UNCLASSIFIED

AD NUMBER

CLASSIFICATION CHANGESTO:FROM:

LIMITATION CHANGESTO:

FROM:

AUTHORITY

THIS PAGE IS UNCLASSIFIED

ADC022345

UNCLASSIFIED

SECRET

Approved for public release; distribution isunlimited.

Controlling DoD Organization: Naval OceanSystems Center, San Diego, CA 92152.

ISCAP ltr dtd 6 Dec 2016; ISCAP ltr dtd 6 Dec2016

r~·-.

,-,·:· . .,~,~:.­,, ~-;:t-:o·.:

DECLASSIFIED UNDER AUTHORITY OF THE IN'fERAGENCY SECURITY CLASSIFICATION APPEALS PANEL, E.O. 13526, SECTION 5.3(b)(3) I

;I I

ISCAP APPEAL NO. 2006..033, document no. 1 L DECLASSIFICATIO;\l" DATE: December 02,2016 t

N Technical Report 422

PROJECT COMBO: REVIEW AND RECOMMENDATIONS (U)

_.., ,.;__~~~~ ~~ci!N~.rx- ~~~~~O:A ~~lao4 Dtaalosur• SUbJoot to Cr1m1anl ....\111111• .. :..-----

WC Cummings, RL Seeley, PO Thompson, RA Johnson

July 1980

Prepared for: Naval Sea Systems Command

and Naval Electronic Systems Command

OTIC

SELECTS: AUG 28 1980

NAVAL OCEAN SYSTEMS CENTER SAN DIEGO, CALIFORNIA 92152

SECRET """"'"' .,,. s~o __ -'-"-"s------2.R.Ls_.......s ...... l_,._7 ....... _ ..... , .............. ·"""''"·---..... ··-"--'*--&iffi-· #AI

Best Available

Copy

'

iP»S!Ppw^:?spps

«^-«MMiurtna <ma ■iiMiwLMiini.Hi^w^MmCWWWMBgmBBHWI ■ -,.-I,...I-STI-"-;

SECRET

(This pcrgs fe uncles;-:v-j

NAVAL OCEAN SYSTEMS CENTER, SAN DIEGO, CA 92152

AN ACTIVITY OF THE NAVAL MATERIAL COMMAND

SL GUILLE.CAPT, USN HL BLOOD Commander Technical Director

ADMINISTRATIVE INFORMATION (U)

(U) This work was performed under sponsorship of Naval Sea Systems Command and Naval Electronie Systems Command.

(U) We acknowledge the assistance of J. F. Fish and R. S. Gales of NOSC through- out Project COMBO. We are grateful for the help of M. li. Dahlheim, A. Dibelka. D. Ferrin. J. D. Finch. M. D. Flaska. S. J. Kennison. LT M. Marich USN. D. H. Marsh. J. L. Stewart and H. Wrench at NOSC. for the support of I). C Hoffman and C. D. Smith at NAVSFA and M. L. Parker. Jr. and I. L. Smietan at NAVFL1.X. and for the assistance of 11. B. Stone in OP c)87 and CDR D. R. Moyer USN. CDR G. B. Lowe USN and LCDR A. J. Dietzler USN at NMAT 03.

(U) J. I". Fish and G, A. Turton were the technical reviewers. W. A. Friedl assisted with technical editing and final report organization.

Released by JM STALLARD, Mead Bioacoustics & Bionics Division

Under authority of HO PORTER, Head Biosciences Department

SECRET (This page is unckissined)

i■iä*i2S■'@5E^■"'. " ' qurecv**

CTh1l paqe fs CONFIDENTIAL SECR[i

<;ECU l~IT Y CL ASSI FICATION OF T HI S P AGE (When Dele Enter ed)

REPORT DOCUMENTATION PAGE READ INSTRUCTIONS BEFORE COMPLETING FORM

(jj) ~O~~jrR-422 r tA;OVT ACCESSION NO. 3. REC IP I ENT~S CATALOG NUMBER

-Co'J:l 3y::;--~ T ITL E (.nd Subtitle)

' 5 . T YP E OF REP ORT Ill P E RIOD COVERED --- - - -·-c PROJECT COMBO: REVIEW AND RECOMMENDATIONS (U) ~ 1'----Final VJu'1 71 - Sep 78__; 1 ---- , I -- 1'- . MtH~<IRG. R E P ORT NUMB ER --

' 7 . AU T HOR(t) . I 8 . _£.Q!:!!.ftACL~T NUMB ER( •)

I f .- -- /( I

WC CUMMINGS, RL SEELEY, PO THOMPSON, RA JOHNSON l,lf- . --~ I • I . . '· PE R FO RMING O RG ANI Z ATI O N NAM E AND ADORE!; I 10.~~~r?~l,fi(E~E~T~JECT, TA~r: Naval Ocean Systems Center, Hawaii Laboratory

_61721N:_f21221 SF2128Hn~ ( P. 0 . Box 997 Kailua, Hawaii 96734

XF21222701 fNSEimOm-:9-r,---" N&L&X 310

11 . CONTROL L I NG OFFI C E NAME AN D ADDRESS ---- 12. REPORT D A T E

Naval Sea Systems Command (NSEA 63R33) ___./, 'JL. July 1980

Department of the Navy I 13. NUMBER OF PAGES

Wamin~on,D.C.20362 44 14. MONITORI NG A G ENCY N AM E Ill ADD RESS(II diffe rent from Controllln' Office) 15. SECURITY CLASS. (of thl t report)

/" ) vf ( ' . n' I./ . 1 \ I SECRET I

I ; .j . '· / ·, 15e. DEC L ASSIF1c;}i._01;DOW::£W+Nf SCHEDULE ·~ .

Review _on : 16. DISTRIBUTION STATEME NT (of thlt R eport)

17 . DISTRIBU TIO N STATEMEN T (of the e bttrtcl tnle red In Bloclt 20, It diffe rent from R eport)

18. SU PPLEMENTARY NOTES

19. KEY WO RDS (Continue on rever•• e lde If nee•••., M d Ide ntity by block number)

Bandwidth Large Whales Pinnipeds Source Levels Bioacoustics Low-Data Rate Signal-to-noise Ratio Sunmarine Communications Marine Mammals Signal Synthesis Transducers Covert Measure Fonnat Small Whales Waterfall Display lACS ~"" Pilot Whale Sonogram Whistles 20. ABSTR"fT (Cont inue on re ver,. t ide II n e cu,.ry lind Identity by blocl< nu..,ber)

(C) PrcJeCt COM BO produced the elements of a covert communications system that conveys coded messages using selected marine animal sounds in a natural sound environment. Project progress includes development of a coding tee!- .ique, a ;omputer program for elemental synthesis, and a recognizer/decoder instrument that is compatible with fleet low-data-rate communications equipment. The report also discusses properties of available t ransduccrs and characteristics of marine animal sounds as possible elements of covert communication systems.

I

-DO FORM

I J AN 73 1473 EDITION oF 1 No v u 1s oas~ page LCCONFlDENTIAL SI N 01 02-L F -01 4-6601 ---~~~~~:-:-::-=:-::7.~:-:::777:-::::-::::-=:--:--;:-:::-:--::-

SECURITY CLASSI P'ICATION OF THIS P AGE (lt'he n D ete •ntered)

SECRET

FOREWORD (U)

(C) Project COMBO* uses coded marine animal sounds fo r a covert underwater communications system with ranges to 50 nmi. The Project COMBO concept o riginated in 1959 at the Navy Electronics Laboratory. Collection and analysis of sounds suitable fo r ul­timate use in Project COMBO began in 1965. We recorded marine animal sounds. mea~ured sound source levels. observed animal behavio r and initiated acoustic analyses. In I <no. the Defense Advan~ed Research Projects Agency ( DARPA) spo nsored analysis and d upli catio n of sonogram patte rns o f marine mammal sounds we had collected and the Naval Shi p Systems Command (NA VSHIPS) sponsored the Project COMBO communications applica­tion of the sounds.

(S) During Project COMBO, we develo ped a coding techniq ue that uses temporal and frequency patterns to convey messages. We also develo ped a recognizer/decoder instru­ment. the COMBO Signal Recognizer (CSR). and produced the improved CSR-11 fro m it. We developed metho ds fo r compute r sound synthesis and investigated communica tio ns ap­plicat io ns o f entire sound sequences. Labo rato ry and sea tests used six demo nstrat ion mes­sages based o n coded pilo t whale sounds: messages were; rece ived correct!}' underwate r o ut to 50 nmi.

(C) The Project COMBO effo rt was coo rd inated with the Low Data Rate. Quick Response Program of the Integrated AcoustiC Communi~.:a tion s System plan. In this report we review Project COMBO and re late the project concept to advanced develo pment and flee t use. We present background info rma tio n and d iscuss the Project COMBO p lan.

I AcceSJsion PloT

NTIS GRu\.1.:1 i IIDC T.AB UnannoWlced Jus t ifico tio:J

*(U) COMBO is not an acronym

By

DistributicnL

Ava ilabil i ty Code s

Avail aud /01· Diet., spec ial

q

OTIC

SELECTE~ AUG 29 1990u

0

SECRET

K t itrtH t ' Uti}

SECRET

SUMMARY (U)

OBJECfiVE (U)

(C) Develop a covert communications system with a bioacoustical message format using animal sounds as message elements that are recogmzable with ex isting flee t signal pro­cessing equipment.

RESULTS (U)

1. (C) A coding technique, a computer program for elemental sound synthesis and a recognizer/decoder instrument that is compatible with fleet low-data-rate communications equipment were developed.

2. (C) Pilot whale sounds were synthesized and coded to convey six demonstra­tion messages. Messages were received correctly underwater out to 50 nmi.

RECOMMENDATIONS (U)

1. (C) Coordinate progress and resu lts from Project COMBO into the Integrated Acoustic Communication System Plan, especially the Low Data Rate/Quick Response Program.

2. (C) Develop techniques and equipment to synthesize large whale sound!' and small whale screams and to process wideband clicks and frequency-swept signals.

3. (C) Modify the COMBO signal rece:ver system to utilize more message frequen­cies. Increase the CSR processing bit rate to 0.0~ bits/sec. Use automatic gain control and a wider filter pass band on the receiver system. Adapt parallel coding/decoding to the CSR ; update the input pattern every 50 msec.

4. (C) Perform an error analysis on the triplerredundant message format.

5. (C) Develop and maintain a reference collection of recorded marine animal sounds. Collect information on th t! distribution and habits of bioacoustic source animals.

2

SECRET

■K

" 'fff*?vnir>niwi i

PWRfWfWP^SB^^S^IBSIP^S^^S^P psw^P^pspii

INTRODUCTION . , . page 7

BACKGROUND ... 8

CONCIiPT ... 8 AUTIli-NTICITY , . . UTILITY. . . 8

SECRET

CONTENTS (U)

PROGRhSS 8

HARDWARH CONFIGURATION SIGNAL SKCURITY . . . 9 LNVIRONMHNTAL FACTORS . . SYSTFM COMPONFNTS. . . 10

10

SIGNALS 10

LARGF VVIIALFS ...11 SMALL WIIALFS ... 13 PINNIPFDS . . . 15

SIGNAL DITLCTION ... lo

OPI.RATING I I ATURI.S ... 16 SI-ATI SIS ... 19 RI-COMMI-NDATIONS IOR DFVLLOPMFNT 25

SIGNAL SYNTIIFSIS ... 25

ANALYSIS OI; MARINI-: MAMMAL SOUNDS ... 25 TLCIINIOUF Ol- COMPUTLR SYNTIIFSIS ... 26

Murinc Mammal "Screams" ... 26

"RANSDUCFRS . 3

BACKGROUND.. . 31 i-LFFT PROJi.CrOR.S ... 31 OTIIFR PROJFCTORS. . . 32 LOW-I-RIOUI^'CY PROJICTORS ... 37 FLITT RIXIIVINCi TRANSDUCFRS ... 37

SECRET

:_, ■

-"rr"" ■;--■:"-—. i^to'&iti|g>saf'J^**^^^«^-'-^a^'^^'1*^^ ■T«. - ^..'

SECRET

DL-PLOYMHNT SHCURITY ... 37

TRANSDUCl-RS ... 37 SOURCIi LliVIiLS ... 37 SIGNAL SYNTHHSIS ... 39 SIGNAL OCCURRLNCi: ... 40

LLHHTIMPACT ... 40

RI-:I:I-:RLNCI-:S ... 41

RHLATHD DOCUMLNTS ... 44

ILLUSTRATIONS (U)

1. COMBO system concept . . . page 7

2. Sonograms of sound with COMBO applications... 12

3. COMBO Signal Recognizer. CSR-II, elements and configuration ... 16

4. Sonogram of natural pilot whale sounds used in Project COMBO ... 17

5. Received level of COMBO signals measured in CSR-I tests ... 22

(•>. Time-frequency waterfall plot of a simulated killer whale sound ... 27

7. Time-frequency waterfall plot of a natural killer whale sound ... 28

8. Waterfall display of computer-synthesized, pilot whale sinusoidal whistles ... 29

9. Kxpanded waterfall display of the middle set of computer-synthesized whistles... 30

10. Sonograms of blue whale sounds re-recorded directly from the playback tape recorder . . . 34

1 I. Waterfall displays of killer whale spectra ... 35

I 2. Response of IIX-182A projector in series with a 1 10-to-200 llz band rejection filter . . . 36

■

SECRET

.<«*Ii«A£lvl. kiiMiämMlMiäm^M^^~~^-

m*

ip«?gip!!aii8qp^^^

SECRET

•

TABLES (U)

1. Signal-to-noise ratios (dB) from Project COMBO transmission tests. September lc)73 ... 21

2. Signal-to-noise summary. April lc)75 ... 23

3. Received signal-to-noise ratios/I I/, in dB . . . 24

4. Fleet sonars considered for COMBO-likc applications . . . 33

5. Non-fleet projectors considered for COMBO-like applications . . . 33

(i. Low-frequency projectors surveyed for possible COMBO-like applications

7. Fleet receivers suitable for possible COMBO-like applications. . . 3l)

38

5/6

SECRET

. ä ^ffl iMiiliii^^iffii^Wftrf'rtri

mUi*

W&w&igmt&iB^&mmmgm'n pjfpispswsfi^

SECRET

INTRODUCTION (U)

((') Project COMBO was u four-year research ami development effort to produce a covert-ln-disguise underwater comnumications system using coded underwater sounds of marine animals. The coded sounds are generated at one platform and received and decoded at another (figure 1). Marine animal sounds are common and prominent components of the natural underwater sonic environment. Military communications based on natural animal sounds confound detection by even informed enemy surveillance because the messages are but a small portion of the total biological chorus. As part of Project COMBO we character- ized marine animal sounds, devised a distortion-free coding scheme, identified signal process- ing techniques, developed a method to project, detect and decode signals, and synthesized sounds with computer algorithms. Project development can be traced in references 1-4.

COMBO RECEIVER

IN THE FLEET TO BE DEVELOPED

r LDR/QRP

COMBO SIGNAL RECOGNIZER

DECODED 1 MESSAGES SPECTRAL

PROCESSING

SHIP SONAR ( '

MESSAGE FORMAT

SYNTHESIZED ANIMALSOUNDS

(CONFIDENTIAL)

(U) Figure I. COMBO system concept.

(S) Project COMBO relates to the Integrated Acoustic Communications System (IACS) Plan (reference 5). especially the Low Data Rate. Quick Response Program (LI)R/ QRP) (reference 6). The project provides a system for submarines to communicate LDR messages up to a 50 nmi range without revealing their location. Such LDR messages may be useful in Identification Friend or Foe (IFF) functions, during escort or to coordinate sup- port, trail, barrier or fire control activities. The system also provides submarines a covert means to transmit sweep finds to surface units and to contact other submarines prior to conventional communication. Use of natural underwater sounds makes the COMBO system a relatively covert solution to FDR communication tasks.

SECRET

■•■•.':. ■

^MBMBMHHW

Hi

tjpgtigi^.j»wifeam^^

SECRET

BACKGROUND (U)

CONCEPT (U)

(Si The project's objective is to develop ;i covert communicalions system with u bioacoustic message format using animal sounds as message elements. Messages can be received with existing signal processing equipment in the fleet. The system will confound interceptors by presenting an overwhelming task of analyzing all marine mammal sounds encountered in order to identify possible message elements.

AUTHENTICITY (U)

((') fidelity of transmitted soundsdepends on signal-to-noise ratio and transmission bandwidth. Sounds recorded for communications applications must have a high signal-lo- noise ratio or undergo subsequent noise removal. Transmitters must maintain sufficient bandwidth to preserve the sounds' natural qualities.

(U) Improved fidelity factors result from better recording techniques, improved sig- nal processing and synthesis and advanced transducer development.

((') The transmissions' temporal and geographic consistency depends on detailed knowledge of distribution and habits of bioacoustic source animals.

UTILITY (U)

(S) TheCOMBO communication systemutilizes receivingand processing equipment, such as the W(,)('-5, already in fleet use. System implementation requires message recogni- tion equipment, transducer optimization and signal generation capability. Tape recordings or real time signal-synthesizing computers can generate the signals.

PROGRESS (U)

HARDWARE CONFIGURATION (U)

(U) The configuration of the Project COM BO equipment is described in this section.

(S) The waveform generatoi is a magnetic tape recorder that plays cassettes contain- ing coded analog recordings of pilot whale phonations in natural choruses.

( U) A specially-designed 1-kW amplifier drives selectable-impedance transducers at lull power in the frequency band from 20 11/. to 10 kHz.

(S) Acoustic sources included a free-flooded cylindrical transducer to reproduce pilot whale sounds. Critical characteristics of the transducer were broadband response, high

SECRET

■ . .

; .....■.■■.-....■.■..,... -.. .T..^_---.-

MUU HÜ

J!*S#?."i1-''f#S««!SW •^^^i^iii^^*i;i^faf^'"^r''

SECRET

power capacity, depth insensitivity, small size and light weight. Parametric transducers were also used in test pools and at sea in cooperation with personnel from the Naval Underwater Systems Center.

(S) Channels up to50nmi long were established during sea tests. Coded pilot whale sounds in the 1.5-to 4.()-kH/ band were projected at levels up to 200 dB re Ipl'a at I m. Pro- jector and receiver depths varied from near surface to I 2?. m. The water column was well- mixed.

(C) Acoustic sensors are cable-connected, calibrated hydrophones or military sono- buoys. Modified slot buoys provide an ejeclable transceiver to extend the sjrface or airborne communication range without increased transmitted acoustic powei.

(S) The coiling employs frequency shift keying (1'SK) using natural tonals inherent in the signals within a chorus of animal sounds. Triple-redundant transmission is used to overcome interference. A message is repeated three times in three different ways in a %- second interval. The message format is deterministic and provides a choice from six different messages with assigned meaning.

(S) The message detector is tne modified COMBO Signal Recognizer. CSR-I1. The CSR-II has an opeiator-sel threshold calibration, error-correcting logic and external circuit testers. The instrument shifts automatically to account for missing up to two of the three signals in the first signal set. The CSR-II also accounts for missing signals and for random signal occurrence within a specific time interval. The instrument recognizes multiple qualify- ing signals preceding messages and extracts the messages from the transmissions.

(C) The signal band is not limited in this plan. Analog voltages input to the CSR-II may represent any preprocessing algorithm and thus do not confine the encoding-decoding method to tonal preprocessing. The decoding scheme accommodates broadband animal sounds with rapid I'M sweeps and rich harmonic structure. The CSR-II uses standard spectral processing instruments related to tue AN/BQR-20/22/23 equipment and to the \\'0C-5 (LDR/QRP) communication system (reference b).

SIGNAL SECURITY (U)

(S) Covert use of the COMBO communication scheme requires computer synthesis of animal sounds. Marine animal signals recorded in nature are inseparably imbued with noise and cannot be used directly as a signal waveform. Non-uniform spatial distributions in the noise field projected with natural animal sounds would destroy the communications' covertness. The code format requires a unique, ever-changing sequence of background sounds to maintain the code's natural quality. In the COMBO scheme, all sounds would be synthesized.

(S) The COMBO plan provides six possible messages which can be assigned arbitrary meaning that can be changed at any time. A '"code of the day" is possible with the plan. The number of messages can be increased by using different equipment configurations and other signal processing techniques. The CSR-II bit rate is about 0.035 bits/sec and can be doubled to 0,07 bits/sec by using sweep limit control (reference 7).

SECRET

il^apjpw'^^-1^^

SECRET

(S) Transducers and amplifiers used in tlie COMBO scheme must be compatible with fleet equipment. We surveyed transducers in fleet use and other low-frequency trans- ducers that could be used in the COMBO system (reference 8. ')). With minor modifications, fleet transducers are marginally acceptable for COMBO transmissions in certain frequency bands. For example, the UOC-2 transducer is the best fleet transducer for pilot whale sounds. The fleet transducers surveyed have limited low-frequency capability. The extended low-frequency response of the other transducers surveyed would enable transmission of more diverse signals over greater ranges.

ENVIRONMENTAL FACTORS (U)

(C) Covertness of COMBO communication depends upon signal fidelity and the naturalness of the animal sounds. Ihe sounds' physical characteristics, such as source level and tonal components, must be proper. Preferred sounds are from marine mammals, such as pilot whales or killer whales, with a cosmopolitan occurrence. The sounds must be used in an appropriate context. For instance, mating signals transmitted in a non-mating season could arouse the suspicion of a knowledgeable monitor. The sounds from a single species may vary regionally. Thus, well-known dialects should not be used out of context. These speculations presume a knowledgeable interceptor.

(C) The COMBO plan requires natural sounds to provide a "cover" for communica- tion signals. In some areas certain species of whales have become scarce through overharvest- ing. but conservation measures have allowed other populations to increase. For example, the southern right whale is nearly extinct but the grey whale population is increasing. Population size is not necessarily directly related to acoustic presence. For example, the humpback whale population is small but the species is particularly soniferous. Similarly, the killer whale (Orciiuis omi). is frequently heard even though its population is depleted. Thus, the use of COMBO sounds in the ocean requires relatively few whales, and conservation measures should guarantee populations at least as large as present ones.

SYSTEM COMPONENTS (U)

IX (U) The following sections present six aspects of the COM BO system: Signals. Signal

clion (CSR), Signal Synthesis, Transducers. Deployment Security, and Fleet Impact.

SIGNALS (U)

(C) Three categories of marine mammals produce sounds with potential COMBO application: large whales, small whales ( :.eluding porpoises), and pinnipeds (seals and sea lions). The listed sounds are common and contain tonal components. Within the whale cate- gories, animals are discussed in order of probable potential applicability to the COMBO communications scheme from most to least likely. Pinnipeds are not so ordered, because most occur in non-overlapping geographical areas, and nearly all are very vocal.

SECRET

pppjlfpjplp^

V^jVfre^WHJW»—iBi WJiJi»WMa7«Er».mi.-. » ■y-v>fi-(<.-^>lTnt^ 'f

SECRET

(C) Large whales produce low-frequency sounds with high source levels. The low- frequency sounds require special transducers. The sounds could be used in a COMBO com- munications scheme with an intended range beyond 50 nmi or for shorter distances under very noisy conditions. Sounds from small whales and from pinnipeds are higher in frequency and therefore easier to reproduce with transducers. In a COMBO communications scheme, sounds from small whales and pinnipeds are limited to ranges below 50 nmi unless source levels are at least 20 dB over natural. The most suitable large whale sounds for COMBO are from humpback, right, and finback and from pilot, killer and beluga among the small whales (see figure 2).

LARGE WHALES (U)

(U) The humpback whale. Mcgaptcni iionwcingliac. produces a variety of under- water calls, including howls, moans, grunts, cries, yelps, and low-frequency pulses. Funda- mental components cover frequencies from 20 to 2000 Hz. with harmonics as high as 5000 Hz. Sound duration varies from 0.2 sec to over 5 sec, and source level may exceed 180 dB re I ^Pa. Humpback whale phonations are often frequency modulated.

(U) When not migrating, humpback whales occur coastally and are distributed in all oceans. The whales occur in lower latitudes during winter and spring and in higher latitudes, including polar seas, during summer. In the breeding areas around Bermuda, the West Indies, Hawaii, and New Zealand, humpback whale sounds are grouped into "songs" that last up to 30 min, contain sections of repetitive, stereotyped phrases, and may be repeated for hours (reference 10).

(U) The right whale, I'.uhalacua gknialis. produces sounds similar to humpback whale sounds: moans, grunts, and bellows intermingled with pulses and rasping sounds (reference I I ). Line components vary between 50 and 1500 Hz, with durations from 0.2 to at least 4 sec. Most of the energy in the phonations occurs below 500 Hz. Sound source level is over 180 dB re I juPa. Right whales are widely distributed in small numbers, in coastal waters, in latitudes poleward of 25 degrees. Right whales migrate to lower latitudes during the winter.

(U) The finback whale, Balaciioptcm physalus. produces pulses and moans in the range from 16 to 100 Hz, with less than I sec duration. Some finback whale sounds are fre- quency modulated (references 12, 13). Finback whales produce pulse trains centered at 20 Hz, with regular inter-pulse intervals of a few seconds (reference 14). The 20-Hz pulses occur in doublets with intra- and inter-doublet intervals of 10 to 25 sec. The pulse trains continue for hours, and constitute well-known interferences at SOSUS stations. Another finback whale sound is a 3-sec moan composed of a (18-llz, 1.8-sec component followed by a 34-llz component varying in length up to 1.8 sec. Source levels of finback whale sounds range up to 180 dB re 1 pPa (reference 13). The finback whale occurs in all oceans, espe- cially in polar seas during summer months.

(C) Off Chile, the blue whale, Ihilaeiiopwra innsciilus, produces 37-sec sequences of three moans. The interval between the first two moans varies up to 2 sec, and between the second and third from 1.5 to 3.5 sec. A 390-Hz tonal pulse occurs just before die last moan.

11

SECRET

: i ■ a"—- ■--'^ ^'-" -■-:- -"-■ ■ ""^-^-~ —- - ■

V~ ^awfe^." y.

mmmm^m m^^w^w^^^sm^mim1^'

4

3

2

N I

> o z UJ D a DC

3

2

1

0 , 0

5

3

2

1

o-y

x\

-> V

14

12

10

8

6

4

2

0

(S) I'iguro 2 D. clicks

0.1

'i ., HuJi.i.,

.t..u.:.: .„„.I. L; 'h

0.2 0.3 0^6 0.4 0.5 0.6 0.7 0.8 TIME, sec

(SI;CRI;T)

Sonogiams of sound will) COMUO applications. A. beluga. B. killer whale, C. pilot whale, of each, respectively. L-ffeclive analyzing filter bandwidth, 20 11/ (A.B.C.) 800 Hz (D).

12

SECRET

•■ ■■ ^-jV^-K. ■-■''-

SECRET

Tlie blue whale moans average 188-dl3 source level, and are amplitude-modulated al 3.1-) and 7.8/sec. The strongest components occur at 26 H^. and others extend to about 200 \U Sequences typically repeat every 1.7 min. but the silent interval may he as long as 3.3 min (reference 1 5). In the eastern North Pacific, blue whales produce similar moans of two t\ pev with the most energy near 20 Hz. Blue whales occur in all oceans. They concentrate at high latitudes during summer months and off California in autumn. The long. 20-11/ phonalion^ arc well-known at SOSUS stations from southern California to the Bering Sea (reference I 2 i

(C) Thesperm whale. Pliysctcr ca lotion, produces trains of short clicks. At slo\\ re- petition rates, the sperm whale click trains are called "carpenter fish" sounds by Nav\ sonar- men. The clicks have a source level as high as 1 75 dB and most energy between 200 and 10.000 Hz. The click repetition rate ranges from I to 100 per second. Short click trains maj have distinct rhythmic patterns, and may be repeated (references 16, 17). Sperm whales are usually found in groups that produce clicks in chorus. The sperm whale is distributed in all oceans, especially in tropical and sub-tropical waters near the Azores. Peru, the Galapagos, and off Central California. Because of their ubiquity, high source level and common occui rence, sperm whale sounds are familiar to sonannen of all nationalities and are excellent sounds for COMBO application.

(S) The click structure of sperm whale signals is incompatible with present COMBO methods of narrow frequency band processing and coding and requires instead, wideband threshold detection and simple temporal coding patterns. Nevertheless, sperm whale sounds are ideal for biologies communication applications.

SMALL WHALES (U)

(C) The pilot whale. (Uobicephalu nwlaena. (1. sca/nnioni, and (i. unun)rliviului produces a variety of sounds, including clicks, whistles, squeals, warbles, and grunts (lelei ences 18. 1^ and 20). Some pilot whale whistles are pine-tone, and others are rkh in har- monics. A whistle may remain at a single frequency for up to a second, or sweep over sew ul kHz in the same time. Whistle and squeal energy is distributed from 500 11/ to 15 kll/ it ;i source level of al least I 75 dB. A large proportion of whistles and squeals are hclou > I.I!/ Hxcept for polar seas, pilot whales occur in every large body of marine water. Pilot whale sounds are suitable for COM BO transmissions of 25 to 50 nmi.

(Ü) in the eastern North Pacific, killer whales,O/r/m/.v o/rtf. emit nanowband cluks of 10 to 25 msec duration (reference 2 I ). The fundamental resonant click frequenc\ is in the 250- to 500-11/. range. The clicks occur in short bursts of 10 to 1 5. or in much lungci. high-repelilion-rate, raucous screams. Screams occur in two pails, one at a repetition rale about 0.5 kHz, and theolherat about 2 kHz. Bolh scream pails have many liannonics. In Antarctic waters, killer whales emit click trains similar to those of northern animals (icl'ci ence 22). Antarctic killer whales do not produce the stereotyped screams characteristic o, their northern coimlerparls. Killer whales occur worldwide, especially near coasts at highei latitudes and in polar regions.

13

SECRET

. ttelaa&a&ite^äiiaääai^ ... 2«äSa ^--1—,-^^-^":^,'i''f'|fÜ^if'i''1'' ',';'?g' ^

SECRET

( LI) Belugas. Delphiiiapienis Icuvas. may be tlie most soniterous sea mammals (rel- eivi'ce 23). Belugas emit a variety of sounds, including whistles, warbles, chirps, clicks, buz/cs. cries and moans at frequencies up to 100 kil/.. Beluga sounds typically contain energy in narrow frequency bands. Belugas occur mainly in the Arctic Ocean. The whales occur seasonally as far south as 50ÜN. and are abundant in the Davis Stiaits. Hudson Bay, (lulf of St. Lawrence, and along the coasts of Alaska, Norway, and the Soviet Union. Belugas do not occur in the southern hemisphere.

(LJ) False killer whales, Psemlorca crassiücns. produce whistles in the 2.5- to .S-kll/ frequency band, with some energy up to 8 kll/. False killer whales also emit clicks and buzzes (reference 20). Fxcept in polar regions, false killer whales occur in open seas through- out the world.

(U) The bottlenose porpoise, Ttirsiops inuicalus. clicks at frequencies up to 200 kHz and whistles in narrow bands between 4 and I 5 kHz (reference 24). Bottlenose porpoise click trains can sound like barks, yelps, squeaks, squawks, or rusty gates. The bottlenose porpoise occurs widely in the Atlantic and Indian Oceans and is common along the coasts of the Americas and Furope. Related species ( T. gilli and T. niitianu) occur in warmer parts of the Pacific, particularly in the China Sea and in coastal American waters.

( LJ) Atlantic spotted porpoises, Sienella plagioilon, produce broadband clicks and whistles in the 5- to 15-kHz frequency range (reference 20). The spotted porpoise occurs along the Atlantic coasts of Furope and America, and in the Indian Ocean. A related species, Sienella gmffimmi. also called "'spotted dolphin", occurs in the eastern tropical Pacific.

(LI) Pacific white-sided porpoises, Lagc'iiorliyiiclius ohlic/iiidens, may occur in herds of up to 2000 animals. Pacific white-sided porpoises produce broadband clicks and whistles. The whistle may occur in pulse trains or as a raucous bleat with independent dual compo- nents (reference 25). Distinct line components extend as low as 1.5 kHz in Pacific white-sided porpoise sounds. Pacific white-sided porpoise occur from Baja California to the Aleutian Islands and in Japanese waters. A related species, Lagenoiiiyiicliiis acutus. occurs in colder waters of the North Atlantic, and Lagciiorhyiuliiis ohscunis. occurs in the South Pacific ami South Atlantic oceans. The different species probably produ e distinct sounds.

(U) The saddleback porpoise, Dclphinus delplus. produces broadband clicks and narrowband whistles that have been recorded in the 10- to 20-kllz band and as low as 5 kll/. (reference IH). The saddleback porpoise occurs in large groups in warmer waters of the Atlantic, Indian, ami Pacific oceans.

(S) All small whales produce some sustained narrowband signals that could be used in a C OMHO commuivi ition scheme for ranges less than 10 nmi. The sounds' high frequen- :ies are subject to high attenuation. Small whale sounds might have utility for short-range commr.nication at low levels to limit the chances of distant interception.

14

SECRET

. ■ .' i : ■ _.., i ■■..---'■-

SECRET

PINNIPEDS (U)

(U) The Wedclell Wdl.Leptonyc/ioies wedüclli. emits narrowband pulses at repetition rates from above 140 sec to one in several seconds (reference 26). The sounds' frequency range is from 50 to 10.000 Hz. The pulse trains typically start at I to 2 kHz with a last repetition rate and end at about 0.1 kHz with slow repetition. Individual pulses in a train are frequency modulated from 2 kHz at 10 kHz to 50 Hz at 1ÜÜ Hz. Weddell seals also produce howls, with fundamentals ranging from 50 to 4000 Hz and with strong harmonic structure, and ringing sounds in the I - to 3-kHz range that last up to 10 sec. Weddell seals occur in the southern hemisphere around the Antarctic continent and its adjacent pack ice. Weddell seals migrate as far as 30oS to South America, to Australia, and to New Zealand.

