Sound Testing

Sound Testing
Author

Benedict Thekkel

✅ Overview: What You’ll Do

  1. Record audio from the mic
  2. Visualize the waveform
  3. Play it back
  4. (Optional) Run basic analysis (volume, noise floor, frequency)

🧰 1. Install Required Python Libraries

pip install sounddevice numpy scipy matplotlib ipython

These tools use your system’s microphone input and work across platforms.

🎙️ 2. Record Audio from Microphone in Jupyter

import sounddevice as sd

print("Available audio input devices:")
devices = sd.query_devices()

for i, device in enumerate(devices):
    if device['max_input_channels'] > 0:
        print(f"[{i}] {device['name']} — {device['hostapi']}")
Available audio input devices:
[0] UM02: USB Audio (hw:1,0) — 0
[5] HD-Audio Generic: ALC897 Analog (hw:3,0) — 0
[7] HD-Audio Generic: ALC897 Alt Analog (hw:3,2) — 0
[8] UGREEN camera 2K: USB Audio (hw:4,0) — 0
import sounddevice as sd
import numpy as np
from scipy.io.wavfile import write

# Recording parameters
samplerate = 44100  # Hz
duration = 5        # seconds

print("Recording...")
recording = sd.rec(int(samplerate * duration), samplerate=samplerate, channels=1)
sd.wait()
print("Recording complete.")

# Save to WAV file
write("mic_test.wav", samplerate, recording)
Recording...
Recording complete.

✅ You now have a 5-second recording from your default microphone.

📊 3. Visualize the Audio Waveform

import matplotlib.pyplot as plt

plt.figure(figsize=(10, 3))
plt.plot(recording)
plt.title("Mic Input Waveform")
plt.xlabel("Sample")
plt.ylabel("Amplitude")
plt.grid(True)
plt.show()

Look for spikes (speech), flatlines (silence), and background noise.

🔊 4. Playback in Notebook (Optional)

from IPython.display import Audio

Audio("mic_test.wav")

🎧 This plays your mic test right in the notebook.

🧪 5. Basic Mic Quality Metrics

RMS Volume:

rms = np.sqrt(np.mean(recording**2))
print(f"RMS Volume: {rms:.4f}")
RMS Volume: 0.0079

Noise Floor Estimate (first second):

noise_sample = recording[:samplerate]  # First 1 second
noise_rms = np.sqrt(np.mean(noise_sample**2))
print(f"Estimated Noise Floor: {noise_rms:.4f}")
Estimated Noise Floor: 0.0082

🧠 Advanced Options (Optional)

Task Tool / Method
Plot frequency spectrum Use FFT via numpy.fft.fft
Record stereo or longer Change channels=2 or duration=10
Test multiple mics Use sd.query_devices() and device=...
Clean noise Use scipy.signal, librosa, or noisereduce
Transcribe speech Use speech_recognition or whisper

📋 Summary Workflow

1. 🎙️ Record with sounddevice
2. 📦 Save as WAV
3. 📊 Plot waveform
4. 🔊 Playback with IPython
5. 🧪 Measure volume / noise

Would you like: - A downloadable Jupyter notebook template for mic testing? - A version that tests and compares multiple microphones side-by-side?

Let me know how deep you want to go — I can add frequency analysis or even ML-based speech quality scores!

Back to top