Members

GROUP 9

• BI12-076 Mai Hải Đăng (Leader)
• BI12-074 Đoàn Đình Đăng
• BI12-074 Trần Hải Đăng
• BI12-040 Trần Ngọc Ánh
• BI12-099 Nguyễn Thanh Đức

Denoising Techniques

• All techniques are stored in the folder "techniques"
• All original audio files are stored in "audio_files"
• All clean audio files are stored in "audio_files/audio"
• All noise files are stored in "audio_files/noise"
• All noisy files are stored in "audio_files/noisy_audio"
• All denoised files are stored in "filtered_audio_files"
• All plotting images are stored in "plotting_image"

Spectral Subtraction

BI12-074 Đoàn Đình Đăng TO DO:

• Add noise to piano song
• Use Spectral Subtraction for denoising
• Return new filtered audio files after denoising
• Plotting spectrum for comparison

Denoising

Run "spectral_subtraction_method.py" will do these task:

1. Get audio file from "audio_files/audio/" and add 3 noises form 3 files in "audio_files/noise" then return the noisy files in folder "audio_files/noisy_audio".

2. Read the noisy files in "audio_files/noisy_audio/" and start denoising using "Spectral Subtraction" techniques -> Return filtered signal_array.

3. Write the filtered signal_array to new files and store in folder "filtered_audio_files" in this form "ssm_file_name.wav".

Plotting

Run "spectral_subtraction_plot.py" will do these task:

1. Read the noisy audio in "audio_files/noisy_audio/" and filtered audio

2. Plot noisy audio with corresponding filtered audio in 3 forms:

   • Amplitude - Time
   • Frequency - Time
   • Spectrum - Time