ok guys.. time to show how much you know.. T&L vs T-buffer...
- Thread starter WaterB
- Start date
Now wouldn't that be a matter of opinion? And wouldn't that get people started on the Nvidia vs. 3dfx camps? Nah, probably not. We're all big kids
Let's take a look at what each has to offer. The T-buffer has full-screen anti-aliasing (get rid of those nasty edges in 3D games), real time motion blur, depth of field blur, soft shadows, and soft reflections. What this all boils down to is a better looking 3D game. Something 2D like Fallout or Fallout 2 won't really be affected by this feature.
Now T&L, or texture and lighting, is something different. Where this is going to make a difference is with static geometry. In layman's terms - landscape. You'll be able to have more trees, bushes, and buildings with less load taken by the CPU. On top of this, static geometry is stored in local video memory. What does this mean? It means AGP bandwidth is conserved and thus CPU cycles can be off-loaded to other uses, like improving the AI to better whup ya. Now T&L cannot take care of things like vertex blending (smoothing joints on player models), so those are still run by the CPU, since T&L would have to pull all the model information across the AGP bus, more or less negating the advantages gained in static geometry.
That's just a very brief gloss of the advantages. I could go into more detail, but this post is already too long
Anybody else want to add anything?
Let's take a look at what each has to offer. The T-buffer has full-screen anti-aliasing (get rid of those nasty edges in 3D games), real time motion blur, depth of field blur, soft shadows, and soft reflections. What this all boils down to is a better looking 3D game. Something 2D like Fallout or Fallout 2 won't really be affected by this feature.
Now T&L, or texture and lighting, is something different. Where this is going to make a difference is with static geometry. In layman's terms - landscape. You'll be able to have more trees, bushes, and buildings with less load taken by the CPU. On top of this, static geometry is stored in local video memory. What does this mean? It means AGP bandwidth is conserved and thus CPU cycles can be off-loaded to other uses, like improving the AI to better whup ya. Now T&L cannot take care of things like vertex blending (smoothing joints on player models), so those are still run by the CPU, since T&L would have to pull all the model information across the AGP bus, more or less negating the advantages gained in static geometry.
That's just a very brief gloss of the advantages. I could go into more detail, but this post is already too long
torvus> I know you only did the brief version, but it should be pointed out that what you descrived is hardware T&L. Most 3D games use T&L, but it's rendered in software by the CPU. In the case of the new cards coming out, the T&L calculations and processing is done in hardware on the graphics card, and thus offloaded from the CPU.
Software T&L:Hardware T&L as software DVD decoding:hardware DVD decoding.
They basically look the same, but you get better peformance when it's done in hardware.
------------------
From another BBS:
"hmm... Cyrix III, huh? You can put syrup on sh*t, but that don't make it pancakes."
Software T&L:Hardware T&L as software DVD decoding:hardware DVD decoding.
They basically look the same, but you get better peformance when it's done in hardware.
------------------
From another BBS:
"hmm... Cyrix III, huh? You can put syrup on sh*t, but that don't make it pancakes."
