Update sonogram_utility.py

sonogram_utility.py

@@ -3,6 +3,7 @@ import random
 import copy
 from pyannote.core import Annotation, Segment
 import numpy as np
+import soundfield as sf
 import torch
 import torchaudio
 import pandas as pd
@@ -158,8 +159,8 @@ def loadAudioCSV(sampleCSV):
     return [], prediction
     
 def splitIntoTimeSegments(testFile,maxDurationInSeconds=60):
-
-    waveform, sample_rate = torchaudio.load(testFile, backend="soundfile")
+    data, sample_rate = sf.read(testFile, dtype="float32", always_2d=True)
+    waveform = torch.from_numpy(data.T)  # shape: [channels, samples]
     audioSegments = []
     
     outOfBoundsIndex = waveform.shape[-1]