PSOrvia

Glossary

Technical terms from the learning guides, briefly explained.

Bass
The low range of a sound. In these guides everything below about 250 Hz: bass drum, bass line, the warmth of the sound.
Bit depth
How many steps a digital sound has for each value: 16 bits are 65,536 steps, 8 bits only 256. The fewer the steps, the larger the rounding error.
Bitrate
How much data per second a lossy format uses for the sound, measured in kbit/s. The lower, the smaller the file and the more is missing.
Compressor
A tool that makes loud passages quieter. Afterwards the whole piece can be made louder without overloading. Dynamics are lost in the process.
Correlation
Measures how similar the two stereo channels are: +1 identical, 0 independent, −1 opposing. Strongly negative values mean that parts cancel out in mono.
Crossfeed
On headphones, mixes a delayed, duller copy of each channel under the other, as happens by itself with speakers.
Distortion
An error that follows the sound and adds components that were not there before. It sounds rough or grainy, not like even noise.
Dither
A breath of noise added before the rounding. It turns the rounding error into even noise, so quiet sounds stay audible.
Dynamics
The distance between the quiet and the loud passages of a piece. If a piece is compressed, it shrinks.
FLAC
An open, lossless audio format: the file is smaller than uncompressed and still returns every value.
Frequency
How often an oscillation repeats per second, measured in hertz (Hz). Low tones have low frequencies, high tones high ones.
Highs
The high range above about 2.5 kHz. It gives shine and clarity, but carries little energy.
LUFS
The unit of loudness: Loudness Units relative to full scale. The closer to 0, the louder; a quiet piece sits at −23, a very loud one at −9.
Lossless
A format that returns the sound as it was, value for value, for example FLAC. The file is smaller than uncompressed, but not small.
Lossy
A format that leaves out parts that the ear hears poorly, for example Opus. The file becomes much smaller, but what was left out does not come back.
Loudness
How loud a sound arrives in the ear. It is not measured at the peak, but averaged over the whole piece and weighted according to hearing (ITU-R BS.1770).
Mids
The range between bass and highs, here from about 250 Hz to 2.5 kHz. Most instruments and speech lie here.
Mono
Both ears hear the same signal. A piece that also sounds good in mono is mono-compatible.
Noise shaping
Moves the dither's noise out of the frequencies where the ear is sensitive into higher ones. Overall there is more noise, but it disturbs less.
Opus
An open, lossy audio format (RFC 6716) for music and speech, good at low bitrates. PSOrvia has its own encoder and decoder.
Polarity reversal
One stereo channel is flipped: it swings against the other. Audible as a missing anchor in the center; in mono, identical parts cancel each other out.
Quantization
The rounding of every value to the nearest step of the bit depth. The quieter the sound, the more audible the rounding error.
Spectrum
Shows how strongly which frequencies occur in a sound: low on the left, high on the right, loud at the top.
Stereo
Two channels, one per ear. Differences between them create the impression of width and space.
TPDF dither
Dither with a triangular distribution, the sum of two random values, about one step in size. Its noise is independent of the sound.