(U) The bearded seal, lirigiiathus harhanis. emits a whistle with a tremolo charac- teristic caused by fast frequency modulation of a few hertz. The whistle is modulated further over several hundred Hertz at rates of about I/sec to one in several seconds as the whistle drops 3 to 4 kHz to 0.2 kHz over a period of a minute (reference 27). The bearded seal is circumpolar in waters of the Arctic Ocean and the adjacent Atlantic and Pacific regions.

(U) The ringed seal. Ptisa hispichi. produces a number of sounds, including a steady discrete-component, downward frequency sweep imbued with harmonics (reference 28). The sounds last from about 1 to 4 sec, range from 5 kHz to 0.2 kHz and have at least four strong harmonics which may extend above 10 kHz. This seal also makes a broadband ringing sound in the frequency range from below 2 kHz to about 6 kHz, with durations up to 15 sec. The ringed seal occurs widely in the Arctic Ocean and on the north coasts of Europe. Greenland. North America and the Soviet Union.

(U) The ribbon seal, Phoca (llistriophoca) fasciala. produces two principal under- water sounds coincident with spring reproductive activities: an intense 1- to 5-sec downward frequency sweep in the 7- to 0.1-kHz region and a 0- to 5-kliz breathing-type sound of al- most 1 sec duration but at 20 to 25 dB lower in pressure level (reference 2'-)). This seal is distributed from the northeast Bering Sea to the Sea of Okhotsk and into the northern Sea of Japan. Ribbon seals also occur on the coasts of Korea and northern Japan.

(U) The California sea lion. Zaloplms calijornianus. emits a variety of underwater sounds including barks, groans, growls, and bleats (references 28 and 30). The barks' main components occur between 0.5 and I kliz. Harmonics extend up to about 5 kHz. Most of the other California sea lion sounds occur in the frequency range of 0.1 to 1 kHz. The Cali- fornia sea lion occurs along the California coast, Baja California, the Calapagos Islands and the coast of Japan. Bulls occur as far north as Washington in the wintertime.

(U) Pinniped sounds usually occur within lOnmi of coasts, ice packs or drifting ice.

15

SECRET

—: _____ ___ . : ^ -.. ' /C:'. ■.... .•■-■*-in,,,- ' ' \^J '■':^Mi«i^M&^äii>^^it^

■v^^tjUfoiM^WMsr.-^ \ _^ . ' ■ ■ -'m

SECRET

SIGNAL DETECTION (U)

(C) The COMBO Signal Recognizer. Second Generation. CSR-1I (figure 1). was designed and assembled, and then tested both in tiie laboratory and at sea. The signal recog- nition device can decode properly even if seven of the nine signals are lost. Information extraction logic uses read-only memory units ( ROMs). The CSR-II can fill in missing infor- mation elements, average and weigh results, and display the best possible output. The word '■BAD" appears in light-emitting diode (LHD) displays when data are insufficient or do not conform to one of six message formats. The CSR-li accommodates broadband-processed bioacoustic signals. That is. given a lechnk]ue to process sweeping signals, the CSR-II type decoding would also suffice for broadband data.

TEST DATA

INPUTS -/)tf\-

'OUTPUT DISPLAY FORMAT

TIME FRAMES

I

GROUP II

III

(UNCLASSiril-D)

(U) Tigurc 3. COMBO Signal Recognizer, CSK-II, elements and eonfiguration,

1 2 3

4 5 6

7 8 9

OPERATING FEATURES (U)

(S) The CSR-II signal format is three mutually exclusive elements, each transmitted three limes at different frequencies. The CSR-li produces six possible messages, using nine different transmitted frequencies. To compensate for frequency shifts during message trans- mission, the receiver accepts frequency "bands'" centered on the possible Iransmission fre- quencies. The nine frequency bands are adjusted to match natural components in nine dif- ferent pilot whale sounds, preselected according to the code of the day. The frequency bands are randomly arranged between 2.5 and 3.5 kll/ and the CSR-II recognizes nine animal sounds spaced throughout a "natural" chorus of phonations ( figure 4).

16

SECRET

A _. 1 -..:■■.:■ ■,' ■ ^ '. .CS.:..;.,iU.\\

lLIllllJJIlJI!U^fllWUiWIJl;il!lll.il!l »pi|fflH«B^ippp*p?^äips«pi»SP^ .,

SECRET

>

)

if. »

, ir^»'

i,..».

^ i

in

y

Ü ^ -

Ü ^ ^r

s rz -

— C j_

r ~ o ^■. ~ d

V— '-> •.

5" /: r> u E <*

Ü ■■.;

< ■_' OJ

(1) c 0 (A P £X 1

E UJ UJ E '-» 2

Q -a E c

1- X s. ~3

ra

o — J

s. (J >. C

..,.. ^ S.

o v.

a ^Z

"n o

ra ~ -n •-• o c •o E IS,

73

Zh or — z: -n T3 " o ^ y. r: - t y

r.J

o _^ -3

I 00

I o

u

(ZW) A0N3nO3dd

17

SECRET

_ ifit .jfji t^ ■..- MftM. ^tt^HHÜ tl»- ^ •,*.»*.^.-'K«t¥

\.. ••• , ■ ■-■■•

"i

SECRET

(S) Pre-recognition processing in the CSR-II is tonal analysis with a real-time spectrum analyzer tiiat is sweep-limited at 2.45 to 4 5 kHz and with an ensemble ave-ager that uses eight averages. The analyzer is a Spectral Dynamics SI) 301 and is the same analyzer used in the WQC-5 (LDR/QRI'I fleet communication system. The averager is a Spectral Dynamics 30).

(S) The entire spectrum in a selected format of animal sounds is projected into the water. Received data are first bandpassed between 1.25 and 5 kHz. We also use a three- dimensional waterfall display to characterize the received signals and ambient noise and to compare signals to natural sounds as they would appear to sonarmen on the l3(,)R-20 or 22.

((') Hach of the nine frequency bands is represented by an analog voltage from Spectral Dynamics Line Order Trackers that monitor the power spectra within a 36-Hz band about the center frequency. Analog voltages are compared to preset values and separated into a sequence of nine time frames divided into three time groups. Incoming data are dis- played in I.I D displays corresponding to each time frame if the data coincide with the proper time group, lor example, data shown by LI.D's 1.2.3 are obtained during time (Iroup I. Ll-D's4. 5. 6 during (iroup II. and LI.D's 7. 8, 9 during Ciroup 111 (figure 3).

(C) lime frame generation begins when qualifying data in (iroup i are acquired. Additional qualifying inputs are sequentially slroKd into the proper LLD display and stored in memory for subsequent majority/complementary logic analysis. A (iroup II data acqui- sition during time (iroup I indicates that some previous data were missed and an automatic shift occurs to move the display to the proper frame.

((') Timing frames for which no data are acquired display an "H" for empty. Mul- tiple acquisitions during a time frame are mdicaled with a point It) the left of the number displayed in the frame. The first number received is displayed and the multiple acquisition is also stored in memory for later message extraction. After time frame (), ROM passes hori- zontal, complementary logic in sequence on the rows of stored data and attempts to fill sin- gle voids accompanied by two bona fide inputs. Lor example, if the data for (Iroup I were 31-1 in time frames 1. 2 and 3. respectively, a 2 will replace the L during the initial comple- mentary logic pass to produce mutually exclusive message elements. By similar processing, other single voids or multiple acquisitions will be changed, e.g.. 321- to 321 or .22 I to 32 I. Multiple voids remain unchanged and serial multiple acquisitions display empty, e.g., 3.1.3 becomes 3LL. Majority logic scans each column of processed data sequentially and places the majority character for each column in a final best-message matrix, lor example, if the first column contains 3L3 in frames 1. 4. and 7. respectively, a 3 will occur in the first position of the 1 •.•si-message matrix. Similarly. 2.32 in the second column produces a 2 in the second position of the best-message matrix, and LLL in the third column produces an 1 in the third position. A second complementary logic pass on the best message matrix at- tempts to fill voids before results are-displayed in LLDs. If the final message is uncertain be- cause of insufficient or conflicting data, the letters "BAD" are displayed in the LLD instead of a final message. Single-bit. even parity is maintained throughout the logic passes. A point displayed in a final message frame indicates an inoperative ROM. After final message assem- bly, a decoder network activates solenoids or tape recorders to announce messages.

18

SECRET

^"T " : L • ; "•"'""■-.__ ■ - .', _-^ .._ ., .■•ffSHI

SECRET

(S) i wcnly hums of bioacoustic recordings with high signal-to-noise ratios were played into the (SR-li without false alarm. The recordings included many pilot whale

sounds because pilot whale soundsconstituted the coded category of sounds. At times during the test, a single time window would qualify, but 'he message display was always "BAD".

I ower signal recognition or triggering thresholds will increase incorrect message acceptance or "■BAD"' display, but a slow attack and slow release automatic gain control prevents this problem. The tests indicate that the chances of receiving spurious signals of the required

frequencies in the right order and at the right times are infinitesimal. The time conditional codingof the COMBO scheme suppresses false alarms. Actual false alarm rates(lAR) depend upon equipment configurations for specific applications and conditions.

SEA TESTS (U)

(Si The COMBO concepts, systems and processing schemes were tested in the labo- ratory and at sea to improve transmission fidelity and measure performance. Initial tests used

the CSR-1. A total of 76 consecutive repetitions of three different COMBO messages were COITCCIIN recogni/ed and decoded by the 5-windowed CSR-1 in a test at the Transducer 1 valuation Center on Point 1 oma. Sounds were played from a free-flooded cylinder to an omnidirectional hydrophone and a tape recorder at source levels up to 20U dB re 1 /jl'a. Sig- nals were synthesized pilot whale sounds superimposed on a chorus recorded at sea. The CSR-1 is described and discussed in references 31 and 32.

(C) In July l('73. preliminary sea tests occurred off Catalina Island. The sounds,

projector, receiver, processing technique and decoding method were the same as used in the TRANSDIC tests. The projector and receiver were at a depth of 150 ft and at ranges of O.Q 1. Ü. I 3, 0.25. 0.5. I and 2 nmi. The water column was well-mixed. Source levels were

approximately 175 dB re 1 /jPa. All recorded transmissions were decoded in the laboratory without error.

(S) In September 1973. the same COMBO message format was projected at levels to 200 dB re I juPa at sea to ranges of 5. 10. and 20 nmi. Instruments were (he same as for

previous tests. The receiver was a Wilcoxon hydrophone. Mll^OA and the projector was a free-flooded cylinder. ITC 2014. Both instruments were at 250 ft. A specially designed 1-kVV amplifier projected synlhesi/ed pilot wlujje sounds at source levels 25 dB above natural levels. Preprocessing and decoding were the same as in the previous field tests. The five signals were distributed through the frequency band from 4.4 to 6 kll/..

(C) The CSR-1 correctly detected all recorded messages (coded animal sounds) at

the 5- and 10-nmi ranges and c'0 percent of the messages at 20 nmi. Ambient noise condi- tions were near maximum Weiu shipping noise levels. Signal loss occurred at two of the five frequencies transmitted at 20 nmi.

■ ■-.■■ Üifc^ ..^^.Aktoj*.:^. 1-"-

19

SECRET

mm. _a-^- T-—■.•-■;; ; ■;- . - "" V - ' -" .: '. / ^ " .." " --

laiw, •I'm; H'>««»-TWI»;.-

SECRET

(S) Signal-to-noise ratios were detennined on the real-time spectrum analyzer 11 ihic 1 i Al 5 and 10 runi. the noise was nearly all ehorus sounds projected In this system. At 2U nnn. amhienl soa noise above 1 kHz masked the signals. Received level as a function of frequency is shown in figure 5. I'or reference, the source level of the 4.4-kll/ signal was 200 dB re 1 /jl'a. Our measurements fit Urick's attenuation loss model (reference 33).

IS) In April ]i'n5. a ''0-sec tape-recorded sequence of nine rOMBO pilot whale sig- nals was projected at ranges of 2. ^ and Id nmi. The natural tonals occurred in a fiequencv band from 2.5 to 3.5 kil/ as part of a realistic chorus. Instrumentation was the same as for previous tests. Source levels and transducer deptiis are given in table 2.

(Cl four COMBO sequences and a 3-k'lz calibration lone were transmitted at each range.

(S) Recorded sequences were played into the CSR-II system after processing with a real time spee'rum analyzer and line-order trackers. Messages were recognized perfectly at all ranges ;ind signal-to-noise ratios I table 21. The received sign a I-to-noise ratios agree with the prediction for 3 kHz (table 3).

(S) In the CSR-II system, both the bit rate and the signal-to-noise ratio required for message detection arc frequency dependent, lor example, assuming processing gain equi- valenl to that of the CSR-II system and a l'0-percent probability of detection, a Id-dB signal-to-noise ratio for the spectrum level is required for a pilot whale whistle at 3 kli/. that lasts 0.7 sec I or a 20()-ilB source level, the signal is theoretically detectable by CSR-II out In 100 mm under average ambient noise anil propagation conditions (table 3). With the same processing techniques, the signal-to-noise ratio decreases with frequency, but the data rale is reduced because more time is required for recognition at lower frequencies. I'or exam- ple, whale sounds below 100 11/ require a signal-to-noise ratio of —1 dB. but take 100 limes longer to be detected than the 3-kll/ pilot whale signal. Thus, the bit rate is reduced by a '.iclm ol 100.

IS) Tests also were conducted to determine the utility of sonobuoys. especially the SSO-l I A, as receivers. The only limiting factor was radio range, which is line of sight. COMBO signals received and relayed by sonohnoy were processed, recognized and decoded normally. Successful tests of CSR-II were made in the laboratory using data projected up to 3 nmi to sonobuoys 5 to 7 nmi offshore.

|( I Tests with parametric sources were made in cooperation with personnel of the Naval Underwater Systems Center (NUSC). A primary frequency of (>6 kHz was mixed with ( ()MH() sounds and projected in the NUSC quarry at a source level of 1 73 dB. We decoded Ihe recorded signals with the CSR-I with ()()-pcrcent probability of correct detection. Sounds above 1 kHz were received with high fidelity, but 20-11/ to 1-kilz sounds lacked fidelity. In particular, killer whale and pilot whale sounds transmitted well, whereas blue whale and Wedell seal sounds were garbled and degraded by noise.

20

SECRET

. :a^aiaBga!»^-^^-~'-V^mf^-^ir^

^S^^r^mf^ni^^m^mi-'m^M V-^^^p*!* !{mi4m^^^B»SiV^^^i/^^^^m^km»*^ ■^^r^^w'^

SECRET

o r i

>r. ir. o 'r. O "-.' d r--' ö d ri

—

-

— X

p

r - r i iy-, r i oc

3 r i

C

r 1

o

h-

rr, I - (■<, 1/-,

o r j

O X r | —

q

r

E

s.

1-

N.

o -

—

-

c 0 M

C) 3 a J^ **.

■>■

D C u

s w •_i

Zz , o i~ LT. o

ir. CL

r *J c o j;

c X. '" ~ £ c -a

ra 0 c

. „ 0

p s — c C 6IJ r l ^

o —

o t o o O _ o 2

SIJ

u o

*+ 'r. <r, c

X. 3 00 S

21

SECRET

iSJfe. ^k^^ilil&^tittttHMfll *': ' ■ ""

^jifeuiu.^.f.^w.iia..,f'..,^-'^

SECRET

180

170

o

^^ 120 (T3

Q. a. T-

0) k.

m ■o SM^

_l 110 UJ > UJ -J

Q UJ >

in Ü 1ÜÜ

Ul oc

90

5 NM

80

PREDICTED

O OBSERVED

FREQUENCY (kHz) (CONFIDENTIAL)

(Li) Figure 5. Received level of COMBO signals measured in CSR-I tests and predicted levels Ironi Urick (reference 33). l-.ach plotted measurement is an average from 6 COMBO signal transmissions

SECRET

v "'"■ - * ■ ■'- ' "'■■

^|^SfSf<W^!StJö)^|il!|J|^p^4W!SLu 'f1 Ml- ..^7wi^-T..^™^i^y5s»t«!^R»?

\ T^rifi '2*r?zr™ ■

— llj ■a j^ ~ o o t_

f-^

rv u vA Cij n O '^ •)

■ —

1—

C ^^ X -', 'r, i/-.

O C

o i_ c^ o

r3 ?• (U r- -;

'-» o S r3

u c^ j — (»

^ -j , Slj

»r, 1^ ■■J ^ o •" n -^ c

'.j _^

ÜI 'j o <• en. < C

r3 /■ U u-

TJ ,.. L/; ^_

c; _^ '_j

3 i/-, ra n ■—.

r r*-. 3. ♦J

o — ^- r3

n c C i— r3 u. ^ % ir. 02

t- o

0 r- > "" o O

' '-/ y o —,

sz '-» ' o *—' '-» U- ^ O -J n X ™ (IJ *-; —- V. &ij

c = P

^ - r I

23 T3

o N X — r 1 t

c C '3

— — ca l/-, ~ i/;

oo c y

c o -r

TZ _

03

C

n. <

23 x CJ -' LD L«

C Q c , LL u o 3 '~ c s Cl,

'_/ oo

o r i

~ c r u:

fei: ■ fHv

c I- -^ 3 r-

X y. "3 o c~ 1— u

u. ■-»

1—;

n ■y. J ^

'-j ^-

M» ^ a: ■y y,.-

■^ Ei' u ^ 1J

,—. .^. C— ^ o .■J — j—,

o '^— —' ■ —

'J 0 '_J ,i

y. —

—; V ^

23

•— r i

23

SECRET

.•äM*ama>HiMMkkMWM — """"^ --^-ZZ

tsiiiSaäsSÄiaÄS&ÄÖÄiii;

SECRET

— r 1 r' r | i -

X — I r I ir. r i i

X r i

r i r- i- I r l

'■'. t

C? C C i"

r I c

r i r«', r~ r~ oc i/", i/-. i/", — NC r- ri r| r \ r \ r \ ( [ r \ r \ —

XOOOC ~C — ir, CTrfX l/*j w sC 'y* 'r, lA, Tf rv, rr( r j

r I t (••-, r-, c- 00 OO T)- O- O rrt rrl r^, f-, r ] r 1 r 1 r i .— .—

O t-- r- r- '<'. 0~ ", y~. r i tr, -t -)■ '•-, r<-,

t r^, r i x

r i —

X I/-, T r i ~ •r. 3- r 1 ^ . , X X X X r- r^ C T. i/',

'■'", O^ X i/-, w i/-, i/-, ir,

1- X

i- (•i

(U c > o _o

■i o U— t^i

^ '5 c

rj c

»-■ £ ■u E on r3

cc ■N

T3 C o

o '^ C r i c n 0 ^ •a 0 5J C-

*o C3 CJ

.!_> or

-5 c o O

Q D- UJ •^5

CQ c u. T3 c CJ

on C o-. < N 'c _J X c o üfl *-, Z 0 c

(U ~~J

r: Xi E

ii» « 0 c < 6 o Y 0«

■H a c o Wj >

73 'iTi •TD C

1» 1 X 1U ^ rTj

Di

ro ■ — c o 1- i,

O i- jo Td ^

1- => t:

i/-, r i O O

O >r, •+ rrl r \

24

SECRET

,■_ LJ— ■ ' L^_ -

SECRET

((') In June !c)74. COMBO sounds were Iransmitled between ilie submarine Dolphin and a reeeiving ship 0,5 to 4.5 nmi away. All COMBO signals were recognized by the CSR-II. In a long-range test, the signals were too weak to be automatically detected at 50 nmi.

RECOMMENDATIONS FOR DEVELOPMENT (U)

(C) CSK-il is a test system to evaluate message rates, frequency constraints, ranges, and covertness available through the COMBO concept. The CSR-II requires modification to increase message content, processing speed and coding variability to be compatible with fleet systems.

(S) The number of messages can be increased by using more signal frequencies, for example, increasing the number of signal frequencies in a transmission from l) to 21 would increase the number of triple-redundant signals from 3 to 7 and the number of possible mes- sages from h to 5040.

(S) Doppler tolerance can he improved by using wider filter pass bands. Assuming differential speeds of dO knots and pilot whale signals, more than 21 adjacent filters of in- creasing bandwidth can be used, increasing bandwidth is needed to accommodate greater doppler shifts at higher frequencies. Reduced ambient noise levels at the higher frequencies nearly compensate for the increased noise passed by wider filters.

(S) Automatic Cain Control (ACC) would eliminate the manually adjustable gain. Noise varies with frequency: thus a separate ACC should be used for each frequency band.

(S) Transmission time can be reduced by using parallel coding and decoding. The CSR-II system transmits serially one 10-sec window at a time and requires at least c)0 sec of acoustic data for one message. Parallel coding would update an input pattern every 50 msec and require about 20 sec to transmit 21 signals, or one of 5040 possible messages with triple redundancy. Actual transmissions will be longer than 20 sec to add the realism provided In randomi/ed amplitude variations of natural choruses.

SIGNAL SYNTHESIS (U)

ANALYSIS OF MARINE MAMMAL SOUNDS (U)

(S) Synthesis of naturally-occurring sounds depends upon the analysis of recordings from the sea. Proper analysis includes real-time spectral estimation, digital and analog wave- form studies, digital filtering, and instantaneous amplitude and zero-crossing estimation. The sounds" spectrum contains information about the sounds" frequency range and harmonic content. The waveform suggests source generation mechanisms and. hence, possible synthe- sis procedures. Operator-interactive digital filtering isolates individual sounds and enhances signal-to-noise ratios.

25

SECRET

mmgmmmg^mm

TECHNIQUE OF COMPUTER SYNTHESIS (U)

MARINE MAMMAL •'SCREAMS1' (U)

(C) Marine iiKimnuil sounds called "screams" and ■'bu//es"" can he composed of pulses with varying repetilion rales and varying frequency content. Small whales and

killer whales produce such sounds.

(C) A time-lreciuency waterfall plot of a simulated killer whale sound is shown in

figure (i. The basic pulse was generated by low-pass filtering the harmonics ol a rectangular pulse, with a repetition rate that varied from 1000 to 200 Hz. The frequency harmonics of an actual killer whale sound have absolute frequency components that are different from

the simulated scream because the pulse shapes differ (figure 7).

(Cl Marine mammal sounds include long sinusoidal oscillations, called "whistles",

as well as short-duration pulses. Whistle synthesis uses procedures different from pulse

synthesis. Sinusoidal sounds have been synthesized (reference 34); synthesis of pulses ami clicks remains unfinished.

((') Pilot whale whistles are sinusoidal, i.e.. the sound's instantaneous time wave-

form can closely resemble a sine wave. Only the instantaneous amplitude and frequency vary over the duration of the sound. The amplitude is likely to rise smoothly over about 50

msec to an average level, to remain at the level, and to decrease smoothly during the last 50 msec of the sound. The sounds lack transient onsets or offsets that are humanly discernible.

The sounds contain both frequency sweeps am! constant-frequency portions.

