arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.02774cs.SDeess.AS
收藏

用于远场关键词识别的更真实的房间模拟研究

A study on more realistic room simulation for far-field keyword spotting

Sonos Inc(Sonos公司);Line Corporation(LINE公司);École Polytechnique Fédérale de Lausanne(洛桑联邦理工学院)

Eric Bezzam, Robin Scheibler, Cyril Cadoux, Thibault Gisselbrecht

摘要:

我们研究了更逼真的房间模拟训练远场关键字定位系统的影响,而无需对域内数据进行微调。为此,我们研究的影响,将以下因素在房间的脉冲响应(RIR)的生成:空气吸收,表面和频率依赖系数的真实材料,和随机光线追踪。通过消融研究,唤醒词任务用于测量这些因素的影响,与实测RIR的地面实况集进行比较。在一组保持干净和嘈杂的远场条件下的重新记录,我们证明了$35.8\%$相对于常用的(单吸收系数)图像源方法的改进。Pyroomacoustics软件包中提供了源代码,允许其他人将这些技术融入他们的工作中。

英文摘要:

We investigate the impact of more realistic room simulation for training far-field keyword spotting systems without fine-tuning on in-domain data. To this end, we study the impact of incorporating the following factors in the room impulse response (RIR) generation: air absorption, surface- and frequency-dependent coefficients of real materials, and stochastic ray tracing. Through an ablation study, a wake word task is used to measure the impact of these factors in comparison with a ground-truth set of measured RIRs. On a hold-out set of re-recordings under clean and noisy far-field conditions, we demonstrate up to $35.8\%$ relative improvement over the commonly-used (single absorption coefficient) image source method. Source code is made available in the Pyroomacoustics package, allowing others to incorporate these techniques in their work.

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。

↑