Nasrallah v R; R v Nasrallah [2015] NSWCCA 188 | Case summary

Ad hoc expert voice identification evidence

The first ground of appeal is: the trial judge erred in admitting the evidence of Federal Agent Succar as ad hoc expert voice identification evidence.
It is helpful to begin with an analysis of the way in which the issue of the proposed expert evidence emerged at the trial.
At a pre-trial directions hearing a week before the trial, the Crown raised a number of issues that would have to be determined by the Court before a jury was empanelled. One was an issue raised by the appellant as to the admissibility of the 21 DHL calls and, in particular, whether the Crown could establish that they had been lawfully recorded. There was no suggestion at that stage that if those recordings were admitted there would be expert voice identification evidence comparing the voice of the caller in those calls with the recordings of the appellant in his calls from the gaol. On the contrary, although the Crown foreshadowed tendering the calls from the gaol for that purpose, she said the comparison would have to be made by the jury, explaining "there's no witness. It wasn't a case of a long-running investigation" (see transcript of 22 October 2013, T3.34).
On the Friday before the trial was due to commence, the trial judge held a voir dire to determine the pre-trial issues including the admissibility of the DHL calls. During the hearing on the voir dire, the Crown began to lead oral evidence from Federal Agent Succar as to his comparison of the 21 DHL calls with the calls from the gaol. The barrister then appearing for the appellant objected and the judge noted what the Crown had said on the earlier occasion (that it was going to be a matter for the jury to compare the voices). The Crown evidently indicated her intention at that point to use Federal Agent Succar as a voice identification expert. The transcript records that the barrister appearing for the appellant had not been informed of that intention until that morning. He had accordingly not had any opportunity to prepare to argue the admissibility of the alleged expert evidence. For that reason, the argument had to be stood over until the following Monday.
Before the continuation of the voir dire hearing on the Monday, the appellant's barrister noted that the Crown had not served any supplementary statement from Federal Agent Succar setting out the proposed expert evidence. The matter was stood down for that to occur. A statement was prepared during a short adjournment and was tendered on the voir dire (statement dated 28 October 2013, identified in the transcript as exhibit 6 on the voir dire but described in the appeal books as exhibit E on the voir dire).
In his supplementary statement, Federal Agent Succar stated that he is "an Australian with a Lebanese background". He described his experience of the Arabic language, both at home and in the performance of his duties as a police officer. He also said that he had studied Arabic as an elective subject in 2004 when he completed a degree in business and computing. It should be noted, however, that none of the disputed calls was in Arabic. All of the calls are in English. The Crown's contention appears to have been that Federal Agent Succar had relevant specialised knowledge in the identification of the accent of Arabic speakers when speaking English.
In his original statement, Federal Agent Succar had evidently purported to identify the voice of the appellant on the DHL calls after listening to only four of the calls from the gaol (T32.25). In his supplementary statement, he said that, since making the original statement, he had listened to the DHL calls and about 142 of the gaol calls "on multiple occasions". He said: I can say that I recognised that the voices heard in the 21 calls to be that of [the appellant]. The pronunciation and the use of some words, such as "youse" occurs in come (scil: some) of the DHL calls as well as some of the 142 calls from NSW Corrective Services".
The Crown expressly relied on s 79 of the Evidence Act, which provides: "If a person has specialised knowledge based on the person's training, study or experience, the opinion rule does not apply to evidence of an opinion of that person that is wholly or substantially based on that knowledge."
The reference to the opinion rule is a reference to s 76 of the Act which provides: "Evidence of an opinion is not admissible to prove the existence of a fact about the existence of which the opinion is expressed."
The Crown did not rely on s 78 of the Evidence Act (relating to lay opinions).