(S) Whistle synthesis requires real-time, high-speed computational capability. The COMBO synthesis work used a PDP-l 1/40 minicomputer. High-speed digital-to-analog (D/A) conversion of the generated sounds required a first-in. first-out (I-'II'O) memory-based D'A

converter. Smooth sine waves were generated by limiting the output of the I)/A band with an active filter. The synthesis software program has a self-contained lookup table that con- tains the microstructure of the sound to be synthesized (reference 34).

((') The microstructure table can represent various wave shapes. One table generates

pure sine waves, and others superimpose second and third harmonics. Program execution

speed limits the real-time high-frequency to 7 kHz. The frequencies to be produced are stored separate from the generation program. Thus, a sound can be generated repetitively or as a

sequence. The generation program constructs the pat'ern and fills the I'il-'O memory of the D/A converter with the output samples. The I'll O is clocked at a fixed M-/JSCC interval. In

addition, the onset and offset of the sounds are program controlled. The program is not limited by lil'O memory size. Between clock intervals, new digital samples are loaded into

the I'll'O memory, figures X and l) show time-frequency waterfall displays of computer- synthesized pilot whale whistles.

26

SECRET

,. r-.. ■■■ - -■ ■ ;"__^ i '-■■ — —-

. ... . . . ■ . . ■ ' - - : , -'.^■■/>>•.■ ■••■. ;>VM^j{

vwrnw*** ■ >j,ii!iiiiamP€W'^WW^IW^^^

0.8 -

SECRET

FREQUENCY (kHz)

(I'NCI ASSII II l)|

(T) Figure 6. TiniL'-rrcqucncy waterfall plot ofa siimilatecl killer whale sound. Time bar et|iials 100 msec.

! '—- ~~ . . i '> ■ ^ T •- ■ , . ' ' . . ^

SECRET

^.U.--,,. ^j^gfeBSJ^Ksfe^^^^Jslm

^g!«^JBW^P^ _ ;^--.T---- ..— p^,^,^;,,^^ ■ :.:7^™_

SECRET

0.8-

FBEQUENCY (kHz)

(I \( I VSSII II m

(III iguie 7. I niR'-liociuoncs walerfalj phil ol ;i luiltinil killoi whale somul. I imc hai ci|iials I 00 msec.

SECRET

/i^^^i^js^^tiäx^^i^r.-^iüt^i^ ;."t'- > ' " ' "• HI ■ ^..-IIJII. ■> ii. i—^

jlippgp|j;g»}^w^

SECRET

_ o

> —

LU --■ D _ O w UJ / tr _

(MS) 3|Alli

2()

SECRET

_ ■4 'is-f'Sw&Mrii^S =a-7is ̂ g^B^g^J JBI^flipJiP

pppl^^ ipBSII^ppap^WFTSSS^^p--" - ---^,WJ5B

SECRET

'

_ o

I CO

o (3aS) 3IA1I1

I o

D/l

E

X <*- ^a c •—■ — > O 2 s: a

\r. T3 LU <; -n ■D I

O u UJ X o cc ;_j tr

-a

30

SECRET

^'•■■' .-.:.. ... i^lr ■:-/ —^- T3!fc&m**&^ ;Äa^ÄS

SECRET

(S) The synthesis of pilot whale sounds requires an ab ility to describe the sounds' microstructure and to choose the generated frequencies. For applicatio ns no t requiring re:.tl­time generation , sounds can be tape recorded . Recorded sounds would have an upper fre­quency limit about 7kHz . Me thods of sound synthesis invo lve signal generaTion tech niq ue~

a nd computer software but the uppe r frequen cy of synthesized so unds is limi ted by hardware.

TRANSDUCERS (U)

BACKGROUND (U)

(S) Transducers for transmitting and receiving COMBO low-da ta-ra te m essages should have the fo llowing characte ristics:

o A frequency range o f 20 Hz t o I 0 kHz. o A source level of I 60 to 200 dB re I pPa a t I m . and o A projecto r signal-to-no ise ratio of a t least 55 dB.

Different transducers may be required to transmit d ifferent types of sound

(U) We surveyed '2 7 transmitting and 9 rece iving systems to ident ify tra n~ducers

acceptable for COMBO use.

(S) Sound projectors transmi t most e ffect ively in a relatively narrow frt•quenc) band . For some projec tors. the e ffec tive band is wide e no ugh fo r COMBO app lication with certain sounds. but no projecto r is suitable for all COMBO sounds. Fur exa mple. COMBO codes to nals in pilot whale phonations from I to 3kHz. but the animal's sonic repertoire ex­tends from 300Hz to I S kHz. The COMBO scheme is directed toward use of fl eet sonar equ ipment. but othe r transduce rs we re surveyed because tleet equipment does not ful fi ll all

COM BO requirements.

(C) No tleet sonar system projects effect ively over the freque ncy I and required for a ll low d ata rate transmissions in the COM BO scheme. so we conside red projectors wi th r,·. gard to the kind of sounds that could be projected. The sound 'aninwl sources were L'lassi­fied by frrquency limits. High-frequency animal sources p roduce sound in 1 he 500- to 8000-Hz range. Small whales and porpo ises. spec ifically beluga. ki lle r and p ilo t whales. arc high· frequency sources of COMBO sounds. Components o f sma ll whales· signals m ay exceed R kHz. but attenuatio n limit s communica tio n utility o f the signals' high ~.. r frequencies. Mid­frequency animal sources produce sounds in the 70- to 3000-Hz rang,·. Some large wha k:-.. s•.1 ch as humpbacks. and most pinnipcds arc m id-frcq ucncy COMBO so urces. Lo w-fr,•q uenL y

animal sources produce sounds in the '20- to '200-Hz ra nge. Large whale!>. such a · b lue. finba .. : h. and gray whales, are low-frequency COMBO sources. Because the frequency limi ts ove rl ap. some animals are sources in two catego ries.

FLEET PROJECTORS (U)

(S) F lee t sonars applicable to COMBO applica ti o ns are listed in tab le -l . lnadcqti.Jt , low freque ncy response limits most of the un it s. Only GNATS sa tisfies a comple te suh t;1,k ca tegory. namely the mid-freq uency segm ents. The GNATS syste m uses t hree tran~duc,· r~

and each transducer o pe rates in a d istinct freque ncy range (reference JSl.

3 1

SECRET

'"i ■■■ I s »fifPSp«ip^^^i^.W^i!tpi^WIJ^^^

SECRET

(S) The WQC-2 low-band projecting transducer system could be equalized to pro- vide adequate response from 1.1 to at least 5 kHz. at a source level of approximately 190 dB re 1 ßPi\ at 1 in. Alternatively, sounds could be preequalized tor use with the system. Use of the \VQC-2 driving amplifier in a COMBO scheme would require auxiliary input from a Hat response circuit. The WQC-2 is installed on submarines and surface vessels for communica- tions.

(S) The frequency response of the WQC-2 is too nanow for the entire range of the high-frequency category of sounds, but is adequate for COMBO transmissions of killer whale, false killer whale, beluga, and bottlenose porpoise sounds.

(S) The frequency response of the SQO-23 with an auxiliary amplifier and of the BQS-13 in a communications mode nearly equals the response of the WOC-2. The BOS-13 is limited below 2.5 kHz and is thus limited to transmitting porpoise-like sounds in a COMBO commuaications application.

(S) The CiNATS system frequency range is suitable for transmitting high- and mid- frequency category animal sounds, but the system has a maximum 160-dB source level ca- pability. COMBO sounds require 1X0- to I^O-dB source levels to achieve adequate communi- cations range. Another possible disadvantage to COMBO application of the CiNATS system is that only 10 (JNATS sets were built, and the availability of units is questionable.

(C) Our survey indicates that fleet sonar projectors are most suitable for projecting COMBO animal signals in the high-frequency category.

OTHER PROJECTORS (U)

(S) Other projectors surveyed meet certain COMBO criteria. Several projectors have potential use for acceptable COMBO transmissions (table 5). Except for the Honeywell HX 182-A, the listed projectors are small and lightweight. The NOSC PHD IV is the only transducer surveyed that can project broadband sound at frequencies as low as 20 11/ with no significant distortion (figure 10). Disadvantages of the NOSC PHD IV for COMBO communications applications include an upper source level limit of 1 56 dB. the absence of active depth compensation, and a minimum frequency limit that increases with depth.

(C) The ITC 2010 and 201 1 projectors cover a broad frequency band and operate al an adequate source level. The ITC units are small and omnidirectional in the plane normal to their axes. Together, the ITC projectors have been successfully used in COMBO sea tests.

(S) The parametric system from the Naval Underwater Systems Center (NUSC) is broadband above 1 kHz and is suitable for projecting COMBO sounds of porpoises and small whales (figure I I ). The parametric system is very directional and may project unwanted output al primary frequencies that will fall within the detection range of COMBO signal recognition instruments.

(S) The Honeywell HX 1S2-A projector operated satisfactorily at frequencies from 60 to 1 100 Hz (figure The projector is inadequate for COMBO signals in the low- or

■ '.:_

en

m

nw

:::U

w

~

Sy~h

.'m

I. \VQC-~ (

Lo

w B

and)

, SQO-~

-~

3.

BQ

S-I

J (C

omm

Moc

k)

4.

GN

AT

S

Dcp

loym

t•nt

~·eqth

.!lll'y.

kHz

Sur

face

/Sub

mar

ine

1.5

-5

Su

rfaC

l.' I (

'?)

-8

Sub

mar

ine

2.5

-8

Sub

mar

ine

0.0(

1 -

6

(CO

NFI

DE

NT

IAL)

Max

imum

U

ndis

tort

ed

Sou

rce

Lev

el.

dB f

l.' I

JJPa

at

lm

19

0

200+

200+

160

+

(U)

Tab

le 4

. Fl

eet

sona

rs c

onsi

dere

d fo

r C

OM

BO

-like

app

licat

ions

.

rrcq

uc

ncy

Ope

rati

ng

Res

pon

se

Sou

rce

Lev

el

Dep

th

App

rox

imah

.' T

rans

d uc

l!r

kH

z dB

rc

I JJ

Pa.

lm

Lim

it.

rn

C on

l"ig

u ra

t ion

S

izc.

m

--

---

·-----

-----

NU

C P

HD

IV

0.

01 -

0.1

15

6

No

dat

a C

ylin

d<"r

0

.4 D

ia x

0.5

HX

18

2-A

O.

OC1

-I.

I 19

0 23

0 C

ylin

der

0.6

Dia

x 1

.8

lTC

~010

I -4

18

0+

Un

lim

ited

R

ing

0.5

Dia

x 0

.3

lTC

201

1 4

-9

190+

U

nlim

ited

R

ing

0.2

Dia

x 0

.1

NU

SC

Par

amet

ril·

I -~0

195

+

91

Dis

c 0

.9 D

ia x

0.1

(UN

CL

ASS

IFIE

D)

(LJ)

T

able

5.

Non

-fle

et p

roje

ctor

s cn

nsiu

crcd

for

CO

MB

O-li

ke a

pplic

atio

ns.

en ., n :::u ~

^^^^p^^? mmmmmtm ^mmmwmmiim^Mffi'm-^^ \

SECRET

D Q. Z

II V

a D O

4

o CO

i

us

^

o o o o o

Vo I o

(ZH) AON3n03dd

34

SECRET

a - ~ c

x: "^

E a

^-*N ~u c Ü "3J rt 0) "a — (/) Q > •*—' O

UJ Q

S y: Ö E £ f- X)

(i> —

,,; C ^- ■—'

^ 5 >. .^ z L?) x ,. w u o "a «3 '*- '^

5 -^ S o ■^ c C\J o

3 42

5 s 0 ,*-| -^ (^1 r3 — g tj •« nj .— (^ CJj

rt _>• — r3

C n

C^ l| d , ' *-* — « .Ü rj n <—

a. w 3 C 1) slj

iE C P

^ •1lliilffir•^• SU^^jJg&i— ^^

>^iy'.'M'^JtWW'>»W^V..^^'^^ ^P^«pwsaBWB|pp|PB)|p??sw?«?5fW

SECRET

z o

D a; H- U LU Q.

