naturerecordists
[Top] [All Lists]

Re: artifact in spectrogram created using Raven

Subject: Re: artifact in spectrogram created using Raven
From: "rock_scallop" rock_scallop
Date: Mon Jul 16, 2012 3:18 pm ((PDT))
Hi Laura,
I could not access your SounCloud files for some reason. But nevertheless, =
from what you say below I am willing to bet the problem is in the MP3 compr=
ession. Such visual(but not necessarily heard)artifacts are common with MP3=
. If the sound is loud enough, the quieter frequencies above and below it a=
re somehow removed to save space.

If you can, it is best to record in wav format at a minimum of 44.1kHz/16bi=
t resolution. If you must record in MP3, use the highest bitrate available.=
 Any bitrate less than 128kbps is likely to exhibit audible artifacts.

John Hartog
rockscallop.org





--- In  "laura.hilberg" <=
> wrote:
>
> Thank you all so much for your suggestions (and experiments)! It's a huge=
 relief to have a community of people who really understand this stuff.
>
> I uploaded a short recording of a single song (the same one in the pictur=
e). It is on SoundCloud, and the URL is http://soundcloud.com/laura-hilberg=
/nesp-song-short-clip. I made it public and downloadable, so hopefully it w=
ill work for everyone. I also uploaded the complete, unmodified sound file =
(about 10 minutes), which is at http://soundcloud.com/laura-hilberg/6-12-12=
-bhm-nesp-11.
>
> To try to answer some of your questions:
>
> 1) I am absolutely sure that the automatic gain was turned off. What this=
 meant was that I adjusted it manually for each recording, trying to ensure=
 that the gain was high enough to pick up the sound but not so high that it=
 went over. I'm not sure if this is the same thing as 'fixed gain', but I d=
o know that it was adjusted manually only.
>
> 2) The limiter was turned off.
>
> 3) Is clipping the same thing as aliasing (referred to in the Raven manua=
l)? If so, then I know what you're referring to, but if not then I'm still =
not sure what exactly that is. John, your experiment with the fan sounds ex=
actly like what I am experiencing, with the white shadow eclipsing the back=
ground noises in high-volume (or perhaps it is related to high frequency) s=
ounds.
>
> If it has to do with aliasing (as I understand it, sounds that are over t=
he Nyquist frequency are distorted upside down somehow), then Raven has a l=
ow-pass (anti-aliasing) filter. I'll experiment with that and see if it cha=
nges what I see. I know the low-pass filter was turned off on the recorder,=
 so perhaps this is the problem.
>
> 4) My files were recorded as MP3s, with a 44.1kHz sampling frequency and =
16-bit sampling rate. Perhaps this is the problem... if so, then I am embar=
rassed to say that it's a problem I'm stuck with, as that was the default f=
ile format, no one mentioned that it wouldn't be good enough, and I didn't =
know enough to change it.
>
> Laura
>








<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