*1 191 367 Mass. Lykus. Commonwealth u. Leate,
monwealth v. 452, 352 Mass. Warren (1967). on Homicide 95 (1938). §
We not see do error in statement to the judge’s jury the deceased no current to the threat represented incident, defendant at the time the judge may rule on whether the evidence is properly sufficient demonstrate circumstances particular warranting Commonwealth v. verdict of Campbell, manslaughter. Mc 387, Mass. Commonwealth v. (1967). Cauley, Commonwealth 554, 355 Mass. (1969). Caine, 366, 366 Mass. 374-375 That is ruling the evidence. supported by
We conclude that instructions considered judge’s did not separately together establish error.
9. Consonant with our 278, under G. L. duty c. 33E, have we reviewed the record and § transcript find no order new grounds trial or direct a verdict of a lesser degree guilt.
Judgment affirmed. Stanley Lykus. Edward vs. January Bristol. 1975. March 1975. C.J.,
Present: Tauro, Reardon, Braucher, Quirico, Hennessey, Kaplan, Wilkins, & JJ. Evidence, test; Spectrograph Opinion: expert; Telephone conversa- Witness, Expert. tion. Identification. In deciding whether to admit deduced
from principle, the standard set forth in Frye v. United States, F. 1013 (D. C. Cir. 1923), subsequent Massa- chusetts cases respecting general acceptability is if the satisfied principle generally accepted by expected those who would be to be familiar with its use. [196-203]
Commonwealth *2 indictments, admitting the trial there was no error in At the of the voice' of the opinion of for the Commonwealth that an calls, telephone whispered “tapped” unknown of recorded speaker of concerning missing youth payment the and the part, most a ransom, opinion of the was the voice the defendant where was spectrograms, in visual “voice- part comparison based on a machine, produced by depicting by a prints” and recorded graphic patterns telephone the recorded calls voice, part listening exemplars by of the defendant’s and voices; justified the admission of the was evi- recorded general acceptance spectrography as and dence judicial opinions hearing, by introduced at an extensive voir dire Quirico, J., concurring from other jurisdictions, [205-206], relevant scientific Kaplan, J., in a separate writings. opinion, [205] admissible, voice was not on the concluded that the identification States, 1013, F. Frye standard of v. United ground that the [206-213], had been met 1923), Cir. C. returned found and Superior Three indictments Court, 6, 1973, 8, third two on February June 1973. McGuire,
The cases tried before were J.
The cases were submitted on briefs. for the defendant. Ronald D. Harper Rollins, A. District for the Common- Attorney, Philip wealth. In this of first case impression
Hennessey, defendant Commonwealth, issue single argued by evidence, from an expert concerns the allowing judge’s Commonwealth, as to witness introduced by on voice identification based in visual analysis part referred commonly “voiceprints.” spectrograms, that the evidence was admitted. We conclude properly convicted, trial, on The was after a defendant jury three indictments him with charging kidnapping Cavalieri, from murder of Paul and extortion first degree L. c. victim’s The is under G. father.1 appeal 33A-33G. §§ defendant, specifically charged The extortion indictment authority, forcibly imprison one Paul did confine and
without lawful as follows. Paul case was early years old, house in the Cavalieri, left his thirteen evening home did not return November 1972. He boy’s days night, later the or ever thereafter. Two telephone a said in a call in which voice mother received “you whisper, Although a will receive instructions.” Mrs. Cavalieri had known the defendant for seventeen years, recognize she did not the voice. permission “tap” placed
With Cavalieris, was telephone recording incoming on their calls, for the “trap” placed telephone and a so called also was tracing to assist in the of such calls. *3 days
A that, few after a ransom note was received demanding ensuing days, $50,000. In the several the telephone victim’s father received several calls all payment concerned with his son and the of ransom. He recognize Following did not the caller’s voice. instruc- telephone tions in the note and in calls, the the father left money the money ransom locations; at two different the picked up. was telephone As a result of the third money call, he the left aat new and different location. At 12:10 a.m. on 10, 1972, November a white male was by police picking up money. observed the Sometime away, later, and not far other officers observed a man carrying bag get a white come out the woods and wagon. By tracing into a they red station vehicle, the identity learned the the driver, one Tardiff. From they him, identity learned of the of the defendant as a person pretext, persuaded who, on a had Tardiff to leave highway the place on defendant near the where the pick up ransom been left, had to later, defendant prearranged place. at a police
