hookvision-rk/tools/generate_voices.py

"""使用本机 SAPI 语音合成生成 assets/voices/*.wav"""
import os
import sys

OUTPUT_DIR = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "assets", "voices")

VOICES = {
    "hooking.wav": "挂包中",
    "ready.wav": "挂包成功",
    "releasing.wav": "放包中",
    "arrive_1.wav": "到达1号工位",
    "arrive_2.wav": "到达2号工位",
    "arrive_3.wav": "到达3号工位",
    "arrive_4.wav": "到达4号工位",
    "arrive_5.wav": "到达5号工位",
    "offline_reader.wav": "读卡器离线",
    "offline_cam_left.wav": "左侧摄像头离线",
    "offline_cam_right.wav": "右侧摄像头离线",
}


def main():
    import struct
    os.makedirs(OUTPUT_DIR, exist_ok=True)

    import win32com.client

    speaker = win32com.client.Dispatch("SAPI.SpVoice")

    SSFMCreateForWrite = 3
    SAFT16kHz16BitMono = 4

    for filename, text in VOICES.items():
        path = os.path.join(OUTPUT_DIR, filename)
        stream = win32com.client.Dispatch("SAPI.SpFileStream")
        af = win32com.client.Dispatch("SAPI.SpAudioFormat")
        af.Type = SAFT16kHz16BitMono
        stream.Format = af
        stream.Open(path, SSFMCreateForWrite)
        speaker.AudioOutputStream = stream
        speaker.Speak(text)
        stream.Close()
        with open(path, 'rb+') as f:
            data = bytearray(f.read())
        fmt_chunk = b'fmt '
        fmt_pos = data.find(fmt_chunk)
        if fmt_pos == -1:
            print(f"  跳过音量放大（无 fmt chunk）")
        else:
            channels = struct.unpack_from('<H', data, fmt_pos + 10)[0]
            sample_rate = struct.unpack_from('<I', data, fmt_pos + 12)[0]
            bits_per_sample = struct.unpack_from('<H', data, fmt_pos + 22)[0]
            data_chunk = b'data'
            data_pos = data.find(data_chunk)
            if data_pos != -1:
                data_size = struct.unpack_from('<I', data, data_pos + 4)[0]
                sample_start = data_pos + 8
                sample_end = sample_start + data_size
                if bits_per_sample == 16:
                    fmt = '<' + 'h' * (data_size // 2)
                    samples = list(struct.unpack_from(fmt, bytes(data[sample_start:sample_end])))
                    gain = 3.0
                    samples = [max(-32768, min(32767, int(s * gain))) for s in samples]
                    data[sample_start:sample_end] = struct.pack(fmt, *samples)
                elif bits_per_sample == 8:
                    samples = list(data[sample_start:sample_end])
                    gain = 3.0
                    samples = [max(0, min(255, int((s - 128) * gain) + 128)) for s in samples]
                    data[sample_start:sample_end] = bytes(samples)
                with open(path, 'wb') as f:
                    f.write(data)
                print(f"生成: {path} (增益 {gain}x, {sample_rate}Hz, {bits_per_sample}bit, {channels}ch)")
            else:
                print(f"生成: {path} (无 data chunk)")


if __name__ == "__main__":
    main()