Merge pull request #46 from benfmiller/working

Adding normalize field in BaseConfig

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Change Log
 
+## [1.2.2] 2023 - 02 -05
+
+### Added
+
+- normalize option in BaseConfig, Turns off all normalization if False
+
 ## [1.2.1] 2023 - 01 - 08
 
 ### Changed

audalign/__init__.py

@@ -24,8 +24,9 @@
 from audalign.config import BaseConfig
 from audalign.recognizers import BaseRecognizer
 from audalign.recognizers.correcognize import CorrelationRecognizer
-from audalign.recognizers.correcognizeSpectrogram import \
-    CorrelationSpectrogramRecognizer
+from audalign.recognizers.correcognizeSpectrogram import (
+    CorrelationSpectrogramRecognizer,
+)
 from audalign.recognizers.fingerprint import FingerprintRecognizer
 from audalign.recognizers.visrecognize import VisualRecognizer
 
@@ -224,11 +225,15 @@ def fine_align(
             strength_stat=recognizer.config.CONFIDENCE,
         )
         paths_audio = filehandler.shift_get_files(
-            recalc_shifts_results, sample_rate=recognizer.config.sample_rate
+            recalc_shifts_results,
+            sample_rate=recognizer.config.sample_rate,
+            normalize=recognizer.config.normalize,
         )
     else:
         paths_audio = filehandler.shift_get_files(
-            results, sample_rate=recognizer.config.sample_rate
+            results,
+            sample_rate=recognizer.config.sample_rate,
+            normalize=recognizer.config.normalize,
         )
 
     max_lags_not_set = False
@@ -265,6 +270,7 @@ def fine_align(
                 new_results["names_and_paths"],
                 write_extension,
                 write_multi_channel=write_multi_channel,
+                normalize=recognizer.config.normalize,
             )
         except PermissionError:
             print("Permission Denied for write fine_align")
@@ -280,6 +286,7 @@ def write_processed_file(
     destination_file: str,
     start_end: tuple = None,
     sample_rate: int = BaseConfig.sample_rate,
+    normalize: bool = BaseConfig.normalize,
 ) -> None:
     """
     writes given file to the destination file after processing for fingerprinting
@@ -290,13 +297,14 @@ def write_processed_file(
         destination_file (str): file path and name to write file to
         start_end (tuple(float, float), optional): Silences before and after start and end. (0, -1) Silences last second, (5.4, 0) silences first 5.4 seconds
         sample_rate (int): sample rate to write file to
-
+        normalize (bool): if true, normalizes file when read
     """
     filehandler.read(
         filename=file_path,
         wrdestination=destination_file,
         start_end=start_end,
         sample_rate=sample_rate,
+        normalize=normalize,
     )
 
 
@@ -549,6 +557,7 @@ def _write_shifted_files(
     write_extension: str,
     write_multi_channel: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     Writes files to destination_path with specified shift
@@ -560,6 +569,7 @@ def _write_shifted_files(
         names_and_paths (dict{str}): dict with name as key and path as value
         write_multi_channel (bool): If true, only write out combined file with each input audio file being one channel. If false, write out shifted files separately and total combined file
         unprocessed (bool): If true, writes files without processing. For total files, only doesn't normalize
+        normalize (bool): if true, normalizes file when read
     """
     filehandler.shift_write_files(
         files_shifts,
@@ -568,6 +578,7 @@ def _write_shifted_files(
         write_extension,
         write_multi_channel=write_multi_channel,
         unprocessed=unprocessed,
+        normalize=normalize,
     )
 
 
@@ -592,6 +603,7 @@ def write_shifted_file(
     destination_path: str,
     offset_seconds: float,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     Writes file to destination_path with specified shift in seconds
@@ -602,9 +614,14 @@ def write_shifted_file(
         destination_path (str): where to write file to and file name
         offset_seconds (float): how many seconds to shift, can't be negative
         unprocessed (bool): If true, writes files without processing.
+        normalize (bool): if true, normalizes file when read
     """
     filehandler.shift_write_file(
-        file_path, destination_path, offset_seconds, unprocessed=unprocessed
+        file_path,
+        destination_path,
+        offset_seconds,
+        unprocessed=unprocessed,
+        normalize=normalize,
     )
 