The defendant was arrested. The warned him rights of his constitutional in terms that are not now adequacy, by contested, as to defendant. Cavalieri, his against with intent to cause him to be held to service will, thereby. money things with intent extort or other valuable Mass. 191
Commonwealth v. up picked police he had defendant admitted package, done it for that he had but he contended drug business. He pay traffic from a man of $500 telephone having the Cavalieris. calls to made denied money leaving where he had The defendant told by him. The who had hired the man been instructed place money (not told to police later at the recovered the by defendant), result of but rather as a them Delaney, from a man named received information confided the location the defendant had whom bags, money. as the same package identified A of black bag been recovered the ransom had kind as the which police, vehicle. The in the defendant’s was found personal were effects and his automobile defendant’s put dye been ransom that had stained money. body April Cavalieri was 12, 1973, the of Paul
On Attleborough. Death in North in a wooded area found by gunshot to the In addition wounds. had been caused a .38 above, it shown at trial that outlined was by police in the been found Colt revolver had caliber apartment. testified that An defendant’s body been fired could have in the victim’s bullets found *4 by rifling impressions gun, on the as shown from the bullets there were bullets, to oxidation on but due microscopic Ammuni- for identification. marks sufficient gun defendant’s discovered in the the was also tion for found in apartment ammunition, like the bullets and type. addition, five body, of a rare victim’s was the and staples were examined the ransom note found in size with in manufacture and be consistent determined to by About the stapler staples defendant. the owned disappearance, defendant was the victim’s time of the where, place seen admittedly victim was last the near the near several times car was seen The defendant’s alive. money “drop” first place was of ransom where the the response money, to tele- the made, the time when at that location. phoned instructions, been left at had Commonwealth Eight listened the to one of the witnesses at trial telephone recorded calls. had Seven these witnesses speak telephone prior heard the defendant on the occasions. Six of them the as of the identified voice eighth prior The defendant. witness had not heard telephone defendant; talks of the he identified the also voice as that of the The were defendant. witnesses particularly identifying definite which certain words whisper, clearly. Only said, were not in a but one eight witnesses, issue, called as to this could make no voice identification. “tapped” telephone
All of the recorded calls the whispered, part, Cavalieris had been at least the most by exemplars the caller. Several of the defendant’s voice reading previously telephone the words of the recorded including exemplars calls, whispering in the defendant’s given police. voice, telephone tapes the were exemplars were sent to Lt. Ernest W. Nash Michigan police. State appeals:
We turn now to the issue raised these admission in the voice identification spectrograms. based on voice In an extensive voir dire hearing admissibility to the evidence, Dr. Oscar professor Michigan Tosi, described as “a at State Uni- versity Speech Hearing and director Research Laboratory,” Michigan and Lt. Ernest W. Nash of police, State described as “of Voice Identification Unit, which is a unit of the Scientific Criminal Labora- tory, Lansing, Michigan,” stationed at East both intro- length sup- duced Commonwealth, testified port admissibility of the evidence. Of the two subsequently jury, Lt. Nash testified after before judge expert opinions had ruled that his were admis- sible. Dr. Louis Gerstman, introduced the defend- *5 general spectrogram analysis ant, testified voice spectrograms time, unreliable and that, at this voice should be used as a in a means identification qualifica- forensic situation. No issue is raised as Mass. v. Commonwealth indeed, Nash, or, of the three wit- tians of Lt. any expert evi- nesses. The sole issue relates whether expert here been to be dence of Lt. Nash offered has shown reliable to be admissible evidence. sufficiently Fatalo, 266, In 346 Mass. Commonwealth v. Juvenile, v. Mass. A (1963), 421, concerned with the admissi- 425-426 cases (1974), evidence, we the standard adopted bility polygraphic dis- as to based scientific admissibility expert opinions States, in the case v. United coveries as first stated of Frye scien- 1013, 293 F. C. Cir. when a 1923): “Just line tific crosses the between principle discovery is difficult to define. and demonstrable stages experimental zone force of in this the evidential twilight Somewhere must and while courts will be recognized, principle deduced admitting go long way scientific or discovery, from a well-recognized principle be the deduction is made must from which thing established to have gained general acceptance sufficiently in the in which it belongs” particular (emphasis field added). case,
In we said: the Fatalo supra, “Judicial can occur of a scientific or instrument theory acceptance the com- when it follows a general acceptance by sub- involved. supported by of scientists When munity stantial scientific establishing reliability, authority court the benefits of science.” accept has not hesitated 176- See also Commonwealth Stappen, Devlin, Mass. and Commonwealth (1957), 149, 153, fn. 3 (1974). articulated accordance with standards matter, we, consider case,
Fatalo as a preliminary involved. Like some nature machine devices, machine signifies other scientific a spectrographie on a paper, its graphic patterns results by producing ex- analysis are then the which subject patterns main three parameters aminer. The spectrogram portrays axis), frequency (vertical axis), time speech: (horizontal *6 367 Mass. Lykus. o. shading intensity (degree amplitude in of or
and relative time/frequency regions). Identification is the different operator attempted by means. The aural and visual both recordings compares by listening known and voices of high-speed by visually analyzing voices, unknown spectrograms If he finds sufficient of each voice. sound points similarity, voice ex- he will indicate that two by person. emplars made same were not from thus draws his conclusions examiner comparison. It can be from an aural a visual but also aspects considering the that, said at least in these reasoning by importance subjective relative process poly- examiner, resemblance to bears a closer graph testing fingerprint handwriting it does than analysis. important respects, however, In some support polygraph claims made are funda- mentally spectrograph, different from those made for part reasoning and in those distinctions lies of our treating differently. important the two Most is the urged reading breadth of the inference from the respective Relying part machines. on voice character- by spectrograph, and measured istics demonstrated compare no more than seeks to do the examiner there reflected contrast, the measurements voices. In from polygraph, extrapolates the examiner then to arrive at judgment something directly measured credibility person machine, is, examined. doing, purported polygraphic evidence, In so with its ability testimony, may constitute, to discern truth in any jury’s case, intrudes far into the most a force which important determining credibility functions of of wit- finding nesses and are facts. We aware that proof may posture mystic in some “a instances assume infallibility laymen.” eyes jury United of a 1974). Addison, States v. C. Cir. 498 F. cited, For the reasons we have that consideration does not lead us to exclude the voice identification impose here nor to so restrictive a standard of admis-
Commonwealth e. sibility applied polygraphic as we evidence in Com- Juvenile, v. A monwealth reaching our conclusion that the evidence was
properly admitted, we have examined decisions relevant jurisdictions; writings from other relevant in scientific journals produced sources; and other and the evidence general acceptance the trial as to and of the principle. concepts We with the start that neither infallibility acceptance principle nor unanimous of the proved justify need be its admission evidence. judicial acceptance spectrography, As it is undoubtedly holdings more useful to consider the year prior various courts since the 1971 than thereto. The years laboratory conclusions from two of extensive tests supervised by Dr. Tosi were first made available in 1971. comparable study No was made before or has been made since this one. appellate
Of the
courts which have considered the
admissibility
published,
issue since these results were
(Minnesota,
California)
Florida,
courts
three States
have ruled that such voice identification evidence is
Hedman,
ex
Minn.
admissible: State
rel. Trimble v.
291
(1971) (voiceprints
probable
442
admissible to establish
purposes); Worley
State,
cause
corroborative
v.
(Ct. App.
1972)
properly
(voiceprints
Commonwealth o. which ruled have since the Tosi have decided that study Addison, evidence is inadmissible. United States v. F. G. Law, Cir. 1974). People Cal. 3d 69 App. Law case no further goes than to that such evidence say is not admissible where the voice to be identified was disguised mimicked.
Literature
of the field does not
reflect unanimous
approval
who,
on the
of those
from
technique
part
their
claim to be
writings,
Favorable
qualified
speak.
comments in
have,
course,
strong
been written
support
of the studies
reports
directed
Dr. Tosi. While
some scientists have written in
others
approval,
appear
from their
to have serious
writings
reservations as to the
issue of admissibility.
*8
Appeal, in deciding
case,
the Worley
supra,
Cary
commented on the
cases,
and King
supra, as follows: “Both of these cases are over three
years
then,
old. Since
impressive scientific data has been amassed as
to the voiceprint’s reliability.
that,
hypothesize
We
based
changes in Dr.
alone,
Tosi’s testimony
Cary
King
and
[the]
[cases]
would be decided differently today.
reasons,
For these
we decline to
follow these cases.”
