完成语音识别功能（win测试，arm未知）

tools/amplify_wav.py

@@ -0,0 +1,45 @@
+import argparse
+import numpy as np
+import wave
+
+
+def read_wav_int16(path: str):
+    with wave.open(path, "rb") as wav:
+        sample_width = wav.getsampwidth()
+        if sample_width != 2:
+            raise ValueError(f"{path} is {sample_width * 8}-bit PCM; expected 16-bit")
+        sample_rate = wav.getframerate()
+        channels = wav.getnchannels()
+        frame_count = wav.getnframes()
+        raw = wav.readframes(frame_count)
+    samples = np.frombuffer(raw, dtype="<i2").astype(np.int16, copy=False)
+    return samples, sample_rate, channels
+
+
+def write_wav_int16(path: str, samples: np.ndarray, sample_rate: int, channels: int):
+    with wave.open(path, "wb") as wav:
+        wav.setnchannels(channels)
+        wav.setsampwidth(2)
+        wav.setframerate(sample_rate)
+        wav.writeframes(np.asarray(samples, dtype="<i2").tobytes())
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Amplify a WAV file.")
+    parser.add_argument("input", type=str, help="Input WAV path")
+    parser.add_argument("output", type=str, help="Output WAV path")
+    parser.add_argument("--gain", type=float, default=5.0, help="Gain factor")
+    args = parser.parse_args()
+
+    data, sr, channels = read_wav_int16(args.input)
+    print(f"Original max abs: {np.max(np.abs(data))}")
+
+    amplified = np.clip(data.astype(np.float32) * args.gain, -32768, 32767).astype(np.int16)
+    print(f"Amplified max abs: {np.max(np.abs(amplified))}")
+
+    write_wav_int16(args.output, amplified, sr, channels)
+    print(f"Saved: {args.output}")
+
+
+if __name__ == "__main__":
+    main()