AM

Share this post


Link to post
Share on other sites