naturerecordists
[Top] [All Lists]

Re: Newbie's Guide to Post Processing.....

Subject: Re: Newbie's Guide to Post Processing.....
From: "Avocet" madl74
Date: Fri Nov 25, 2011 11:11 am ((PST))
Marc,

I've had a good play with your track using Audacity V1.3 and
have pushed it a bit so you can see what can be done and the
snags.

DR000938michisssample.wav

On track box down arrow - select Waveform (dB) display

I've included the scripts below to add to the Audacity file
\Audacity\EQCurves.xml or just copy the GIFs. If you alter
the code, save a backup in case you goof up. :-)

Apply Effects - Equalisation as in
www.stowford.org/sounds/hpf800-2-4-6.gif
This is my standard bass cut which limits low frequency
artifacts when using noise reduction. The numbers are the
slopes in dB per octave.

Select sample at 00:26.700 for 1 sec

Effects - Noise Removal - Get Noise Sample.

Select all and apply Effects - Noise Removal - 9dB; 500Hz;
Attack 1.0secs
I usually use 500Hz as it gives fewer LF artifact.)

This lowers the general HF b/g noise

Select sample 01:03.400 for 1 sec. This is a distant cicada
sample.

Again, Effects - Noise Removal - Get Noise Sample

Select all and again apply Effects - Noise Removal.
Note: a second noise reduction only works with a different
frequency sample.
The general cicada noise is now lowered. Dynamic range is
too high with HF peaks near the end.

Reduced dynamic range with a "limiter" setting of the
compressor:
Effect - Compressor; 16dB; 60dB; 10:1; 0.1sec 1.0sec; make
up gain "no"; Compress based on peaks.

Now the quieter sounds are more prominent.

Listening to whole track, at 01:20.000 the high frequency
peaks tend to cause distortion.
Applied Equalization - see
www.stowford.org/sounds/step4000to10kminus6.gif
Not enough - do  it again.

This version is on
www.stowford.org/sounds/dr000938michisssamplenrnrcompstep.mp3

There is a little burbling from the double noise reduction
and a steeper bass cut would probably reduce this but remove
the pigeon as well.

I tried boosting the pigeon frequencies after finding them
using the spectogram display and applying the bandpass
filter:
www.stowford.org/sounds/bpf300to400plus6.gif
but it brought up the burble as well.

The answer to the low level burble is to mask it with a low
level parallel track of the original after rolling off
frequencies above 100Hz.

Have fun. :-)

David

David Brinicombe
North Devon, UK
Cogito cogito ergo cogito sum - Ambrose Bierce


Scripts to add into the file: \Audacit\EQCurves.xml
Save a backup first. :-)

 <curve name=3D"HPF800-2-4-6">
  <point f=3D"20.000000000000" d=3D"-26.000000000000"/>
  <point f=3D"200.000000000000" d=3D"-6.000000000000"/>
  <point f=3D"400.000000000000" d=3D"-2.000000000000"/>
  <point f=3D"800.000000000000" d=3D"0.000000000000"/>
  <point f=3D"22049.000000000000" d=3D"0.000000000000"/>
 </curve>
 <curve name=3D"step4000to10Kminus6">
  <point f=3D"21.000000000000" d=3D"0.000000000000"/>
  <point f=3D"4000.000000000000" d=3D"0.000000000000"/>
  <point f=3D"10000.000000000000" d=3D"-6.000000000000"/>
  <point f=3D"22049.000000000000" d=3D"-6.000000000000"/>
 </curve>
 <curve name=3D"BPF300to400plus6">
  <point f=3D"21.000000000000" d=3D"0.000000000000"/>
  <point f=3D"150.000000000000" d=3D"0.000000000000"/>
  <point f=3D"300.000000000000" d=3D"6.000000000000"/>
  <point f=3D"400.000000000000" d=3D"6.000000000000"/>
  <point f=3D"800.000000000000" d=3D"0.000000000000"/>
  <point f=3D"22049.000000000000" d=3D"0.000000000000"/>
 </curve>







<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