Ääniditheringissä

Ääniditheringissä is a Finnish term referring to the process of dithering audio. Dithering is a technique used in digital audio processing to reduce the audible artifacts that can arise when quantizing (reducing the bit depth of) a digital audio signal. When audio is converted from a higher bit depth to a lower one, the amplitudes of the samples are rounded to the nearest available level. This rounding introduces a form of distortion known as quantization error or quantization noise.

Dithering works by adding a small amount of random noise to the audio signal before it is quantized. This intentionally introduced noise has a random distribution and effectively masks the structured quantization errors, transforming them into a more broadband, less noticeable hiss. While it might seem counterintuitive to add noise to audio, the resulting random noise is generally perceived as less objectionable and more natural-sounding than the harmonic distortion or "steps" that would otherwise be present.

The amount and type of dither can vary. Common dithering algorithms include triangular-probability distribution (TPDF) and noise-shaping dither. Noise-shaping dither attempts to push the dither noise out of the audible frequency range and into higher frequencies, where it is less likely to be perceived. Dithering is particularly important when reducing bit depth for final mastering, such as from 24-bit to 16-bit for CD audio, or when preparing audio for streaming services that may have lower bit depth requirements. Without dithering, the audible consequences of bit-depth reduction can be significant, especially on quiet passages or sustained notes.