python/demo.py

#!/usr/bin/env python3
# coding: utf-8
"""GCRN AX650 语音增强示例：noisy.wav -> enhanced.wav"""

import argparse
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent))

from gcrn_sdk import GCRNDenoiser


def main() -> None:
    ap = argparse.ArgumentParser(description="GCRN speech enhancement on AX650")
    ap.add_argument("--model", default="models/model.axmodel", help="axmodel (或 onnx)")
    ap.add_argument("--input", default="test_audio/mix.wav", help="16kHz mono noisy wav")
    ap.add_argument("--output", default="enhanced.wav", help="输出 wav")
    ap.add_argument("--report", default="gcrn_report.json", help="性能报告输出")
    args = ap.parse_args()

    denoiser = GCRNDenoiser(args.model)
    report = denoiser.enhance_file(args.input, args.output)
    denoiser.save_report(report, args.report)
    print(f"input={args.input}")
    print(f"output={args.output}")
    print(
        f"elapsed={report['elapsed_seconds']:.3f}s "
        f"audio={report['audio_seconds']:.3f}s "
        f"rtf={report['real_time_factor']:.4f}"
    )


if __name__ == "__main__":
    main()