The argument before the trial judge proceeded on the common assumption that Federal Agent Succar's supplementary evidence was evidence of an opinion that the voice of the caller in the 21 DHL calls was the voice of the appellant. The evidence was sought to be admitted to prove the existence of that fact. In accordance with s 79, the evidence was accordingly not admissible unless Federal Agent Succar had specialised knowledge based on his training, study or experience and the opinion was wholly or substantially based on that knowledge.
On the present state of the law, it is accepted that, for the purpose of s 79 of the Evidence Act, an expert witness can be one who has acquired his or her specialised knowledge ad hoc (that is, for the very purpose of the legal proceedings in which the opinion evidence is relied upon): R v Leung & Wong [1999] NSWCCA 287; 47 NSWLR 405 at [40] per Simpson J (as her Honour then was); Spigelman CJ and Sperling J agreeing at [1] and [67] respectively. The correctness of that view was accepted in Irani v R [2008] NSWCCA 217 at [14] per Hoeben J (as his Honour then was) at [14]; McClellan CJ at CL and Harrison J agreeing at [1] and [34] respectively.
The trial judge ruled that the evidence was admissible under s 79, relying on those decisions and the authorities cited therein. Her Honour remarked that the activities undertaken by the witnesses in those cases to acquire their ad hoc expertise "do not provide a minimum test for the admissibility of voice identification evidence".
The trial judge also cited the remarks of Wood CJ at CL to like effect in R v Madigan [2005] NSWCA 170 at [92], where his Honour said "there are no preconditions for the admissibility of voice identification evidence, other than relevance under s 55 of the Evidence Act" (citing R v Riscuta and Niga [2003] NSWCCA 6 at [34] per Heydon JA). However, the remarks of Heydon JA in Riscuta (and the decision of R v Adler (2001) 52 NSWLR 451 cited by his Honour as the authority for that proposition) were not concerned with voice identification evidence given by an ad hoc expert. They were cases in which the identification in question was made by a person who had some prior familiarity with the voice the subject of the identification.
In the argument in the present case, the appellant had also submitted that, if admissible under s 79, the evidence should be excluded under s 137 of the Evidence Act. The judge rejected that submission but noted that the jury should be given appropriate warnings in accordance with ss 116 and 165 of the Evidence Act.
The appellant submits that the Court's recognition of ad hoc voice identification expertise is wrong but no notice was given that a bench of five should be convened in order to consider whether Leung and Irani were wrongly decided. In light of those authorities, the submission must be taken to have been put only formally in this Court. I am of the view, however, that the evidence of Federal Agent Succar was not admissible under the principles stated in those cases.
It is important to be clear about the circumstances in which the witnesses in Leung and Irani were accepted as having ad hoc specialised knowledge (within the meaning of s 79) so as to warrant admitting evidence of their opinions concerning the voices in question.
In Leung, police had covertly recorded conversations between three men while they opened packages imported from Bangkok. Parts of the conversations were in foreign languages. At the trial, the recordings of the conversations (referred to as the DAT tapes) were admitted, together with a transcript of a translation prepared by a qualified interpreter. During the preparation of the transcript, the interpreter had discerned three different male voices, which he called M1, M2 and M3. His basis for distinguishing the three different voices was described in the judgment of Simpson J at [18] as follows: He reached this conclusion by evaluating the pitch and the volume of the voices, and the speed of the language used. He said the speaker he nominated as "M1" was a Cantonese speaking person with a medium to high pitched voice and spoke relatively quickly. [In the evidence he gave before the jury, he described this voice of as medium pitch]. "M2" was a person who spoke accented Cantonese as though that language was not his mother tongue, also in a medium to high pitched voice, but at a relatively low volume. This speaker spoke some sentences in Mandarin. "M3" was a Cantonese speaker with a relatively high pitched voice.
At a later point in time (a week before the trial), the interpreter was asked to listen to tape recordings of conversations between each of the three appellants and Federal police and to compare the voices on those tapes with those on the DAT tapes in order to express an opinion as to the identity of the voices on the DAT tapes. The interpreter concluded that the voice he had nominated as "M1" was the voice of Leung and the voice he had nominated as "M3" was the voice of Wong.
