naturerecordists
[Top] [All Lists]

6. Re: FR-2LE & AT4022

Subject: 6. Re: FR-2LE & AT4022
From: "Paul Jacobson" thebrunswicktwitcher
Date: Mon Mar 15, 2010 6:26 pm ((PDT))
On 16/03/2010, at 8:35 AM, vickipowys wrote:

> I might add that I don't hear this hiss in other recordings I have
> from Andrew (recorded with SASS), nor from my own SASS recordings.  =

> Also Paul says that the hiss does not occur in later parts of the
> (long) recordings that they made, so it must be simply the fact that
> Andrew's mics were pointed slightly upwards towards the tree canopy
> and the SASS has picked up leaf noise when there has been a slight
> wind movement.

Hi Vicki and all,


I've gone back to listen to a segment I've chopped out for Rob and I'm now =
not so certain this is foliage noise.  To get to the levels of the files we=
 submitted to you for inclusion in the journal and which uploaded to the we=
bsite I think the SASS recordings have been boosted by around 12dB+.  My HS=
PBA recording has had additional 6-9dB gain applied to match the SASS files=
.  When I match and boost another quiet section a similar level of hiss is =
apparent and I am fairly sure this is just the shaped noise floor of the MK=
H mics becoming audible due to the amount of post-processing gain applied.=


Raimund Specht's noise floor testing illustrates the shaping   used in the =
MKH mics perfectly - http://www.avisoft.com/test/noisefloors.pdf  - the ris=
ing noise floor of the MKH60 above 10khz is clearly illustrated.  You could=
 consider this as the "dirty little secret" of the MKH series which is tail=
ored to give good A-weighted noise figures. Compare the a-weighting curve s=
hown on  http://www.tonmeister.ca/main/textbook/node330.html with the risin=
g noise floor of the MKH-60 shown in the noisefloors.pdf and you can see pr=
ecisely why the MKH-60 has a good a-weighted noise spec. There is obviously=
 less happening above 10khz in most nature recordings so we can often knock=
 back 10Khz+ safely with a HF Shelf filter but I know I got a shock when I =
first heard an MKH-416 compared to a ME66 - the HF hiss was really very obv=
ious in quiet recordings.

Warning - this deviates into a bit of a diatribe on noise measurements...

Doing a bit of research on the weight curves I found that "A-weight" was no=
t designed to be used for measuring electronic equipment. The CCIR/ITU_R 46=
8 weighting is apparently more appropriate for this use: http://en.wikipedi=
a.org/wiki/ITU-R_468_noise_weighting

"It is important to realise that the 468 specification covers both weighted=
 and 'unweighted' (using a 22 Hz to 22 kHz 18 dB/octave bandpass filter) me=
asurement and that both use a very special quasi-peak rectifier with carefu=
lly devised dynamics (A-weighting uses RMS detection for no particular reas=
on[citation needed]). Rather than having a simple 'integration time' this d=
etector requires implementation with two cascaded 'peak followers' each wit=
h different attack time-constants carefully chosen to control the response =
to both single and repeating tone-bursts of various durations. This ensures=
 that measurements on impulsive noise take proper account of our reduced he=
aring sensitivity to short bursts. This quasi-peak measurement is also call=
ed psophometric weighting.
This was once more important because outside broadcasts were carried over '=
music circuits' that used telephone lines, with clicks from Strowger and ot=
her electromechanical telephone exchanges. It now finds fresh relevance in =
the measurement of noise on computer 'Audio Cards' which commonly suffer cl=
icks as drives start and stop."

The same situation really applies to digital field recorders that are often=
 subject to low level electronic noise that is smoothed over by a-weighting=
 and rms readings of EIN.  There seems to be some promising developments in=
 that Robin Whittle has proposed to begin writing a free software implement=
ation of a CCIR 468-4 filter during 2010.

http://realfield.com/anm/software/

"The program will be useful for applications including reliable measurement=
 of:
        =95 Pre-amp and ADC noise.
        =95 Microphone noise, as equivalent input weighted SPL noise.
        =95 Background noise in any electronic device."

I'll be very interested to see how this kind of filtering will reshape our =
ranking of recorders and external preamps :)

cheers
Paul









<Prev in Thread] Current Thread [Next in Thread>
  • 6. Re: FR-2LE & AT4022, Paul Jacobson <=
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