Sound analysis

Signal Analysis Tool I ↗

The standard entry: about 50 audio features over time (meyda.js) – RMS, energy, spectral centroid (brightness), zero-crossing rate, spectral flatness, MFCC and more.

Input

wav/mp3, also many files

Output

features per time frame, absolute or z-standardised (CSV)

In Healthy Strings

Loudness and brightness per task; comparing players or recordings.

Watch out

Keep microphone and distance constant – otherwise you compare rooms and phones, not playing.

Signal Analysis Tool II ↗

A second, independent feature library (JS-Xtract) with partly different descriptors (tristimulus, irregularity, inharmonicity, spectral crest).

Input

wav/mp3

Output

features per time frame (CSV)

In Healthy Strings

Cross-checking results of Tool I.

Timbre analysis with many global values: attack time, loudness, dynamic complexity, dissonance, inharmonicity, pitch salience (essentia.js).

Input

wav/mp3, batch

Output

values per file and over time (CSV)

In Healthy Strings

Describing the tone of a task or excerpt.

Fundamental frequency over time with the CREPE neural network (ml5.js).

Input

wav/mp3, one voice

Output

F0 per time frame (CSV)

In Healthy Strings

Intonation curves of scales and the repertoire excerpts.

Watch out

Monophonic only; double stops confuse it.

Violin Intonation (VIP) ↗

Checks intonation while playing: tuning, temperament and timbre, live or from a file.

Input

microphone or audio file

Output

deviations in cents, displays

In Healthy Strings

Teaching and self-checks; intonation tasks of the protocol.

Analyses vibrato, tremolo, jitter and shimmer and saves them as numbers (and can transfer them to other sounds).

Input

wav/mp3

Output

modulation rate and depth (CSV)

In Healthy Strings

Vibrato rate and depth in the vibrato task (T13).

Pitch & Tempo Analyser ↗

Polyphonic pitch, chords, key and tempo (essentia.js), MIDI export.

Input

wav/mp3

Output

pitch, chords, BPM, MIDI

In Healthy Strings

Checking tempo and key of the excerpts.

Pitch, tempo and note onsets (aubio.js).

Input

wav/mp3

Output

onset times, F0, BPM (CSV)

In Healthy Strings

Finding note and bow-change times; checking the metronome grid of part 1.

Multi-band rhythm analysis that estimates the tempo as humans perceive it.

Input

wav/mp3

Output

BPM and rhythm curves

In Healthy Strings

Tempo of the free-timing excerpts (part 2).

AI sample slicer: cuts recordings into events or notes, with automatic pitch detection, and downloads them in a batch.

Input

wav/mp3

Output

one wav file per segment

In Healthy Strings

Cutting a protocol recording into tasks – check every cut.

Sound pressure and loudness from calibrated recordings: dB SPL from dBFS, dBA/dBB/dBC, LAeq, LUFS, sone, phon.

Input

wav from a calibrated microphone

Output

level values (Excel)

In Healthy Strings

Comparing dynamics only when the microphone was calibrated.

Watch out

Phone microphones are not calibrated – use relative values.

Loudness after ISO 532B (Zwicker).

Input

wav

Output

loudness in sone over time

In Healthy Strings

Perceived loudness of dynamics tasks.

Separates harmonic and noisy parts of a sound and computes the harmonic-to-noise ratio.

Input

wav/mp3

Output

HNR over time (CSV)

In Healthy Strings

Noise in the tone, e.g. bow pressure and contact point.

Aligns the perceived loudness of many files for listening experiments.

Input

wav/mp3 files

Output

normalised files

In Healthy Strings

Preparing clips for listener ratings.