Let's talk... PSP graphics + Design
- Thread starter Paul
- Start date
Uhmmm no.
I wanted a official PS3 topic to begin with as PS3 posts flood the board, PSP topics don't which is why I posted this one.
Can I control what other people post about PS3? No, so how would me only practicing my example prove that it could work if there was still ps3 topics around because of no moderator to delete them? Oh and PSP isn't "always" on the first page, Cryz bumped it when it was on like the 3rd page if you haven't noticed.
Show me where I said everyone should only post psp stuff in a psp topic. Your not "asking questions" your just trying to start trouble, which is something a 3rd grader would do. Typical.
Go ahead respond to this, because your not getting a response from me, nor do you deserve one at this point.
Panajev2001a
Veteran
PC-Engine said:
Is Flipper able to avoid rendering at all times any kind of scene in multiple rendering passes ?
No.
Hence what I did say still holds true.
marconelly!
Veteran
No idea because no one knows exactly what PSP would be capable of, but to say that screen size is the limiting factor to graphcis quality is wrong. If you play DVD on a 4.5" LCD you will sure as hell be able to spot the difference between the movie quality and a Dreamcast quality graphics.
Panajev2001a said:
Sure, but why would you need more than 8 texture layers displayed on a tiny 4.5" LCD?
I'm sorry but that is a flawed analogy.
You do not compare a DVD movie using 35mm film to 3D graphics...DUH of course you're going to see a difference. You're probably thinking, "what about FFTSW?". Well I'm sorry to burst your bubble but you won't be seeing too many details on that little 4.5" LCD
All of those millions of polygons in FFTSW that make up those little details won't be noticeable like they were on a good sized television set. The only thing you're going to notice are LARGE objects and characters near the camera. All the stuff in the background will be too small to see the little details.
You think developers are gonna spend a lot of time on high resolution textures and high polygon models so that nobody can notice them on that puny little screen???
zidane1strife:
Seems like another imbalanced design. Newer technologies are inevitably the future no doubt, but only when their efficiencies outweigh the economies of the old technology. With regard to memory architecture, a solution like PowerVR is so efficient it can compete using inexpensive, external RAM, allowing for enough that it'll have the resource advantage in the end. That'll mean less compromise between image quality and texturing.
That's a parallel to the symptomatology being displayed by some here. They claim the PS2 and its VUs had unmatched graphical performance in its price class for its time of release, yet the system wasn't even the performance leader. The same architecture family was still ahead: Naomi 2, with its Dreamcast chipset (not even clocked higher) and its year-later addition of ELAN T&L clocked at only 100MHz and coupled with a duplicate DC renderer to rebalance the load, sustains greater complexity and capability - 10 million polys/sec with 6 light sources (up to 16 lights maximum). Again, its efficiency enabled it to compete without using double data rate SDRAM, so its advantage was packing in the SDR SDRAM in large quantities.
Panajev2001a
Veteran
notAFanB said:
THANK YOU
Lazy8s,
Unbalanced design ?
Call me when the Dreamcast supports HOS surfaces tesselation in HW.
The ELAN, ONLY clocked at 100 MHz...
Do I need to remind you that the Elan basically only did T&L and had 10 MTransistors ?
That is almost the same amount of transistors the EE has ( ~12-12.5 ): unless you convince me that VU1 + VU0 took alone 10 MTransistors your point is pretty much moot: also consider that Elan is pretty much an Hardwired Fixed Function solution while the VUs are beyond Direct X 9 as far as degree of programmability is concerned.
Show me one pocket machine announced to come out within next 6months that can match even 50% of these specs. Hell, show me one that actually has 3d acceleration of any kind.Lazy8 said:
Name me one product that uses this?
It's not about displaying "more" layers. It should help to alleviate algorythms that are multipass by nature - and there's a couple that are becoming pervasively used in recent software.PCEngine said:
notAFanB said:
That wasn't my point. My point was Flipper can do 8 texture layers in a single pass before having to resend geometry. What effect requires multiple passes on Flipper if you're not looking for more than 8 texture layers? Are there other ways to do the same effect without extra passes on Flipper?
PC-Engine said:
my understanding is that PSP will not require to resend geometry either and texturing is thus as cost efficient (in theory) as flipper.
as to your question on what 'effects' require multiple passes I am not to sure myself, cubemaps? stencil operations? can anyone clarify?
EDIT:
anyone else concerned with the (lack of) memory here?
Panajev2001a
Veteran
Ok, PC-Engine:
Any effect that requires multi-pass rendering on Flipper does hit the T&L chip too because you have to re-transform an re-send the geometry to the Rasterizer.
With this P-buffer, the T&L engine does not need to re-transform the geometry that needs multiple rendering passes.
The P-buffer does not use part of the VRAM.
Any effect that requires multi-pass rendering on Flipper does hit the T&L chip too because you have to re-transform an re-send the geometry to the Rasterizer.
With this P-buffer, the T&L engine does not need to re-transform the geometry that needs multiple rendering passes.
The P-buffer does not use part of the VRAM.
Fafalada said:
