JONES, Chief Justice.
Appellant, Adam Andrew Topa, was tried before a jury and convicted of murder in the first degree. From his conviction he filed a motion for new trial delineating fifteen reasons in support thereof. The court en banc denied appellant's motion. On this appeal, appellant urges us to reexamine six issues which he believes warrant reversal and command a new trial. Five of these arguments relate to the use made by the Commonwealth of a tape recording of a telephone call received by the police from an individual who identified himself as "Roger Ferretti" and claimed to have committed the murder for which appellant was convicted. Because we find that the trial court erred in allowing expert testimony concerning spectrograph, or voiceprint, analysis of the recorded telephone call, we reverse the judgment of sentence and remand for a new trial.
The five arguments are: (1) that the trial court erred by allowing a recording of a telephone call made to the Scranton police and recorded without the caller's knowledge or consent to be introduced into evidence at appellant's trial; (2) that the trial court erred by requiring appellant to make a second recording for the purpose of comparing his voice with the voice of the person who called the Scranton police; (3) that the trial court erred by allowing expert testimony based upon voice print comparison tests; (4)that the trial court improperly limited cross-examination of a Commonwealth witness concerning a recording of appellant's voice made without his knowledge or consent while he was being interrogated by state police; and (5) that the trial court erred by denying appellant's motion for a mistrial when it appeared that the Commonwealth had twice deceived both the trial court and defense counsel concerning the existence of the interrogation recording.
The third argument is meritorious. Because we reverse on the basis of that allegation we need not address the remainder. However, the factual allegation that the Commonwealth deceived both the defense and the trial court is verified by the trial court's opinion dismissing on the merits, post-trial motions which raised all of the issues concerning the recordings urged here. Appellant argues that he was prejudiced by the Commonwealth's refusal, despite the trial court's mandate, either to reveal the existence of the interrogation tape or to turn over to it the first, inconclusive voiceprint comparison test. This claim has substantial appeal, for the interrogation recording which was withheld by the Commonwealth might have made a powerful exhibit for the cross-examination of the voiceprint expert. We need not reach this issue, however, for the reason stated above.
This case begins on Sunday, October 22, 1972, at about 12:40 p. m. with the discovery, by two boys hiking through a remote area near St. Patrick's Cemetery in Blakely, Pennsylvania, of the body of a woman. One of the boys ran to his grandmother's house to alert his father who was a medical doctor. Upon their return, the doctor made a cursory examination of the body and determined that the person had expired. Shortly thereafter, several police officers arrived as did the coroner of Lackawanna County. The coroner identified the decedent as Mary Ellen Walsh.
The following is a summary of the evidence of the crime introduced at trial. It is presented in the light most favorable to the Commonwealth as verdict winner. See Commonwealth v. Robson, 461 Pa. 615, 625, 337 A.2d 573, 578 (1975); Commonwealth v. Boyd, 461 Pa. 17, 334 A.2d 610 (1975); Commonwealth v. Murray, 460 Pa. 605, 334 A.2d 255 (1975):
(1) Four witnesses saw appellant with the victim on the evening before her body was found. Two of the witnesses described the victim's clothing and identified her coat in court. One of the witnesses identified appellant's jacket in court.
(2) The victim's body was discovered in an area which was accessible only by means of a rutty, lightly traveled road composed of mud, stone, rock and shale.
(3) On the day the victim's body was discovered, appellant was observed, by several policemen, driving past the crime scene. When he saw the officers he rapidly left the area.
(4) About a quarter of an hour later the Scranton police received a telephone call from "Roger Ferretti". This call was taped in the normal course of police operations. This caller claimed to be telephoning from a local VFW post, and also stated that he had just stabbed a woman in the cemetery. An expert testified that, on the basis of voiceprint comparison tests he had performed, appellant was the caller.
(5) Mr. Roger Ferretti testified that he did not make the call to the Scranton police, and further testified that at about 1:30 p. m. on the day the call was made, appellant was in the VFW post staring at him strangely. Ferretti stated that when he asked appellant why he was staring at him, appellant replied that he was mad at Ferretti.
