Talk:Z-buffering

Precision Disambiguation

Hmmm.. I agree that precision needs to be disambiguated [1], but I don't think accuracy and precision is the right meaning. In that sense, precision is the repeatability of measurements, or degree of mutual agreement between values... But what is meant in the z-buffer article is more about the ability to *distinguish* between two close values. This seems more related to resolution, actually, but that article doesn't seem to have the precise (heh) meaning that's called for (in fact, it links back to accuracy and precision).... hrm. PenguiN42 14:47, 26 Mar 2004 (UTC)

Mathematics section

I think 'Mathematics' is a poor name for the section heading. Perhaps 'Implementation' would be better? Also, I find the Javascript implementation at the end of the section inappropriate. It is basically the formula rewritten in code; supposedly Javascript code, but it is indistinguishable from 'pseudocode'. The <canvas> tag mentioned has no link or explanation, and isn't explicitly used in the code. --GrimRC 86.4.53.107 21:58, 6 July 2006 (UTC)

W-buffering merge

There is very little information at W-buffering, and there is not much hope for expansion as it is a very simple extension of Z-buffering. It should be merged here. - Rainwarrior 18:33, 14 August 2006 (UTC)

Most of the W-buffering is already in this article. Merging would probably be best. AdShea 13:35, 28 August 2006 (UTC)

Done. - Rainwarrior 17:44, 28 August 2006 (UTC)

vertex / fragment shaders don't solve the Z-fighting problem

In the article it is claimed, that with vertex or fragment shader no Z-fighting would appear and clip planes were obsolete. This is definitely wrong. At best a fragment shader can change the z value, but the depth test is still performed and suffers from aliasing for similair numbers i.e. Z-fighting. For this reason I removed that part from the article. —The preceding unsigned comment was added by 88.217.40.168 (talk) 02:59, 30 December 2006 (UTC).

Mistake in Z-culling?

"there is no canonical to pick a point on a triangle (e.g.: no matter whether one sorts triangles by their centroid or closest point or furthest point, one can always find two triangles A and B such that A is "closer" but in reality A should be drawn first)." Shouldn't it say that A is "closer" but B should be drawn first? 69.72.86.104 (talk) 02:49, 8 October 2008 (UTC)

It sure seems that way. Harutsedo2 (talk) 23:41, 17 October 2010 (UTC)

z-trick

At the start of a new scene, the z-buffer must be cleared to a defined value

However, below that is mentioned that you could also use a one frame positive, one frame negative trick. Is this the trick that Quake used? 82.139.86.96 (talk) 02:38, 24 March 2009 (UTC)

Yes, you can actually read the implementation here https://github.com/id-Software/Quake/blob/bf4ac424ce754894ac8f1dae6a3981954bc9852d/WinQuake/gl_rmain.c#L995 — Preceding unsigned comment added by 91.164.6.153 (talk) 22:49, 15 July 2018 (UTC)

Multiple Z-Buffers

The Smallgantics technique uses up to 8 z-buffers to achieve a visual effect of big scenes looking small. Write that to the article? --Fabrício Kury (talk) 20:16, 9 March 2012 (UTC)

The example image provided is wrong

The grayscale image provided as example of a Z buffer is not correct. The objects in the z-buffer are not represented with solid shades of gray, but rather with gradients corresponding to how far each individual point is from the image plane.

Even if we decided to consider the given z-buffer as a simplified one where each object in the 3D scene is represented as being wholly at a single point in distance from the image plane, then the floor itself wouldn't be a gradient either, but rather solid grey, too. --Fabrício Kury (talk) 20:29, 9 March 2012 (UTC)

Also, in the buffer demonstration image, the cube is much darker than the ground underneath. If the image were correct, both of them should have nearly the same shades of gray instead, don’t you think? Gamerhuy (talk) 12:03, 2 June 2020 (UTC)