concurring in part, and dissenting in part.
The holding of Evans v. State, 177 So.3d 1219, 1229 (Fla. 2015), is clear: the “testimony that a witness recognizes the voice of the accused is inadmissible ... unless the testifying witness (1) was an eyewitness to the crime, (2) has some prior special familiarity with the voice of the defendant, or (3) is qualified as an expert in identification.”5 Because I conclude the majority misapplies this holding, I respectfully dissent.6
The legal system relies heavily on witness identifications for investigating and prosecuting crimes. However, identification errors are the leading cause of false convictions. Cindy E. Laub, Lindsey E. Wylie & Brian H. Bomstein, Can the Courts Tell an Ear from an Eye? Legal Approaches to Voice Identification Evidence, 37 Law & Psychol. Rev. 119, 119 (2013). Because reliability is the linchpin in determining the admissibility of identification testimony, Manson v. Brathwaite, 432 U.S, 98, 114, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977), and the role of witness identification is so important, especially when there is no physical evidence, the Florida Supreme Court limited the witnesses able to give opinion testimony about the identity of a recorded voice in Evans. See C. Ehrhardt, Fla. Evidence § 901.6 (2016 ed.).
To ensure the reliability of opinions presented to the jury, Evans raised the standard for a voice identification by requiring the witness offering an opinion on the identity of a speaker to be either an eyewitness or a witness with prior special familiarity with the voice.7 As a result, Evans demands more than did the earlier cases relied on by the majority. In an effort to satisfy the Evans construct, the majority contends that Agent Scovel either was an eyewitness to the crime (a position *654not advanced by the State) or had sufficient prior special familiarity with Johnson’s voice to be able to reliably identify it on the recordings. I disagree with both contentions.
First, with respect to the majority’s position that Agent Scovel was an eyewitness, an eyewitness is someone who personally sees an event and can describe it later. Eyewitness, Black’s Law Dictionary (10th ed. 2014); see also Earwitness, Black’s Law Dictionary (10th ed. 2014) (defining earwitness as witness who testifies about something he or she heard, but did not see). I do not dispute that Agent Scov-el listened to the tapes of the recorded telephone calls or that the tapes were properly authenticated. But the question here is not one of authentication; it is whether Agent Scovel is able to reliably opine—either as an eyewitness or as a witness with prior special familiarity with Johnson’s voice—that he recognized Johnson’s voice on the tape.8 Agent Scovel was not present when the calls were recorded, and did not observe Johnson speaking on the phone while simultaneously listening to the calls. Had that occurred, arguably, he could be considered an eyewitness. But, with the exception of one brief personal conversation with Johnson, Agent Scovel was in no better position to identify the voice on the recordings than anyone else who listened to the tapes and the recording of Johnson’s pretrial testimony.
The majority’s reliance on Mack v. State, 54 Fla. 55, 44 So. 706 (1907), Martin v. State, 100 Fla. 16, 129 So. 112 (1930), and Macias v. State, 673 So.2d 176 (Fla. 4th DCA 1996), is misplaced. Each of those cases involved the victim of the crime testifying as to what the perpetrator said while the crime was being committed. The victims in those cases were prototypical eyewitnesses to the crimes and as such, satisfied the Evans requirements. Just as the officers in Ruffin v. State, 549 So.2d 250 (Fla. 5th DCA 1989), and Alvarez v. State, 147 So.3d 537 (Fla. 4th DCA 2014), were not eyewitnesses to a crime merely because they later watched a recorded video of the crime, Agent Scovel is not an eyewitness to the conspiracy merely because he later listened to the recorded phone conversations discussing it.
Perhaps the strongest support for the majority opinion can be found in Alvarez, which held that “[e]ven non-eyewitnesses may testify as to the identification of persons depicted or heard on a recording so long as it is clear the witness is in a better position than the jurors to make those determinations.” 147 So.3d at 542. Because Agent Scovel had a brief, in-person conversation with Johnson, he was in a slightly better position than the jurors to determine if Johnson’s voice was on the recorded telephone calls. However, this “better position than the jurors” language—which the majority refers to as the “helpfulness” standard—cannot be found in Evans, which postdates Alvarez.9 It was apparently adopted from Federal Rule of Evidence 901 and numerous federal cases cited by the majority, but it is premised on what appears to be a lower standard for the *655admissibility of opinion testimony in federal courts.
