In response to Maximus_Alex2003
Maximus_Alex2003 wrote:
Let's be honest, the Bills will never have a year.

There's a reason I refer to them as perennially disappointing in my author bio.

Also, which fair did you go to? I've gone to the Genesee and Niagara county. I'm avoiding Erie county for a while since that whole tornado.

You're thinking of county fairs. I haven't been to a county fair since I was a kid. I'm talking about the Great New York State Fair.
County Fairs are the best though. Especially if you live in a really country area.
Except you forgot about the pig races, tractor pull/races, motherfuggin animalss, shitty local art, and filthy 60+ year old rides that still have the puke stain from when you were 6!
Yut Put wrote:
we have all those things as well. do u have the demolition derby and inevitable cow shit under your shoe after leaving

i've lived on a farm my whole life. Not only is cow shit integrated into my body on a molecular level, but my brain associates the smell with positive nostalgia so cow shit smells really fucking good imo. (but yes, yes i do)
I've lived within about 20 minutes of the state fairgrounds pretty much all my life. The Fair is in my blood. Also in my belly, butt, arms, thighs, and face.
In response to Lummox JR
Lummox JR wrote:
I've lived within about 20 minutes of the state fairgrounds pretty much all my life. The Fair is in my blood. Also in my belly, butt, arms, thighs, and face.

By 'The Fair' you mean cow poop, right?
In response to Lummox JR
Lummox JR wrote:
There's a reason I refer to them as perennially disappointing in my author bio.

Fair enough. It's odd being a Colts fan in a Bills area. Everyone around here mocks me, haha.

You're thinking of county fairs. I haven't been to a county fair since I was a kid. I'm talking about the Great New York State Fair.

Syracuse is like 3 hours away, not worth for me going.
In response to Maximus_Alex2003
Maximus_Alex2003 wrote:
Fair enough. It's odd being a Colts fan in a Bills area. Everyone around here mocks me, haha.

As well they should. Back when you had Peyton at least you had an excuse.

Of course my cousin recently moved back to Syracuse after spending over 20 years in Indy, so he became a bigtime Colts fan. But he gets a pass.
In response to Lummox JR
Lummox JR wrote:
As well they should. Back when you had Peyton at least you had an excuse.

Of course my cousin recently moved back to Syracuse after spending over 20 years in Indy, so he became a bigtime Colts fan. But he gets a pass.

Peyton, even retired, still doesn't wanna quit completely.

The Bills just can't seem to make up their minds about players and coaches, or whether they should play all running or passes.

I was born in Indy, funny enough.
I'm trying to figure out how doable an emboss filter is--in terms of something that would add a soft bevel, since a raw emboss is near-useless in a simple filter stack.

Motion blur should be easy (since it falls out of the Gaussian blur filter) except it will be bidirectional; that has some drawbacks from a gaming standpoint, though it might still be worth including. Single-direction blur is something pretty much nobody covers. The math to do it such that long blurs can be done by successive calls apparently doesn't exist.

I've got animation working for filters now.
Emboss would be really nice! I've really appreciated using emboss for videos, and I bet it would find great uses in games.
The main problem I'm having with an emboss concept is that unlike, say, Photoshop's emboss, a regular emboss filter does not have a size component. Therefore some kind of filter would be needed as a next step, or as part of the first, that pushed values toward their extremes, did a blur, and pulled them back; and if I created a special blur to do that, I suspect it is not separable. If I do the move-to-extremes part and it's not separable, I need to be sure that the buffer retains the actual floating point values I want and doesn't try to encode them as clipped colors.

My thinking is something like this:

Step 1: Emboss (alpha only)
Step 2: Calculate c' = (c-bias)power for every pixel
Step 3: Horizontal blur
Step 4: Vertical blur
Step 5: Calculate c' = c-power+bias for every pixel
Step 6: Apply result as highlight/shadow to image

If 2-3 and 4-5 can be combined into single steps, that'd be ideal. The question is whether the math would still be correct, because if the buffer doesn't retain its values as floating point--which might differ from system to system for all I know--very high or very low values from the power filter will be cut off and the result will be awful.
Hmm, and some Googling tells me my concerns were warranted.

The shader model I'm currently using will clip to 0-1 regardless. If I go up to a newer shader model, which seems pretty reasonable at this point, it will depend on the render target and I could, in theory, setup a render target that uses a fully floating-point format as a temporary buffer between 2-3 and 4-5.

I suppose it might be mathematically correct to do it this way:

Step 1: Emboss (alpha only)
Step 2a: Calculate c' = (c-bias)power for every pixel in blur
Step 2b: Horizontal blur
Step 2c: Calculate c' = c-power+bias for result
Step 3a: Calculate c' = (c-bias)power for every pixel in blur
Step 3b: Vertical blur
Step 3c: Calculate c' = c-power+bias for result
Step 4: Apply result as highlight/shadow to image

Maybe that's what I should aim to do.
Page: 1 2