The voir hearing case was and comprehensive. in this lengthy judge and reflected the qualifications Evidence Nash, Dr. and Lt. but also appraised Tosi have written of those who lack same) qualifications (or in voice to the use spectrograph opposition a substantial showing identification. At the hearing, the Tosi labora- was made from the results of reliability Addison, 744-745, fn. C. Cir. v. 498 F. 2d at United States 80-81, Law, fn. 12 People App. and 40 Cal. 1974), opinion (United Court in the Addison case States the District made Supp. 1972]), judge F. D. C. the trial Raymond, 337 [D. Ladefoged’s findings judgments as to Dr. following significant changed study substantially field: “Dr. Tosi’s has community as to the of voice expressed by the scientific striking voice. A identifying as a means of an unknown spectrograms Ladefoged, of Dr. Peter of this can be seen in the case example Ladefoged C. L. A. Dr. was co-author Professor of Phonetics at U. conclusions, and leading study which criticized the Kersta of a article Vanderslice, and ‘Voiceprint’ Mystique, supra, Ladefoged spectrograms admission of into expert against even as an testified case, study, examining the Tosi supra. evidence in the Trimble After however, spectrograms he now believes that Ladefoged Dr. stated identification, as a reliable method voice have been established in the case at spectrograms admission testified favor added). bar” Id. at 644-645. (emphasis Bolt, analysis Cooper, includes critical of Literature Denes, Stevens, Pickett, David, Identification Speaker Speech *9 Reliability Legal Purposes, A Scientists’ View of its Spectrograms: others, Speaker (1970); Soc. Am. Bolt and 47 Acoustical 597 J. Observations, Some Further by Speech Spectrograms: Identification Black, Lashbrook, But see (1973). Acoustical Soc. Am. 531 54 J. Tosi, Nash, Identifica- Truby, Reply “Speaker and Oyer, Pedrey, Observations,” Further 54 Speech Spectrograms: Some by tion J. Hazen, Differing See also Effects (1973). Acoustical Soc. Am. 535 Identification, Speaker Spectrographic Contexts on Phonetic J. disguised voices). (1973) (i.e., Soc. Am. 650 Acoustical — Ahead, See, 11 Am. Crim. Danger Voiceprints generally, Jones, Identification, Note, Georgetown Voiceprint L. Rev. 549 (1973); Comment, Admissibility Spectro Evidence: (1972-1973); L. Identification, Com Rev. 1235 (1972). 56 Minn. L. graphic Voice — Thomas, Eyes It), Have Myth (The or Miracle Voiceprint pare use), Valley (1974) (opposed L. Rev. 15 of San Fernando 3 U. — Comeback, of San Boren, A 3 U. Staging Voiceprint with use). favor of Valley (1974) (in L. Rev. Fernando
Commonwealth tory experiment which culminated 1971. In that study approximately a total of 35,000 trials of identifica- tion speakers, were conducted. The 250 males, all and English speaking, all were selected at random from a homogeneous population Michigan of students at State University. spectrograms Visual examinations of were by twenty-nine conducted trained Results, examiners. published, per indicated an error of six cent false identifi- approximately per cations and twelve cent false elimina- tions. Other correlated and detailed as recorded by supported the examiners a that, conclusion if the express examiners had been able to no when in doubt, errors of per identification would be about two cent.
From the of Dr. Tosí and Lt. Nash the trial judge could well experience conclude that forensic major witnesses, since laboratory study, had reduced the chance of findings error in negligible their to a First, amount. it was shown that the “real life” examiner listens to the visually voices addition to inspecting spectrograms, and it is the combination of both comparison aural and visual which enhances relia- bility.4 Second, experience the forensic examiner has training audiology speech beyond science that laboratory of a usual professional worker. Third, telephone Identification voices witnesses familiar with the voice of person the identified long has permitted by been the law of Bonomi, Commonwealth. Commonwealth v. 335 Mass. Vehicles, Chartrand Registrar Motor (1963). Expert testimony as to voice identification may also be Thus, admitted. Gerstman, this case Dr. an expert introduced defendant, testified that he compared the two aurally using voices tests, three sorts of not including spectrogram. He further testified he reached firm conclusion that the defendant was not telephone speaker. regard could it be said experts bring, examinations, even to aural knowledge experience in voice identification superior to those lay persons, and to this extent no great departure from traditional evidentiary principles is made when a voiceprint expert’s opinion as to aural analysis, apart from his *10 examination, This, course, is admitted. of depends on the degree to which placed reliance is independent aural
Commonwealth v. worker, can take all examiner, unlike laboratory decision, he can examine he to make his time wants to be necessary, the words he believes and reexamine all Fourth, available for study. if these words are examiner, examiner, unlike the laboratory professional of decisions one of several alternative types can select any to no free to choose to come and, he is most important, testified decision in any given case. Lieutenant Nash 3,300 voices examined forensically that he had them; he had that positively of identifying hope them, had 1,900 of more than eliminated slightly them. more than 500 of identified only slightly positively a third to no decision He thus came almost cases. evaluations on the bearing
