Recordings

The trouble with spectrograms is that some qualities of speech we perceive are extremely subtle. For example we probably have evolved to read peoples emotional state from their voice. However, one could not read this from a spectrogram. Partly this is because the spectrogram loses information, and partly because our perception of sound is far more detailed in these specialized domains than our visual perception.

We will need to listen back to recordings of ourselves, because we do not hear ourselves properly when speaking. If you’re here then you probably find listening to a recording of yourself very unpleasant. If you could hear yourself this way when speaking you probably never would have started speaking badly in the first place.

By watching or listening back to recordings, we will create a memory of what we sound like which we can then use to develope an accurate perception of what we sound like in real time.

To make watching and listening back more pleasant you should mix audio of the instrumental accompaniment in with your voice. You can make your own with various software from the original audio. I’ve tried spleeter but it doesn’t work that well. Adobe Audition and iZotope RX work well but they are not free.

If you have a subfolder in the folder you picked your audio files from called Accompaniments with files with the same names and file extensions as the files you picked, then they will be mixed into the output audio in the Music/Speech School Recordings folder.