Multimedia Evidence: Admissibility and Expert Testimony

Authentication is the threshold step. It is the process of proving, to the court's satisfaction, that the exhibit is what the party claims. For a knife, this is typically done by the officer who collected it testifying that this is the knife from the scene, identified by the exhibit tag. For a video file, it is more complex.

Under FRE 901(b)(9), a process or system can authenticate evidence by showing that it produces an accurate result. For CCTV footage, this means evidence about the recording system: Was it functioning at the time? Was the clock accurate? Was the recording continuous? Who exported the footage and how? Hash verification is increasingly accepted as strong authentication evidence, because it provides mathematical proof that the file in court is identical to the file at the point of hash computation.

In England and Wales, the key authentication document for digital evidence is now the chain-of-custody record and the forensic integrity report. In practice, defence will often not challenge authentication of routine CCTV if the chain is clear. Challenges arise when the chain has gaps, when the footage comes from an unusual source such as a private phone or a cloud platform, or when there are metadata inconsistencies.

Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993) established that US federal judges must evaluate the reliability of expert methodology before allowing testimony. The Supreme Court identified four non-exclusive factors for this assessment.

• Testing: has the theory or technique been tested? A method that has never been validated on a reference dataset with known ground truth cannot establish its reliability.
• Peer review: has it been subjected to peer review and publication? A proprietary tool used as a black box without published evaluation fails this criterion.
• Error rate: is there a known or potential error rate? For face comparison evidence, NIST FRVT benchmarks provide error rates for automated systems. For examiner-based morphological comparison, the absence of validated error rates has led to Daubert challenges.
• General acceptance: is the method generally accepted in the relevant scientific or professional community? SWGDE and OSAC documents, and the forensic science literature, inform this question for multimedia methods.

Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999) extended Daubert to technical and experience-based expertise, not just scientific experts. This is significant for multimedia forensics because many practitioners come from law enforcement backgrounds and their methods may be more craft-based than laboratory-validated. A traffic accident reconstructionist or a DVR analysis specialist is subject to Daubert scrutiny just as a molecular biologist would be.

English and Welsh courts do not apply Daubert directly, but the underlying question is similar. Expert evidence must be based on reliable methodology and the expert must stay within their area of expertise. The leading case on admissibility of novel scientific evidence in England and Wales is R v. Reed and Reed [2009] EWCA Crim 2698, which confirmed that expert evidence is admissible only if it is based on a sufficiently reliable body of expertise.

For facial mapping evidence, the Court of Appeal's position was set out in R v. Atkins and Atkins [2009] EWCA Crim 1876. The court held that morphological facial image comparison by a suitably qualified expert is admissible, but noted that the weight of such evidence must be carefully assessed by the jury and that experts must not use language conveying a higher degree of certainty than the method supports. This case is the direct UK parallel to the Daubert face-comparison decisions in US courts.

In Australia, the admissibility of expert evidence follows the National Evidence Act test: the opinion is based on specialised knowledge held by the person by reason of training, study, or experience, and the opinion is wholly or substantially based on that knowledge. The Makita principles (Makita (Australia) Pty Ltd v. Sprowles [2001] NSWCA 305) require that the scientific or other intellectual basis of the opinion be explained. This effectively imports a reliability assessment similar to Daubert without using that framework explicitly.

In adversarial systems, experts are called by a party and paid by that party. This creates a structural pressure toward advocacy. The antidote, in most jurisdictions, is a formal declaration that the expert's duty is to the court, not to the instructing party. In England and Wales, CPR Part 35 and the accompanying Practice Direction require an explicit statement to this effect in every expert report.

For multimedia evidence, overstatement typically takes one of two forms. The first is claiming a definitive identification from evidence that supports only a class-level match. An expert who says this is the same person based on face comparison evidence that the literature shows has a meaningful false-positive rate is overstating. The second is omitting or downplaying the limitations of a method, such as failing to mention that the facial recognition system used had a significantly higher false-match rate for the demographic group the suspect belongs to.

The obligation to disclose limitations is not a concession of weakness. Courts expect it. An expert who acknowledges the resolution limits of the source footage, the error rate of the comparison method, and the factors that could produce a false result is more credible, not less. An expert who claims certainty that the method cannot support will be destroyed in cross-examination when those limitations are put to them.

Research on juror decision-making consistently shows that visual evidence, particularly video, carries substantial persuasive weight. Jurors tend to believe what they see on a screen, even when what they are seeing is low-resolution, partially obstructed footage from a camera whose clock may be incorrect and whose lens may introduce distortion.

This is the CSI effect operating in reverse: not the expectation of perfect evidence, but over-trust of imperfect evidence once it is played in court. A blurry clip of a figure in a dark jacket can create an impression of identification that the underlying image quality cannot support. Contextualising the evidence for the jury is a core obligation of the multimedia expert.

• Explain resolution limits: tell the jury how many pixels cover the face in the best frame, and what details cannot be resolved at that resolution. Use annotated exhibits that show the pixel count, not just the image.
• Explain conclusion scales: if the comparison conclusion is 'limited support for the proposition that the person depicted is the defendant', explain what that means in plain terms. Jurors are not familiar with FISWG or IAFPA verbal scales.
• Explain what was not found: if the analysis found no features that exclude the defendant, say so clearly. This is not the same as saying it is the defendant.

Defence experts serve a crucial function here. A defence expert who reviews the prosecution's multimedia analysis and identifies methodological weaknesses, undisclosed limitations, or overstatements provides the jury with the counterweight they need to evaluate the evidence properly. In cases where the only identification is from CCTV, defence counsel should seriously consider instructing a multimedia forensics expert.

Three cases have shaped multimedia forensic evidence practice more than any standards document.

• R v. Clare and Peach [1995] 2 Cr App R 333 (UK): established that police officers familiar with a scene could testify as to what CCTV footage depicted. This opened the door for lay identification from CCTV but was later qualified by cases requiring expert analysis for disputed quality footage.
• R v. Atkins and Atkins [2009] EWCA Crim 1876 (UK): confirmed the admissibility of expert facial image comparison evidence but required that experts not claim more certainty than the method supports. The court endorsed a verbal scale approach to expressing conclusions. This case is the benchmark for facial comparison expert evidence in England and Wales.
• Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993) and Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999) (US): established the federal gatekeeping standard for expert methodology. Kumho's extension to technical expertise brought multimedia forensic methods within the gatekeeping framework. Cases following these decisions have excluded facial comparison experts, image authentication experts, and audio analysis experts where methods lacked validated error rates.

Practitioners must be familiar with the jurisdiction-specific case law because courts cite these decisions when admissibility is challenged. Knowing that Atkins governs the conclusion scale for face comparison in England and Wales, and that Daubert requires a known error rate for admission in federal courts, determines what an expert report must say to survive challenge.