Lossyprosessi

Lossyprosessi, or lossy process, refers to a process in which information is irreversibly discarded during encoding, so that the original cannot be perfectly reconstructed from the output. The term is commonly applied to lossy compression techniques used in audio, image, and video data, where reduced data size is prioritized over perfect fidelity. The key idea is to map the input data to a lower-dimensional or more compact representation by discarding components judged to be less important to the intended use or perceptual quality.

Lossy processes are characterized by a rate-distortion trade-off: increasing compression lowers data rate but increases distortion of the reconstructed signal. The extent of loss is typically controlled by a quantization step, perceptual models, and entropy coding. Common approaches include transform coding with quantization, subband coding, and perceptual coding using psychoacoustic or psychovisual models. Distortion is measured with metrics such as mean squared error, peak signal-to-noise ratio, or perceptual quality scores.

Practical examples include JPEG image compression, MP3 and AAC audio, and MPEG video. These techniques are designed to exploit redundancy and perceptual masking to remove information that is less noticeable to human senses. The advantages are significantly reduced storage and bandwidth, enabling efficient streaming and distribution. The main drawback is the potential for visible artifacts, audible distortion, or loss of fine detail, particularly at high compression levels.