Pull voice from noise
) Firstly a big thanks for letting me join.
I have loads of questions about this this and other topics, computer forensics has always fascinated me!
For now, I would like to know what is involved in extracting meaningful and understandable voice from a sound clip that is virtually all noise?
I do a lot of HF radio and sometimes I wished I could clean up a file to hear whats being said.
I use a programme called Goldwave that allows bandpass and stop and other filters, but what is out there that can do a far better job?
If you can get a small clip of it and email it to me i'll have a go at it.
Thnaks I'll do that, but I would like to know the methods of doing it and what software I could employ )
If you do a lot of HF radio (HAM-radio?) probably you can try some of the software DSP suites to filter the data. On freshmeat.net you can find some of these, and on the other hand using sox on linux can clean-up quite a lot. But the commandline isn't really fun.
I should first try narrowing the bandwidth to somesing similair as our natural voices, say 800 - 2400 Hertz and run a noice-killer together with it.
The software to do audio is just as expensive as the software for CF.
I use Sound Cleaner, DC Live Forensics, Audition, etc. The equipment is just as important, if you don't have a good SnR and or are importing straight into a pc, you will be reintroducing more noise into your original recording.
Can you say which might be the best software to try.
I am happy with the hardware here.
In my current prog, I can do noise filters, pass and stop filters etc.
But I am aware there are much more capable things out there.
OK I downloaded the trial version of DC- Millennium, seems very capable but without knowing how to proceed
and what filters to try I don't know how I will get on.
Once I get some time off work I will read all the manuals etc.
In the meantime I would be interested to see how others get on.
This link is for the sound clip I am trying to clean up.
So far all I can hear is the word 'everybody'
All I have done to this clip is to normalise as the original was very low volume.
If you do take up this challenge could you note down how you proceeded so i can learn from in -)