naturerecordists
[Top] [All Lists]

Re: Archiving and Preserving Sound Files for Fifty Years Hence, Hund

Subject: Re: Archiving and Preserving Sound Files for Fifty Years Hence, Hund
From: "Dan Dugan" dandugan_1999
Date: Wed Aug 6, 2008 8:47 am ((PDT))
Martin Krieger, you wrote,

> How do you archive and preserve sound files?

At a minimum, I use .wav files and keep them on two hard drives, one a
mirror of the other. For long-term storage I make audio and data CDRs.

> Calibrate for the long term?

That's a challenge I'm concerned about, and I'm not to a satisfactory
solution yet. Here is the calibration procedure I'm currently
recommending for volunteers helping with soundscape monitoring in
national parks:

*** start quoted text

Recording Level
             To establish a starting point, hold the microphone at
arms length, sensitive side facing you. Speak normally to a person
near you. Set the recording gain so that the peaks don=92t exceed a
level of =966dB on the recorder=92s meter. This gain will work for quiet =

environments. Loud sounds will require lower gains.

Calibration
             An affordable scheme for calibrating field recording
outfits is being sought. Here is one proposal:
             Terminology: =93Recording Gain=94 is the setting of the
recorder. =93Recording Level=94 is the reading of the recorder=92s peak met=
er.
             Equipment: A digital audio workstation with speakers, a
sound level meter, the recording outfit to be calibrated.
             Procedure:
1)    Prepare a one-minute stereo (or more channels if you have them)
track of pink noise. If possible, make the channels uncorrelated by
recording the noise on each channel separately.

2)    Filter the track with a high-pass at 200Hz and a low-pass at
Message: 2000Hz.

Subject: 
3)    Play the filtered track on each speaker, adjusting levels so the
SPLs of all individual channels are equal. Turn on all channels and
read the sound level meter while adjusting the monitor level to 64dBA
(slow) at your =93sweet spot=94 head position.

4)    Place your mic precisely at the same position and adjust the
recording gain so that the recorder=92s peak meter reads =966dB (this is =

dBFS, meaning =93decibels full-scale=94).

5)    Note or mark your recorder=92s gain setting. Use this as a
standard gain for recording in quiet places. Recordings should be
annotated =93Calibration: 0dBFS =3D 70dBA=94

6)    Make a one-minute recording of this calibration noise. Save it
to calibrate transfers later.


Slates
             Each take should have a voice announcement (slate) at the
head or tail including:
                         Recordist=92s name
                         Date
                         Time
                         Location
                         Weather and wind conditions
                         Microphone(s) model number
                         Type of stereo array
                         Type of windscreens
                         Recorder model number
                         Recorder gain setting
                         Optional description of the scene and the
sound sources you can identify, observed wildlife behavior.

*** end quoted text

I'm dreaming about making a field calibrator. Version 1 would be a
little bullhorn with a portable-radio antenna the extends .5m out in
front. It would be used by holding it on-axis to the mic, touching the
tip of the antenna to the tip of the windscreen, and holding the
trigger to make a 5-second tone. On to the next mic.

It would have to have its output stabilized for battery level and
temperature changes. It would put out a 1K tone at 64 dBSPL at the tip
of the extended antenna. Hopefully the .5m (19 inch) distance would
make the difference between the tip of the windscreen and the mic
diaphragm small enough in proportion to be within the accuracy of the
system, again hopefully +/- 1 dB.

Version 2 would play a programmed sequence when you pulled the
trigger: sound level calibration tone, voice announcement of time,
weather data, GPS location, and the same data low-speed modem-encoded
so it could be machine read.

-Dan Dugan


<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