-r

/

3WI1 > u z LU

o UJ

SECRET

■ . : - -a>:-.. ^s»^:.: aa^tAUa±£ia^aiaa^B._^M

^mmmm^m^m

SECRET

LXJ > 111

180

170

(8) 160

cc Q.

0) 150

CO •a

140

LU o QC 130 D O (/)

120

110 10

HX-182A

100 1000 10000

FREQUENCY (Hz) (UNCLASSIFIED)

(U) Figure 12. Response ofHX 182-A projector in series with a 110-to-200 Hz band rejection filter. The input signal was 100 V mis at 400 Hz. The source depth was 62 m.

36

SECRET

■ ■: ■..'. ■ ■ ....

ilMfffilTtiYllillliiiTi

SSSPPS^iS^^^^giiSISf^SflPliS^^'W^^p^piSg^S^BS« KW^^.'W^"!-.»mi*»i«v. f

SECRET

mitl I'rcq unu \ categories. A lice I soii.ii used in conjuiu linn u illi ;iii i l\ I N.VA .uul a l'lll)-l\' woukl he cipahle olCOMIU) Irailsiiiissions ,il I ie(|iic:icies lioin -M) lo 10.001) 11/

LOW-FRIQUI NCY I'ROJI (TORS (U)

(U) SIN lou-lieiiuency projcclors siincycil arc ilcscnhcil in lahle '■ I lie liNiliaulu mechanisms of 1 he liydioacimstic ile\ ices listetl picuiiue unacceplahle ha rm on k di si oil ion and ex Ira neons com pone ills. I he IIX 254-H uses a diHeienl prcssuie com pen sal ion s\ stem

than the IIX 1 S2-A. All projectors list cd in table d ha\e ;JU alei source levels and less ile pill seiisili\ii\ than the IMII) IV.

FLEET RECEIVING TRANSDUCERS (U)

(Si Oui siir\e\ indicates that receivers meeting < ()M BO ciitena are a\ailahle m the

Heel (lahle '" I I he receh inj: icspoiise ol tlie W(,)( -2 is narrowest ol the surve\ ed units lor cMsiuiü iiansdiu eis. ihe SOO-.1.J is lu'st lor (OMIU.) application lor surlai-e ships and the BUR-7 is optimal lor submarines. I owed arras s LMVC exleiuU d low t ici|ucnc\ cmeia^e. but

then use is resiiictive. A broailband hydrophone will be necessary loi surlace ships to receive COMBO sounds I he B(,)K-7 enables submarines to receive COMBO sounds at rrequencies as low as 50 11/

(S) Surface ships can also use sonobuoys to receive COMBO siimals in the entire

20- to 10.000-(1/ frequency ranue A sonobnov tleployed awav Irom a surface ship would likely receive less own-shi|i noise than liull-mouiiled leceivers. Antenna heijihl limits sono

buoy radio range. Depeiklini; on sea slate, a surface ship movmi; a! 20 kt should teccive COMBO siimals from a sonohuov |(>r 1 5 lo 30 min.

DEPLOYMENT SECURITY (U)

IS I ( overt and secure application of the COMBO communications scheme requires

that the hioacoustic sijmals have ailcciuale nalmal (|ualilies. Acoustic criteria loi signal nat- uralness include lulehu . source level, frequencv content, waveform, amplitude and duration. Specific signals also must occur at limes and locations proper for source animals.

TRANSDUCERS (U)

(S) No underwater sound projecting equipmenl can transmit bioacoustk signals with a'nsolute fiilelits. Signal analysis can reveal the perturbations thai transducer svsiems

impose on transmissions ami can identify unnatural bioacoustic signals, I ransducers that project hioacoustic signals for comnumicalions nuisl limit the transmilteil perturbations si)

the signal variability occurs within normal limils for a knowledgeable listener

SOURCE LEVELS (U)

iSl Comiminicalion applkalions ol hioacoustic signals may requue highcr-lhaii-

na I ura I source levels to ove iconic propagation loss I actors and to achieve sullicienl coinmu- nications langes I he source levels ol plionalioii'- bv largi ..hales are adequate lor most

37

SECRET

- - — / _*L ^^^^mmm

SECRET

3 SO

O

o .c -

■C V c

'-I n _t Q

o 3. '; u. ' u

u.

c/: as

-_i c 2 O N

u. a:

^ c "-

oil -^

v; o _3 o C— ^ ,_ 'J o

O n

1) X c 73 cz

o

a* -D

o

H

c 30

o

o 30

M ^

t r 1 X

O ■C 1-

o X)

r,ij 11/ .^: _^ O

ir, c >r, cc

ir, in

73 "a o 2

1/1

-a ir, u y.

^^ 33 h r^

O O o 00 c- o •— —< r i

EC

r J

O

33 r3 3Q 33

r3 o 6 ^ Q Li-

^:A X c a 0

o 2 UU y-

^ r i

5 ■i"

"2 > n Si

u

CQ

r3

o o

o r3

^ X .— (U

5: <

> •v.

38

SECRET

■ _

TjtMVftf^ffi.jMiiwiafa • - '*"'' " r._:^

0

rz 'J

a a n aj ^

o ca S o o

QJ

JO w (/I c a

_l i^

< Ü H •a ?. >, UJ ■u Q > a. 3

in y. (^ o O u t j

— r 1 C

3 CT

73 ?

c C N ir, o O '1 o ■T- O o o 1/-.

1/-) r I

5 V. UJ

O ir, o

-t 1

o 1

o r 1 r3

I-

^M^^iapjii^

SECRET

System

SOO-23

BOR-7

BQS-6, 13

S0R-17(TASS)

BQR-15(STASS)

Deployment

Surface

Submarine

Frequency, kHz

0.5 - (>

0.1 - 7

Submarine 0.5 - cS

Submarine 0.01 - i.l

Submarine 0.05 - 2.5

(SLCRUT)

(U) Table 7. Flejt receivers suitable for possible COMBO-like applications.

COMBO applications. Communication signals that use phonations of other manne mammals may have to exceed natural source levels by up to 30 dB. Source level is a function of the transmission's coded frequency and intended range. Use of unnaturally high source levels should be limited to short-duration transmissions and low frequency sounds.

SIGNAL SYNTHESIS (U)

(S) Signal synthesis has reproduced portions of pilot whale sinusoidal waveform whistles. The sounds of large whales have not been synthesized. Adequate synthesis of bio- acoustic signals for covert communications requires quality recordings of natural sounds. Synthesized signals are constructed from and compared with the natural templates. Synthe- sized signals must appear natural on sonardisplays and must sound natural to knowledgeable listeners.

(S) The formats presenting bioacoustic communications signals must also be natural. The increase in critical bandwidth with frequency in animals' auditory function implies thai bandwidth of animal sounds also increases with frequency. Thus, synthesized animal signals that sweep across frequency should have more frequency variance in the high-frequency por- tion. Some formats, such as the amplitude rise and decay of choruses, change gradually and vary randomly. Other formats, such as humpback whale songs, occur in distinct rhythmic sequences that are repealed as stanzas.

(S) l-.nsembles upon which coded signals are superimposed must be randomly con- structed. The library of component synthesized sounds must be large enough that long-term analog recordings are nonrepeating.

(S) During sea trials of COMBO equipment, transmission of synthesized pilot whale sounds attracted pilot whales to the transmitter. The whales' phonations added to the sonic environment without degrading the communications effort.

39

SECRET

: l~.~l.l.

-.V--.- ifTMirririfiniit"—-- -; ■■fe^--- i,aküJJJ.;.,^-..i!

SECRET

SIGNAL OCCURRENCE (U)

($) Bioacoustic communication signals must occur at the proper time and place . No marine mammal is pan-oceanic and many distinct phonations occur o nly at ce rtain times or places. The frequency with which a bioacoustic communication will be used depends upo n the communicators' activities. More common signals should be used for mo re frequent com­munications.

(S) A guide to biologics has been produced to familiarize sonarmen with bioacousti c signals. bioacoustic sources. and marine mammal distribution ( reference 12).

FLEET IMPACT (U)

(S) A six-week study of Project COM 80 and of the e ffec ts of bioacous tics on fleet operations concluded that both the U.S. and the Soviet Union would have the capability to apply bioacoustics to communica tio ns tasks after 1975 (reference 36). Covert active sonar using marine animal sounds requires changes in fleet sonar hardware and operating doctrine. Re liable. re lat ively cove rt communications possib le through Project COMBO also will affec t submarine tactical deployme nt.

(S) Poor communications disrupt coordination in submarine escort or ASW opera­tions. Even a limited covert communicatio n ability will enhance coord ination between sur­face ships and submarines. Projec t COM 80 provides covertness unavailable with presently­deployed two-way communication systems and thus enhances coordination in criti cal scenarios.

(S) U.S. Sonarmen typically tune ou t biologics as interference. but Soviet use of bioacoustics for communications is likely . so fleet sonannen must become more famili ar with bioacoustic signals. Past biologics classification was based on NAST AD training tapes. au ral memo ry and the sonannen's experience. Anno tated BQR-20/ 22 displays and bio­aco ustic indices (refererv:c 12) now augment the classifica tion e ffort.

(S) Marine animal sounds arc ubiquito us (refe rence 37) and like ly we ll-known to Soviet sonarmcn . Signal processing techniques will aid and advance unde rwater military communicatio ns. but U.S. fo rces may face an increasingly costly race to stay bdow pre­sumed det ectio n thresholds. Covert communicatio n by natural disguise is a psyd10logica ll y confou nding alte rnative approach that can convey consid erable advantage to the adver. ary first employing the technique.

40

SECRET

,-„-,,„„ „,„r,,,-,, ^-^-^.mp^j

SECRET

REFERENCES (U)

1. Cumniings. VV. C, R. L. Seeley, P. 0. Thompson, S. J. Kennison. Unique Communica- tions System (U). Patent Disclosure No. 59. 978. 1975. (Si.CRHT)

2 Seeley. R. L.. W. C. Cummmgs, S. J. Kennison. Communication Channel i'ading Compensator (U). Patent Disclosure No. 58.979. 1975. (SIXRI. 1)

3. Johnson. R.A.. R. L. Seeley, W. C. Cumniings. S. J. Kennison. Communications Irror Correction Device (UI, Patent Disclosure No. 58.980. 1975. (SliCRLT)

4. Seeley, R. L. W. C. Cummmgs. S. J. Kennison. Synchronizer lor Communication Initialization (U). Palent Disclosure No. 59.542. 1975. (SICRI I I

5. Ail Hoc Study Croup. NSIA Ad Hoc Study Acoustic Communication Mnal Report ill. National Security Industrial Association, 1972. (C()NI;ll)i:NI'lAL)

(). Howard, I.. F, lACS/QRP Operational l-'easibility and lilTectivcness Study: Operator Coded Message Inputs (U). NUSC TM No, LAIX-Cl 5-74. 1974. (CONl-IDI NTIAI. I

7. Wrench. H. 11. Jr. increased Signal Processing Speed Using Sweep Limit Control (l). Unpublished NUC Report. ( UNCLASSIIIi;!))

