there is an audiobook which CIA censored many names from the memoire, and the publisher pasted an exact audio phrase on the whole file repeatedly many times and listening to it is frustrating experience.
So I need to cut this 30seconds long warning phrase with a simple beep (or simply cut it out from the whole audiobook altogether ).
Can anyone guide me to how to do such task?
Thank you in advance.
cutout an specific spoken phrase repeatition from WAV a file
Re: cutout an specific spoken phrase repeatition from WAV a
For what it is worth, I had to do something similar some time back. I had several audio files of WW2 airmen's stories. These files contained dozens "ah-em" , "you know" coughs and other annoying sounds. I ended up by copying each file and going through each copy carefully and deleting the unwanted events one by one. Highlight each unwanted section and delete it. You can check how it sounds after each deleted clip by using the "Play the unselected section" option (Hit the F11key and choose the"Playback" drop down item). It will take forever but worth doing. By working with a copy means that if you mess up big time, you can always start again! Now save the copy with another name.
Re: cutout an specific spoken phrase repeatition from WAV a
very strange, copying a 5second exact phrase into clipboard in a sound app and searching/highlighting of the existing instance of it within the wave file automatically/sequentially, seems not an impossible task to me.
I'm wonder why no one take care of it!
I'm wonder why no one take care of it!
-
- Posts: 1171
- Joined: Fri Mar 11, 2005 11:15 pm
- Location: Washington DC Metro Area
- Contact:
Re: cutout an specific spoken phrase repeatition from WAV a
Because it sounds simple...but it's not.
There have been things attempted in the past...the ability to do such a task is still not easy. Converting speech to text isn't 100% accurate.
The other problem is the fact once it goes through compression...things change. You can't say "find me every match for these 1,000,000 samples"...for starters you'd need a LOT of resources to compare every million samples...not to mention to get accurate you'd have to generate your scan window for every sample. So scan a million samples, move one, scan another million, move another.
There have been things attempted in the past...the ability to do such a task is still not easy. Converting speech to text isn't 100% accurate.
The other problem is the fact once it goes through compression...things change. You can't say "find me every match for these 1,000,000 samples"...for starters you'd need a LOT of resources to compare every million samples...not to mention to get accurate you'd have to generate your scan window for every sample. So scan a million samples, move one, scan another million, move another.