Computer Graphics

Image resolution describes the detail an image holds. It has three different meanings: spatial resolution, dynamic range, and chroma sampling.

Spatial resolution[]

When the meaning is spatial resolution, it refers to the amount of spatial detail in the image. The term is often used in relation to digital images, but is also used to describe how much detail can be recorded on photographic film, on analog videotape, by an eye, etc. Higher resolution means more image detail, i.e. small details are apparent. For digital raster-images, the convention is to describe the image resolution with the set of two positive integer-numbers, where the first number is the number of pixel-columns (width) and the second is the number of pixel-rows (height). The second most popular convention is to describe the total number of pixels in the image (typically given as number of megapixels), which can be calculated by multiplying pixel-columns with pixel-rows. Other conventions include describing resolution per area-unit or resolution per length-unit such as pixels per inch. A more technical convention is described under optical resolution. Below is an illustration of how the same image will appear at different resolutions.

File:Resolution illustration.png

In sampled images or other sampled data, the resolution should usually not exceed the sampling rate, or there will be aliasing.

The spatial resolution may be described by the number of resels (resolution elements).

Dynamic range[]

When the meaning is dynamic range, it refers to the amount of contrast available in each pixel. Dynamic range capability is limited by both capture and display capability in terms of the contrast range. Dynamic Range is often erroneously linked to bit-depth capability, which only determines the shades available to map the range from black to white, but does not inherently describe the dynamic range that those codes can represent. Codes can be allocated linear to light, or gamma-corrected. Bit depth can impact dynamic range capability if you include the caveat of banding or contouring issues. If you have an insufficient bit-depth for a given dynamic range, the discrete steps between one level and the next can become distinctly visible, and gradations in the image will not appear smooth; instead they will have visible banding problems. But the bit-depth itself does not inherently determine the dynamic range capability, and may often see erroneous information such as that 8-bit content has a dynamic range of 255:1.

Chroma sampling[]

In color digital systems, the luminance and chrominance may be sampled at different intervals. These give rise to the 4:2:2 and other sampling conventions, especially in relation to high-definition television. Greater chroma subsampling gives rise to greater spatial resolution when color is a discriminator.

Resolution in various media[]

Image display[]

Resolution can be a measurement of the total number of pixels displayed but this applies mainly to static images and can be affected by how it is displayed. To calculate the total number of pixels in an image, the number of pixels in the horizontal axis can be multiplied by the vertical axis. On monitors used for television diplay the longest axis is the horizontal axis and the shortest is the vertical axis. The ratio of the two is the aspect ratio (image).

  • Standard-definition television is most commonly the 480i system. The 480i system uses 640 horizontal lines x 480 vertical lines giving 0.31 megapixels per frame.

The format depends on which international television standard (ITU-R, IEC/ISO MPEG, DVB, SMPTE and ATSC) or computer display standard (VESA) is being used. In analog systems there are no pixels on a picture tube display and resolution is affected by the frequency response of the video amplifier. In digital systems it may be expressed as the number of lines and the number of picture elements on each line. For example, a 525 line NTSC component Y, Cb, Cr, picture in a digital studio format would have an "active picture" of 720 pixels on 480 interlaced lines, which usually is displayed in a 4:3 aspect ratio. Progressively scanned — this results in "square pixels" which are easier to manipulate in digital effects processors. In the USA, the SD part of the ATSC digital standard specifies the number of horizontal pixels reduced by 16 to 704, mainly to avoid problems with MPEG compression of edge of picture errors.

In systems that use the 625 line system (usually referred to as PAL), the digital studio format also has an "active picture" of 720 pixels, but on 576 interlaced lines (576i). In Europe, Asia and Australia, those pictures are increasingly produced and seen in 16:9 aspect ratio. This effectively reduces the hoizontal resolution because the same number of pixels are spread over a longer line. In fact, the CCIR, now the ITU-R in Recommendation 601 specified that a 16:9 picture should have 960 pixels per line but this was virtually ignored because it required extensive modification to the TV studio digital equipment. Fortunately at 720 pixels per line, the horizontal resolution is higher than what can be obtained by the NTSC or PAL analog off-air transmissions, so the effect is not so noticeable.

