Score:1

software for detecting voices on audio recording

ph flag

I record audio for hours on my smartphone for security purposes, personal safety in shared accommodations, and need to prepare evidence but it takes forever to listen to these recordings. VLC really is terrible for this as even the play faster, slower are hidden in the menus. What I want is a player that skips the no-voice or "silent" ie. irrelevant parts of recordings and stops when it detects human voices or other anomaly sounds like loud crashes and bangs. Then I can take over from there, listen and make note of the times when voices and sounds to be used as evidence. Is there anything like this? I saw this Python: https://gist.github.com/analyticsindiamagazine/c1ebc56e2dece9629e7c67e5406281e8#file-comparingsongs-ipynb but for the record I'm on Ubuntu 20.04 LTS on Gnome desktop.

user.dz avatar
ng flag
Btw, you can enable shortcuts for VLC play speed
Score:2
ng flag

Audacity

Not really what you want but it has some features that may help. For such advanced feature, you may look for similar audio editing tools.

  1. Open file > Select track
  2. Analyze > Label sounds ...
  3. Set threshold measurement and level then press OK
  4. New track created with labels.

You can play one by one, delete trivial labels then File > Export rest as list if you want. Below an example with (-20dB Peak) and exported list.

screenshot of example

0.250567    4.830930    Sound 01
8.469161    12.778912   Sound 02
18.401633   25.688118   Sound 03
27.963265   33.826531   Sound 04
37.033787   42.425986   Sound 05
45.663311   54.793968   Sound 06
57.369796   64.856735   Sound 07
69.156463   89.201814   Sound 08
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.