feat(audio): 重构音频处理模块、音频流重采样测试成功

caption-engine/audioprcs/__init__.py

@@ -1 +1 @@
-from .streamchnl import mergeStreamChannels
+from .process import mergeChunkChannels, resampleRawChunk

caption-engine/audioprcs/process.py

@@ -0,0 +1,49 @@
+import samplerate
+import numpy as np
+
+def mergeChunkChannels(chunk, channels):
+    """
+    将当前多通道音频数据块转换为单通道音频数据块
+
+    Args:
+        chunk: (bytes)多通道音频数据块
+        channels: 通道数
+
+    Returns:
+        (bytes)单通道音频数据块
+    """
+    # (length * channels,)
+    chunk_np = np.frombuffer(chunk, dtype=np.int16)
+    # (length, channels)
+    chunk_np = chunk_np.reshape(-1, channels)
+    # (length,)
+    chunk_mono_f = np.mean(chunk_np.astype(np.float32), axis=1)
+    chunk_mono = np.round(chunk_mono_f).astype(np.int16)
+    return chunk_mono.tobytes()
+
+
+def resampleRawChunk(chunk, channels, orig_sr, target_sr, mode="sinc_best"):
+    """
+    将当前多通道音频数据块转换成单通道音频数据块，然后进行重采样
+
+    Args:
+        chunk: (bytes)多通道音频数据块
+        channels: 通道数
+        orig_sr: 原始采样率
+        target_sr: 目标采样率
+        mode: 重采样模式，可选：'sinc_best' | 'sinc_medium' | 'sinc_fastest' | 'zero_order_hold' | 'linear'
+
+    Return:
+        (bytes)单通道音频数据块
+    """
+    # (length * channels,)
+    chunk_np = np.frombuffer(chunk, dtype=np.int16)
+    # (length, channels)
+    chunk_np = chunk_np.reshape(-1, channels)
+    # (length,)
+    chunk_mono_f = np.mean(chunk_np.astype(np.float32), axis=1)
+    chunk_mono = chunk_mono_f.astype(np.int16)
+    ratio = target_sr / orig_sr
+    chunk_mono_r =  samplerate.resample(chunk_mono, ratio, converter_type=mode)
+    chunk_mono_r = np.round(chunk_mono_r).astype(np.int16)
+    return chunk_mono_r.tobytes()

caption-engine/audioprcs/streamchnl.py

@@ -1,22 +0,0 @@
-import numpy as np
-
-def mergeStreamChannels(data, channels):
-    """
-    将当前多通道流数据合并为单通道流数据
-
-    Args:
-        data: 多通道数据
-        channels: 通道数
-
-    Returns:
-        mono_data_bytes: 单通道数据
-    """
-    # (length * channels,)
-    data_np = np.frombuffer(data, dtype=np.int16)
-    # (length, channels)
-    data_np_r = data_np.reshape(-1, channels)
-    # (length,)
-    mono_data = np.mean(data_np_r.astype(np.float32), axis=1)
-    mono_data = mono_data.astype(np.int16)
-    mono_data_bytes = mono_data.tobytes()
-    return mono_data_bytes

caption-engine/main-gummy.py

@@ -8,7 +8,7 @@ elif sys.platform == 'linux':
 else:
     raise NotImplementedError(f"Unsupported platform: {sys.platform}")
 
-from audioprcs import mergeStreamChannels
+from audioprcs import mergeChunkChannels
 from audio2text import InvalidParameter, GummyTranslator
 
 
@@ -26,13 +26,13 @@ def convert_audio_to_text(s_lang, t_lang, audio_type):
 
     while True:
         try:
-            data = stream.read_chunk()
-            data = mergeStreamChannels(data, stream.CHANNELS)
+            chunk = stream.read_chunk()
+            chunk_mono = mergeChunkChannels(chunk, stream.CHANNELS)
             try:
-                gummy.send_audio_frame(data)
+                gummy.send_audio_frame(chunk_mono)
             except InvalidParameter:
                 gummy.start()
-                gummy.send_audio_frame(data)
+                gummy.send_audio_frame(chunk_mono)
         except KeyboardInterrupt:
             stream.closeStream()
             gummy.stop()

caption-engine/requirements.txt

@@ -1,5 +1,6 @@
 dashscope==1.23.5
 numpy==2.2.6
+samplerate==0.2.1
 PyAudio==0.2.14
 PyAudioWPatch==0.2.12.7 # Windows only
 pyinstaller==6.14.1

caption-engine/sysaudio/linux.py

@@ -34,7 +34,7 @@ class AudioStream:
 
         音频样本块大小：{self.CHUNK}
         样本位宽：{self.SAMP_WIDTH}
-        音频数据格式：{self.FORMAT}
+        采样格式：{self.FORMAT}
         音频通道数：{self.CHANNELS}
         音频采样率：{self.RATE}
         """

caption-engine/sysaudio/win.py

@@ -65,7 +65,7 @@ class AudioStream:
     def printInfo(self):
         dev_info = f"""
         采样设备：
-            - 设备类型：{ "音频输入" if self.audio_type == 0 else "音频输出" }
+            - 设备类型：{ "音频输出" if self.audio_type == 0 else "音频输入" }
             - 序号：{self.device['index']}
             - 名称：{self.device['name']}
             - 最大输入通道数：{self.device['maxInputChannels']}
@@ -76,7 +76,7 @@ class AudioStream:
 
         音频样本块大小：{self.CHUNK}
         样本位宽：{self.SAMP_WIDTH}
-        音频数据格式：{self.FORMAT}
+        采样格式：{self.FORMAT}
         音频通道数：{self.CHANNELS}
         音频采样率：{self.RATE}
         """