Evaluate speech enhancemnt model performance
Project description
Speech Enhancement Model Evaluation
This is a python package to evaluate your speech enhancement model performance.
Installation
pip install model-evaluation-777 -i https://pypi.org/simple
PESQ and STOI Evaluation
def evaluate(ref_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/clean/", deg_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/noisy/", fs=16000):
'''
Args:
ref_dir: reference signal directory
deg_dir: degration signal directory
fs: sample rate
Returns:
pesq_score_average, stoi_score_average
'''
from model_evaluation import stoi_pesq
pesq_score, stoi_score = stoi_pesq.evaluate(ref_dir=ref_dir, deg_dir=deg_dir, fs=16000)
Plot SNR vs PESQ
def plot_snr_pesq(out_dir, enh_dir, ref_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/clean/", deg_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/noisy/", fs=16000):
'''
Args:
out_dir: directory to save figure
enh_dir: enhanced signal directory
ref_dir: reference signal directory
deg_dir: degration signal directory
fs: sample rate
Returns:
snr vs pesq score data
'''
The SNR vs PESQ figure will save to out_dir automatically.
from model_evaluation import stoi_pesq
data = stoi_pesq.plot_snr_pesq(out_dir=out_dir, enh_dir=enh_dir)
Plot SNR vs STOI
def plot_snr_stoi(out_dir, enh_dir, ref_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/clean/", deg_dir="/media/youwei/Chauncey's/VERSO_Dataset/test/noisy/", fs=16000):
'''
Args:
out_dir: directory to save figure
enh_dir: enhanced signal directory
ref_dir: reference signal directory
deg_dir: degration signal directory
fs: sample rate
Returns:
snr vs stoi score data
'''
The SNR vs STOI figure will save to out_dir automatically.
from model_evaluation import stoi_pesq
data = stoi_pesq.plot_snr_stoi(out_dir=out_dir, enh_dir=enh_dir)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Close
Hashes for model-evaluation-777-1.0.2.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 39900dee529e829b69e9ce70b1d34f006c7496234cf4c33953e02256e7223895 |
|
MD5 | 3e8fd2eada762272f4f6246ea8fb42bb |
|
BLAKE2b-256 | fba083f677b277bb8e2e6eb4f960877a7e8e098edd10f18250be52682084468b |
Close
Hashes for model_evaluation_777-1.0.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | beb56e9f9f3b05d7a5c9b1340b4df700f000603d10285e366661936cbd30339e |
|
MD5 | 5f9cffc0098d4ac11fdbc5413c45c23a |
|
BLAKE2b-256 | e90bc51d80acc936e8e53b43bcfbf5b5a9e5b6374355e1a0a81b3e76ad3ea077 |