Sound Recording in the Field

This tutorial suggests recording equipment and describes some of the most common problems and errors that may occur during sound recording in the field.

Microphone Choice

There is a wide range of microphones available that could be used for recording animal sounds in the field. Well-tried and reliable microphones are available from Sennheiser. The model K6/ME66 provides a good compromise of size, sound quality and price. The more expensive Sennheiser MKH series provides slightly lower self-noise levels at low frequencies (below about 10 kHz) and is more robust in humid environments. However, the noise performance at frequencies between 15 and 20 kHz is better in the ME series (the MKH series microphones have been optimized for best noise characteristics to human listeners, while the ME series provides an equally distributed noise floor across the entire frequency range).

A parabolic reflector in conjunction with an omni-directional microphone (e.g. Sennheiser K6/ME62) would provide a significantly higher degree of directionality. Sounds originating from off-axis directions will be very effectively attenuated. This property of a parabola can dramatically improve the spectrographic analysis of recordings made in dense forest environments (a high level of short-term echoes (reverberation) from other directions can disable the meaningful analysis of rapidly modulated sounds). A parabola also provides a higher microphone output level (it is an acoustic amplifier). Unfortunately, a parabolic reflector also distorts the frequency response of the entire microphone system (see Parabolas). Due to its limited size, the reflector is only effective at wavelengths shorter than the diameter of the dish. Another drawback is its large size.

Note that all directional microphones (certainly shotgun and parabolic microphones) have a non-linear frequency response (off-axis colouration) when the subject is not located on-axis (they are more directional at higher frequencies). Consequently, these microphones should be pointed exactly at the vocalizing animal.

The following table shows the suitability of the various microphone models for various recording distances.

Microphone modelshort range (< 1 m)mid range (< 10 m)long range (> 10 m)
Sennheiser K6/ME62 (omni-directional)++++
Sennheiser K6/ME64 (cardioid)+++++
Sennheiser K6/ME66 (super-cardioid / short gun)+++++++
Sennheiser K6/ME67 (lobar / long gun)+++++++
Sennheiser K6/ME62 + Parabola+++++++

Further information on microphone theory and operation can be found in the book Microphones – Methods of Operation and Type Examples by Bore and Preus, Georg Neumann GmbH, Berlin.

Recorder Choice

Recorder modelhandlingruggednessweightsound qualitybandwidthP48price
Marantz PMD561++++++++353g++++++++low
Marantz PMD620++++++110g++++++low
Marantz PMD660+++++++500g+++++++low
Marantz PMD661MKII++++++++353g+++++++++low
Marantz PMD661MKIII++++++410g++++++++low
Marantz PMD670++++++++1.3kg+++++++mid
Marantz PMD671++++++++1.3kg+++++++++low
Tascam DR-1++++++++203g++++++mid
Tascam DR-2d++++++++140g++++++mid
Tascam DR-05+++++116g++++++low
Tascam DR-07+++++125g+++++++low
Tascam DR-07MKII+++++127g+++++++low
Tascam DR-08+++++56g++++++low
Tascam DR-70D++++++++530g++++++++++low
Tascam DR-100++++++++290g++++++++mid
Tascam DR-100MKII++++++++290g+++++++++mid
Tascam DR-100MKIII++++++++370g+++++++++++mid
Tascam DR-680++++++++1.2kg+++++++++++mid
Tascam HD-P2++++++++1.3kg+++++++++++mid
Tascam HS-P82++++++++3.7kg+++++++++high
SoundDevices 722++++++++1.2kg+++++++++++high
KORG MR-1000++++++++1.0kg+++++++++mid
Fostex FR-2++++++++1.5kg+++++++++++high
Fostex FR-2LE++++++++0.8kg++++++++++low
M-Audio MicroTrack II++++++150g+++++++low
Edirol/Roland R-09HR++++++174g++++++++low
Roland R-26++++++++390g+++++++++mid
Olympus LS-3/LS-7+++++++90g++++++++low
Olympus LS-5, LS-10, LS-11+++++++165g++++++++low
Olympus LS-11/12+++++++168g++++++++low
Olympus LS-100++++++++280g+++++++++mid
SONY PCM-M10+++++++187g++++++++low
SONY PCM-D50++++++++365g++++++++mid
SONY PCM-D100++++++++395g++++++++high
New MiniDisk (Sony HiMD) *+++120g+++++++very low
Standard MiniDisk (Sony, Sharp) *+++120g++++very low
Standard MiniDisk (HHB Portadisk)++++++++1.8kg+++++high
DAT professional (TASCAM DA-P1, SONY TCD-10) *+++++++1.2kg+++++++high
DAT consumer (SONY TCD-D8, TCD-D100) *+++++low++++++mid
Sony WMD-6C (analog compact cassette) *+++++low++++mid

