Matplotlib - Audio Visualization

---

---

Audio visualization is a way to see sound. When you listen to music or any other sound, there are patterns in the sound waves. Audio visualization takes these patterns and turns them into something you can see, like colorful shapes or moving lines.

It helps you understand the rhythm, pitch, and intensity of the sound by representing it visually. Imagine watching your favorite song as it plays, and seeing colorful patterns dancing along with the beat - that is audio visualization −

Audio Visualization in Matplotlib

Audio visualization using Matplotlib is a way to see and understand sound in a graphical format. Just like how we can draw pictures to represent data, we can use Matplotlib to create visual representations of audio, making it easier to analyze and interpret.

Imagine you are looking at a graph where one side represents time, and the other side represents how loud or quiet the sound is at each moment. When you play a song or speak into a microphone, the graph will show peaks and valleys that correspond to the changes in volume over time. This graph is called a waveform, and it gives us a visual understanding of the sound's intensity and duration.

Matplotlib does not have built-in functionalities for audio visualization. However, you can use it in conjunction with other libraries such as NumPy, librosa, and scipy to perform audio visualization tasks, allowing us to see the shape of the sound and how it changes over time. By examining these visual representations, we can identify patterns, compare different sounds, and even spot anomalies in the audio data.

Audio Visualization: Mel Spectrogram

In Matplotlib, an audio visualization Mel spectrogram is a graphical representation of the frequencies in an audio signal over time, using a Mel scale. The Mel scale is a perceptual scale of pitches that approximates the human ear's response to different frequencies. It divides the audible frequency range into a set of perceptually uniform intervals.

The Mel spectrogram displays time on the x-axis, frequency on the y-axis, and the intensity of each frequency at a specific moment as a color or grayscale shade. Brighter colors or lighter shades usually indicate higher intensity or amplitude of frequencies, while darker colors or shades represent lower intensity.

Mel spectrograms are commonly used in audio processing and analysis, particularly in speech recognition and music processing applications, as they provide a more accurate representation of how humans perceive sound.

Example

In the following example, we visualize the Mel spectrogram of an audio file using librosa and Matplotlib. We load the audio file and compute its Mel spectrogram using librosa's melspectrogram() function. We then use Matplotlib to display the spectrogram, with the color intensity representing the power in decibels −

import matplotlib.pyplot as plt
import librosa.display
import numpy as np

# Loading audio file
audio_file = "audio_file.wav"
y, sr = librosa.load(audio_file)

# Calculating the mel spectrogram of the audio
S = librosa.feature.melspectrogram(y=y, sr=sr)

# Plotting mel spectrogram
plt.figure(figsize=(10, 4))
# Displaying the mel spectrogram
librosa.display.specshow(librosa.power_to_db(S, ref=np.max), sr=sr, x_axis='time', y_axis='mel')
# Adding a colorbar to the plot with dB format
plt.colorbar(format='%+2.0f dB')
plt.title('Mel Spectrogram')
# Displaying the plot
plt.show()

Output

Following is the output of the above code −

Audio Visualization: Chromagram

In Matplotlib, an audio visualization chromagram is a graphical representation of the pitch content in an audio signal over time, focusing on musical notes or harmonics.

The chromagram displays time on the x-axis and musical pitch classes on the y-axis, usually represented as semitones or octaves. Each point in the chromagram indicates the presence or strength of a particular pitch class at a specific moment in time, typically shown as a color or grayscale shade.

Chromagrams are commonly used in music analysis and audio processing to identify musical patterns, chords, and key signatures within audio recordings.

Example

In this example, we are visualizing the chromagram of an audio file using librosa and Matplotlib. We load the audio file and compute its chromagram, representing pitch classes over time using librosa's chroma_stft() function. Then, we use Matplotlib to display the chromagram, with a color bar indicating intensity −

import matplotlib.pyplot as plt
import librosa.display

# Loading audio file
audio_file = "audio_file.wav"
y, sr = librosa.load(audio_file)

# Calculating the chromagram of the audio
chromagram = librosa.feature.chroma_stft(y=y, sr=sr)

# Plotting chromagram
plt.figure(figsize=(10, 4))
# Displaying the mel chromagram
librosa.display.specshow(chromagram, x_axis='time', y_axis='chroma')
# Adding a colorbar to the plot
plt.colorbar()
plt.title('Chromagram')
# Displaying the plot
plt.show()

Output

Following is the output of the above code −

Audio Processing: Beat Tracker

In Matplotlib, an audio processing beat tracker is a tool used to detect and visualize the rhythmic structure or tempo of an audio signal.

The beat tracker generally analyzes the audio signal to detect recurring patterns or accents in the sound, indicating the beat positions. It then generates a visualization, often as a plot, showing the detected beats or tempo changes over time.

Example

In here, we visualize the beat positions detected in an audio file using librosa and Matplotlib. We load the audio file, and compute its tempo and beat positions using librosa's beat_track() function. We then use Matplotlib to plot the audio waveform with beat positions marked as vertical dashed red lines −

import matplotlib.pyplot as plt
import librosa.display
import numpy as np

# Loading audio file
audio_file = "audio_file.wav"
y, sr = librosa.load(audio_file)

# Calculating tempo and beat positions
tempo, beat_frames = librosa.beat.beat_track(y=y, sr=sr)

# Plotting beat positions
plt.figure(figsize=(10, 4))
# Plotting the audio waveform with transparency
plt.plot(librosa.frames_to_time(np.arange(len(y)), sr=sr), y, alpha=0.5)
# Adding vertical lines at beat positions
plt.vlines(librosa.frames_to_time(beat_frames, sr=sr), -1, 1, color='r', linestyle='--', label='Beats')
plt.xlabel("Time (s)")
plt.ylabel("Amplitude")
plt.title("Beat Tracker")
plt.legend()
# Displaying the plot
plt.show()

Output

Following is the output of the above code −