arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2212.04532eess.AScs.LGcs.SDeess.SP
收藏

逐帧WaveGAN:具有极低计算复杂度的时域高速对抗声码器

Framewise WaveGAN: High Speed Adversarial Vocoder in Time Domain with Very Low Computational Complexity

Amazon Web Services(亚马逊云科技);University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)

Ahmed Mustafa, Jean-Marc Valin, Jan Büthe, Paris Smaragdis, Mike Goodwin

摘要:

GAN声码器是目前用于构建高质量神经波形生成模型的最先进方法之一。然而,他们的架构大多需要数十亿次浮点运算每秒(GFLOPS),以产生语音波形的采样方式。这使得GAN声码器在没有加速器或并行计算机的普通CPU上运行仍然具有挑战性。在这项工作中,我们提出了一种新的架构,GAN声码器,主要依赖于经常性和全连接的网络直接产生的时域信号在帧的方式。这导致计算成本的显著降低,并且能够在GPU和低复杂度CPU上非常快速地生成。实验结果表明,我们的Framewise WaveGAN声码器实现了显着更高的质量比自回归最大似然声码器,如LPCNet在一个非常低的复杂度为1.2 GFLOPS。这使得GAN声码器在边缘和低功耗设备上更加实用。

英文摘要:

GAN vocoders are currently one of the state-of-the-art methods for building high-quality neural waveform generative models. However, most of their architectures require dozens of billion floating-point operations per second (GFLOPS) to generate speech waveforms in samplewise manner. This makes GAN vocoders still challenging to run on normal CPUs without accelerators or parallel computers. In this work, we propose a new architecture for GAN vocoders that mainly depends on recurrent and fully-connected networks to directly generate the time domain signal in framewise manner. This results in considerable reduction of the computational cost and enables very fast generation on both GPUs and low-complexity CPUs. Experimental results show that our Framewise WaveGAN vocoder achieves significantly higher quality than auto-regressive maximum-likelihood vocoders such as LPCNet at a very low complexity of 1.2 GFLOPS. This makes GAN vocoders more practical on edge and low-power devices.

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。

↑