Meadowbrook
Alternative meyda cli
Allows extracting features from a segment of audio and store this information in a json document with large data blocks encoded as base64 strings
Example
meadowbrook horse.mp3 all --o=out.json
Usage
title: meadowbrookexample: meadowbrook [options] <INPUT_PATH> <...FEATURES> [options]type: objectproperties: INPUT_PATH: title: Audio file to read type: string FEATURES: title: Features to extract type: array description: > List of features to extract seperated by space, If 'all' is in the list extract all items: enum: - all - rms - energy - zcr - amplitudeSpectrum - powerSpectrum - spectralCentroid - spectralFlatness - spectralSlope - spectralRolloff - spectralSpread - spectralSkewness - spectralKurtosis - loudness - perceptualSpread - perceptualSharpness - mfcc "--o": title: Output path description: > File to write data too, if not specified prints to stdout type: string example: --o=<outputPath> "--bs": title: Buffer size example: --bs=<bufferSize> defaut: 512 type: number "--w": title: Windowing function example: --w=<windowingFunction> defaut: hanning enum: - blackman - sine - hanning - hamming "--f": title: Frames per second example: --f=<fps> type: number description: > How many frames of feature data should be captured per second of audio defaut: 60 "--format": title: Output format example: --format=<format> description: Structure of output json default: array enum: - array - object "-h": title: Print help type: boolean
Structure
Json schema for output json file
description: Extracted data outputtype: objectproperties: extracted: type: object description: Name:data pairs for extracted features patternProperties: "/^[a-zA-Z]+$/": oneOf: - type: array description: > Data separated by channel, starting with channel 0 in audio source items: type: array description: > frames of extracted audio feature data, representation varies by feature, check meyda docs - type: object description: > Data separated by channel, starting with channel 0 in audio source patternProperties: "/^[0-9]+$/": type: object description: > frames of extracted audio feature data, representation varies by feature, check meyda docs fps: type: number description: > Number of frames of data extracted for each second of audio default: 60 sampleRate: type: number description: Sample rate of audio source default: 44100 duration: type: number description: Length in seconds of audio source numberOfChannels: type: number description: Channels extracted from audio source default: 2 bufferSize: type: number description: Size of buffer used when extracting default: 512 windowingFunction: type: string description: Windowing function used when extracting default: hanning outputFormat: description: How collections are organized in this file default: array enum: - array - object