长时间静默录制流音频 [英] recording streaming audio with huge periods of silence

问题描述

我在Linux盒子上使用darkice( http://darkice.org )来捕获音频提要从消防部门的无线电系统.它运行良好，我可以将流转发到Icecast2( https://icecast.org )服务器，以便消防员可以听现场广播.

I'm using darkice (http://darkice.org) on a Linux box to capture an audio feed from a fire department radio system. It works great and I can forward the stream to an Icecast2 (https://icecast.org) server so the firefighters can listen to live radio transmissions.

我的下一个目标是实际将无线电传输记录到文件中.

My next goal is to actually record radio transmissions to file.

消防部门并不总是那么忙，所以我正在捕捉的溪流有大量的寂静(嘶嘶声).我的目标是以某种方式捕获和记录仅存在真实的人类语音传输的时间段，而不会浪费大量的硬盘空间来记录嘶嘶声.

The fire department isn't always that busy, so the stream I'm capturing has huge periods of silence (hiss). My goal is to somehow capture and record to file only the periods where there are real, human voice transmissions and not waste huge amounts of hard disk space recording hiss.

对可能可以解决这个问题的工具有何想法?

Any thoughts on the tools that might be able to conquer this?

谢谢！

推荐答案

您应该可以为此使用FFmpeg及其 silenceremove 过滤器.

You should be able to use FFmpeg for this, and its silenceremove filter.

未经测试，但请尝试以下操作:

Untested, but try something like this:

ffmpeg -i <Icecast URL> -af silenceremove=stop_periods=-1:stop_duration=1:stop_threshold=-30dB output.webm

这篇关于长时间静默录制流音频的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！