From e636166b8523f1bfb0ea4688721f4337071b712e Mon Sep 17 00:00:00 2001
From: Chenghao Mou <mouchenghao@gmail.com>
Date: Fri, 4 Apr 2025 06:04:56 +0100
Subject: [PATCH] fix: floor (#293)

---
 musetalk/utils/audio_processor.py | 44 +++++++++++++++----------------
 1 file changed, 22 insertions(+), 22 deletions(-)

diff --git a/musetalk/utils/audio_processor.py b/musetalk/utils/audio_processor.py
index 1c41ceb..7601c1b 100755
--- a/musetalk/utils/audio_processor.py
+++ b/musetalk/utils/audio_processor.py
@@ -1,16 +1,17 @@
-import os
 import math
+import os
+
 import librosa
 import numpy as np
 import torch
-
 from einops import rearrange
 from transformers import AutoFeatureExtractor
 
+
 class AudioProcessor:
     def __init__(self, feature_extractor_path="openai/whisper-tiny/"):
         self.feature_extractor = AutoFeatureExtractor.from_pretrained(feature_extractor_path)
-        
+
     def get_audio_feature(self, wav_path, start_index=0, weight_dtype=None):
         if not os.path.exists(wav_path):
             return None
@@ -19,11 +20,11 @@ class AudioProcessor:
         # Split audio into 30s segments
         segment_length = 30 * sampling_rate
         segments = [librosa_output[i:i + segment_length] for i in range(0, len(librosa_output), segment_length)]
-        
+
         features = []
         for segment in segments:
             audio_feature = self.feature_extractor(
-                segment, 
+                segment,
                 return_tensors="pt",
                 sampling_rate=sampling_rate
             ).input_features
@@ -32,13 +33,13 @@ class AudioProcessor:
             features.append(audio_feature)
 
         return features, len(librosa_output)
-    
+
     def get_whisper_chunk(
-        self, 
-        whisper_input_features, 
-        device, 
-        weight_dtype, 
-        whisper, 
+        self,
+        whisper_input_features,
+        device,
+        weight_dtype,
+        whisper,
         librosa_length,
         fps=25,
         audio_padding_length_left=2,
@@ -48,30 +49,30 @@ class AudioProcessor:
         whisper_feature = []
         # Process multiple 30s mel input features
         for input_feature in whisper_input_features:
-            audio_feats = whisper.encoder(input_feature.to(device), output_hidden_states=True).hidden_states                
+            audio_feats = whisper.encoder(input_feature.to(device), output_hidden_states=True).hidden_states
             audio_feats = torch.stack(audio_feats, dim=2).to(weight_dtype)
             whisper_feature.append(audio_feats)
-            
+
         whisper_feature = torch.cat(whisper_feature, dim=1)
         # Trim the last segment to remove padding
         sr = 16000
         audio_fps = 50
         fps = int(fps)
         whisper_idx_multiplier = audio_fps / fps
-        num_frames = math.floor((librosa_length / sr)) * fps
-        actual_length = math.floor((librosa_length / sr)) * audio_fps
+        num_frames = math.floor((librosa_length / sr) * fps)
+        actual_length = math.floor((librosa_length / sr) * audio_fps)
         whisper_feature = whisper_feature[:,:actual_length,...]
-        
+
         # Calculate padding amount
         padding_nums = math.floor(whisper_idx_multiplier)
         # Add padding at start and end
         whisper_feature = torch.cat([
-            torch.zeros_like(whisper_feature[:, :padding_nums * audio_padding_length_left]), 
-            whisper_feature, 
+            torch.zeros_like(whisper_feature[:, :padding_nums * audio_padding_length_left]),
+            whisper_feature,
             # Add extra padding to prevent out of bounds
             torch.zeros_like(whisper_feature[:, :padding_nums * 3 * audio_padding_length_right])
         ], 1)
-        
+
         audio_prompts = []
         for frame_index in range(num_frames):
             try:
@@ -86,7 +87,7 @@ class AudioProcessor:
                 print(f"num frames: {num_frames}, fps: {fps}, whisper_idx_multiplier: {whisper_idx_multiplier}")
                 print(f"frame_index: {frame_index}, audio_index: {audio_index}-{audio_index + audio_feature_length_per_frame}")
                 exit()
-                
+
         audio_prompts = torch.cat(audio_prompts, dim=0)  # T, 10, 5, 384
         audio_prompts = rearrange(audio_prompts, 'b c h w -> b (c h) w')
         return audio_prompts
@@ -97,5 +98,4 @@ if __name__ == "__main__":
     audio_feature, librosa_feature_length = audio_processor.get_audio_feature(wav_path)
     print("Audio Feature shape:", audio_feature.shape)
     print("librosa_feature_length:", librosa_feature_length)
-    
-    
\ No newline at end of file
+