* discontinued recorders

The column “handling” depicts the ease of use in the field. The smaller recorders have often very small buttons that are more difficult to use than the bigger ones in the larger recorders. The “ruggedness” mainly depends on the type of microphone connectors. Tiny 3.5 mm mini jack connectors are less reliable than large XLR connectors. CompactFlash recorders have no moving parts and are therefore more reliable than analog compact cassette, DAT, MiniDisk and Harddisk recorders.

The sound quality depends on the self-noise level of the preamplifier (though, in conjunction with sensitive condenser microphones (50mV/Pa), the inherent self noise level of most recorders is not an issue for practical applications in the field (see also Understanding microphone and recorder/preamplifier noise specifications and Microphone Input Noise Comparision) and artifacts caused by lossy audio compression formats (MiniDisk or MP3). The bandwidth depends on the maximum available sample rate (44.1, 96 or 192 kHz) and the recording format (ATRAC/MP3 compression). The potential artifacts that can be caused by ATRAC/MP3 compression depend on the bandwidth of the sounds to be recorded (see also Audio Compression Effects). P48 indicates whether the unit is capable to provide 48V phantom power for microphones. Phantom power is required by high-quality microphones (e.g. Sennheiser MKH series). The Sennheiser K6/ME series can also be powered via P48 (the battery is then not required).

The optionally available 24 bit recording mode that is availably in many recorders will usually not provide any advantage for common sound recordings in the field. The 16 bit format already provides a dynamic range (theoretically 96dB) that exceeds the dynamic range provided by the microphone. Even a very high sound level of 90 dB (that might occur when the microphone is placed very close to a loudly vocalizing animal) recorded with a very low-noise microphone (self-noise level of 5dBA) would only provide a dynamic range of less than 85 dB (= 90dB – 5dB).

Supported Sample Rates

Matching the recorder’s sample rate to the bandwidth of the sounds to be recorded (or to the frequency range you intend to investigate) will reduce the required storage space of the .wav files. Also note that spectrographic analysis is easier when the sample rate matches the frequency range you are interested in. The maximum signal frequency should be less than about 40% of the sample rate (except at 174.4 and 192 kHz, where most recorders provide less steep anti-alias filters).