All these conclusions have of use those writers who oppose the criticism leveled by most For the identifications. of spectrography judicial error will that the percentage contend part they the laboratory from depart increase when conditions a the extent that world. To considered in the real are case, made, profes- in this that as case convincing than the reliable results more will achieve sional examiner extent examiner, then to that opposition laboratory can be discounted. and Fatalo cases of as in the Frye requirement, assures in the scientific community
general acceptance to assess the validity those most general qualified voice. See the determinative method will have a scientific Addison, C. 743-744 United States 498 F. rule Cir. Additionally, application 1974). who a reserve experts by assuring the parties protects be needed to testify. may en- are both of these goals be
It can argued claim limited number persons when dangered Thus, us in case would issue before in one sense the comparison. to voice identifica- expert an be stated as whether part relied in on a if that received in evidence may tion be reach his conclusion. comparison to as well as aural spectrogram *11 203 191
Commonwealth v. Lykus. knowledge appears special field, in the to be the case spectrography. though with Limited in number the experts may Frye requirement be, the the rule general acceptability opinion, satisfied, is in our if the principle generally accepted by is those would be who expected admitting to be familiar its with use. evi- procedure dence a it called Nalline test was said: experts “No were called defendants and testimony as above summarized stands uncontradicted exception: People’s the record with this Each of experts did admit on cross-examination that medical profession generally is unfamiliar with the use of Nalline truthfully and therefore it cannot be said that the Nalline general acceptance by test has met with the medical profession general acceptance being whole, as a at present specialized limited to those few in a field who problem. deal with the narcotic Should fact render inadmissible? We believe All not. testimony points medical to the It test. generally accepted by has been those be who would expected age to be with its use. In this familiar specialization required” more (emphasis should not be added). People Williams, App. Supp. v. 164 Cal. (1958). 858, 861-862 suggested requirement
It general has been that the acceptance, Frye cases, as in the should be Fatalo modified or McCormick, abandoned.5 See Evidence p. (2d 1972), § 203, said, ed. it where “‘General acceptance’ proper taking scientific is a for condition judicial notice facts, of scientific but not a criterion for admissibility Any of scientific evidence. relevant con- supported by qualified expert clusions which are Commonwealth, essence, Frye suggests Fatalo rule may already holdings have been modified our Devlin, and Common (1974), Gilbert, wealth We Mass. 18 make no comment Devlin the Gilbert case has this time as to application whether in any circumstances other the precise principles than involved these cases. Mass.
Commonwealth o. other unless there are reasons should be received witness suggestionsby and others exclusion.” The this author Mager, Worley (see concurring opinion J., supra, [1972]) urge case, So. 2d that the qualified expert of a should be received and that *12 Frye expressed the considerations similar to those the to and Fatalo cases should be for fact finder as weight opinions. the and value of general no of the
There is need modification uphold principle Frye to of the or cases in order Fatalo judge’s ruling (1) the of the in case. Examination admissibility presented judge, to evidence as before the opinions jurisdictions, (2) judicial (3) from other and writings convincing provides proof to scientific relevant justify of the evidence. The considerable admission proved experiment, greatly reliability by the the Tosi application the of further added induced experienced working the under forensic skills examiner totality conditions, and the of the evidence received hearing dire tended minimize the the voir importance which weight skeptical writings and of or all adverse support general acceptability as serve to a conclusion of required by Frye the rule of the Fatalo and cases.6 in expressed conclusion agree do with the we not While Justice discord” sound the Kaplan’s separate opinion that “turbulence acceptance spectrographie of community’s measure certainly is uniform and total agree we there not analysis, accurately separate opinion of acceptance the method Yet the several in the field. Fatalo expresses experts the doubt of view, unanimity only general Frye require of standard does view is inevitable. divergence of scientific of degree acceptance; experts greater weight to those disposed give In this case we are field of experience empirical who have had direct and empirical have had Among such scientists who spectrography. noted, are, previously of method as experience supportive and are University Phonetics at Dr. Professor of Ladefoged, Peter Black, Speech Professor of Angeles; W. California at Los John Henry Truby, formerly of Dr. University; Sciences at Ohio State Florida; Miami, Oyer and Charles Herbert University of Drs. University, participated Tosi Michigan who Pedrey of State. Mass. 191 o. We hold there error in the admission of was no Nothing opinions the contested Nash. that we Lt. the dis- have said in this is intended to narrow cretionary judge appraising function trial excluding qualifications expert expert witness, of an or in judge’s for other reasons within the traditional area discretion. We add that the admission analysis subject as be should judicial scrutiny, particularly any to the closest case where there is absence an of evidence of voice identifica- voiceprint where, tion other that of than but for voiceprint, there would be insufficient evidence to any guilt. warrant inference of And, the defendant’s voiceprint course, traditional, as is once admitted jury may give weight they in evidence the it such proper. deem
