Seich
/
mayavoz
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

rmv sr filtering

This commit is contained in:
shahules786 2022-09-29 22:38:26 +05:30
parent 192b8ffa7b
commit fccbd88ba2
2 changed files with 13 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
enhancer/data/dataset.py
View File

@ -66,15 +66,13 @@ class TaskDataset(pl.LightningDataModule):
train_clean = os.path.join(self.root_dir,self.files.train_clean) train_clean = os.path.join(self.root_dir,self.files.train_clean)
train_noisy = os.path.join(self.root_dir,self.files.train_noisy) train_noisy = os.path.join(self.root_dir,self.files.train_noisy)
fp = Fileprocessor.from_name(self.name,train_clean, fp = Fileprocessor.from_name(self.name,train_clean,
train_noisy,self.sampling_rate, train_noisy, self.matching_function)
self.matching_function)
self.train_data = fp.prepare_matching_dict() self.train_data = fp.prepare_matching_dict()
val_clean = os.path.join(self.root_dir,self.files.test_clean) val_clean = os.path.join(self.root_dir,self.files.test_clean)
val_noisy = os.path.join(self.root_dir,self.files.test_noisy) val_noisy = os.path.join(self.root_dir,self.files.test_noisy)
fp = Fileprocessor.from_name(self.name,val_clean, fp = Fileprocessor.from_name(self.name,val_clean,
val_noisy,self.sampling_rate, val_noisy, self.matching_function)
self.matching_function)
val_data = fp.prepare_matching_dict() val_data = fp.prepare_matching_dict()
for item in val_data: for item in val_data:

26
enhancer/data/fileprocessor.py
View File

@ -1,12 +1,13 @@
import glob import glob
import os import os
from re import S
import numpy as np import numpy as np
from scipy.io import wavfile from scipy.io import wavfile
class ProcessorFunctions: class ProcessorFunctions:
@staticmethod @staticmethod
def match_vtck(clean_path,noisy_path,sr): def match_vtck(clean_path,noisy_path):
matching_wavfiles = list() matching_wavfiles = list()
clean_filenames = [file.split('/')[-1] for file in glob.glob(os.path.join(clean_path,"*.wav"))] clean_filenames = [file.split('/')[-1] for file in glob.glob(os.path.join(clean_path,"*.wav"))]
@ -18,16 +19,15 @@ class ProcessorFunctions:
sr_clean, clean_file = wavfile.read(os.path.join(clean_path,file_name)) sr_clean, clean_file = wavfile.read(os.path.join(clean_path,file_name))
sr_noisy, noisy_file = wavfile.read(os.path.join(noisy_path,file_name)) sr_noisy, noisy_file = wavfile.read(os.path.join(noisy_path,file_name))
if ((clean_file.shape[-1]==noisy_file.shape[-1]) and if ((clean_file.shape[-1]==noisy_file.shape[-1]) and
(sr_clean==sr) and (sr_clean==sr_noisy)):
(sr_noisy==sr)):
matching_wavfiles.append( matching_wavfiles.append(
{"clean":os.path.join(clean_path,file_name),"noisy":os.path.join(noisy_path,file_name), {"clean":os.path.join(clean_path,file_name),"noisy":os.path.join(noisy_path,file_name),
"duration":clean_file.shape[-1]/sr} "duration":clean_file.shape[-1]/sr_clean}
) )
return matching_wavfiles return matching_wavfiles
@staticmethod @staticmethod
def match_dns2020(clean_path,noisy_path,sr): def match_dns2020(clean_path,noisy_path):
matching_wavfiles = dict() matching_wavfiles = dict()
clean_filenames = [file.split('/')[-1] for file in glob.glob(os.path.join(clean_path,"*.wav"))] clean_filenames = [file.split('/')[-1] for file in glob.glob(os.path.join(clean_path,"*.wav"))]
@ -38,11 +38,10 @@ class ProcessorFunctions:
sr_clean, clean_file = wavfile.read(os.path.join(clean_path,clean_file)) sr_clean, clean_file = wavfile.read(os.path.join(clean_path,clean_file))
sr_noisy, noisy_file = wavfile.read(noisy_file) sr_noisy, noisy_file = wavfile.read(noisy_file)
if ((clean_file.shape[-1]==noisy_file.shape[-1]) and if ((clean_file.shape[-1]==noisy_file.shape[-1]) and
(sr_clean==sr) and (sr_clean==sr_noisy)):
(sr_noisy==sr)):
matching_wavfiles.update( matching_wavfiles.update(
{"clean":os.path.join(clean_path,clean_file),"noisy":noisy_file, {"clean":os.path.join(clean_path,clean_file),"noisy":noisy_file,
"duration":clean_file.shape[-1]/sr} "duration":clean_file.shape[-1]/sr_clean}
) )
return matching_wavfiles return matching_wavfiles
@ -54,12 +53,10 @@ class Fileprocessor:
self, self,
clean_dir, clean_dir,
noisy_dir, noisy_dir,
sr = 16000,
matching_function = None matching_function = None
): ):
self.clean_dir = clean_dir self.clean_dir = clean_dir
self.noisy_dir = noisy_dir self.noisy_dir = noisy_dir
self.sr = sr
self.matching_function = matching_function self.matching_function = matching_function
@classmethod @classmethod
@ -67,23 +64,22 @@ class Fileprocessor:
name:str, name:str,
clean_dir, clean_dir,
noisy_dir, noisy_dir,
sr,
matching_function=None matching_function=None
): ):
if name.lower() == "vctk": if name.lower() == "vctk":
return cls(clean_dir,noisy_dir,sr, ProcessorFunctions.match_vtck) return cls(clean_dir,noisy_dir, ProcessorFunctions.match_vtck)
elif name.lower() == "dns-2020": elif name.lower() == "dns-2020":
return cls(clean_dir,noisy_dir,sr, ProcessorFunctions.match_dns2020) return cls(clean_dir,noisy_dir, ProcessorFunctions.match_dns2020)
else: else:
return cls(clean_dir,noisy_dir,sr, matching_function) return cls(clean_dir,noisy_dir, matching_function)
def prepare_matching_dict(self): def prepare_matching_dict(self):
if self.matching_function is None: if self.matching_function is None:
raise ValueError("Not a valid matching function") raise ValueError("Not a valid matching function")
return self.matching_function(self.clean_dir,self.noisy_dir,self.sr) return self.matching_function(self.clean_dir,self.noisy_dir)