8. Thompson. P. 0. and W, C. Cumniings. Critical Survey of Certain Underwater Sound Transducers (U). Unpublished NUC Report. (UNCLASSIIIhD)

9. Christian, J, Operating Characteristics of J15-3. IIX-I«:. and ll.\-l,s: in IIX-IS4 Hull l.ow-l'requency Transducers (U). Unpublished NUC Report. 1975. ( UNCLASSIIIID)

Payne. R. S, and S. McVay. Songs of Humpback Whales. Science, 173;585-597, 1971.

Cumniings, W. ('., ,1. I", I'ish and P. O. Thompson. Sound Production and Other Behavior of Southern Right Whales, luihalacna ghicialis (U). Transactions of the San Diego Society of Natural History 17(1): 13 pp. 1972.

Cummings. W. C. and P. O. Thompson. Sonarmaii's Cuide to Biologies (U). NOSC ID 1X1, 77 pp, July 197S.(CONIII)l-.NTlAL)

Cummings. W. C. and P. O. Thompson, Sounds from Bryde's and 1'inback Whales in the Cult" of California (U). NOSC unpublished report, 2() pp. ( UNCLASSllllDi

SchevilLW. H.,W. A. Watkins.and R. 11. Backus. The 2()-Cycle Signals and IhiUwiiopicra (1'in Whales) (U), Marine Bio-Acoustics; Proceedings of a symposium held at the Lerner Marine Laboratory, Bimini, Bahamas, April 1 1-13. 19(i3. Macmillan Company. Neu York. 19^4. (UNCLASSILII D)

41

SECRET

IU.

12.

13.

14.

^ia^^M^^^Afa^ ..._,,*-- /w^t^r^».^-.,^-.. - ^, .■^■-^^J^«r^^.-J^«^^ . ^I^T TTJB^SSSSSjfiS^l

SECRET

IS . Cummings. W. C. and P. 0. Thompson. Unde rw3te r Sounds from the Blue Whale. Balaenoptera musculus (U). J . Acoust. Soc. Am. 50( 4): I 193-1 198. t Q7J.

16. Backus, R. and W. E. Schevill. Physeter C'licks (U). In Whales, Dolphins. and Porpoises. edited by K. S. Norris. University of California Press, Berkeley and Los Angeles, 1966.

17. Watkins, W. A. Acoustic Be havior of Sperm Whales ( U). Oceanus (Woods Hole Oceano­graphic Institution), 20(2): 50-58, 1977.

18. Busne l. R. G. and A. Dziedzic. Acoustic Signals of the Pilot Whale Globiceplwla melaena and of the porpoises Delphinus de/phis andPiwcoena plwcoena (U)./11 Whales. Dolphins. Porpoises, edited by K. S . Norris. University of California Press. Berke ley and Los Angeles, 1966.

19. Fish , J. F. and C. W. Turl. Acoustic Source Levels of Four Species of Small Whales. NUC TP 547. 14 p., Dece mber 1976. (UNCLASSIFI ED)

20. Schevill , W. E. a nd W. A. Watkins. Whale and Porpoise Voices. A Phonograph Record. Woods Hole Oceanographic Institution. 24 p. phonograph disk , 1962.

2 1. Sc hev ill. W. E. and W. A. Watkins. Sound Structure and Directionality in Orcinus (Killer Whale) ( U). Zoologica. 51(6): 76-76, 1966.

22. Kooyman, uupublished recording of killer wha le sounds in Antarctic. Scripps Institution of Oceanography.

23. Schevill. W. E. and B. Lawrence. Underwate r Listening to the White Porpoise (Delplli­llapterus leucas) ( U). Science. 109(2824) : 143-144, 1949.

24. Evans. W. E. and J . H. Prescott. Observ C:t tions of the Sound Production Capabilities of the Bo ttlenose Porpoise: A study of Whistles and Clicks (U). Zoologica. 47 : 12 1-1 28. 1962.

25. Cummings and Thompson. unpublished recording of Pacific White-S ided Porpoise. (UNCLASSIFIED)

26. Schevill. W. E. and W. A. Watkins. Underwater Ca lls of Leptonyclwtes (Weddell Seal) (U). Zoologica. 50( I) : 45-46. 1965.

27. Ray . C .. W. A. Watkins. and J. J. Burns. The Underwater Song of Erignatllus (Bearded SeC:tl) (U) . ZoologicC:t. 54(2): 79-83, 1969.

28. Schevill , W. E., W. A . Wa tkins. and C. Ray. Underwate r Sounds of Pinnipeds (U). Science. 141(3575): 50-53 . 1963.

29. Watkins. W. A. and C. Ray. Underwater Sounds from Ribbon Seal. Plwca ( Histriopltoca) fasciata (U). Fishery Bulle tin. 75(2): 450-453. 1977.

42

SECRET

"^ -h .W^.IA^MfSJ^,^^vlP."«.^l!lJR'JIW»-li'«,l'il'J'-;J ««iipijlliijjkiiiiii ..(piujpijp.u;Ul>,H \

SECRET

30. Scluisterman. R. J., R. Cicntry. ami J. Sdimook. Underwater Sound Production by Captive California Sea Lions. Zaloplnis vulifoniUiniis (U). Zoologica. 52: 21-24. \i)^l.

3 1. Cummings. W. C. Project COMBO: Progress Report tor liscal Year 1973 ( U). Naval

Undersea Center TN 1009.22 pp. 1973. (SliCRirD

32. Cummings, W. C. Progress Memo. Project COMBO. 2S December 1()73 (U). Naval Undersea Center Memo to I). C. llotTman, NSI1IPS PMS 3024. 17 pp.

(CONFIDHNTIAL)

33. Urick. R. J. Principles ofUnderwater Sound for Ingineers. McCiraw-llill. San I'rancisco. 342 pp. 1967. (UNCLASSIFIHD)

33. Johnson. R. A, and A. \V. Dibelka. Analysis and Synthesis of Natural Sounds (U). NUC TP 533. 3d pp. 11)7(,. (CONMDI-NTIAL)

35. Bell. T. (i. and C C. Connally, Jr. Future Requirements for (leneral Noise and Tonal Systems (C.NATS) (U). NL1SC TR 4613.84 pp. 1973. (SHCRHT)

36. Marich. M. NUC Memorandum. Ser 4013/WCC.S1 1/75. Report of Activity During Temporary Additional Duty (U). 15 pp. 1()75. (STX'RFT)

37. Cummings. \V. ('.. W. T. Leapley. (I. B. T'arquhar. C. L. Brown, Jr.. VV. A. Triedl,

A. L. Brooks, Bioacoustic ITfects on Naval Operations: Report of Status and Recom-

mendations (U). Naval Undersea Center IP 407, 30 pp. 1974, (CONFIDHNTIAL)

i

43

SECRET

, " - —7—: —r~:.~~—."-—rr—r—t.—;—,■;;.: .:;■:■• r —v—r-1—:—:——.-■.■• ■---^■,fc;._ -!■■_ j^-'j,.'-^ •r^itiynr/j

Sj^A^Wi^tbiPPipf^^ s4£ä£äieä&0m&^

mtzpme. -

SECRET

RELATED DOCUMENTS (U)

1. ( mnimiiiis. U. ('. \Jv;iiKi'd Systems ( OIKCJ-)! COMliU (U). Mciiiü. CüMMK lo ( OMXSIII'S, 0101: IK .Minn. .V^K), Sei OIOl-üOlO. 2 pp. I()71. (SI (KhT)

2. ( unmiiiiüs. W. ( . jiul .1. I . Iish. Covcil ( omimmiciitioii Tcchni(|iic (U), NUC Memo lo I). A. Wilson. M ( Code 04. 505 \V((:eJli. Sei 505-0012 71.5 pp. 1^71. (Si ( Rl.l )

v ( in in nines. \\. ( . Projeel ( ()\IIU) In lorn ml ion Response to Brieling to Dr. Ileilmeier. DDRI d) Memo (OMMC lo (OMNSIIPS (I'MS M)1A). 4 pp. |W7.v (COM 11)1 Ml Al l

4. ( mnmiiiL's. \\, ( ..null I . Iish Snpplementiiry Inroniuition on I'rojeet COMBO (U). M ( Menu. Report lo KM A I 0.M5. ( DR Donald Moyer, 505 1 WCC .di. 5()5-()()2.:!4- 7.C 4 pp. lw7.v (SI ( kl 1 )

5. ( nminmes. W. C. Adwnieed System Coneepl lor COMBO (I ). Memo COMNUC lo COMNSIII'Sd'MS-.^O:). 0101 I I'Cmie 5400. Ser 01 0 1-005. 2 pp. Il)75. (SI (Rl I )

d. Nav\ Scienlilie and leehine.il Intelligenee Bulletin. OM-(p-l 7-5-73. 7 pp. |U7,C (SI CRi I )

7. Munson. .1 ( Aeoiislk ( omiminieations lor Mobile Plallornis IV 75-77 I xjiloratoiA' Development: A Slud\ and Keeommeiulations (11). Report lo MOS I ( ommillee. NRI. M()()-.s::.ICM iet. I nelosures. pp. 25. lh. 37-40. 42. A-2. A-(). A-7. 1<'74. (SICRI.I )

■S Do.imienl OSI-V\ -S-.U) 14 (U). p. 3. ll)74. (SI (Rl/I I

44

SECRET

...... .. .. ■

MEMBERS

DEPARTMENT OF DEFENSE Garry P. Reid

DEPARTMENT OF JUSTICE Mark A. Bradley

DEPARTMENT OF STATE Amb. Larry L. Palmer

OFFICE OF THE DIRECTOR OF NATIONAL INTELLIGENCE

Jennifer L. Hudson NATIONAL ARCHIVES AND RECORDS ADMINISTRATION

Sheryl J. Shenberger NATIONAL SECURITY COUNCIL

John P. Fitzpatrick, Chair

December 6, 2016

Dear

Interagency Security Classification Appeals Panel

c/o Information Security Oversight Office 700 Pennsylvania Avenue, N.W., Room 100

Washington, D.C. 20408 Telephone: (202) 357-5250

Fax: (202) 357-5907 E-mail: iscap@nara.gov

EXECUTIVE SECRETARY

William A. Cira, Acting Director INFORMATION SECURITY OVERSIGHT OFFICE

Please be advised that the Interagency Security Classification Appeals Panel (ISCAP) has concluded its consideration of the mandatory declassification review appeal2006-033 filed by you and that the 60-day period during which an agency head may appeal an ISCAP decision to the President has expired. Enclosed is a copy of the document and a chart that outlines the IS CAP decision on the records under appeal. With the exception of any information that is otherwise authorized and warranted for withholding under applicable law, we are releasing all information declassified by the IS CAP to you. If you have questions about this appeal, please contact William Carpenter of my staff at (202) 357-5250.

Sincerely,

I~ IJ,. /.-. /},. (,ll./1,(£~12 tr,_ // ~ t~t'~"'

WILLIAM A. CIRA Executive Secretary

Enclosures

cc: Mr. Mark M. Langerman, DoD Liaison to the ISCAP

Ms. Gaye Evans, Deputy DON/ AA

ISCAP DECISION ON THE MANDATORY DECLASSIFICATION REVIEW APPEAL FILED BY

·IDENTIFYIN~ ' :NllMBttH$

, document No. 1

ISCAPNo. 2006-033

NOSC Technical Report 422

Project Combo: Review and Recommendations

July 1980

48 pages

Secret

ACTION·

DECLASSIFIED THE DOCUMENT IN ITS ENTIRETY