公开(公告)号：US20150256141A1

公开(公告)日：2015-09-10

申请号：US14719296

申请日：2015-05-21

Applicant: Broadcom Corporation

Inventor： Nambirajan Seshadri ,  Jeyhan Karaoguz ,  Xuemin Chen ,  Chris Boross

IPC: H03G3/20 ,  H04R1/08

CPC classification number: H03G3/20 ,  H04N5/23229 ,  H04N5/357 ,  H04N13/122 ,  H04N13/25 ,  H04N13/261 ,  H04R1/08 ,  H04R2499/11

Abstract: A monoscopic camera comprising one or more image sensors and a depth sensor may generate video based on two-dimensional image data captured via the one or more image sensors and corresponding depth information captured via the depth sensor. The camera may process corresponding audio for the generated video based on the captured depth information. The audio processing may comprise mitigating noise in the corresponding audio, enhancing voice quality in the corresponding audio, and/or enhancing audio quality of the corresponding audio. The camera may be operable to determine, based on the captured depth information, one or more sound paths between a source of the corresponding audio and a microphone utilized to capture the corresponding audio emanating from the source. The processing of the audio may comprise removing portions of the captured audio arriving at the microphone via one or more reflection paths.

Abstract translation: 包括一个或多个图像传感器和深度传感器的单视镜相机可以基于经由一个或多个图像传感器捕获的二维图像数据和经由深度传感器捕获的相应深度信息来生成视频。 相机可以基于捕获的深度信息处理所生成的视频的相应音频。 音频处理可以包括减轻相应音频中的噪声，增强相应音频中的语音质量和/或增强对应音频的音频质量。 摄像机可操作以基于所捕获的深度信息确定相应音频的源和用于捕获从源发出的对应音频的麦克风之间的一个或多个声音路径。 音频的处理可以包括经由一个或多个反射路径去除到达麦克风的捕获音频的部分。