Making animal sound recordings

An introduction to the science-- and art--of recording and annotating bird and other animal sounds

by Richard Bradley•

Introduction Nearly all early work in bioacoustics, as well as much of the current literature, deals with three major groups; insects, anuran B IOACOUSTICS,has come intothe activestudy of focusanimal relatively sounds, amphibians,and birds. A quick tally of papers recently. Although some valuable work was dealingwith acousticcommunication in birds done on a variety of species prior to the demonstratesthe rapid growth in this field 1940s, most of the research in this field (Fig. 1). With the tape specimensin hand, beganwith the introductionof easilyportable workerscan carefullyand objectivelyanalyze soundrecording equipment in the mid-1950s. acoustic signals used by animals for com- munication. Currently, work in bioacoust•cs isexpanding rapidlyinto almost everytaxon 300 - and most environments. New recording equipment has opened up the underwater acoustic environment as well as the ultra- sonic and subsonicranges. Amateurs as well as professionals are 200- producingan increasingnumber of new re- cordings,many of which have great poten- tial value. Unfortunately, few workers are careful to include all pertinent data neces- sary to complete a specimen. This paper 100- is meant to serve as a guide for both the serious student and the casual recordist to insure that their recordings are valuable scientificspecimens. Many people are familiar with recordings 44-48 49-53 54-58 59-63 64-68 69-74 of animal sounds now available on commer- YEARS cially publishedphonograph discs. Although

Figure 1. Growth in bird song research as indi- cated by the number of technicalpapers published x Associate in Natural Sciences, Florida State since 1944 Museum, Gainesville, Fla 32611

Volume 31, Number 3 279 these discs perform a useful service, they stantly changingtheir machinery and tech- are of secondaryimportance to scientific niques to conform to the flux in the "state investigators. Most of them are produced of the art" in electronics. This does not from recordingshoused in various private mean that others limited to less sophisti- and institutional sound archives, and are cated equipment cannot produce valuable usually presentationsof typical examples of recordings. On the contrary, excellent re- well documented sounds. Perhaps the most cordings can now be made by the serious famous sound archive, established by Peter amateur on equipment that is readily avad- Paul Kelloggand Arthur A. Allen, is housed able on the retail market for under $300. at the Laboratory of Ornithology, Cornell Two basic types of tape recorders are in University. Many other fine archives exist, use today. One type uses nalTOW(¬ in.) and scientificanimal sound specimencollec- magnetic tape enclosed in a sealed unit tions have mushroomed in recent years. A called a cassette (not to be confusedwith preliminary list of sound specimencollec- a cartridge). Another type usestape threaded tions is presentedby Boswall (1974). Sound between two separate plastic reels, and is specimensare cataloguedin the samegeneral usually referred to as open reel or reel- way as in otherscientific collections. Complete to-reel recorder. I will discuss the relative recordingdata are kept with the specimens, characteristics of these two groups below. including notes on the specific behavioral Although investigatorsoccasionally make re- context in which the vocalization was given. cordings of their subject in a laboratory, The material in sound archives is used this paper deals with equipment and tech- by investigatorsin a variety of fields in- niques for work in the field. In addition cluding ethology, ecology, taxonomy and to the many different tape recorders, an evolutionary biology. Although hundredsof equally confusingalTay of microphonescan field recordists have recorded miles of tape be used in bioacousfics. The specific type specimens,much basic information remains of recorder and microphone appropriate for to be documented.In fact, sound specimens any given situation depends upon the type have probably been obtained for fewer than of organism being recorded as well as Its half of the known speciesof birds, the best surroundingenvironment. documentedtaxon. The vocal performances Naturally most workers would like to pro- of myriads of tropical insects, amphibians duce "clean" recordings of an organism's and mammals are totally undocumented. sound, i.e. without extraneous background Recorded specimensfrom aquatic animals noise. However, some noise is usually un- are barely represented.Many of these forms avoidable. In fact, weak recordings of a have complex and interesting acoustic be- rare phenomenonare often more valuable havior. Amateurs can perform a valuable than beautiful clear tapes of more common service by contributing well documented sounds. The use of modern laboratory audio recordings of animal sounds to scientific equipment can often transform so-called collections. "useless" recordings into valuable speci- [Definitions of technical terms italicized mens. m the text may be found in the glossary, Field Notes page 285].

General Methods UTTINGperhapsDATAthe singleWITH most a important specimen step is in bioacousticrecording. It is critically im- RACTICALLYwork in bioacousticsSPEAKING involvesALL MODERNrecord- portant that as much data as possiblebe lng of animal soundsonto magnetic tape. recorded onto the tape at the time of re- The great variety of tape recording equip- cording.In this way the actual specimenand ment now available ranges from barely associateddata are physically linked and suitable to highly sophisticated.The type of confusion or transcription errors are far equipmentused dependstipon convenience less likely. Some identificationinformation of operation,portability as well as monetary should be included on the tape to avoid hnutations. Professional recordists are con- confusion, although notes put on a tape may

280 American Birds, May 1977 often be supplementedby written field notes. TapeRecorders Objections to this technique involve "wast- lng" tape for voice comments by the re- HE TWOBASIC TAPE systems used •n cordist, but experience with reels of "no bioacoustics are the cassette and open data" recordings has convinced me of the reel systems. Cassette recorders have been importance of putting data on the original shunned for many years for high quahty field tape. recordings because of several inherent prob- The minimum data that should be re- lems. corded on the tape includes: the name of Early cassettemachines had several d•sad- the organism, the location, date, time and vantages. The most important was the slow weather data where possible. Additionally, tape speed which is used for all cassette it is important to add notes on the context recorders--1.5 inches per second (ips). The of the vocal behavior. In some cases a upperfrequency responsewas usually limited statement like "singing from the top of a to about 8000 Hz. Tape speed control was bare tree" is sufficient. In other cases a also poorer than open reel machinesoperat- description of a particular display or inter- ing at higher speeds. The thin tape used •n action may be more appropriate.The more cassette recorders often resulted in a poor relevant data that are included, the more signal-to-noise ratio. These two problems valuable is the specimen. Most recordists have been largely eliminated in cassette will not wish to •'talk on top of" their technology,and modem high quality portable recordings. Unless one has the facility cassette decks are available that y•eld for recording a separate voice track simul- superior results. These are often more com- taneously, it is customaryto wait for a pause pact and convenient than the standard open •n or the end of a recording for notes. A reel machines. short note giving the name of the organism One important factor in this improvement •s often inserted near the beginning of a has been the development of chromium particular cut. This will aid during the dioxide cassettes. The chief complaint re- cataloging procedure. maining is that cassette tapes are difficult Use of an organism's scientific name is to edit compared to open reel tapes. Open preferred over its common or vernacular reel machines are usually heavier and less designation. If the vocalizer is not seen, convenient to operate. The finest open reel th•s should be noted because, at least in machines still maintain somewhat superior the case of birds, imitation is often a pos- quality to the best cassettemachines. sible source of misidentification. The loca- When selecting a tape recorder one must non should be described as the distance and balance quality and convenienceagainst eco- d•rection from a town or city, or other nomic factors. In general, inexpensive open geographic landmark, or coordinates. The reel machines give better performance than country, state and county should be in- inexpensive cassette machines. RelaUvely cluded where appropriate. Weather data inexpensive cassette recorders such as the should indicate sky and wind conditions Sony C-108 (ca $150) are probably the and air or water temperature. In the case best bet for the occasional recordist. For of amphibians it is particularly important more serious investigators who wish to have to take temperature readingsfrom the water good quality at a reasonable price, the ff the subject is partly submerged. Uher 4200 (ca $675) and Tandberg II (ca For conveniencea shorthandnotation may $800) are popular open reel models. F•ne be used. The location, date, and weather professionalquality portable open reel umts data may be recorded once for a given such as Nagra IV-SD and Stellavox SP-7 field tape, often at the beginning. The time are available for those who can afford to of each recording is then added for each cut spend around $4000. For an excellent rewew as well as any notes on changes in the of the available equipment (recorders, tapes weather. With a little practice, full data can and microphones)consult the 1976tape eqmp- be added to a specimenusing only a few ment buyersguide publishedby High Fidehty feet of tape. magazine(Tynan 1975).

Volume 31, Number 3 281 Open Reel Recorders Cassette Recorders

PENbewilderingREEL TAPE varietyRECORDERS of styles.come Onein ofa •ORTABLErange in CASSETTEprice from TAPE underRECORDERS $100 to the •mportant œeaturesone must select when elaborate portable professional decks that choosingany tape machine is the tape head may cost $400 or more. The cheapestcas- configuration. Monophonic recorders using sette machines are generally not suitable standard V4 in. tape now come with two for bioacoustics. These machines are pri- basic configurations. The first, full track, marily designedas dictatingmachines and are refers to machines that use the entire V4 in. useful for taking field notes. They suffer tape width for the signal. With this sort of from rather poorfrequency responsecharac- machine one gets the best possiblerecording teristicsand excessivewow andflutter. They characteristics. Since there is only one usually have a poor signal-to-noiseratio channelthere cannotbe any crosstalk. Ob- Medium priced units are available that are viously one can use only one side oœthe sufficientfor many bioacousticapplications, tape, since if one turned the tape over to with adequatefrequency responsefor most record the back side, it would erase the uses. Such machines can be used with an previously recorded signal. The other stand- auxiliary microphoneand are very light and ard type is called «-track. These machines compact.The new high quality portablecas- use halt of the V4 in. tape width. This sette decks offer excellent recordingquality leaves the other half free for recording a and are fairly compact. Examples of these second track by turning the tape over and include the Uher CR-134 and JVC CD-1635 recording on side 2. There is a minor loss of quality due to the narrower signal track. Stereo V4 in. tape machines come in several Tape head configurationsalso. The first type, 2- track, is essentially the same as the « track mono system except that both tracks OTHrecordersCASSETTE use thinAND plasticOPEN tape REEL coated tape may be recorded simultaneouslygiving two with a metallic oxide that can be magnetized s•gnal tracks in the stone direction. In to store the signal. Although there are many b•oacousticsthis can facilitate recordingfield tape manufacturers and equally as many notes. One can record clean sounds of the varieties of tape, there are a few basic target organism on one track while putting conceptsabout tape that can be usedto select a narration or voice track simultaneously an appropriatekind. on the other track. The type of film used for the tape is In addition, other stereo machines come called the backing. Most modern tape is w•th 4 separate tracks. These Va-track made with a polyester plastic backing often machines may record two tracks in one referredto by its Dupont trade nameMylar d•rection (stereo) then the tape is reversed Some recording tape is still manufactured and stereo is also available on side 2. Some using cellulose acetate ("acetate") backing machines also allow for 4 separate tracks This is wearer and more likely to break to be recorded in the stone direction in than polyester tape but becauseacetate tape quadraphonic. This could be valuable For does not stretch the break will be clean multiple mike recordings,but it lies outside and can be mended. With Mylar tape acci- the realm of amateur recording. dental stress will cause the tape to stretch Open reel machinescan recordat a variety and it cannot be repaired without removing of tape speeds. Common speedsare pre- a section of the recording. The newest sentedin Table 1. The standardcompromise Mylar tapes are tensilized or pre-stretched betweenrecording quality (faster speed)and so that this is not likely. Long periods of tape economy is 7.5 ips or 19 centimeters storage can result in warping, drying or per second (cms). Most moderately priced wrinkling of the tape backing. These effects open reel units can be operated at this occur more often with acetate-backedtape speed. Open reel and cassettetapes come in three