We have reviewed the entire record in accordance with our duties G. 278, 33E, under L. c. § and we observe nothing justice modify which in indicates that we should jury. the results reached the
Judgments affirmed. (concurring). In Quirico, Commonwealth v. A J. expressed Juvenile, (1974), I the dissenting opinion, view, polygraph in a that the art of testing yet point had not advanced the that the polygraphic admission of should, under imposed by majority limitations court, be left study; Steer, and Dr. Max D. of Speech Professor Sciences at Purdue University, who as acted consultant the Tosi study. We are further buttressed in our decision fact “judgment” of experts quoted six some length separate in the opinion as critical of the study conclusions claimed in the Tosi consists in substance of a theoretical of the Tosi report review is in the merely page form of a four letter in 54 Soc. Acoustical Am. 531 Thus, See supra. fn. find presented we the evidence in support the'reliability of voiceprints, expressed in particularly as Dr. Tosi’s study, sufficiently persuasive outweigh the criticism expressed by certain other scientists in field of acoustics.
Commonwealth o.
predi
judge. This
to the discretion of the trial
view was
large part
that,
because of “an
cated
reason
infallibility”
impenetrable
sur
almost
aura of scientific
grave
rounding
polygraph machine,
risk
there was “a
regard
[polygraphic] opinion
jurors
that the
will
. .
.
testimony
resolving
question
the ultimate
of the de
considering
guilt
polygraphic
innocence,
fendants
testimony
contrary
they
and,
evidence as
do other
weighing
instructions, not in
it with all
to the court’s
fact
presented,”
(Quirico,
the other evidence
ment to the was part on visual evidence the voice identification based 207 Commonwealth v. conclusion can rest
analysis speech spectrograms. My on the v. Frye standard simply proposition States, United Cir. 293 F. 1014 C. has 1923), not been met: method the Tosi used spectrogram by cannot, case group employed my present view, be said derived to be from a scientific principle established to have ac- “sufficiently gained general in the field which it ceptance particular belongs.” ' Some is useful here. Before Dr. Tosi’s chronology work, had not been speech spectrography sufficiently validated to its results for forensic use.1 qualify Technical Committee on Communication Speech Acoustical had America six Society authori- requested ties in this in the field of acoustics to country study — In 1970 the question. Bolt, R. F. E. experts Cooper, — David, Denes, P. Pickett, K. Stevens their gave J. answer in an article published Journal It was Society. negative. concluded “the They available results are to establish the inadequate of voice identification by spectrograms. We believe conclusion shared by most scientists who are knowl- about edgeable hence, speech; of them many are deeply concerned about use of spectrographie evidence in the courts.”2
Dr. Tosi carried out his after this experiments unfavor- able about the state of art. He and co- judgment his Lashbrook, workers W. (H. Oyer, Nicol, C. Pedrey, J. and E. their results in Nash) published 1972.3 Thereafter Carey, People King, Cal. State 1 See 56 N. v. 266 (1970); J. Wright, see United States App. 2d 437 But (1968). U. A. (1967). S. C. M. 2Speaker Identification A Speech Spectrograms: Scientists’ View of Reliability Legal Its Acoustical Am. Purposes, Soc. J. criticism, ones, challenged This all (1970). subsequent like not the technique recording but speech spectrograms rather validity spectrograms conclusions drawn the examiner of upon viewing and comparing them. Identification, Experiment on Voice Acoustical Am. Soc. *15 v. identifications in evidence admitted number of courts a Dr. Tosi method. appraised based spectrograms See in these cases. witness himself was principal Tosi Hedman, Minn. rel. Trimble v. ex (1971); State Fla. Ct. State, 2d 613 App. 263 So. (Dist. v. Worley State, So. 2d Alea v. corroboration); 1972)4 (use Hodo Worley)’, Fla. 1972) (following Ct. App. 96 (Dist. Contra, Court, Cal. 3d App. v. Superior Marin 4516, Cty., Ct. Chapter, Super. No. People however, said, that be It can Cal., fairly 1973. July community decided the cases were when the work and Dr. Tosi’s time to study had sufficient had not over pre- advance as to its possible reach conclusions Law, 40 Cal. field. See People work in the vious reflected thus The decisions 81-82 (1974). App. community the relevant scientific less a consensus an absence than was acceptable, the Tosi method be based one could informed opinion on which an study the other. or way case, the scientif- time of the trial present
theBy
to me
react,
it
appears
had begun
ic community
Thus
negative.
