Skip to contents

Introduction to ReVAMP

Source: vignettes/introduction.Rmd
introduction.Rmd

Overview

ReVAMP provides an R interface to the Vamp audio analysis plugin system developed by Queen Mary University of London’s Centre for Digital Music. Vamp plugins are widely used for Music Information Retrieval (MIR) tasks including:

  • Tempo and beat detection
  • Onset detection
  • Pitch tracking
  • Spectral analysis
  • Audio feature extraction
  • Chord detection
  • Key detection

Installation

Installing Vamp Plugins

ReVAMP requires Vamp plugins to be installed separately. Popular plugin collections include:

See the Vamp Paths vignette for details on plugin installation and search paths.

Basic Usage

Discovering Available Plugins

First, check what plugins are available on your system:

# List all available plugins
plugins <- vampPlugins()
head(plugins)

# View plugin information
str(plugins)

The dataframe contains:

  • library - The plugin library name
  • id - The plugin identifier
  • name - Human-readable plugin name
  • description - Plugin description
  • maker - Plugin author
  • category - Plugin category

Running a Plugin

The main function is runPlugin(), which executes a plugin on audio data and returns results as a named list of data frames.

# Load audio file
audio <- readWave("path/to/audio.wav")

# Run amplitude follower plugin
result <- runPlugin(
  key = "vamp-example-plugins:amplitudefollower",
  wave = audio,
  params = NULL,      # Use default parameters
  useFrames = FALSE   # Return timestamps in seconds
)

# Result is a named list of data frames (one per output)
names(result)
str(result[[1]])

Each output data frame contains:

  • timestamp - Time of the feature (in seconds or frames)
  • duration - Duration of the feature (if applicable)
  • value - Feature value(s) as a list column
  • label - Text label (if provided by plugin)

Plugin Keys

Plugins are identified by a key in the format library:plugin:

# Examples of plugin keys
"vamp-example-plugins:amplitudefollower"
"vamp-aubio:aubiotempo"
"qm-vamp-plugins:qm-tempotracker"
"nnls-chroma:chordino"

Working with Plugin Parameters

Many plugins accept parameters to customize their behavior.

Discovering Parameters 

# Get parameters for a plugin
params_df <- vampPluginParams("vamp-aubio:aubiopitch")
print(params_df)

The parameters dataframe shows:

  • identifier - Parameter name for use in code
  • name - Human-readable name
  • description - What the parameter controls
  • unit - Unit of measurement
  • minValue, maxValue - Valid range
  • defaultValue - Default setting
  • isQuantized - Whether it has discrete values
  • quantizeStep - Step size for quantized parameters
  • valueNames - Named values (for quantized parameters)

Setting Parameters

Pass parameters as a named list:

# Run pitch detection with custom parameters
result <- runPlugin(
  key = "vamp-aubio:aubiopitch",
  wave = audio,
  params = list(
    maxfreq = 800,      # Maximum frequency in Hz
    minfreq = 100,      # Minimum frequency in Hz
    silencethreshold = -70  # Silence threshold in dB
  ),
  useFrames = FALSE
)

Getting Help

References

  • Cannam, C., Landone, C., & Sandler, M. (2010). Sonic Visualiser: An open source application for viewing, analysing, and annotating music audio files. Proceedings of the 18th ACM International Conference on Multimedia, 1467-1468.
  • Vamp Plugins: https://www.vamp-plugins.org/