Justice Simpson's conclusion that the interpreter could give evidence as an ad hoc voice identification expert was subject to an important qualification expressed at [44] to [45] of her Honour's judgment, as follows: 44. Voice comparison is not necessarily a question for expert evidence, although it may be. If the two sets of tape recordings in the present case had been in English, it would have been open to the Crown to have left it to the jury to make their own comparison and assessment of whether the voices on the DAT tapes (or any of them) corresponded to either of the voices on the police tapes. That course theoretically remained open but would have left the jury with a task immeasurably more difficult, given the reasonable assumption that no member of the jury understood either of the Chinese languages involved. The jury would, truly, have been comparing voices only, without the intrusion of language and speech patterns that are part of voice identification. 45. There is another aspect to the task undertaken by Mr Fung that should be mentioned. He was not asked to compare the voices until a few days before the commencement of the trial. There was no specific evidence as to the instruction he was given, but it is an obvious inference that he would have approached his task on the assumption that the two voices on the police tapes were in fact the same as two of the voices on the DAT tapes, and that his role was to determine which voice on the police tapes corresponded to voices identified as M1, M2 or M3 on the DAT tapes. This is a quite different task to determining whether either of the voices so corresponded. If there were any real basis to doubt the assumption, the manner in which Mr Fung was asked to perform the comparison might raise real questions of propriety. The situation is analogous to physical identification by photographs or by a police lineup, in which care must be taken not to suggest that a particular person is the suspect. However, for reasons which will appear below, I am satisfied that in this case the assumption was a valid one. It was therefore proper for Mr Fung to approach his task on the basis that two of the voices on the DAT tapes did in fact correspond to the two voices on the police tapes, and his function was to determine which was which.
Although the decision in Leung stands as authority for the proposition that a person who has repeatedly listened to voice recordings can give evidence as an ad hoc voice identification expert, that conclusion was heavily qualified by those remarks. Further, as explained at [42] of the judgment, the task undertaken by the interpreter in that case was highly specialised, invoking the interpreter's considerable specialised knowledge in foreign language and accent.
The judgment of Simpson J provides a helpful analysis of the provenance of the notion of an "ad hoc expert" which Her Honour noted appears to have originated in the New Zealand decision of R v Menzies [1982] 1 NZLR 40. No question of voice identification arose in that case. Evidence had been admitted in the trial of tapes undoubtedly recording the voice of the appellant. The issue with which the relevant ground of appeal was concerned was the admissibility of a transcript of those tapes prepared by a police officer. The trial judge had initially refused to admit the transcript. However, after listening to two tapes, the judge changed his ruling because, on being listened to only once or twice, the tapes were unintelligible. The police officer, who had listened to the tapes "over and over again", was accepted by the trial judge to have acquired "a special expertise in their interpretation". The judge allowed the jury to have the transcript. The Court of Appeal upheld that ruling. A premise of the Court's decision was that, in the case of a recording in a foreign language or where deficiencies in the recording "make it necessary to play tapes more than once to enable a better understanding", while there should normally be at least one playing of the tapes to the jury, "the evidence of an expert should be admissible as an aid". It was on that premise that the Court said "he may be a temporary expert in the sense that by repeated listening to the tapes he has qualified himself ad hoc."
Justice Simpson noted that the idea of an ad hoc expert was subsequently endorsed by the High Court in Butera v DPP [1987] HCA 58; (1987) 164 CLR 180. Again, that was a case concerned with the admissibility of a transcript of a recording rather than with voice identification evidence. The appellant was convicted of conspiracy to traffic heroin. The Crown case included a tape recording of a conversation among some of the alleged co-conspirators which was mostly in Punjabi but partly in English and partly in Thai or Malay. Parts of the conversation were muffled and could only be made out after being listened to repeatedly by a person familiar with those languages. Applying Menzies, the High Court held that the transcript was admissible as a means of assisting the jury to understand the tape: at 187 per Mason CJ, Brennan and Deane JJ. In a separate judgment, Dawson J also cited Menzies as authority for the proposition that "an ad hoc expertise may be acquired by a witness by playing and replaying a tape so as to become more familiar with its contents than could be done by playing it only once or twice."
In Leung, Simpson J noted that each of those decisions was before the Evidence Act 1995 but drew comfort from two cases decided after the introduction of the Act in which the same principle was applied: Eastman v R (1997) 76 FCR 9 and R v Cassar; R v Sleiman (No 17) [1999] NSWSC 436.
In Eastman, the Full Federal Court held (applying the decisions in Menzies and Butera) that the trial judge had not erred in admitting a transcript of indecipherable tapes prepared by a police officer who spent "literally thousands of hours" listening to the tapes. It was not in dispute that the voice on the tapes was Mr Eastman. Although the case was decided after the introduction of the Evidence Act, there was no consideration of the admissibility of the evidence under s 79 of the Act. The Court did make passing reference to s 48 of the Act.
Cassar was a decision during a trial. The issue was the same as in Eastman; the Crown sought to tender transcripts of indistinct tapes of conversations recorded by a covert listening device. There was no issue as to the identity of the voices. Justice Sperling admitted the transcripts, relying in part on s 48 of the Evidence Act. His Honour did not consider s 79, citing Menzies and Butera as authority for the following proposition: Where a tape is indistinct, a transcript made by an "ad hoc expert", being a person qualified only by having listened to the tape many times, may be used for this purpose. That is particularly so where the tape needs to be played over repeatedly before the words uttered could be made out unaided: Menzies at 49 cited in Butera at 188;
Although in Leung the use of the ad hoc expert went beyond an opinion as to the words said on the tape and extended to identifying the person speaking, that was in the context of the important qualification at [45] of the judgment to which I have referred.
In a number of decisions since Leung (including Irani) voice identification evidence has been admitted from a police officer who became familiar with the voice of the accused during the course of the investigation. The evidence has not been dealt with in those cases as being that of an ad hoc expert but rather simply as identification evidence given by a person familiar with the relevant voice.
On any analysis, those cases are very different from the present case. In Irani, the appellant was convicted of four counts of supplying cocaine. In each case the sale of the drug was alleged to have been made to an undercover informant and recorded by police. The detective who gave voice identification evidence had spent "about two months" listening to recordings of the drug sales to the undercover informant. The detective had played the recordings many times for the purpose of preparing a transcript of them and had spent about four days replaying them to the undercover informant for him to confirm the accuracy of the transcripts.
The principal issue in the trial was whether the undercover informant was lying when he identified the appellant as the person involved in the sale of the drugs. The detective's evidence was that, when the appellant and his co-offender presented themselves at the police station, he "immediately identified their voices". The evidence was voice identification evidence which the detective was able to give because, before hearing the appellant's voice at the police station, he had become very familiar with it from repeatedly listening to the tapes. The evidence may well have been admissible without resort to s 79 of the Evidence Act.
In any event, the appellant submits that the present case is distinguishable from Irani since, in that case, the detective had spent two months repeatedly listening to recordings for the purpose of transcribing them. It was that lengthy undertaking, which could not practicably be replicated by the jury, from which he acquired the experience by reason of which he could properly be classified as having ad hoc expertise.
The evidence in the present case was admitted on a different premise. The fact in issue was whether the voice on the DHL tapes was the voice of the appellant. Federal Agent Succar professed no prior familiarity with the appellant's voice and had acquired no specialised experience beyond that which the jury could itself acquire. His only familiarity with the appellant's voice was that which he acquired for the purpose of constituting himself an ad hoc expert to prove the fact in issue. He did so by listening to the very evidence the Crown intended to put before the jury.