Other Features

Recorder modelpre-record
buffer
boot-up
time
mono
mode
computer
interface
mediaremarks
Marantz PMD6203 s+USB 2.0 miniSD
Marantz PMD6602 s8 s+USB 1.1CFrechargeable batteries don’t fit
Marantz PMD6612 s3 s+USB 2.0SD
Marantz PMD6702 s8 s+USB 1.1CF
Marantz PMD671up to 4 s8 s+USB 2.0CF
Tascam DR-12 s8 sUSB 2.0 miniSD
Tascam DR-2d2 s5 sUSB 2.0 miniSDdual recording mode with different gain settings, automatic recording
Tascam DR-052 s5 sUSB 2.0 minimicroSD
Tascam DR-072 s6 sUSB 2.0 miniSD
Tascam DR-07MKII2 s5 sUSB 2.0 minimicroSD
Tascam DR-082 s7 sUSB 2.0 minimicroSDautomatic and timer-controlled recording
Tascam DR-1002 s6 sUSB 2.0 miniSDlow-noise built-in UNI mics,
Tascam DR-100MKIII2 s3 s+USB 2.0 microSDlow-noise built-in UNI mics, automatic recording
Tascam DR-6802 s8 s+USB 2.0SD6 microphone inputs, automatic recording
Tascam HD-P2up to 10 s10 s+FireWireCF
Tascam HS-P82up to 5 s+USB 2.0SD8 microphone inputs
SoundDevices 722up to 10 s5 s+FireWireHDD/CF
KORG MR-10004 sUSB 2.0HDD
Fostex FR-2up to 20 s+USB 1.1CF
Fostex FR-2LE2 s6 s+USB 2.0CF
M-Audio MicroTrack II18 s+USB 2.0 miniCFbattery not replaceable!
Edirol R-09HR10 sUSB 2.0 miniSDthe (recording level) display is very difficult to read when the sun is shining
Roland R-262 s7 s+USB 2.0 miniSDtouch panel
Olympus LS-53 s+USB 2.0 miniSDV-Sync option, USB audio interface mode
Olympus LS-3/LS-72 s4 s+USB 2.0 minimicroSDTimer/V-Sync option, USB audio interface mode
Olympus LS-104 sUSB 2.0SDV-Sync option
Olympus LS-112 s+USB 2.0 miniSDV-Sync option, low-noise built-in mics, USB audio interface mode
Olympus LS-12/142 s4 s+USB 2.0 miniSDvery long battery life, USB audio interface mode
Olympus LS-1002 s8 s+USB 2.0 miniSD
Sony MZ-RH1USB 2.0 miniHi-MDpotential risk to lose audio data after a power failure or a shock while recording
SONY PCM-M105 s6 sUSB 2.0 miniMemory Stick Micro, microSDvery long battery life, very low-noise (17dBA) built-in mics
SONY PCM-D505 s8 sUSB 2.0 miniMemory Sticklow-noise (20dBA) built-in mics
SONY PCM-D1005 s1 sUSB 2.0 microMemory Stick, SDlow-noise (19dBA) built-in mics

A pre-record buffer allows to safely capture unexpected or sporadic sound events. The boot-up time (the duration from powering the device on to ready-to-record) should be as short as possible. If stereo recording is not required, the recorder should support a mono recording mode in order to save CF card storage space. A fast computer interface (USB 2.0 or FireWire) will accelerate the data transfer to the PC (In any case, it is possible to use a separate USB 2.0 card reader interface for a quick file upload).

Clipping

Adjust the recording level (gain) in such a way that clipping does not occur (there should be a head-room for covering unexpected louder sound events). If the gain was too high, louder sounds would exceed the available dynamic range of the recorder and would be clipped. Clipping (overmodulation) can also be caused by very low-frequency but high-amplitude wind noise. Wind noise can be minimized by using a basket windshield with a hairy cover for the microphone and/or by activating a high-pass filter either on the microphone (the Sennhheiser K6 powering module has an integrated switchable high-pass filter) or on the recorder. Another way to reduce the influence of wind is to hold the microphone close to the ground, where the wind speed is lower.

Note that directional microphones are usually more sensitive to wind and handling noise than omnidirectional microphones. Handling noise can be reduced by using a suspension/pistole grip

  • Overlayed low-frequency wind noise causes clipping that may produce spurious harmonics.
  • Protecting the microphone against the wind or activating a high-pass filter prevents the distortion shown above.

Noise and Reverberation

In order to get satisfying signal-to-noise ratios, the microphone should be placed as close as possible to the vocalizing animal. If possible, keep away from any disturbing noise sources and acoustically reflective surfaces such as vegetation. Subsequent automated analysis procedures will benefit from higher signal-to-noise ratios (high signal level of the vocalizing animal and low noise levels). If it is impossible to approach the subject close enough, a more directional microphone would help to attenuate noise and short-term echoes (reverberation) from other directions.

Ideally, the microphone should pick up only the direct sound (solid red arrow). Any other noise and reflected sounds should be minimized.