Archive: bug in texer?


22nd June 2003 22:58 UTC

bug in texer?
today i was working on some images to use them with unconed's texer ape. i figured there was always the very right pixel-column missing. not sure if i'm making something wrong or if it's the ape. please help.


23rd June 2003 07:15 UTC

maybe you're using more particles than specified in the ape.


23rd June 2003 11:14 UTC

i tried it with various scopes and i played around with texer's settings


23rd June 2003 15:32 UTC

I recognized this too, thought it was the scope.


23rd June 2003 18:39 UTC

just make a border around the image of a couple pixels.. it should work out fine


23rd June 2003 21:17 UTC

not in the texer image.

look at this:

x=-1 works, but x=1 doesn't


23rd June 2003 21:51 UTC

Now try this :
(no texer)
x=-1;
y=i*2-1;

Good, now x=1.

Good, now do the same with switched x and y.

It's an AVS bug.
I knew this for some time, thought it's a minor thing...
Maybe the positive sides are shifted one pixel due to the origin point.


23rd June 2003 22:44 UTC

Known issue, due to optimizations. Texer processes two pixels horizontally at a time, so if an image extends beyond the right edge and the visible part has an odd amount of pixels, then the last column will be missing. It happens on the left side as well, except that the missing column is on the right side of your particle (hence you can avoid it there by adding a black column of pixels to your texer image).

A fix inside AVS would be to align your dots at even pixels...

x=((((x+1)*.25*w)%w)/w)*4-1

or something like that.

I didn't think it would matter much because of the inherent limitation in the texer system that a particle disappears as soon as its center goes off the screen (hence, its dot would be gone). Using texer particles at the edges tends to be ugly.

I *could* fix it by reversing the drawing direction on the right side (so the cutoff is on the opposite side) and extend the image when loading, so that it has a black column on both sides.

Don't have time for that now though. If I update it I'll also add a scaling feature (optional of course) where it scales the dot positions so the particles move completely off-screen before disappearing.


24th June 2003 02:17 UTC

Whilst we are here could we also have different blend modes for the texer? Mainly max blend, i would like to use the texer for creating 'point sprite' type 3d models... but it is impossible at the moment.


24th June 2003 04:41 UTC

jheriko...

effect list (replace/Max)
Texer

you should get the maximum blend effect


24th June 2003 16:38 UTC

I think he means the mode for blending the particles with themselves. I'll see what I can do ;). Basically it's a lot of copy/pasting of assembly code from colormap.

I have a colormap update sitting here as well by the way, including the requested (R+G+B)/3 index mode, rather than (R+G+B)/2 which is the current 'brightness' (the old mode still works though).

Don't have time to wrap it up and publish it though.


25th June 2003 10:02 UTC

nic: yeah i noticed that too.. something to do with the way w and h are made into -1>1.
It seems to be just with the superscope though.
The low limit is always slightly less than -1, and the high limit is always slightly less than 1, so any points at 1 on x or y will never show up.
The actual difference changes based according to resolution and whether or not pixel doubling is on.
Try this: make the window pretty small, and then work out the lowest y value that the scope will work on. (eg. y=-1.00389105; x=2*i-1).
Then make the window nice and big again, and the line should disappear!!)


30th September 2003 12:58 UTC

lalala...:D


30th September 2003 13:25 UTC

I noticed another texer bug:

The track bar wont restore saved settings below the initial one.

Try saving a texer with one particle and reload the preset. :weird: