20.1: Invited Paper: What Is A Pixel?

Most computer graphics researchers would agree that the pixel is the fundamental atomic element of imaging. But what is a pixel? The simple notion of a hardware displayable pixel becomes far more complex when it is examined more closely. Early 2D windowing systems considered a pixel to be a little s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:SID International Symposium Digest of technical papers 2000-05, Vol.31 (1), p.285-287
1. Verfasser: Blinn, James F.
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Most computer graphics researchers would agree that the pixel is the fundamental atomic element of imaging. But what is a pixel? The simple notion of a hardware displayable pixel becomes far more complex when it is examined more closely. Early 2D windowing systems considered a pixel to be a little square. A more enlightened signal processing approach thinks of a pixel as a point sample of a continuous function. Many signal processing problems occur since physical pixels can only have positive values. Hardware pixels are usually given with 8 bits of resolution. But the precise translation of these bit patterns into physical intensities is often not specified. More careful treatment of intensity encoding takes into consideration gamma correction and dithering, as well as the intensity sensitivity of the human eye. Color pixels are generally thought to have Red, Green and Blue components. A closer examination leads to the concept of a pixel as some projection of a full spectral distribution that must take into account the response of the eye and the physics of the display emitters. In addition, some modern displays, like LCD panels, have a predictable geometric offset between the color emitters. This has implications on the proper spatial filtering of an image in addition to any color matching problems. Compositing and blending images leads to the creation of a fourth channel, called alpha, to represent an amount of geometric coverage. The exact meaning of this transparency channel has also been a subject of debate. Furthermore, the traditional single alpha channel is often used to represent colored object transparency. This is another approximation to a more accurate approach that progresses through usage of three alpha channels per pixel to, ultimately, a whole spectral distribution for transmission as well as emission. Finally, there are reasons to specify still more quantities, such as depth information, on a pixel‐by‐pixel basis. Ultimatly, like the blind men and the elephant, a pixel means different things to different people. In fact, when we consider the proper handling of color, geometry and transparency we realize that an ordinary hardware pixel is simply a bad image compression technique.
ISSN:0097-966X
2168-0159
DOI:10.1889/1.1832938