(6) Approximately one hour after the call was made, appellant was seen by the state police in the VFW post with what was suspected to be a bloodstain on his jacket sleeve. He was taken into custody and interrogated.
(7) Blood on appellant's jacket sleeve was inconsistent with his own blood, but was consistent with the victim's blood.
(8) Wool fibers taken from the bloodstain on the jacket were identical with fibers taken from the victim's coat. An expert testified that the fibers were located in the stain in such a way that they could only have been deposited in the stain when the blood was still wet.
(9) Appellant was missing a button from his jacket. A button found near the body was identical to the remaining buttons on appellant's jacket.
(10) The victim's body revealed thirty-five stab wounds on the right side of her trunk, one of which punctured a main vein of the body. These wounds were the immediate cause of death.
As set forth above, on October 22, 1972, a person called the Scranton Police Department, identified himself as "Roger Ferretti" and made a report saying he had stabbed a woman and indicating where her body could be found. This call was taped by the Scranton police. Roger Ferretti was located and denied making the phone call. After appellant's arrest, the Commonwealth petitioned the lower court to require the appellant to make a recording via the telephone wherein he would repeat some of the same words as in the October 22 call to the police. The purpose of this request was to enable the Commonwealth to submit both tapes to a voiceprint expert in order to determine whether the appellant was the original caller.
The court below granted the Commonwealth's petition and another tape recording of appellant's voice was produced. This recording, along with the recorded October 22 telephone call and a tape recording of the known voice of Roger Ferretti, was then subjected to voiceprint analysis by the Commonwealth's expert in the field of spectrography, Lieutenant Ernest W. Nash of the Michigan State Police. From a comparison of spectrograms made from the tapes, Lieutenant Nash concluded that the telephone call into police headquarters was definitely not made by Roger Ferretti, but was definitely made by the appellant.
On the witness stand, Lieutenant Nash testified that he was employed by the Michigan State Police as the officer in charge of the Voiceprint Identification Unit of the Scientific Crime Laboratory stationed at East Lansing, Michigan. He was selected in 1966 to be trained as an analyzer of spectrograms because of his background in electronics and his training in fingerprint analysis. He studied spectrography in 1967 under Mr. Lawrence Kersta, the acknowledged innovator in the field, and since the fall of 1968, has studied audiology and speech sciences at Michigan State University. His academic advisor and teacher there has been Dr. Oscar Tosi who has conducted extensive experiments with the sound spectrograph and testified as an expert on voiceprint analysis in many cases. At the time of the trial, Lieutenant Nash had been analyzing spectrograms for upwards of six years, had examined more than three thousand voices and had analyzed in excess of 180,000 voiceprints.
Spectrograms, or voiceprints, are the product of the sound spectrograph machine which was described by Lieutenant Nash in his testimony:
"The sound spectrograph, for one part of it, it is a regular tape recorder. It has the ability to record and play magnetic tape recordings. There also is a bank of filters and a variable filter with a pickup head that analyzes a 2.4 second segment of the tape at one time, and during the analysis cycle, the magnetic information on the tape is picked up by the scanning head and it is transformed into electrical energy which is passed through the bank of filters and brought out to a stylus, or a piece of wire, that is writing on the spectrogram paper during the analyze cycle, and as the paper revolves in front of that piece of wire, wherever energy is present, there is an electrical fire or spark that comes off the wire that burns the pattern onto the paper
making a permanent record." Notes of Testimony, p. 405.
Spectrograms are compared, in much the same way as are fingerprints, for points of similarity which identify the speaker. The theory behind voiceprint identification was also explained by Lieutenant Nash:
"As each one of the ridges on your fingers or on the palm of your hand differ from each other, so do all of the other parts of your body. They are unique to you and . . . include your voice mechanisms. So the theory is that if you use unique voice mechanisms to produce the sounds of your voice, then the sounds will also be unique. Then by taking a recording of your voice and subjecting it on one occasion to analysis by the sound spectrograph, and then doing the same thing on another occasion with your voice, subjecting it to the same process, that those uniquenesses will be displayed on the pieces of paper or the voice prints."