has been prevailingly
the reaction
work,
Dr. Tosi’s
reviewing
the 1970
the authors of
paper,
difficulties5 and
several
troubled by
declared
were
they
follows:
their
as
summed up
judgment
Worley cases were
decided before the
Although the Trimble
results,
his
Dr. Tosi
testified
about
of Dr. Tosi’s
publication
cases,
courts relied on that
and both
in both
completed experiment
comparisons.
in-
accepting
comparisons
increase in error rate
mentioned substantial
They
times,
comparisons
against
different
as
taken at
samples
of voice
when
increase in error
rate
substantial
samples;
contemporaneous
comparison,
against
were used for
from sentences
words taken
isolation;
problems
to consider
failure
spoken
words
levels,
voices,
changes of voice
mimicking
disguising
rate as
increase of error
speaker;
state
changes of emotional
to unknown
which match
increased from
samples were
number of
be made.
voice was to
*16
Commonwealth o. study improved understanding “The Tosi has our problems some spectro- of the of voice identification from grams by indicating important the influence of several accuracy variables on the of identification. In uncover- ing factors that tend to errors, increase identification study given however, the has not us a definitive answer question: reliably person ‘How can a be identified by examining spectrographie patterns speech of his laboratory sounds?’ Under certain conditions and for sample population, some probability selected making an error in identification can be stated. But for the less-than-ideal conditions encountered in forensic situations, the probability indications are of error substantially. will increase Further needed, studies are particular with attention to the examiner’s decision speaker criteria, population, the selection of the time lapse samples, background-noise between voice condi- tions, psychological and the speaker. condition of the lawyers,
“As scientists rather judg- than we offer no ment speech as to spectrograms whether or to what extent should be used for identification in the courts. We wish point present out that methods for such use lack adequate an estimating basis for many practical laboratory situations and that evaluations of these increasing methods show errors as the conditions for evaluation move toward hope real-life situations. We explanations that our of some of the factors that affect speaker provide legal profession identification will helpful with judg- information on which to base its own concerning admissibility ments method.”6 group, seeking reply
Dr. Tosi and his to the criticism comparisons that forensic would tend to be less accurate aseptic than experimental those made under the more pointed conditions, to some factors which in their 6Speaker Identification Speech Spectrograms: Some Further Observations, 54 Acoustical Soc. Am. 533-534
Commonwealth v. direction.7 The dispute worked opposite quite not resolved. the Tosi from outside recent scientific writing
More — voir below at not available to the writing judge group — method,8 at although is unfavorable to the Tosi dire better has become Dr. P. Ladefoged, least one authority, first to it after being negative.9 disposed *17 “Speaker by Speech Spectrograms: to Identification Reply See Observations,” Soc. Am. 535 (1973). Some Further 54 Acoustical J. as forensic Lt. Nash was allowed to make aural situation voices, longer himself a comparisons well as visual to allow allowed examiners in of time to make a decision than was period felt, conclude, if he so that he could not experiments, the Tosi facts, together with fact that Lt. Nash reach a decision. These training experience spectrographie had much more had study, used in the Tosi have led Dr. identification than the examiners critics, assert, comparisons by his that forensic reply Tosi to less, more, than training a of Nash’s are accurate person Nash or error except But for that the experimental comparisons. when in express as examiner is allowed to no rate declines doubt, an appear experimentally does not to be Tosi’s assertion Moreover, depends Nash on aural supported. to the extent conclusions, should be accorded weight for his comparisons analysis, which more “scientific” accompanying, his large jury, questionable. could have influence with a becomes present Dr. voir dire in the case that if confronted Tosi stated at in a compare whispered telephone with a need to a conversation — noisy recording quiet laboratory in a environment with later — “rely very my in our case he would much on actual situation ears.” LaRiviere, Speaker from Turbulent Portions C. Identification Hollier, Fricatives, H. Peculiar Case of (1974); 29 Phonetica “Voiceprints,” Acoustical Soc. Am. 210 56 J. generally opposed review is also to the use commentary Recent law Identification, note, speech Compare Voiceprint spectrograms. — Danger Voiceprints Georgetown (1972-1973); Jones, L. Thomas, Ahead, Voiceprint (1973); 11 Am. Crim. L. Rev. 549 — Valley Fernando Myth Eyes It), Miracle Have 3 U. of San (The — Boren, Voiceprint with (1974) (opposed use), L. Rev. 15 Comeback, (1974) L. Rev. 1 Valley of San Fernando Staging a 3 U. favor of (in use). Addison, 2d 741 498 F. in United States McGowan Judge as position Ladefoged’s below), took Dr. 1974) (discussed C. Cir.
