naturerecordists
[Top] [All Lists]

Re: artifact in spectrogram created using Raven

Subject: Re: artifact in spectrogram created using Raven
From: "mipartitus" mipartitus
Date: Mon Jul 16, 2012 12:43 pm ((PDT))
Very nice, John. It indeed seems that I spoke too soon when I said clipping=
 probably had nothing to do with it. Inspired by your experiment I did the =
same but then purely digitally. Turns out that if you add a high amplitude =
tone to a white noise background, the noise is attenuated somewhat when the=
 digital signal is clipped. Makes sense perhaps, because the clipped peaks =
contain background noise that is simply removed as all clipped values are e=
ffectively set to a fixed value. Clipping should be pretty strong though, s=
o I don't know if it can account for Laura's phenomenon. Best, Gabriel

--- In  "John Crockett" <> wrote:
>
> Hi Laura,
>
> Do all of your artifacts correspond to clipped signals? I just tried a li=
ttle experiment with an ME66 and both the PMD-670 CF and MZRH1 MD recorders=
. I turned on a fan to create some good background noise, and then whistled=
 loudly. First I had the record gain low enough so the whistle signal did n=
ot clip. Then I turned up the gain so it did clip. In the resulting sonogra=
m, a white (low volume) "shadow" appears in the frequencies below the whist=
le signal wherever it is clipped, eclipsing the fan noise, just as it appea=
rs in your sonogram. The worse the clipping, the more pronounced the "shado=
w."
>
> In the unclipped samples, the fan noise continues uneclipsed.
>
> Is there something about clipping that overloads the preamp so it cuts ou=
t other frequencies besides the signal that is causing the clipping?
>
> John Crockett
>
> --- In  "laura.hilberg" <laura.hilberg@>=
 wrote:
> >
> > I was using a PMD 680 recorder and a Telinga microphone with a parabola=
. I was using manual gain (no automatic), but I had some difficulty getting=
 close enough to the birds for a clear signal and then having them shift po=
sitions so they were facing me more and being blasted away. And this is my =
first time really recording, so inexperience definitely plays a role (and t=
hankfully, they get better as they go along).
> >
> > Can you say what you mean by the gain needing to be set at a fixed leve=
l so that the sound doesn't clip? It was fixed, so I don't think that would=
 be an issue.
> >
> > And sorry about the sound sample - it wouldn't let me post it (464kb) b=
ecause it said there wasn't room.
> >
> > Thank you so much!
> >
>








<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU