Skip to main content
Participating Frequently
Jul 29, 2020 5:13 AM UTC
Answered

Can I Simultaneously Delete All Plosives?

Hi,

 

I'm wondering if there is a way to simulteneously eliminate all instances of a given frequency or sound that recurrs throughout a recording. I have included an exerpt from a podcast I worked on the other day where the repetive sound was kind of like a plosive but generated by some electronic glitch. I would typically scrub out plosives as its more precise but I'm wondering if there is a way to create a noise print of a problematic sound & apply that process across a whole file without distorting everything else too much? It would take forever to scrub each instance of this sound out in this case as its every second or so but the sound is too short to eliminate using the Noise Reduction process. Is ther another way?

 

Thanks in advance for any insight!

This topic is closed to new replies. Start a new post to keep the conversation going.
SteveG_AudioMasters_
SteveG_AudioMasters_Correct answer
Community Expert
Jul 29, 2020 8:19 AM UTC

Yes there is, but how well it will work on this material is a little dubious... It's called the Sound Remover. In order to teach it the sound, you will have to reduce the FFT size in the module*, because you are only learning something very short. I picked the example after the word 'the' to use. You have to select it in the spectral view (you won't see it on the waveform) so you do this, and then click 'Learn Sound Model'. If you now select the whole file, you can test this - and you may well need to experiment with the settings quite a bit. I think it's something of an improvement, but you will have to EQ the sound quite a bit afterwards because doing this alters the tonal balance quite a lot.

 

*If you start by making a small selection and the app wants a longer one, you end up in a vicious circle where it won't open the app to let you change the FFT. So you have to make a larger selection to let the app open, and then go back and make your small one again. Ho Hum...

2 replies

SteveG_AudioMasters_
Community Expert
SteveG_AudioMasters_Community ExpertCorrect answer
Community Expert
Jul 29, 2020 8:19 AM UTC

Yes there is, but how well it will work on this material is a little dubious... It's called the Sound Remover. In order to teach it the sound, you will have to reduce the FFT size in the module*, because you are only learning something very short. I picked the example after the word 'the' to use. You have to select it in the spectral view (you won't see it on the waveform) so you do this, and then click 'Learn Sound Model'. If you now select the whole file, you can test this - and you may well need to experiment with the settings quite a bit. I think it's something of an improvement, but you will have to EQ the sound quite a bit afterwards because doing this alters the tonal balance quite a lot.

 

*If you start by making a small selection and the app wants a longer one, you end up in a vicious circle where it won't open the app to let you change the FFT. So you have to make a larger selection to let the app open, and then go back and make your small one again. Ho Hum...

Translate
soundmindAuthor
Participating Frequently
Jul 29, 2020 5:08 PM UTC

Thanks so much for the reply Steve. I shall now go and study the Sound Remover! Cheers

Translate