differences in jit.gl.render between 1.6 & 1.7?

Aug 13, 2008 at 11:48am

differences in jit.gl.render between 1.6 & 1.7?

Hello all.

I’ve been noticing *something* different in some of my gl patches as
of late, and I think I found it. I am not sure if it has been
mentioned so far.

In max Max 4.6/Jit1.6, when using [jit.gl.render @blend_enable 1
@erase_color .xx .xx .xx .01] (or some variant in which alpha is < 1.
, much more noticeable as it approaches 0.0), a jit.window will draw
as expected, and when switching to fullscreen, the color xx. xx. xx.
will present itself as the background.

In Max 5, running the same patch, the window will render the
background color as expected, but when made fullscreen, the window
draws itself black, then slowly fills in the background color. When
the alpha is set to the lower bounds, < .10 for example, the color
will not fully come up.

I would expect this to behave as it did in Max 4.6/Jit1.6.

I have a simple patch that illustrates the example below (obviously
you’ll need to run it in max 4.6 & 5 to see the difference).

Thanks for any explanation, fix, or clarification.

Macbook Pro 2Ghz, OSX 10.5.4, ATI Radeon X1600, Max5.0.4/Max 4.6.3 Jitter 1.6.4

~scott

#P window setfont “Sans Serif” 9.;
#P window linecount 3;
#P comment 315 267 100 196617 the fullscreen color is identical in 4.6
, but not so in 5;
#P window linecount 2;
#P comment 179 233 100 196617 3) press escape to fullscreen.;
#P comment 177 124 100 196617 2) lookit the color in the window!;
#P window linecount 5;
#P comment 409 36 100 196617 why are the fullscreen colors different
between max 4.6/jit1.6 & max 5?;
#P toggle 76 73 15 0;
#P window linecount 1;
#P message 91 249 68 196617 fullscreen $1;
#P toggle 74 249 15 0;
#P newex 116 224 35 196617 sel 27;
#P newex 74 224 40 196617 key;
#P newex 91 274 82 196617 jit.window huh?;
#P newex 76 181 309 196617 jit.gl.render huh? @blend_enable 1
@erase_color 1. 0.8 0.6 0.01;
#P newex 76 137 58 196617 t b b erase;
#P newex 76 105 51 196617 qmetro 2;
#P comment 99 76 100 196617 1) start the metro;
#P fasten 5 0 6 0 115 241 115 226;
#P fasten 7 0 8 0 90 264 90 249;
#P connect 8 0 4 0;
#P fasten 6 0 7 0 121 246 79 246;
#P connect 1 0 2 0;
#P connect 2 2 3 0;
#P connect 2 0 3 0;
#P connect 9 0 1 0;
#P window clipboard copycount 14;

#39260
Aug 13, 2008 at 4:41pm

http://ccrma.stanford.edu/~pkmital/see/jittererasecolor.avi

The first part shows Scott’s patch in Max5 and Jitter 1.7 and the
second part in Max 4.6.3 Jitter 1.6.3.

MBP3,1 OS X.5.4

#137934
Aug 14, 2008 at 2:38am

i can confirm this, and we’re taking a look.

in the meantime, in case you’re not already doing this, you can just send erase_color with an alpha of one before going full screen, and erase_color with your desired alpha immediately after.

-rob

#137935
Aug 14, 2008 at 1:26pm

Thanks for looking into this.

Anecdotally, and I have no way of making a quantifiable test right
now, it seems when I enter fullscreen mode with OpenGL renders, the
framerate takes a larger hit than it used to. Don’t know if it is
related or not, and if there’s a way I can verify, I’ll pass it along.

~s

On Thu, Aug 14, 2008 at 4:38 AM, Robert Ramirez wrote:
>
> i can confirm this, and we’re taking a look.
>
> in the meantime, in case you’re not already doing this, you can just send erase_color with an alpha of one before going full screen, and erase_color with your desired alpha immediately after.
>
> -rob

#137936
Aug 23, 2008 at 8:35pm

The render-erase-fullscreen issue has been addressed for the next
incremental. Thanks again for the report.
wes

#137937

You must be logged in to reply to this topic.