282 American Birds, May 1977 standardbacking thicknesses (0.5, 1.0 and 1.5 Microphones mil.). Obviously the thinner tapes will allow NADDITION TOTHE choice of a recorder, more time on any particular size tape reel. the aspiring recordist must select a Thin tape has several disadvantages.First, compatible microphone. As with all audio very thin tape (0.5 mil.) is more fragile equipment the range of designs and prices and breaks or stretches easily. Thin tape is wide. Several factors are involved in choos- often allows more print through than 1.0 or ing a mike for bioacoustics. 1.5 mil. tape. A compromisebetween the The first is the choice of a mike that •s convenienceof more recording time per reel matched properly to the recorder. Buying an and the disadvantageof thin tape backing expensive professional quality microphone is medium thickness tape. Most workers for use with an inexpensivecassette recorder now use 1.0 mil. tape in tensilized Mylar. is foolish. The microphone selected must This yields about 24 minutes of recording have the same impedance as the input for time on a standard 5 in. reel or 48 minutes the recorder as well as a sufficient output level. Be sure to check the manufacturers' on a 7 in. reel. These times are based upon the use of only one side of the tape. General specificationson the mike and recorder to procedureis to record only on side 1 because assure proper matching. Proper impedance the tape can be cut and spliced for editing matching means choice of a mike whose and there is no possibilityof crosstalk with impedancerating falls within the range of the recorder to be used. One should not select another channel. a 200 ohm mike for a recorder whose Backing is not the only differencebetween input impedanceis matchedfor 10,000ohms. tapes.The oxide coatingthat carriesthe mag- If the microphone is not properly matched netic signal comes in several types. In gen- it may not record the signal. Even if it does, eral, the best tapes are now made with a the level may be so low that it is nearly black-oxide coating. This is usually less inaudible. The output level of microphones abrasive to the tape recorder heads than varies greatly. The best method of selecting the older red-oxide coatings. Modem tapes a mike is to try it out with your tape also come in low-noise and normal. Low- recorder. If this is not possible, choose a noise tape is more expensive and produces mike with a medium or high output level cleaner recordings with a superior signal- (-57 to -53 decibels) (dB). to-noise ratio. It is especially well suited There are two basic microphone designs to situations where the signal has a great One, the moving coil or dynamic m•ke dynamic range. These differences are most depends upon a signal being produced by a evident with high quality recorders. vibrating coil and a permanent magnet Cassette tapes come in the same three This type of mike does not require an standardthicknesses and one width (• in.). auxiliary power source such as a battery The coatings may be standard ferrous- The second type is called a condenser oxide or chromium dioxide. Standard oxide mike. Condenser mikes depend upon an cassettes are available with the low-noise electrical field that is activated by a sound option. Chromium dioxide coated tapes can modulated capacitor. This second type of be used only with machines designed for microphone requires a power source The these tapes. Such cassetterecorders usually power source may be built into the tape have a switch that indicates the tape type recorder itself or involve a battery pack mode as "std" or "CrO2". When possible located in the microphone body. Many of CrO2 tape is preferred because of its the finest microphones use this condenser superior quality. Cassettescome in several design. For most bioacoustic applications lengths. Small portable cassette machines this distinction is not critical, but condenser often have trouble using the longer cassettes mikes and their associated power systems (C-90 and C-120). For field recording the require additional maintenanceand serwce. C-60 is a suitable compromise. This ler•gth A third factor one should consider when allows about 30 minutes recording time on buying a microphone depends upon the or- one side. ganism to be recorded. In most cases, the

Volume 31, Number 3 283 Table 1. Tape speedsand their uses.

Tape speed

inches/ centimeters/ secondtips) second(cms) use

30 76.2 professionalstudio quality, superiorfrequency response,facilitates ac- curate editing 15 38.1 standardsemi-professional, used by somefield recordists,good editing speed

75 19 good tape economywith goodquality, used in most bioacousticstudies 3 75 9.5 excellent tape economy, compromiserecording quality on all but better machines

I 5 4.7 useful for dictation, minimal quality except finest machines,used by all cassette systems 0 94 (15/16) 2.4 usefulfor field notes or conferencerecordings, poor recordingquality recordist wishes to (or must) remain a con- must be considered. One must choose a site siderable distance from his subject. A direc- that is convenientand accessible,yet not too tional microphoneis then required. The three severelydisturbed by noise.Large highways, basic types of microphones are omnidirec- industrial plants and airports are the worst tional, cardiod and super or hypercardiod. sources of background noise. High wind Omnidirectional mikes are sensitive to sound will cause microphone dropout with most from all directions. Omnidirectional mikes systems. pick up "all of outdoors" including the Timing is often critical when attempting highway behind the recordist. Cardiod mikes to record a particular sound. Certain calls are more sensitiveto soundscoming from in may be uttered only when specific weather front of them and are superior for animal conditionsexist or during one phase of the sound recording. The supercardiodor shot- annual cycle. The presence of a recordist gun mikes mayketedby some manufacturers often has an effect on the anima]'s be- are very convenient,make better recordings havior, and the recordist should remain as since they can be directed accurately at far removed as possible within the limita- the subject,are not sensitiveto background tions of the recording equipment. One tech- sounds even a short distance off mike (to nique, borrowed from nature photography, the side), and unfortunatelyare very expen- is to make preliminary recordingsfrom a sive The SennheiserMK 815 hypercardiod distance,then approachgradually. When this mike retails for nearly $600. Omnidirectional is done the progress should be noted on or standard cardiod microphones can be the tape. Later, in the laboratory one might made directional by mounting them in a be able to determine if progressivemove- curved dish which focuses the incoming ment caused any subtle changes in the sound.The shapeof thesereflectors is usually vocal performance. parabolic. Mikes mountedon parabolic re- Recordists have traditionally used play- flectors yield results comparable to super- back of preliminary recordings or pre- car&od mikes. For most investigators a recorded loops to incite an organism to parabolic reflector with a less expensive approachor vocalize. For many speciesthis car&od mike is the best compromise.One is an effective method of obtaining loud pre-assembledsystem is the Dan Gibson clean recordings. Unfortunately, the char- "sonoscope"parabolic mike (about $150). acter of the vocal performance may be altered by this stimulus. It should always Making Recordings be mentionedin field notes when playback was used to stimulate an animal. In addi- tion, includecareful notes about the specific rHENOf animalPLANNING sounds,FIELD severalRECORDINGS factors responsemade to the recording.