 
@@ -615,6 +632,7 @@ def write_shifts_from_results(
     write_extension: str = None,
     write_multi_channel: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     For writing the results of an alignment with alternate source files or unprocessed files
@@ -633,6 +651,7 @@ def write_shifts_from_results(
         write_extension (str, optional): if given, all files writen with given extension
         write_multi_channel (bool): If true, only write out combined file with each input audio file being one channel. If false, write out shifted files separately and total combined file
         unprocessed (bool): If true, writes files without processing. For total files, only doesn't normalize
+        normalize (bool): if true, normalizes file when read
     """
     if isinstance(read_from_dir, str):
         print("Finding audio files")
@@ -674,6 +693,7 @@ def write_shifts_from_results(
             write_extension,
             write_multi_channel=write_multi_channel,
             unprocessed=unprocessed,
+            normalize=normalize,
         )
     except PermissionError:
         print("Permission Denied for write fine_align")
@@ -684,6 +704,7 @@ def convert_audio_file(
     destination_path: str,
     start_end: tuple = None,
     sample_rate: int = None,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     Convert audio file to type specified in destination path
@@ -694,12 +715,14 @@ def convert_audio_file(
         destination_path (str): where to write file to and file name
         start_end (tuple(float, float), optional): Silences before and after start and end. (0, -1) Silences last second, (5.4, 0) silences first 5.4 seconds
         sample_rate (int): sample rate to write file to
+        normalize (bool): if true, normalizes file when read
     """
     filehandler.read(
         filename=file_path,
         wrdestination=destination_path,
         start_end=start_end,
         sample_rate=sample_rate,
+        normalize=normalize,
     )
 
 

audalign/config/__init__.py

@@ -34,6 +34,9 @@ class BaseConfig(ABC):
     # Decodes audio file to this sample rate
     sample_rate = 44100
 
+    # if true, normalizes all files when read
+    normalize = True
+
     # keys in results dictionaries
     CONFIDENCE = "confidence"
     MATCH_TIME = "match_time"
@@ -64,4 +67,4 @@ class BaseConfig(ABC):
 
     # used if rankings_get_top_num_match is not None. (used in visual)
     # subtracts second value from ranking if num matches is above first value
-    rankings_num_matches_tups: typing.Optional[tuple] = None
+    rankings_num_matches_tups: typing.Optional[tuple] = None

audalign/filehandler.py

@@ -57,6 +57,7 @@ def create_audiosegment(
     sample_rate=BaseConfig.sample_rate,
     length=None,
     unprocessed=False,
+    normalize=BaseConfig.normalize,
 ) -> AudioSegment:
     if sample_rate is None:
         sample_rate = BaseConfig.sample_rate
@@ -73,7 +74,8 @@ def create_audiosegment(
         audiofile = audiofile.set_frame_rate(sample_rate)
         audiofile = audiofile.set_sample_width(2)
         audiofile = audiofile.set_channels(1)
-        audiofile = effects.normalize(audiofile)
+        if normalize:
+            audiofile = effects.normalize(audiofile)
     else:
         sample_rate = audiofile.frame_rate
     if start_end is not None:
@@ -150,6 +152,7 @@ def read(
     wrdestination=None,
     start_end: tuple = None,
     sample_rate=BaseConfig.sample_rate,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     Reads any file supported by pydub (ffmpeg) and returns a numpy array and the bit depth
@@ -167,7 +170,7 @@ def read(
     if os.path.splitext(filename)[1] in cant_read_ext:
         raise CouldntDecodeError
     audiofile = create_audiosegment(
-        filename, start_end=start_end, sample_rate=sample_rate
+        filename, start_end=start_end, sample_rate=sample_rate, normalize=normalize
     )
     data = np.frombuffer(audiofile._data, np.int16)
     if wrdestination:
@@ -524,7 +527,9 @@ def level_by_ave(audiofile_data, index_list, overlap_array, width, exclude_min_d
     return new_audio_data
 
 
-def shift_get_files(results: dict, sample_rate: int = None):
+def shift_get_files(
+    results: dict, sample_rate: int = None, normalize: bool = BaseConfig.normalize
+):
     names_and_paths = results.pop("names_and_paths")
     temp_a = results.pop("match_info")
     temp_rankings = None
@@ -538,6 +543,7 @@ def shift_get_files(results: dict, sample_rate: int = None):
         None,
         sample_rate=sample_rate,
         return_files=True,
+        normalize=normalize,
     )
     results["names_and_paths"] = names_and_paths
     results["match_info"] = temp_a
@@ -553,6 +559,7 @@ def shift_write_files(
     write_extension: str,
     write_multi_channel: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     """
     Args
@@ -563,6 +570,7 @@ def shift_write_files(
         write_extension (str): if given, writes all alignments with given extension (ex. ".wav" or "wav")
         write_multi_channel (bool): If true, only write out combined file with each input audio file being one channel. If false, write out shifted files separately and total combined file
         unprocessed (bool): If true, writes files without processing
+        normalize (bool): if true, normalizes file when read
     """
     _shift_files(
         files_shifts,
@@ -572,6 +580,7 @@ def shift_write_files(
         write_multi_channel=write_multi_channel,
         return_files=False,
         unprocessed=unprocessed,
+        normalize=normalize,
     )
 
