Introduction to ReVAMP

Installing Vamp Plugins

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

• Vamp Example Plugins - Basic examples
• QM Vamp Plugins - Queen Mary plugins
• Vamp Aubio Plugins - Aubio feature extractors

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

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"

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
)