It must be noted that the vertical resolution is affected by the picture capture process and the type of display (i.e. picture tube or flat panel). Interlaced displays which split a picture frame into 2 sequential fields carrying the odd lines and then the even lines, cannot show detail equal to the total number of lines but is typically limited to about 70% — thus a 480i picture has a vertical resolution of about 350 lines and a 576i picture, about 400 lines). Part of the problem with interlaced pictures is that anything moving in the picture will be in different positions on the odd and even fields because the fields capture the movement at different times — 16.7ms in NTSC 60 Hz and 20ms in PAL 50 Hz. If viewed on a progressive display, due to interlace capture time differences, the edge of moving objects is not continued on adjacent vertical pixels and so toothed edges are seen on moving objects. Extensive processing is required to reduce this effect. A progressive scanned picture may show almost the full vertical resolution dependent on any camera or other processing.

A low frame per second progressive scanned picture such as film at 24 frame per second, may be carried in an interlaced format in 2-3 pulldown mode in an interlaced 60 field per second feed, or if speed up by 4% to 25 frames per second, then in an interlaced 50 field system. Such film material if intended to be displayed on an interlace display (standard picture tube), must be preprocessed to reduce vertical resolution or there will be objectional interline twitter/flicker on any fine detail — a typical example is a picture with a multistory building in the background where the lines of the multi-floors may appear to flicker. Most DVDs of films while effectively being progressive scan, have this vertical filtering so standard DVD players with progressive output only benefit by having component analog Y, Pb, Pr or DVI/HDMI digital Y, Cb, Cr connection to the display.

  • High-definition television uses either 1,080 or 720 displayed vertical lines, with a corresponding 1,920 (maximum) or 1,280 pixels per lines in a 16:9 aspect ratio. The 720p standard is defined in SMPTE 296M and 1080i in SMPTE 274M and 295M. 720p has square pixels as does 1080 but only in progressive mode. 1080p in film mode (i.e. at 24 or 25 frames/sec) may be the HDTV standard used in cable television, but could be broadcast with 1440, 1280 or 960 pixels x 1080 lines. A 1,920 x 1,080 frame has a maximum of 2.1 megapixels, and a 1,280 x 720 line, pixel frame has 0.9 megapixels.
  • 35mm motion picture projection film depends on the grain size and the picture area used on the film. It can be digitised in a telecine to 2K pixels and 4K pixels. The final projection resolution depends on the type of film used and the number of intermediate copies between the original camera negative and the final projection print. A camera negative can have 6K and the projection print as few as 2K. From the original camera negative, interpositive prints are made, then a final print, and once approved, internegative prints, and finally distribution prints. Each analog transfer reduces the resolution.

Image capture[]

  • 35mm original camera negative motion picture film captures about 6,000 lines or 6K when using the best lenses and the finest grain film. [1]
  • 35mm motion picture film, as of 2006 is commonly scanned to a digital intermediate (d.i.) format at 2,000 lines, called 2K scanning. The most common "2K" frame size is 2,048x1,556 pixels.
  • 35mm motion picture film can be scanned to a digital intermediate at 4,000 lines, called 4K scanning. [2] This anticipates being used in digital projection or higher resolution, future flat panel displays. The digital cinema initiative standard include both 2K and 4K resolutions for projection. Each 4K frame uses about 48 megabytes of digital storage space. Each 2K frame uses about 12 megabytes. The extra resolution offered by 4K scanning can often prove valuable for chromakey (greenscreen), or other post-production special effects work. The final work is then printed back onto 35 mm film.
  • Sony's and others' digital cinematography cameras as of 2006 capture at nearly 2K. While the lower eschelons of "HD" are reserved for 1080i, and a very few 720p cameras, both Sony and Thomson Grass Valley produce cameras capable of acquiring true 1080p content.

A common misconception is that "HDV" is equivalent to previous high definition broadcast standards (such as HDCAM). In reality, HDV is a consumer standard (with much less inherent quality), and the price of HDV cameras reflects this disparity. One similarity between HDCAM and HDV formats is that although both formats offer the user a 1,920x1,080 signal, that signal is actually upsampled from the 1440x1080 (non-square pixel) format written to tape. HDV is indeed a subset of the ATSC (Advanced Television Standards Committee) HD specification, in both 720p and 1080i formats. 720p is not sampled from non-square pixels, it is a square pixel format. Additionally, while HDCAM and HDCAM SR are square pixel formats at 1920x 1080, these formats too, are upsampled from smaller sized imagers/sensors, just like HDV is. Additionally, the "Lower eschelons" of HD are not relegated to 1920 x 1080, but rather, the specific definition of HD is 1920 x 1080 or 1280 x 720. Anything outside that realm is not HDTV by definition.

External links[]

See also[]

  • Display resolution
  • Raster graphics
  • Optical resolution
  • Chroma subsampling