Is there a way to throw out anything which is not speech? The track has only background noise, silence and speech.
We need to isolate parts of an audio track that have speech. Speech is clearly visible on the spectrogram, but not on the Decibel scale. That's why "Truncate Silence" fails every time with filtering spoken parts that aren't all that loud.
The problem is clearly defined in this sonogram: Kill the left hand box! Remove that which has no spikes! Keep the right one.
I'd cut it out with scissors but the tape is just too long to do it manually.
We need to send hours of audio to a transcriptionist but it has long silences inbetween. We don't want to pay for those.
I have access to a DAW with "Audition", "Sox" and other weird alien software installed. I know basic multimedia editing but Audio is not my field. Been fiddling around with this for over an hour but can't do it.