From Wikipedia, the free encyclopedia
Jump to: navigation, search

Why is this article so complex? Why are vertex transformations mixed with the rasterisation? Rasterisation itself has nothing to do with 3D in the first place! It's just a technical part of the process. Matrix calculations instead of illustrations? Really? If I did not already know how this stuff works I would never understand this article. -- (talk) 20:43, 29 April 2015 (UTC)


Is this a stub? I know stuff can be added, but it has a definition plus additional information. --Taejo 09:23, 8 August 2005 (UTC)

It has been significantly expanded (December 2005), and it's in pretty good shape now.Bryan 14:34, 13 December 2005 (UTC)
I do not think that this article is in a pretty good shape. Rasterization is just the process of converting primitives into pixels (or fragmengs), all the stuff about transformation with matrix and so on should be removed from here. ALoopingIcon 09:08, 7 February 2006 (UTC)
Agreed. The article should focus on algorithms like flood fill, Bresenham (lines and polygon edges), scanline rasterization (convex and concave polygons), and the half-space approach. Circles/ellipses might also be interesting. But not the transformation, clipping and texturing stuff. We could add that rasterization is sometimes mistakenly regarded as the whole process from vector representation to finished pixels, but refer to other articles for techniques that go beyond identifying the pixels to be filled. C0d1f1ed 14:58, 9 September 2006 (UTC)
Definitely. Though it's a good article, I think most of the information in it should more properly be put in the Graphics pipeline article, which is very bare right now. The only part this article needs to be about is the Scan Conversion section (which can be expanded a lot...) Valarauka 09:45, 17 September 2006 (UTC)
I fully agree with you. All these 3D computer graphics stuff should not be here, perhaps only a basic information and a link to another article containing this, like the Graphics Pipeline as you suggested. I don't see anything about algorithms for curves rasterization or so...its still very poor (xissburg 12:32, 9 November 2009 (UTC))
Obvious WP:RANDYism is obvious. Even in the 3D context, rasterisation couldn't possibly refer to the whole pipeline [1], which is basically what most of this article is describing. JMP EAX (talk) 08:54, 9 August 2014 (UTC)

2D vs 3D[edit]

I disagree with this assertion:

In normal usage, the term refers to the popular rendering algorithm for displaying three-dimensional shapes on a computer

I think the term also applies equally to 2D vector graphics, which this article doesn't cover at all except where they overlap. I'm not quite sure whether a split, rewrite or a page move is the best solution though. Thoughts? --Dtcdthingy 20:08, 13 April 2007 (UTC)

I agree and more importantly reliable sources agree [2][3][4]. I've tagged the article accordingly. I'll probably improve it a little if I can find the time. There are so many crap articles on graphics in Wikipedia you wouldn't believe... JMP EAX (talk) 08:15, 9 August 2014 (UTC)
The 2005 version [5] made a bit more sense. I guess nobody gave a shit about WP:V back then. JMP EAX (talk) 08:28, 9 August 2014 (UTC)
Most of the current disaster is actually the work of a single editor Special:Contributions/Mengusfungus, who thankfully didn't write anything else in Wikipedia... JMP EAX (talk) 08:30, 9 August 2014 (UTC)

bmatrix formatting strange[edit]

The first 'bmatrix' in the Transformations sections shows up unformatted on my screen, but I can't find why. Anyone an idea ? Mauritsmaartendejong 17:16, 12 July 2007 (UTC)

Rasterisation vs Rasterization[edit]

It should be decided which of the two words should be used through this article. As far as I know this is a battle of British vs American English, but I don't know which side wikipedia leans to. It seems that it, at some point, was decided to name the article Rasterisation, so I suggest that this method of spelling should apply for the entire article. --MGarde (talk) 18:31, 7 May 2008 (UTC)

WP:ENGVAR. JMP EAX (talk) 08:16, 9 August 2014 (UTC)

clipping after projection from 3d to 2d[edit]

Having the clipping stage after projection is too late as you can't divide by a w of 0 which might be the case if you don't clip against the near plane, dividing by a negative w would also lead to mirroring artifacts. —Preceding unsigned comment added by (talk) 08:39, 15 October 2008 (UTC)

Rasterizing photographs[edit]

This definition of rasterization is quite new and computer centrists.

