Flynn & Anor, R v (Rev 1)

Lay listener voice-recognition evidence may be factual evidence, but its admission under Police and Criminal Evidence Act 1984 s.78 depends upon a realistic assessment of its probative value and prejudicial effect. Poor recording quality, limited familiarity with the suspect’s voice, untested listener ability, short or distorted speech, and unreliable procedures may make the evidence unfair to admit.

Police officers should record their contact with the suspect and the identification process, avoid exposure to another officer’s annotated transcript, and preferably leave the exercise to an independent officer. Jurors must receive careful warnings about the danger of voice misidentification and should be permitted, with proper directions, to compare a recording with voices heard at trial.

The appellants were convicted at Reading Crown Court of conspiracy to rob a computer warehouse and each received 13 years’ imprisonment. The prosecution relied on circumstantial and forensic evidence and on four police officers who said that they recognised the appellants’ voices on a covert recording made in the van used for the attempted robbery.

At trial the judge admitted the officers’ evidence and transcripts of the recording. The appellants appealed against that ruling and against a direction preventing the jury from comparing the recording with their voices when they gave evidence. The central issues were whether the lay listener evidence should have been excluded and whether, despite any errors, the convictions were safe.

1. The appeals were allowed. The convictions were quashed because the police officers’ voice-recognition evidence ought to have been excluded and the jury had also been misdirected.

2. Evidence from a police officer who recognises a voice through familiarity with a suspect is evidence of fact, as recognised in R v Robb [1991] 93 Cr.App.R 161. That classification did not determine admissibility. The decisive question was whether the evidence had sufficient potential probative value to justify its seriously prejudicial effect.

3. The covert recording was poor. Even an expert using sophisticated equipment could not reliably distinguish individual speakers. The officers had only limited opportunities to become familiar with the appellants’ voices. Their identifications also contrasted with the expert evidence, their ability to recognise voices was untested, and one alleged identification concerned a short transmission distorted by a walkie-talkie. Those matters meant that the prejudicial effect far outweighed the probative value.

4. The evidence should additionally have been excluded under Police and Criminal Evidence Act 1984 s.78. The officers had not made adequate records of their contact with the appellants or of the identification exercise. There was an important unexplained inconsistency concerning the preparation and possible annotation of a transcript. Supplying an officer with another officer’s annotated transcript compromises an independent recognition exercise. It is also highly desirable that recognition be undertaken by an officer who is not investigating the offence.

5. The direction forbidding the jury to compare the recording with the appellants’ voices in court was wrong. A jury may make that comparison, provided it receives suitable guidance and warnings. Although the remaining circumstantial case was strong, the court could not say that the jury would inevitably have reached the same verdicts without the prejudicial evidence.

6. By way of guidance, properly qualified expert voice-identification evidence remained admissible. The court did not adopt a rule that auditory expert evidence always requires acoustic analysis. The degree of a lay listener’s familiarity with the voice is central, but poor recordings heighten the continuing risk of error. Judges must give careful warnings wherever voice-recognition evidence is used.

A retrial was directed on a fresh indictment.

• Court of Appeal (Criminal Division): In [2008] EWCA Crim 970, allowed both appeals, quashed the convictions, and directed a retrial on a fresh indictment.
• Crown Court at Reading: On 7 December 2006, the appellants were convicted of conspiracy to commit robbery. On 8 December 2006, each was sentenced to 13 years’ imprisonment.