As already noted, the Crown acknowledged at the pre-trial directions hearing that the comparison of the 21 disputed DHL calls and the calls from the gaol was a task the jury could ("would have to") make themselves. Indeed, it appears at that point that the Crown did not consider Federal Agent Succar to have any relevant expertise. Leaving aside his putative specialised knowledge acquired as "an Australian with a Lebanese background", his assertion of identity was founded on material no different from the material available to the jury. To that extent, his evidence was irrelevant and should not have been received for the reasons explained by the High Court in Smith v R [2001] HCA 50; 206 CLR 650 at [10] to [12].
The Crown submitted that reliance on Smith in the present case is misplaced for two reasons. First, it was noted that Federal Agent Succar had listened to a greater number of calls than were made available to the jury. Of the 142 calls listened to by him, only 52 were ultimately tendered in the trial (the Crown explained that the tender of a smaller number of calls was to ensure that details prejudicial to the appellant were not before the jury). That submission entails hindsight reasoning. As already noted, it was initially acknowledged by the Crown at the trial that the jury could make the relevant comparison itself and that there would be no expert evidence. The Crown's decisions to qualify Federal Agent Succar as an expert, and then to tender fewer calls than he had listened to, came after that concession was made.
Secondly, the Crown submitted that Federal Agent Succar was better placed than the jury to make the voice comparison because of "his knowledge and experience with Arabic persons speaking English". It is necessary to consider whether that cultural experience took the matter any further. In his statement, Federal Agent Succar said: Arabic is my second language. I was born in Australia and attended School and University in the NSW Education System. In 2004, I completed a degree in Business and Computing, whereby I studied Arabic as an elective subject. I can say that I have a moderate proficiency in the Arabic language, whereby I am fluent in conversational Arabic. My Arabic reading and writing are at a very basic level. During the course of my duties as a Federal Agent, I have often been called upon by other AFP members to use my Arabic language skills and to provide language assistance during meetings with Arabic speaking community members who wish to provide information to the AFP about possible criminal activities and community engagements. I have also used by Arabic language skills during other operational duties, such as search warrants. I am familiar with conversational Arabic and the requirement to develop the guttural tones associated with pronouncing letters in the Arabic alphabet, Arabic numbers and Arabic words. I observe that once developed, the guttural tones can transpose into the pronunciations of English words. A person, who has a proficiency in Arabic, may speak English in a way that is consistent with a person who has developed guttural tones as a result from speaking Arabic over an extended period.
The terms of the opinion expressed to be based on that specialised knowledge are set out earlier in this judgment. The opinion went little further in identifying particular features of the voice on the recordings than to note the use of the word "youse". The use of that term is now recognised by the Macquarie Dictionary as the colloquial, albeit non-standard, form of the personal pronoun for the second person (both plural and singular). It is hardly a unique feature of speech.
In his evidence, Federal Agent Succar expanded upon the issue of pronunciation, referring to the guttural tones that are required to speak the Arabic language that "may transpose into the way they speak English". Implicitly, the contention was that, by reason of his upbringing in a Lebanese family in Australia, the completion of a unit of Arabic as an elective subject in an unrelated degree and the use of his Arabic language skills during community engagements as a police officer, Federal Agent Succar had specialised knowledge of the sound of English spoken by an Australian with a Lebanese accent. On its own, that specialised knowledge (if properly characterised as such) would not have qualified him to give evidence of an opinion as to whether the voice on the DHL tapes was that of the appellant but only as to whether it was the voice of an Australian speaking English with a Lebanese accent. It is doubtful whether that is an issue for expert opinion evidence.
In my view, the voice identification evidence of Federal Agent Succar was inadmissible.

