[solved]Depth Buffer configuration

OpenGL OpenGL: Basic Coding
1

Hello all.

I’m now implementing Depth peeling algorithm.

So i check my depth buffer. and i observed a strange comportement.

i try to draw (deferred) this :

     glBegin(GL_QUADS);
glPushMatrix();

            colorConvert(1001);
glVertex3f(-10, 0, -2 );
glVertex3f( 10,0, -2);
glVertex3f( 10,-10,-2); 
glVertex3f( -10,-10,-2); 

glPopMatrix();*/
glPushMatrix();

            colorConvert(1002);
glVertex3f(-15, 1, -3 );
glVertex3f( 15,1, -3);
glVertex3f( 15,-15,-3); 
glVertex3f( -15,-15,-3); 

glPopMatrix();
glEnd();

with this openGL configuration


glEnable(GL_DEPTH_TEST);
	glClearColor(255,255,255,255);

	// S e t t h e p r o j e c t i o n t r a n s f o r m
	glMatrixMode (GL_PROJECTION ) ;
	glLoadIdentity ( ) ;
	gluPerspective( 45, 100/100,1,100 ); 


	// S e t t h e camera o r i e n t a t i o n :
	glMatrixMode (GL_MODELVIEW) ;
	glLoadIdentity();	
	gluLookAt(0,0,5,0,0,4,0,1,0);

and i obtain this results ( only the important part)


row =0 :column =0 :0.865801
row =0 :column =1 :0.865801
...
row =49 :column =98 :0.865801
row =49 :column =99 :0.865801
row =50 :column =0 :0.882877
row =50 :column =1 :0.882877
...
row =50 :column =98 :0.882877
row =50 :column =99 :0.882877
row =51 :column =0 :0.882943
row =51 :column =1 :0.882943
...
row =51 :column =98 :0.882943
row =51 :column =99 :0.882943
row =52 :column =0 :0.883009
row =52 :column =1 :0.883009
...
row =52 :column =98 :0.883009
row =52 :column =99 :0.883009
row =53 :column =0 :0.883075
row =53 :column =1 :0.883075
...
row =53 :column =98 :0.883075
row =53 :column =99 :0.883075
row =54 :column =0 :0.883141
row =54 :column =1 :0.883141
...
row =54 :column =98 :0.883141
row =54 :column =99 :0.883141
row =55 :column =0 :0.883207
row =55 :column =1 :0.883207
...
row =55 :column =98 :0.883207
row =55 :column =99 :0.883207
row =56 :column =0 :0.883272
row =56 :column =1 :0.883272
...
row =56 :column =98 :0.883272
row =56 :column =99 :0.883272
row =57 :column =0 :0.883338
row =57 :column =1 :0.883338
...
row =57 :column =98 :0.883338
row =57 :column =99 :0.883338
row =58 :column =0 :0.883404
row =58 :column =1 :0.883404
...
row =58 :column =98 :0.883404
row =58 :column =99 :0.883404
row =59 :column =0 :0.883470
row =59 :column =1 :0.883470
...
row =59 :column =98 :0.883470
row =59 :column =99 :0.883470
row =60 :column =0 :0.883536
row =60 :column =1 :0.883536
...
row =60 :column =98 :0.883536
row =60 :column =99 :0.883536
row =61 :column =0 :0.883602
row =61 :column =1 :0.883602
...
row =61 :column =98 :0.883602
row =61 :column =99 :0.883602
row =62 :column =0 :0.883668
row =62 :column =1 :0.883668
...
row =62 :column =98 :0.883668
row =62 :column =99 :0.883668
row =63 :column =0 :0.883734
row =63 :column =1 :0.883734
...
row =63 :column =98 :0.883734
row =63 :column =99 :0.883734
row =64 :column =0 :0.883800
row =64 :column =1 :0.883800
...
row =64 :column =98 :0.883800
row =64 :column =99 :0.883800
row =65 :column =0 :1.000000
row =65 :column =1 :1.000000
...
row =99 :column =98 :1.000000
row =99 :column =99 :1.000000

As you can see there is a “first” part from row 0 to 49 where the depth correspond to the first quads (index 1001).

after that there is as increasing part from row 50 to row 65. this correspond to the 2nd quads.
After a first drop of 0,017076.
Each row the depth increase of 0,000066.

so i guess there is a linear depth “blur” or something like that.

I try to find out how to fix that but i cannot find it on internet.

To check my conjecture i’ve tried with only one quads ( the second index 1002) and i obtain the following results :


row =0 :column =0 :0.883838
row =0 :column =1 :0.883838
...
row =64 :column =98 :0.883838
row =64 :column =99 :0.883838
row =65 :column =0 :1.000000
row =65 :column =1 :1.000000
...
row =99 :column =98 :1.000000
row =99 :column =99 :1.000000

so i can conclude my conjecture is good.

so my question is : how to configure the depth buffer not to averaging the value.

thx.

2

What’s odd is OpenGL does not have a “depth blend/blur” function. Are you sure you’re not doing a blur on this buffer yourself?

Do you have multisampling enabled? If so, disable it (i.e. do not allocate a multisample render target and do not enable GL_MULTISAMPLE). Remote possibility that this might be related.

Also, make sure you have:

glDepthFunc( GL_LESS )
glDepthRange( 0, 1 )
glEnable( GL_DEPTH_TEST )
glClear( GL_DEPTH_BUFFER_BIT )

enabled before you draw anything. And just for good measure:

glDisable( GL_ALPHA_TEST )
glDisable( GL_BLEND )

I tried to produce your Z buffer numbers but didn’t, so I would guess your DepthRange wasn’t set to the defaults.

3

I’m not using any multisampling anywhere.

I’ve added what you purpose but not any effects on numbers.

so here is my complete code if you want to try it.

edit : I’ve found why it’s not working.
Only because before calling glbegin(GL_QUADS) , i forget an old glbegin(GL_TRIANGLES)

thx for helping :smiley: