STFTCodec：基于时频域表示的高保真音频压缩
STFTCodec: High-Fidelity Audio Compression through Time-Frequency Domain Representation
AI总结：
本研究提出新型神经音频编解码器STFTCodec，利用STFT构建紧凑频谱表示并引入解包裹相位导数作为辅助特征，采用并行幅相处理分支，在放宽相位重建约束的同时保持相位感知，实现了多比特率下优于现有方法的感知质量，且压缩比调节灵活。
摘要：
我们提出了STFTCodec，这是一种新型的基于频谱的神经音频编解码器，它利用短时傅里叶变换（Short-Time Fourier Transform, STFT）实现高效的音频压缩。与需要大模型容量和大量内存消耗的基于波形的方法不同，该方法利用STFT获得紧凑的频谱表示，并引入解包裹相位导数作为辅助特征。我们的架构采用并行的幅度和相位处理分支，并通过先进的特征提取机制进行增强。通过放宽严格的相位重建约束，同时保持相位感知处理，我们实现了卓越的感知质量。实验结果表明，STFTCodec在多种比特率下均优于基于波形和基于频谱的方法，同时还具备独特的灵活性，无需修改架构即可通过调整STFT参数来调节压缩比。
英文摘要：
We present STFTCodec, a novel spectral-based neural audio codec that efficiently compresses audio using Short-Time Fourier Transform (STFT). Unlike waveform-based approaches that require large model capacity and substantial memory consumption, this method leverages STFT for compact spectral representation and introduces unwrapped phase derivatives as auxiliary features. Our architecture employs parallel magnitude and phase processing branches enhanced by advanced feature extraction mechanisms. By relaxing strict phase reconstruction constraints while maintaining phase-aware processing, we achieve superior perceptual quality. Experimental results demonstrate that STFTCodec outperforms both waveform-based and spectral-based approaches across multiple bitrates, while offering unique flexibility in compression ratio adjustment through STFT parameter modification without architectural changes.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
