A - B   C - F   G - I   J - M   N - P   Q - S   T - Z   AAC Advanced Audio Coding, a technique standardised by MPEG-2 and MPEG-4 for improved audio coding, as a prospective replacement for the popular MP3; see details in lesson 12 absolute hearing threshold the minimum sound pressure level that is audible in silence adaptive quantization a process with quantization parameters being dependent from signal characteristics; see backward adaptation, forward adaptation amplitude spectrum a magnitude of a Fourier transform of a given signal, see details in lesson 5 aliasing a distortion whcih is a result of sampling a signal that is not band-limited; spurious frequency components resulting from interference of the signal content with the sampling frequency antialiasing filter a filter rejecting the spectral components at and above half the sampling frequency thus preventing aliasing; usually a lowpass filter at the front-end of a digital system; see details in lesson 12 arithmetic coding a complicated lossless coding scheme that assigns bitstreams to sequences of symbols aspect ratio a width-to-height ratio, see details in lesson 9 backward adaptation consists in process parameters estimation due to the values of samples already processed; does not need side information binary images (black and white images) images that exhibit only two values: 0 (e.g. black) and 1 (e.g. white); examples: fax pages, raster printed pages blocking effect a type of an artifact that results from block-based processing or compression (like e.g. JPEG), where blocks of image pixels are treated independently and thus affected by different amounts of distortion, see example blurring a result of high frequency loss due to lowpass filtering or many types of compression, blurred image appears unsharp, see example brightness the most common term that describes psychophysiological visual sensations; approximately proportional to logarithm of luminance CCD Charge-Coupled Device color image an image with vector-valued samples: each sample value has three components that correspond to the primaries of some color space color space a representation of a color samples, usually as a mixture of three primaries (red, green and blue for example), see details in lesson 1 convolution a mathematical operation converting two input signals in one output signal, see definition in lesson 5 compression any operation which allows to reduce the amount of data representing the original source, making it possible to transmit the programme with low bit rate (e.g. through Internet, satellite, mobile networks) or store the content in a file which is smaller than required by an uncompressed representation compression ratio a number describing the strength of compression, see definition DCT Discrete Cosine Transform, a tool for signal analysis, see details in lesson 7 DPCM Differential Pulse Code Modulation, a method used in lossless coding, see details in lesson 6 DVB Digital Video Broadcasting, a family of digital television systems (terrestrial DVB-T, satellite DVB-S, cable DVB-C): 576 lines, 720 pixels in each line (Europe) EDTV Enhanced Definition Television: 480 lines, 1280 pixels in each line (USA) entropy the minimal number of bits per input sample needed for coding of a given signal, see details in lesson 6 FIR Finite Impulse Response filter, see details in lesson 3 forward adaptation consists in process parameters estimation taking in account not only past samples but also the current sample; the current process parameters have to be sent to the decoder as side information Discrete Fourier Transform a tool for signal analysis as well as for analysis of signal transmission by linear systems, see details in lesson 5 frequency response a function which describes the way input signal frequencies are altered by a given system / filter gamma correction a point operation that compensates the nonlinearity of the monitor (cathode ray tube) transfer function, see details in lesson 2 Golomb-Rice coding a special version of variable length coding optimum for variables with exponential probability distributions HDTV High Definition TeleVision, a)a part of current Digital Television system (USA); 1080 lines, 1920 pixels in each line ; b)an older Japan system called Hi-Vision; broadcasting analog signal, everywhere else digital technology used hearing threshold the minimum sound pressure level invoking auditory sensation at given context; usually depends on frequency hi-color a color image representation using 16 bit per sample (for example 6 bits for green, 5 for red and 5 for blue component) histogram a function H(x) with the value H equal to the number of samples with the value x, see details in lesson 1 histogram equalization a non-linear point operation processing a given image in such a way, that in the output image histogram the numbers of samples are similar in the individual subintervals of its whole range histogram normalization a point operation that expands a given histogram to its whole range Huffman coding a lossless compression scheme based on variable length coding Human auditory system perception of audio information, at least as complicated as visual perception, see details in lesson 11 HVS Human Visual System, perception of visual information varying between individuals as well changing during lifetime, see details in lesson 1 IHS a color space with Intensity, Hue and Saturation primaries; see details in lesson 1 IIR Infinite Impulse Response filter image binarization a point operation that produces binary images impulse response a response of a given system / filter to a single, ideal impulse indexed color image an image representation where sample value is an index to a table of predefined colors called a palette interlaced sequence a sequence of images consisting two fields, one field consist of even lines, second field - of odd lines inter-frame coding a video coding mode where coded are only differences between two frames, see details in lesson 10 intra-frame coding a video coding mode where given frame is coded independently from other frames, using methods similar to used in static image coding, see details in lesson 10 JPEG ISO/IEC International Standard 10918: Information technology - digital compression and coding of continuous-tone still images, see details in lesson 7 JPEG-LS ISO/IEC International Standard 14495: Lossless and near-lossless compression of continuous-tone still images, see details in lesson 6 JPEG 2000 ISO/IEC International Standard 15444: Information technology - JPEG 2000 Image Coding System; see details in lesson 8 letterbox a method of displaying sequence prepared in 16:9 aspect ratio on a 4:3 display, see details in lesson 9 light radiant energy that produces visual sensations upon stimulation of human eye lossless compression (coding) any