In the '80s rasterization was the name of the process of turning a photograph (on a film) into a printing press ready state (into dots, as in dpi). Please see Raster image processor.

-- (talk) 07:26, 29 December 2008 (UTC)


This article probably frightens many readers with the mathematical matrix and technical explanation. Could someone add a picture that shows what rasterisation is in a simple way? Mlewan (talk) 07:49, 1 February 2009 (UTC)

Ah, Wikipedia. I came here to ask for the very same thing, three years later. I'd do it, but I don't want to fight the image removal bot for the rest of my days. --Steven Fisher (talk) 22:17, 28 July 2012 (UTC)

Rasterization is not Rasterbation[edit]

"The Rasterbator (online tool that allows you to upload and rasterise an image)" is not related to this

rasterbation: —Preceding unsigned comment added by (talk) 02:37, 1 November 2009 (UTC)

Clarifications needed for the matrix section[edit]

  • The way the article works right now, each vertex is represented by a column. The "homogeneous" variable, which is (usefully, always?) 1, is the last element. It would be useful to show an example point or two.
  • Because matrix multiplication is order-sensitive, the current wording of the article is wrong; it require T x V. I'm not sure why the phrase "left multiplication" is used, but Matrix Multiplication only uses the phrase in conjunction with scalars.
  • When I learned this in high school, I learned all the matrices with the directions opposite: with each vertex as a row, and transformations done by V x T. Is there any reference for either being more standard? (talk) 09:25, 23 December 2009 (UTC)

I just found, in the xrandr man page, an example of using the matrix orientation as in the article (but 2D).

Then I found a question (9.005) in the OpenGL [6], that says:

Column-major versus row-major is purely a notational convention. ... You can use any notation, as long as it's clearly stated.
so I guess it officially doesn't matter. There is a note about column-major format being confusing, but that is specific to the OpenGL in-memory format. The best thing to do would be to clearly state which is in use and link to Transformation matrix, which should be expanded. (talk) 21:53, 23 December 2009 (UTC)


"Since all modern displays are raster-oriented, the difference between raster-only and vector graphics comes down to where they are rasterised; client side in the case of vector graphics, as opposed to already rasterised on the (web) server.[1]"

The difference between vector and raster graphics is if they are rendered client side or prerendered server side? or just an excuse to introduce a reference to SVG?

Rasterization is the process of converting vector images to raster images. Has nothing to do with client/side nor web technologies. That rasterization is used in order to draw a vector image(and not only SVG) is not reason enough to intoduce a reference to SVG in this article. —Preceding unsigned comment added by (talk) 12:13, 15 January 2011 (UTC)

Scan Conversion[edit]

I removed links to articles on Scan Conversion, and Hidden Surface Removal, since neither are about scan conversion for rasterization. The Scan Conversion article previously mentioned is about horizontal scanning in television signals. The hidden surface removal algorithm, while relevant for earlier parts of the rasterization pipeline, also does not cover scan conversion. The scan conversion algorithms which should be mentioned are Bresenham's line algorithm, y-intersect scan conversion, etc. -- algorithms which turn a set of projected, vectorized geometry into a set of screen pixels, prior to depth testing. — Preceding unsigned comment added by (talk) 21:03, 26 June 2012 (UTC)

As a point of amusement, mostly[edit]

You can actually do rasterization in software on a [GP]GPU [7]. The (2011) paper does suggest that future GPUs might actually have a fully software-based rasterizer, just like vertex/pixel/etc. shaders have turned other fixed-function bits of the graphics pipeline into programmable ones. JMP EAX (talk) 15:54, 14 August 2014 (UTC)