AudioLocalMeasurements
AudioChannelMix
AudioBlockMap

Audio Processing

RMS Amplitude of Audio Signals

Compute and visualize local measurements of audio signals

Open in Cloud
Download Notebook

Start by down-mixing a audio signal in to a single channel:

In[1]:=

a=AudioChannelMix[ExampleData[{"Audio","MaleVoice"},"Audio"],1];

Compute the

"Novelty"

"HighFrequencyContent"

and

"SpectralCentroid"

of the signal, computed on short-time Fourier data:

In[2]:=

properties=AudioLocalMeasurements[a,{"Novelty","HighFrequencyContent","SpectralCentroid"}]

Out[2]=

NoveltyTimeSeries

Time: 0. to 2.37
Data points: 52

,HighFrequencyContentTimeSeries

Time: 0. to 2.37
Data points: 52

,SpectralCentroidTimeSeries

Time: 0. to 2.37
Data points: 52



Visualize the properties together with the spectrogram of the audio:

In[3]:=

ShowSpectrogram[a],ListLinePlotRescale[#,MinMax[#],{0,10000}]&/@properties,

...

,

...



Out[3]=

	Novelty
	HighFrequencyContent
	SpectralCentroid

Next define a color function to visualize the dB RMS amplitude:

In[4]:=

dbLoudness[partition_]:=20Log10[Sqrt@Mean[partition^2]+10^-6];color[x_]:=Blend-40,

,-35,

,-5,

,x;

Use the function to compute the resulting colors:

In[5]:=

AudioBlockMap[Function[part,color[dbLoudness[part]]],#,{.1,.1,HannWindow}]&[a]["Values"]//ArrayPlot{#},

opts

&

Out[5]=

Next get a collection of audio signals to compute the same properties:

In[6]:=

audios=AssociationMap[AudioChannelMix[ExampleData[{"Audio",#}],1]&,ExampleData["Audio"]〚;;15,2〛];

Visualize the dB RMS amplitude of the collection of audio objects:

In[7]:=

ColumnKeyValueMapLabeledArrayPlot{#2["Values"]},

opts

,Text@#1,Left&,(AudioBlockMap[Function[part,color[dbLoudness[part]]],#,{.1,.1,HannWindow}]&/@audios),

opts

//

legend

Out[7]=

Related Symbols

AudioLocalMeasurements
AudioChannelMix
AudioBlockMap

Publisher Information

Contributed by: Wolfram Staff