Commonwealth t>.
To sum
method;
divided on the Tosi
up, opinion
material
shows
journal
and discord rather
turbulence.
than that “general
which the
acceptance”
case
Frye
lays
down as a
precondition
Nor
it be
can
admissibility.
said that
those
plausibly
with adverse views are either
to have
unqualified
or are
worthy
respect
to the relevant scientific
strangers
“field.”10
A decision of the United States Court of
Appeals
the District of Columbia Circuit
rendered
on June
1974, reviews the current situation and
the results
rejects
of speech spectrogram
analysis
evidence.
United
Addison,
States v.
Commonwealth o. is United States ting speech spectrogram 10, 1974), D. Pa. F. Sample, Supp. (E. June revocation hearing it at a probation which allowed there on could succeed that reasoning prosecution without proof beyond of the evidence” a “preponderance was that no evidence doubt; but it reasonable appears Nash, Dr. of Lt. to counter that the defense presented by coworker. Tosi’s of the specific expert
I have taken account considered below, but, together when especially offered not, it, I make a it does see writings, with the journal I less take satisfies the Frye requirement. case that the results from the fact that than brethren comfort my to the not so close are analysis speech spectrogram results of trial as the of a criminal ultimate probanda are, whence it is suggested analysis typically polygraph is not so the function jury in the former case since I should add that invaded. seriously McCormick’s prop- Professor court, referring while not enter it, upon I do not embrace osition, does admissibility that I would have about reservations thesis were to be accepted if the McCormick case present here. and applied the law might belief my elsewhere indicated
I have a “commission” procedure proceed by preferably methods of new validating handle questions use in a intended for demonstration measurement Juvenile, Av. See court room. *19 on needed were dire further studies at voir that Tosi conceded Dr. critical papers The voices, voices. including whispered disguised point. same have made the Tosi did not case present Nash be observed also that should It only with samples but numerous voice with unknown compare the Nash Although police. by the samples furnished Lykus subject is to his conclusion impartiality, undoubtedly strove assumption any unintended) due to if (even bias possible discount for guilty they had the thought police part his that or realization man. Lumley. Commonwealth 452, 421, Mass. 452-453 (1974) (Kaplan, J., dissenting). seems to me nice case illustration of the present that, it be noted a “commis- point. Lastly may failing sion” it was the trial procedure, open judge, might have been of to him in which he help quandary himself, found to call of whom a experts, independent few with could have been very high qualifications located within a short radius of the court house.
IAs am outvoted on the I question admissibility, need not to answer the difficult whether attempt question the erroneous I (as judge reception it) would reversal of the conviction in require light rest of the proof. Lumley. vs. Clifton January
Suffolk. 1975. March 1975. C.J., Present: Tauro, Reardon, Braucher, Quirico, Hennessey, Kaplan, & Wilkins, JJ. Practice, Criminal, jurors, Examination of Fair trial. Constitutional Law, process Due of law. its decision in Ross, This court reaffirmed Commonwealth v. (1973), Mass. 665 questions to prospective jurors in criminal designed trials possible to discover racial are prejudice constitution- ally mandated when the defendant a “special target for racial prejudice.” [214-216] A motion by the defendant at a criminal trial the judge inter- rogate prospective jurors to possible prejudice by racial specific if, questions granted should be the judge after ascertains that defendant’s decision is knowingly and voluntarily made and with an understanding possible consequences, the defendant insists upon questions. [216-217] A black defendant convicted under an indictment charging robbery of
a white woman was not denied due process under the Fourteenth Amendment United States Constitution the refusal of the judge trial to put questions to the venire on voir relating dire possible case, racial prejudice where the facts of the and the al-