compression method for which it is possible to reconstruct the original programme without any impairments lossy compression (coding) a compression method much more efficient than lossy compression but with the decrease of reconstructed programme quality loudness a perceptual attribute of sound related to its level, see details in lesson 11 luminance luminous flux coming from an unit emitting surface (square meter) into an unit solid angle (steradian), see details in lesson 1 LUT Look-Up Table, a color table (see also palette) used in some point operations macroblock a block of image samples in video coding, see details in lesson 10 magnitude spectrum a part of Fourier transform of a given signal, see details in lesson 5 masking a phenomena which occurs if two or more simultaneous tonal or narrowband noise components are close to each other in frequency domain, see details in lesson 12 monochrome image (gray-level image) an image that exhibit scalar integer sample values, usually interpreted as a brightness levels MOS Mean Opinion Score, a result of the subjective image quality assessment, see details in lesson 4 motion estimation a process of finding for image block the most similar block in referent image, see details in lesson 9 motion vector a position difference between two blocks calculated by motion estimation. MPEG Motion Pictures Experts Group, ISO group responsible for standardization of compression techniques for video and associated audio, also an acronym for standards multispectral image an image with vector-valued samples with arbitrary number of components in various ranges of electromagnetic wavelengths, e.g. infrared, visible (see color image, monochrome image), ultraviolet near-lossless coding any coding scheme which introduces errors that are not percepted by humans normal observer a hypothetical person with some predefined properties of his visual system that correspond to average features of a healthy adult palette a table of predefined colors used by indexed color image; usually the palette contains 256 colors and is addressable by 8-bit indices palettization a vector quantization of color image which produces indexed color image pan and scan a method of displaying a sequence prepared in 16:9 aspect ratio on a 4:3 display, see details in lesson 9 PCM Pulse Code Modulation, digital representation achieved through sampling and quantization of analogue signal pel, pixel Picture ELement phase spectrum a part of Fourier transform of a given signal, see details in lesson 5 photometry the science that considers light measurements using models of the normal observer pillarbox a method of displaying sequence prepared in 16:9 aspect ratio on a 4:3 display, see details in lesson 9 a pitch perceptual attribute of sound related to the frequency of a periodic sound wave, see details in lesson 11 PNS Perceptual Noise Substitution, special technique in MPEG-4 for encoding noisy parts of the audio signal, see details in lesson 12 point operation sample-by-sample copying operation that modify scalar or vector sample values in a predefined way, see details in lesson 2 prediction a universal signal processing method that allows to approximate future samples of the signal using its statistical model and some samples already known; used in predictive coding predictive coding a prediction-based compression technique, used in lossless coding, see details in lesson 6 progressive (noninterlaced) sequence a sequence of images analyzed in video-camera and transmitted line by line PSNR Peak Signal to Noise Ratio, used in objective image quality assessments, see details in lesson 4 psychoacoustic model a model which predicts the behaviour of the human auditory system; usually used for calculation of the hearing threshold, see details in lesson 12 quantization mapping from one set of values to another ("smaller") set, for example from continuous analog values to discrete 8-bits values; an operation similar to mathematical rounding quantization table a set of scaling factors used in quantization of DCT coefficients, see example in lesson 7 radiometry the science concerning radiation measurements recursive filter an IIR filter realized as a circuit/algorithm with memory, where some returning paths from output to input exist (image) resolution defined by number of lines and number of pixels in each line, higher resolution allows representing finer details ringing noise a type of artifact related to the known Gibbs phenomenon, an oscillatory repeating intensity changes in places where originally one sharp transient has been observed, see example RGB a color space with Red, Green ad Blue primaries sample a value of a signal at a given time / position sampling taking samples from the analog signal, usually in equal intervals; in order to represent the input signal without significant loss, the rate of sampling must be at least twice the maximum input frequency (Nyquist law) SBC SubBand Coding; representing the input signal as a series of coded (usually compressed), band-pass filtered and subsampled signals, see details in lesson 8 scalar quantization color image quantization, where each component is quantized independently using its own quantization rule SDTV Standard Definition Television: 480 lines, 720 pixels in each line (USA) separable filter 2-D filter realized as a pair of two 1-D filters, for example first filtering the rows and then the resulting columns SNR Signal to Noise Ratio, used in objective quality assessments; for definition read lesson 11 spectrum a representation of signal in frequency domain; obtained through Fourier transform of the signal, see details in lesson 5 subsampling taking only part (for example only odd-numbered ones) of samples from the previously sampled signal, according to Nyquist law the input signal have to be band-pass filtered first TNS Temporal Noise Shaping, special technique for reducing pr-echo artifacts in transform coding of audio, see details in lesson 12 timbre a complex notion that reflects the structure of the sound event in time and frequency, see details in lesson 11 true color a color image representation using 24 bit per sample (8 bits for each color component) variable length coding a coding method where shorter codewords are assigned to the more frequent symbols vector quantization an extension of quantization to multi-dimensional signals or signals with many components; e.g. a color image quantization, where triples of component values are mapped onto other triples of component values according to a certain rule wavelet coding a coding method based on Wavelet theory; in practical digital systems same as SBC YCbCr a color space with luminance and two chrominance primaries, see details in lesson 1 zigzag scan a special ordering of 2-D DCT coefficients, see details in lesson 7