284 American Birds, May 1977 Glossary Output level--The specific power level of a particular unit. In the case of microphones Automatic level control (automaticgain con- it is measuredin -dB. For example -55 dB trol)--This is a feature which makes an auto- is a higher output level than -65 dB. matic correction for different sound levels. Print through--The tendency for a recorded It is not uõually used for bioacousticsbe- signal to magnetize the adjacent tape it •s cause it tends to increase the background wound next to. This causes an echo effect at the expenseof the signal. since the signalis reproducedwith a fraction Cross talk--The tendency of the signal to of a second delay on the tape. bleed over onto the adjacent signal track Signal-to-noiseratio--The ratio of the s•gnal and become audible as a backgroundnoise level (intentionally recorded sound) to the there. background noise. Background noise here Cut--One continuous recording, from the includesboth the inherent noisesof the tape time the recorderwas engageduntil it is shut and mechanical or electronic noise Intro- off ag•/n. duced by the recorder. A high signal-to- Decibel (dB)--Decibels are a measure of noise ratio is desirable. Such a condition sound intensity, a Changethat is perceived means that the background noise is rela- as twice as loud as equivalent to a change tively low and at a volume that is usually of + 10 dB. The decibel scaleis logarithmic, inaudible. For most recorder-mike systems not linear. a good s/n ratio is about 55 to 60 dB. Ratios Dropout--A blank or intermittently broken lower than 50 dB are consideredpoor. signal due to a poor connection or wind Tensilized--The processby which polyester across the mike face. (Mylar) tape is pre-stretched so that further Dynamic range--The range in loudnessof a stretching is unlikely. Tensilized tape is particular recording. A large dynamic range desirable because the tape is more hkely meansthat some soundsare very quiet while to make a clean break that can be re- others are loud. paired. Flutter--A type of distortion due to poor Warping--The tendency for old tape to speed control where the speed changes change shape and become twisted slightly rapidly so that a constant pitch sound re- from one side then the other. Warped tape sembles a tremolo. will not run smoothly across the tape Frequency--The pitch of a sound in Hz (Hz heads and w/ll not record or playback stands for Hertz and supercedesthe older properly. measure cycles per second [cps]). Wrinkling--This aging effect is similar to Frequency response--The frequency re- warping but only involves the edges of the sponse is given as a range from lowest tape so that a fringe of crinkles appear on to highest in Hz. In addition to the range the tape with the same sort of unsatis- a figure follows such as +2 or +5 dB. This factory results as warped tape. refers to the curve of the frequency re- Wow--A type of distortion caused by •n- sponse. The number of dB variation from correct tape speed control. This is s•mllar a fiat curve is indicated. A unit may have to flutter except that the pitch changes are some sensitivityup to 15,000 Hz, but may slower so that a wow sound is heard. be much less sensitive (-5 dB or more) at that frequency. One shouldchoose a machine which meets or exceedsthe actual frequency Literature Cited range to be recorded within the "flat" curve Boswall, Jeffcry. 1974. A preliminary world list of the recorder and microphone. of wildlife sound libraries. Recorded Sound, $4 Impedance--A measure of the resistance of 276-277. some system to current. For more informa- Burns, Desmond M. and Simon G. G. MacDonald tion on the physics of electronic impedance 1970. Physicsfor biology and pre-medical stu- dents. Add/son-Wesley,London. 614 pp. consult Burns and MacDonald, 1970. Tynan, William (ed). 1975. 1976 buyers guide to Ohms--The units of resistance to an elec- the world of tape. ABC Leisure Magazines trical current. Inc., Great Barrington, Mass. 116 pp.

American B•rds, May 1977 285