Terminology

This page defines the principal terms used in this manual and the product.

These terms describe musical time and the physical or perceptual properties of sound. It is necessary to understand this in order to work with parameters.

A measure, also called a bar, is a structural interval of musical time bounded by bar lines. Its time signature specifies the number of beats in the measure and the note value assigned to one beat. For example, a measure in 4/4 contains four quarter-note beats. A measure does not have a fixed duration in seconds; its duration depends on the tempo.

Tempo is the rate of the musical pulse. It is commonly expressed in beats per minute (BPM). In Amadeus, the BPM value describes quarter-note beats: at 120 BPM, one quarter note lasts 0.5 seconds. A tempo map can assign different values to different positions in the score. This changes the relationship between written musical time and elapsed time, but it does not change the notes’ written positions or durations.

Tone is an ambiguous term. In different contexts it can mean a sound, a whole-step interval, or the character of a sound. For vocal melody, the more precise term is pitch: the perceptual attribute that orders sounds from low to high. A written pitch is the discrete target assigned to a note, whereas a sung pitch is a continuous contour that can move between written targets.

Pitch intervals are commonly expressed in semitones and cents. One semitone equals 100 cents, and one octave equals 1,200 cents. Cents are a logarithmic relative measure, so the same number of cents represents the same musical interval at every register.

Frequency is the physical rate of periodic vibration, measured in hertz (Hz). One hertz is one cycle per second. In a voiced sound, the fundamental frequency, or f0, is the lowest periodic component and is closely related to perceived pitch. The two concepts are not identical: frequency is a physical measurement, while pitch is a perceptual property influenced by the sound’s spectrum and listening context.

The interval between two frequencies can be expressed in cents as 1200 * log2(f2 / f1). Under conventional equal-tempered tuning, MIDI note 69 corresponds to A4 at 440 Hz.

Because Amadeus is a music studio, understanding this terminology is necessary for using its editing and vocal-synthesis features correctly.

A note is a discrete, authored musical event. It has a position, duration, and pitch in the score. In Amadeus, it may also have a lyric slot, a slur relationship, and note-level pitch expression. A note is a notation-level object; it is not the same as a phone, which is a speech-sound unit, or an audio sample, which is a rendered signal.

A slur connects successive notes that are performed as one continuous syllable or legato gesture. In vocal writing, a slur commonly represents a melisma: one lyric-bearing syllable extends across multiple written notes. The notes retain their individual written pitches, while the rendering process may produce a continuous transition between them. In Amadeus, an authored slur is represented by an explicit start-to-stop relationship.

A tie connects notes of the same pitch so that they are heard as one uninterrupted duration. It joins note durations across a notation boundary and does not normally introduce a new syllable or attack. A slur and a tie are therefore not interchangeable: a slur connects musical events, while a tie continues the same pitch.

Lyrics are the textual material assigned to a vocal performance. A lyric token is commonly a word or syllable associated with a note. The active language package interprets the token and produces a pronunciation plan made of phones. By default, one syllable belongs to one note. When one syllable continues across several notes, its lyric belongs to the first note and the following notes form the slurred continuation.

A voice is an independent monophonic vocal part in a score. It contains its own notes and lyrics and can have its own singer model, timing, pitch, and rendering settings. In imported MIDI material, a voice is commonly associated with one MIDI channel, but the musical concept of a voice is the independent part, not the channel number itself.

A stem is the rendered audio output for one voice, normally exported as a mono audio file for subsequent mixing. A stem is a production artifact rather than a score object. If the score or voice settings change, bake again before exporting so that the stem represents the current material.

A file format is a standardized representation used to store and exchange score data. Before importing a score into Amadeus, prepare it in notation or sequencing software that can produce a supported format. For classical, neoclassical, or orchestral music, notation software such as MuseScore is generally appropriate. For modern music, a digital audio workstation (DAW) or notation application is also suitable if it can export MIDI or MusicXML.

The important consideration is not the source application but the information preserved by the exported format.

MusicXML is a notation-interchange format. It represents structured score information and generally preserves more written detail than MIDI, including lyrics, ties, slurs, tempo, and meter. MusicXML is generally preferable when the source contains detailed vocal notation.

MIDI is an event-oriented music-interchange format, not an audio format. It stores timed musical events such as note-on and note-off messages, pitch, velocity, channel, and tempo information. MIDI is useful for transferring pitch and rhythm, but it usually preserves less written notation and vocal-text detail than MusicXML.

These terms are optional to understand. You might find it familliar.

A cache is stored data kept so that an expensive computation can be reused without being performed again. A cache may contain intermediate data or a rendered result. It is reusable only when it corresponds to the current score, voice settings, and processing inputs; the mere presence of cached data does not make it the current result.

An artifact is a concrete result produced by a process. Examples include timing data, pitch data, phoneme labels, and rendered audio. Artifacts may be intermediate results used by later stages or final results intended for playback or export. A stem is an audio artifact associated with one voice.

To bake is to process the current score and voice settings into usable rendered results. A bake may produce intermediate artifacts as well as audio for playback and export. Playback and stem export use the latest completed bake, so changes made afterward require another bake before the result can be considered current.