Speech recognition - Was: Desktop froze

[Ralf Mardorf]

On Wed, 2023-06-21 at 01:00 +0200, Ralf Mardorf wrote:
> A while back I wanted to take a screencapture with audio from a Linux
> session. Almost all programs require pulseaudio. I suspect a similar
> pitfall for speech recognition software, too.
> 
> Terrorist attack on critical infrastructure by red squirrel:
> https://www.nsnews.com/local-news/north-vancouver-power-outage-caused-by-squirrel-6379694

PS: In the end I used Kazam with pulseaudio on Xubuntu 20.04 to record a
session done with the appimage of eSPi https://low-hiss.com/ . The
recorded videos didn't start at the beginning, so I needed to play the
video for around 6 seconds, before I could start to play eSPi. The
sounds of the instrument are immediately played when clicking a pad with
the mouse. If I e.g. played a sound per beat, exactly in time, neither
the sounds in the video were played in time, nor was the recorded sound
in sync to that moment in the video, when I was clicking the pads.

The videos:
File Type                       : MP4
Major Brand                     : MP4 v2 [ISO 14496-14]
Video Frame Rate                : 24
Encoder                         : x264
Audio Format                    : mp4a
Audio Channels                  : 1
Audio Bits Per Sample           : 16
Audio Sample Rate               : 44100
Image Size                      : 1200x954
Megapixels                      : 1.1
Avg Bitrate                     : around 500 kbps

The computer:
HDMI audio was used
Memory Size: 32 GB
CPU with GPU: 6.191.5 "13th Gen Intel(R) Core(TM) i3-13100"