[CM] noise reduction with snd

kelly george hirai khirai@unm.edu
Fri, 23 Aug 2002 10:00:48 -0600 (MDT)


dnoise would be good for removing any fairly constant background noise ie>
hum, hiss and rumble. it compares a sample of your noise to your source
file breaks it down and resynthisizes it using fft. it effectively gates
off very small frequency bands setting thresholds based on the noise
sample file. here is the quick help file:

usage: dnoise [flags] noise_soundfile input_file output_file

flags:
N = # of bandpass filters (1024)
w = filter overlap factor: {0,1,(2),3} DON'T USE -w AND -M
M = analysis window length (N-1 unless -w is specified)
L = synthesis window length (M)
D = decimation factor (M/8)
b = begin time in noise reference soundfile (0)
B = starting sample in noise reference soundfile (0)
e = end time in noise reference soundfile (end)
E = final sample in noise reference soundfile (end)
t = threshold above noise reference in dB (30)
S = sharpness of noise-gate turnoff (1) (1 to 5)
n = number of FFT frames to average over (5)
m = minimum gain of noise-gate when off in dB (-40)
V: verbose - print status info
A : AIFF
W : WAV
J : IRCAM
noise filename must follow all flags

it is a part of the csound package and is free for accidemic use. i hate
to be off  topic but i really feel its the tool for the job. now if your
problem has to do with too much reverberation, you will need some sort of
impulse response filtering. anyone have any suggestions for that?

kelly hirai

On Fri, 23 Aug 2002 rm@fabula.de wrote:

> Let me jump out  of my lurking state here ...
> If i understand the original poster's question right he i
> asking about background noise removal from speech recordings. The tools
> mentioned so far are probably nice to remove scratches and dust noise
> from old vinyl recordings but probably not so good for the task under
> discussion.
> Moritz, you forgot to mention where the noise on the tapes(?) comes
> from. Is it environmental noise or is it noise introduced by the
> probably bad recording equipment. If you just need to isolate the
> speech part to get a clearer understanding of the text you might be
> able to use the ol' telephone trick: human speech stays comprehensible
> even if one removes major parts of the spectrum. You could apply both
> a lowpath and a highpath to the original signal, and, unless your noise
> happens to be in the same area of the frequency spectrum as the speech
> component, get a much more understandable speech signal. It's been quite
> a while since a last worked with/looked at human speech, if you need more
> help send my an email and i'll try to dig up some notes/literature.
>
>  Ralf Mattes
>
> On Thu, Aug 22, 2002 at 07:02:35PM -0600, kelly george hirai wrote:
> > csound comes with an outboard de-noiser called denoise. i have used it
> > with much success to get rid of hiss. check the comand line parms for best
> > results, the default settings are not very good. kelly
> >
> > On Fri, 16 Aug 2002, Michael Edwards wrote:
> >
> > > Hi Moritz,
> > >
> > > Noise removal is a very tricky business, way over my head to be honest,
> > > but I found a nice little app that's idiot-proof (there's only one
> > > slider to fiddle with) and costs only $20.  I'm sure a lot people will
> > > find it too limiting but I've done many a clean-up job with it and swear
> > > by it.  You'll find it below:
> > >
> > > http://www.excla.com/WAVclean/English/
> > >
> > > Good luck,
> > >
> > > 	Michael
> > >
> > >
> > > mlennert@club.worldonline.be schrieb:
> > > >
> > > > Hello,
> > > >
> > > > I am a complete newbie in the field of sound editing, so please tell me if I'm completely out of my league and if this isn't the right place to ask these questions.
> > > >
> > > > I have been conducting interviews for my MSc dissertation and am now in the process of transcribing them. However, my recordings are not always of very good quality and I often have heavy background noise. Since this noise is quite constant and homogenous, I was wondering whether it would be possible to filter it out in order to hear the voices more clearly. I've tried some noise reduction with 'audacity', but the result was not extraordinary. Is it possible to reduce noise in snd ? And if yes, could someone just briefly list the steps I have to follow, or a source of information about this ? I have read through the manual, but since I'm not familiar with the vocabulary of sound engineering this is not always easy.
> > > >
> > > > So any hint would be welcome.
> > > >
> > > > Thanks in advance !
> > > >
> > > > Moritz
> > > >
> > > >   ------------------------------------------------------------------------
> > > >    Part 1.2Type: application/pgp-signature
> > > _______________________________________________
> > > Cmdist mailing list
> > > Cmdist@ccrma.stanford.edu
> > > http://ccrma-mail.stanford.edu/mailman/listinfo/cmdist
> > >
> >
> > kelly, angel, riley & cameron hirai
> > 343 Pennell Circle #3 Tallahasse FL 32310  usa
> > http://www.unm.edu/~khirai
> > "... because writing music is easier than running the world." Jim Atkins
> >
> > _______________________________________________
> > Cmdist mailing list
> > Cmdist@ccrma.stanford.edu
> > http://ccrma-mail.stanford.edu/mailman/listinfo/cmdist
> _______________________________________________
> Cmdist mailing list
> Cmdist@ccrma.stanford.edu
> http://ccrma-mail.stanford.edu/mailman/listinfo/cmdist
>

kelly, angel, riley & cameron hirai
343 Pennell Circle #3 Tallahasse FL 32310  usa
http://www.unm.edu/~khirai
"... because writing music is easier than running the world." Jim Atkins