Id., pp. 427-428.
Although Lieutenant Nash is convinced of the validity of the theory and the reliability of the spectrographic process, his certainty is not shared by all commentators and scientists in the field who have emphasized, for example, the dissimilarities between fingerprints and spectrograms. A concise statement of those distinguishing factors is contained in Note, Voiceprint Identification, 61 Geo.L.J. 703, 725 (1973):
"An individual's fingerprint ridges do not vary. Speech, however, has not been proven to be invariant. Fingerprints cannot be disguised; speech can. Further, fingerprints do not change with age or illness. Whether voices change under these circumstances is uncertain. A further significant difference is that fingerprints are primary evidence. Voiceprints, on the other hand, are graphic representations of a speaker's voice. Thus, unlike fingerprints, a voiceprint can be produced only through the intervention of a mechanical
device, the sound spectrograph. Consequently, the spectrographic pattern, unlike the fingerprint, can be affected by such extraneous factors as distortions of frequency, energy, or the time involved in the transmission of sound."
See also Jones, Evidence vel non, The Non Sense of Voiceprint Identification, 62 Ky.L.J. 301, 318 (1974).
They have also noted the similarities which are present in both spectrograph and polygraph testing.
In both instances the probative value of the expert opinion testimony offered hinges on the competence of the person who administers the test and interprets the results, as well as on the reliability of the mechanical device employed. Lieutenant Nash agrees that the interpretation of spectrographic test results is subjective. Moreover, the process demands the services of individuals with highly developed skills in the areas of audiology and speech sciences and electronics. Thus, as with the lie detector, there is the danger that the trial judge or jury will ascribe a degree of certainty to the testimony of the expert spectrography witness which may not be deserved.
See Jones,
supra, 62 Ky.L.J. 301, 305; Note,
supra, 61 Geo. L.J. 703, 726; Note, 12 N.Y.L.F. 501, 507 (1966).
Dr. Tosi, himself, has observed, while testifying, that voiceprint identification is more difficult than fingerprint identification because of the inability, at this time, to classify voices:
"We do not classify voices. In my dream, it is my dream to have a file of voices and when you have a voice of the offender you go to the file and select ten or fifteen that have the same classification and then try to see if they are known within this or not. The same as fingerprints. In the context, yes, I think more studies have to be done."
Worley v. State, 263 So.2d 613, 618 (Fla.App. 1972) (dissenting opinion).
The results of a polygraph examination are generally inadmissible in Pennsylvania. Commonwealth v. Gee, 467 Pa. 123, 354 A.2d 875 (1976); Commonwealth v. Brooks, 454 Pa. 75, 309 A.2d 732 (1973).
In his testimony, Lieutenant Nash discounted the problems which have been raised in the law review articles noted above. He asserted that mimicry of voices will not affect the accuracy of voiceprint identification, that the voice analysis is made more reliable by the aural examination of the tape recordings in conjunction with spectrogram interpretation, and that there is no possibility of two different people producing the same spectrogram. We do not question the sincerity of Lieutenant Nash's testimony and we respect his considerable expertise in the area of spectrography. But his opinion, alone, will not suffice to permit the introduction of such scientific evidence into a court of law. Admissibility of the evidence depends upon the general acceptance of its validity by those scientists active in the field to which the evidence belongs
"Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery,
the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs."
