Skip to main content
Ezad
Inspiring
January 15, 2020
Answered

Vocal Enhancement Renderings of a Poem

  • January 15, 2020
  • 1 reply
  • 914 views

Hello

I plan to distribute this audio recording a rending of Francis Brabazon's poem via YouTube.
The audio recordings are over four hours.

This link is a sample of the audio.
Before I commit to creating this project using a photo slide show to complement the audio I was wondering if it is possible to improve the vocal.
The fellow's voice in the recording if fine. However I was thinking it may be possible to give his voice a modest change possibly to slightly enchance the sound of his reading.
Let me know if there is a useful method to improve the voice.

Thanks!
Using Audition CS6

https://youtu.be/kkiu5Gil8dQ

This topic has been closed for replies.
Correct answer SteveG_AudioMasters_

This is a classic case of needing to do the Noise Reduction really carefully, using different FFT sizes and much smaller amounts of reduction on each pass. You can lower the noise floor down to about -50dB without any noticeable pumping, with care.

 

At that point, you realise just how bad a recording this really is. The single biggest problem with it is that every 'sss' sound comes out as 'shhh' and there's very little you can do about this, except to reduce some of the energy in it by taking out a narrow band of the sound at 5,500Hz (5.5kHz), and you have to listen to this carefully before you do it. It's also possible at this stage to put a much more gentle noise gate on what's left - but don't remove all of the background noise, because that would sound really weird.

 

Just to give you some sort of idea, I treated this using what I described above. This isn't perfect at all, but it gives you some sort of idea of what you might do. As I said though, the original recording isn't good, and that simply cannot be fixed. This version needs a little more of the sibilance removed, but you get the idea, at least.

 

nothing and everything rework 

 

 

1 reply

SteveG_AudioMasters_
Community Expert
Community Expert
January 15, 2020

It would be a lot easier to suggest what to do with this if we had the original, rather than something that's been rather inappropriately treated with a gate...

Ezad
EzadAuthor
Inspiring
January 16, 2020

This is the audio material that was sent to me. It was from a tape transfer into digital Wav format.

 

https://youtu.be/gARZ7dhXibk

SteveG_AudioMasters_
Community Expert
SteveG_AudioMasters_Community ExpertCorrect answer
Community Expert
January 16, 2020

This is a classic case of needing to do the Noise Reduction really carefully, using different FFT sizes and much smaller amounts of reduction on each pass. You can lower the noise floor down to about -50dB without any noticeable pumping, with care.

 

At that point, you realise just how bad a recording this really is. The single biggest problem with it is that every 'sss' sound comes out as 'shhh' and there's very little you can do about this, except to reduce some of the energy in it by taking out a narrow band of the sound at 5,500Hz (5.5kHz), and you have to listen to this carefully before you do it. It's also possible at this stage to put a much more gentle noise gate on what's left - but don't remove all of the background noise, because that would sound really weird.

 

Just to give you some sort of idea, I treated this using what I described above. This isn't perfect at all, but it gives you some sort of idea of what you might do. As I said though, the original recording isn't good, and that simply cannot be fixed. This version needs a little more of the sibilance removed, but you get the idea, at least.

 

nothing and everything rework