 
@@ -584,6 +593,7 @@ def _shift_files(
     sample_rate: int = None,
     return_files: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     if sample_rate is None:
         sample_rate = BaseConfig.sample_rate
@@ -601,6 +611,7 @@ def _shift_files(
             sample_rate=sample_rate,
             return_files=return_files,
             unprocessed=unprocessed,
+            normalize=normalize,
         )
     else:
         return _shift_write_multichannel(
@@ -611,6 +622,7 @@ def _shift_files(
             sample_rate=sample_rate,
             return_files=return_files,
             unprocessed=unprocessed,
+            normalize=normalize,
         )
 
 
@@ -622,13 +634,15 @@ def _shift_write_separate(
     sample_rate: int = None,
     return_files: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     audsegs = _shift_prepend_space_audsegs(
         files_shifts=files_shifts,
         names_and_paths=names_and_paths,
         sample_rate=sample_rate,
         return_files=return_files,
         unprocessed=unprocessed,
+        normalize=normalize,
     )
     if return_files:
         return audsegs
@@ -690,6 +704,7 @@ def _shift_prepend_space_audsegs(
     sample_rate: int,
     return_files: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     audsegs = {}
     for name in files_shifts.keys():
@@ -698,16 +713,16 @@ def _shift_prepend_space_audsegs(
             audsegs[file_path] = files_shifts[name]
         else:
             audiofile = create_audiosegment(
-                file_path, sample_rate=sample_rate, unprocessed=unprocessed
+                file_path,
+                sample_rate=sample_rate,
+                unprocessed=unprocessed,
+                normalize=normalize,
             )
             if unprocessed:
                 sample_rate = audiofile.frame_rate
             silence = AudioSegment.silent(
                 (files_shifts[name]) * 1000, frame_rate=sample_rate
             )
-            audiofile = create_audiosegment(
-                file_path, sample_rate=sample_rate, unprocessed=unprocessed
-            )
             audiofile: AudioSegment = silence + audiofile
             audsegs[file_path] = audiofile
     return audsegs
@@ -753,13 +768,15 @@ def _shift_write_multichannel(
     sample_rate: int,
     return_files: bool = False,
     unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
     audsegs = _shift_prepend_space_audsegs(
         files_shifts=files_shifts,
         names_and_paths=names_and_paths,
         sample_rate=sample_rate,
         return_files=return_files,
         unprocessed=unprocessed,
+        normalize=normalize,
     )
     if return_files:
         return audsegs
@@ -797,9 +814,15 @@ def _shift_write_multichannel(
 
 
 def shift_write_file(
-    file_path, destination_path, offset_seconds, unprocessed: bool = False
+    file_path,
+    destination_path,
+    offset_seconds,
+    unprocessed: bool = False,
+    normalize: bool = BaseConfig.normalize,
 ):
-    audiofile = create_audiosegment(file_path, unprocessed=unprocessed)
+    audiofile = create_audiosegment(
+        file_path, unprocessed=unprocessed, normalize=normalize
+    )
     sample_rate: int = audiofile.frame_rate
     silence = AudioSegment.silent(offset_seconds * 1000, frame_rate=sample_rate)
     audiofile = silence + audiofile
@@ -808,10 +831,15 @@ def shift_write_file(
 
 
 def get_shifted_file(
-    file_path, offset_seconds, sample_rate=BaseConfig.sample_rate
+    file_path,
+    offset_seconds,
+    sample_rate=BaseConfig.sample_rate,
+    normalize: bool = BaseConfig.normalize,
 ) -> np.array:
     silence = AudioSegment.silent(offset_seconds * 1000, frame_rate=sample_rate)
 
-    audiofile = create_audiosegment(file_path, sample_rate=sample_rate)
+    audiofile = create_audiosegment(
+        file_path, sample_rate=sample_rate, normalize=normalize
+    )
     audiofile = silence + audiofile
     return np.frombuffer(audiofile._data, np.int16)