Frye v. United States, 54 App.D.C. 46,
293 F. 1013, 1014 (1923) (emphasis supplied).
See also Commonwealth ex rel. Riccio v. Dilworth, 179 Pa. Super. 64, 115 A.2d 865 (1955). We agree with those courts which have strictly applied the
Frye standard in considering the introduction of voiceprint evidence that it is not, as yet, admissible,
United States v. McDaniel, 538 F.2d 408 (D.C. Cir. 1976);
United States v. Addison, 498 F.2d 741 (D.C. Cir. 1974);
People v. Kelly, 17 Cal.3d 24, 130 Cal.Rptr. 144, 549 P.2d 1240 (1976);
People v. Law, 40 Cal.App.3d 69, 114 Cal.Rptr. 708 (1974);
People v. King, 266 Cal.App.2d 437, 72 Cal.Rptr. 478 (1968);
State v. Cary, 99 N.J. Super. 323, 239 A.2d 680 (1968),
aff'd 56 N.J. 16, 264 A.2d 209 (1970);
contra, Commonwealth v. Lykus, 327 N.E.2d 671 (Mass. 1975);
State v. Olderman, 44 Ohio App.2d 130, 336 N.E.2d 442 (1975).
But see, Commonwealth v. Lykus, 327 N.E.2d 671, 679 (Mass. 1975) (Kaplan, J., separate opinion).
Strict application of the Frye standard when scientific proof is offered is essential if the defendant is to receive a just and fair trial:
"The requirement of general acceptance in the scientific community assures that those most qualified to assess the general validity of a scientific method will have the determinative voice. Additionally, the
Frye test protects prosecution and defense alike by assuring that a minimal reserve of experts exists who can critically examine the validity of a scientific determination in a particular case. Since scientific proof may in some instances assume a posture of mystic infallibility in the eyes of a jury of laymen, the ability to produce rebuttal experts, equally conversant with the mechanics and methods of a particular technique, may prove to be essential."
United States v. Addison, 498 F.2d at 744.
See also People v. Law, supra, 40 Cal.App.3d at 75, 114 Cal. Rptr. at 712;
People v. King, supra, 266 Cal.App.2d at 461, 72 Cal.Rptr. at 493;
Jones, supra, 62 Ky.L.J. at 306. The testimony of one expert, Lieutenant Nash, cannot satisfy this standard. Furthermore, we must conclude, from our study of those cases and commentaries which have addressed the issue, that the reliability of the sound spectrograph and voiceprint identification has not, as yet, been generally accepted by the scientific community concerned with acoustical science. In 1970, for example, a panel of speech scientists reported that the majority of professionals in their field were "deeply concerned about the use of spectrographic evidence in the courts":
[T] he available results are inadequate to establish the reliability of voice identification by spectrograms."
Bolt, Cooper, David, Denes, Pickett Stevens,
Speaker Identification by Speech Spectrograms: A Scientist's View of its Reliability for Legal Purposes, 47 J. Acoust.Soc.Am. 597, 603 (1970).
In the wake of further voiceprint experimentation, the same panel of experts declined to alter their position:
"Our interpretations of the new data lead us to reiterate our previous conclusion: that the degree of reliability of identification under practical conditions has not been scientifically established."
Bolt et al., Speaker Identification by Speech Spectrograms: Some Further Observations, 54 J.Acoust.Soc.Am. 531 (1973).
These reports are supported by the testimony of scientists who testified as experts in other cases involving voiceprint identification. United States v. Addison, supra, at 745; People v. Law, supra, 40 Cal.App.3d at 80-81, 114 Cal.Rptr. at 715-716; State v. Andretta, 61 N.J. 544, 548-549, 296 A.2d 644, 646-647 (1972); see also Note, supra, 61 Geo.L.J. 703 (1973). Thus, it was error to permit the introduction of voiceprint evidence at appellant's trial.
Trial error is not harmless "unless it is clear beyond a reasonable doubt that it did not affect the result." Commonwealth v. Jones, 233 Pa. Super. 52, 59, 335 A.2d 444, 447 (1975). Because the prosecution's case against appellant was built on circumstantial evidence, we cannot conclude beyond a reasonable doubt that the admission of scientific evidence, the effect of which was to present before the jury a confession to the crime, did not affect the result. The error was prejudicial and necessitates a new trial.
Judgment of sentence reversed. Case remanded for a new trial.
ROBERTS, J., filed a concurring opinion.