per-scanline and per-pixel ?

vrmm

Newcomer
Could someone tell me the difference of the per-scanline and per-pixel model in the rasterization?
I want to know both advantages and shortcomings for them.

Could someone help me ?
 
vrmm said:
Could someone tell me the difference of the per-scanline and per-pixel model in the rasterization?
I want to know both advantages and shortcomings for them.

Could someone help me ?

Only if you help us and tell us what feature your interested in being per scanline or per pixel.
 
I am eager to know their working theroy.

For 3d pipeline, the speed is the key for them especially for soft implementation of the 3d pipeline.
Also, the flexibility of all the state combinations is also important.

I want to have comparsions for both the speed and the flexibility in 3d pipeline.

Thank you very much!
 
vrmm said:
I am eager to know their working theroy.

For 3d pipeline, the speed is the key for them especially for soft implementation of the 3d pipeline.
Also, the flexibility of all the state combinations is also important.

I want to have comparsions for both the speed and the flexibility in 3d pipeline.

Thank you very much!

i'll have to repeat ERP here - exactly what features of the rasterizer are you interested in?

see, in a streight forward scan-converting rasterizer very little is done per-scanline, if anything at all* - the majority of the things are done per pixel, some things are per primitive. but yes, one can come up with various optimisations on a scanline basis. so you need to specify exactly what you're interested in.

* the one thing that immediately comes to mind as per scanline is horizontal sub-pixel error correction, but that's about all that naturally needs to be per scanline.
 
Oh you could use per-scanline options to quickly add fake scanline in emulators LOL.

what do you mean of the fake scanline? Could you explain it ?
Thank you very much!
 
Back
Top