Application of the proviso

The Crown submitted that, even if ground 1 were decided in favour of the appellant, the Court would nonetheless conclude that no substantial miscarriage of justice has actually occurred and would accordingly exercise its power under s 6(1) of the Criminal Appeal Act 1912 (NSW) to dismiss the appeal.
The Crown submitted that, in accordance with the principles stated by the High Court in Weiss v R [2005] HCA 81; 224 CLR 300, this Court must itself decide whether a substantial miscarriage of justice has actually occurred. In Weiss at [41], the High Court said that the appellate court must make its own independent assessment of the evidence (that is, the whole of the record of the trial) and determine whether, making due allowance for the natural limitations of proceeding on the record, the offence was proved beyond reasonable doubt. The Court said: There will be cases, perhaps many cases, where those natural limitations require the appellate court to conclude that it cannot reach the necessary degree of satisfaction. In such a case the proviso would not apply, and apart from some exceptional cases, where a verdict of acquittal might be entered, it would be necessary to order a new trial. But recognizing that there will be cases where the proviso does not apply does not exonerate the appellate court from examining the record for itself.
In Baiada Poultry Pty Ltd v R [2012] HCA 14; 246 CLR 92 at [22], the High Court acknowledged the prospect of cases in which, although persuaded that the evidence properly admitted at trial proved the accused's guilt beyond reasonable doubt, an appellate court should nonetheless hold that the proviso is not engaged. The examples given were the case of a significant denial of procedural fairness or the case of a failure in a trial by judge alone to comply with a statutory requirement to give reasons.
The Crown submitted that there is nothing about the present case or the grounds of appeal that would prevent the application of the proviso. I do not accept that submission. If both the voice identification evidence of Federal Agent Succar and the concession made by the barrister during his closing address are removed from consideration, the independent assessment of the evidence by this Court necessarily requires the undertaking of a task in all likelihood never undertaken by the jury in the present case. The Crown submitted that the appellant's assertion that the jury "quite likely simply relied upon the concession and the evidence of Agent Succar" is "pure speculation". I do not agree. The temptation to do so must have been overwhelming in the circumstances. Whilst care must be taken in reading too much into the length of time of a jury's deliberations, it may be observed that the jury in the present case returned with verdicts of guilty to all charges less than two hours after retiring to deliberate.
It should be acknowledged that, apart from the DHL tapes, the Crown relied on a considerable body of circumstantial evidence but I do not think it was seriously suggested that this Court could consider the proviso on that limited basis. As noted in Weiss, the proviso must be considered by reference to the whole of the record of the trial.
It is clear that, in the present case, the task invited by the Crown would require this Court to listen to the 21 DHL tapes and the 52 gaol tapes tendered in the trial in order to determine a factual issue which the jury was, in effect, told it did not have to determine. That is not a situation contemplated in the principles stated in Weiss. It requires a new factual determination.
The issue whether this Court should undertake that task has some analogy with an issue considered by the South Australian Court of Criminal Appeal, (albeit in a different context) in R v Compton and Barrett [2013] SASCFC 134; 237 A Crim R 177. The appellants were each convicted of the rape of a boy. The grounds of appeal alleged a misdirection (ground 1) and that the verdicts were unsafe and unsatisfactory and could not be supported having regard to the evidence (ground 2). In support of ground 2, counsel for one of the appellants invited the Court to view an audio-visual record of the evidence of the complainant at trial. It was submitted that the appellate court, in considering the unsafe and unsatisfactory ground, should form its own view as to his credibility and reliability.
The Court did not consider it appropriate to embark upon that course (at [158] per Stanley J; Kourakis CJ and Peek J agreeing at [24] and [26] respectively). Justice Stanley said: Exercising the circumspection called for by the Court of Appeal in El Moustafa and approved by the majority in SKA, I do not consider it appropriate that this Court embark on that course, even if empowered to do so. To do so would involve the expenditure of considerable time and resources on the part of this Court. More importantly, the exercise threatens the constitutional role of the jury in circumstances which would not necessarily result in a more accurate assessment of the complainant's evidence than that made by the jury. The obligation of this Court to undertake an independent assessment of the evidence, for the purpose of considering and determining this ground of appeal, as described by the High Court in M v R, does not require that we do so. To do so would be tantamount to disregarding the considerations, identified in M v R, that it is the jury which is the body entrusted with the primary responsibility of determining guilt or innocence, and it has had the benefit of having seen and heard the witnesses. On the contrary, the course pressed upon us involves an invitation to this Court to substitute its own assessment of the reliability and credibility of the complainant based on our assessment of his demeanour, for that of the jury, rather than undertaking an assessment of the evidence that was before the jury as to its sufficiency and quality.
Without viewing the video, the Court rejected ground 2, holding that the matters raised concerning the complainant's reliability did not preclude satisfaction of the appellants' guilt (at [162]). Ground 1 having been upheld (by majority), the Court accordingly ordered a retrial rather than entering a verdict of acquittal.
Although the Court was there concerned with an unreasonable verdict ground rather than the application of the proviso, the concerns articulated by Stanley J reinforce my concerns in the present case. I do not think s 6(1) of the Criminal Appeal Act should be understood to contemplate, let alone dictate, that an appellate court should be, in effect, the first tribunal of fact to undertake the comparison of the voice on the DHL tapes with the voice on the gaol tapes unadorned by the voice identification evidence or the concession of counsel.
For those reasons, the orders I propose are: 1. that the appeal against conviction be allowed; 2. that the conviction be quashed; 3. that there be a new trial; 4. that the Crown appeal be dismissed.

Nasrallah v R; R v Nasrallah

At a glance

Court

Decision date

Before

Source

Judgment (5 paragraphs)

Legislation Cited (2)

Cases Cited (13)

Cited By (5)