Although I agree with the majority that the provisions of the Florida and federal evidence codes regarding lay opinion testimony are similar, one need only compare Evans with United States v. Bush, 405 F.3d 909, 919 (10th Cir. 2005), to see that they have been interpreted differently by Florida and federal courts. Had this case been prosecuted in federal court, Agent Scovel’s opinion testimony would certainly have been allowed. See, e.g., Bush, 405 F.3d at 919 (holding voice identification need only rise to level of minimal familiarity); United States v. Axselle, 604 F.2d 1330, 1338 (10th Cir. 1979) (holding single telephone call, combined with hearing voice in court, is sufficient for voice identification testimony to go to jury). However, I am compelled to follow the interpretation of the Florida evidence code by the Florida Supreme Court, not the interpretation of the federal evidence code by federal courts. Under Florida law, because Evans raised the bar for the admissibility of an opinion on voice identification, I conclude Agent Scovel was not an eyewitness competent to render an opinion on whose voice was heard on the recordings.
Turning to the majority’s position that Agent Scovel had adequate prior special familiarity with Johnson’s voice that qualified him to testify, it seems a matter of common sense that the ability to identify a speaker from his or her voice depends on the number of exposures to the voice in question, the quality of these exposures, and the nature of the identification process. See Lawrence Solan & Peter M. Tiersma, Hearing Voices: Speaker Identification in Court, 54 Hastings L.J. 373, 375 (2003). The Evans court incorporated this common sense approach by creating a standard permitting a witness to identify the voice of the defendant when the witness was previously familiar with the defendant. Evans, 177 So.3d at 1229; see State v. Cordia, 564 So.2d 601, 601-02 (Fla. 2d DCA 1990) (permitting witness to testify who had known defendant “for a significant period of time” and had spoken to defendant in person and over telephone); Hardie v. State, 513 So.2d 791, 792 (Fla. 4th DCA 1987) (holding that police officers who had prior knowledge and contact with defendant could properly testify and identify him).
Here, however, Agent Scovel lacked the requisite prior special familiarity with Johnson. Before initiating the investigation, Agent Scovel had no contact with Johnson. During the course of the investigation, Agent Scovel spoke to Johnson only once, during the search of Howard’s house, and Johnson said little during that conversation. Other than this brief conversation, the only familiarity Agent Scovel had with the voice that he believed to be Johnson’s was through listening to recordings of intercepted phone conversations and a subsequent recording of Johnson’s pretrial hearing. This limited interaction with Johnson is not sufficient to constitute prior special familiarity. Indeed, Evans specifically held that “a police officer investigating a particular suspect’s voice after the investigation is ongoing ... does not constitute the requisite prior familiarity with the suspect.” 177 So.3d at 1230 (emphasis added); see Proctor v. State, 97 So.3d 313, 315 (Fla. 5th DCA 2012); cf. Cordia, 564 So.2d at 601-02; Hardie, 513 So.2d at 792. While the majority contends that the timing of an officer’s interaction with a voice is inconsequential, the Florida Supreme Court held otherwise. It is clear from Evans that an investigating officer does not obtain prior special familiarity with a suspect’s voice if his or her first encounter with the voice occurs after a criminal investigation into the suspect’s behavior has begun, regardless of whether at *656that time the crime has been completed. Evans, 177 So.3d at 1280. Additionally, a trial court’s error in permitting an improper identification “may be exacerbated where the testimony comes from a police officer” and the jury is aware of that position because officers are often regarded as highly credible. Martinez v. State, 761 So.2d 1074, 1080 (Fla. 2000); see also Evans, 177 So.3d at 1230 (finding use of questions that elicit witness’s position as police officer when witness is identifying defendant’s voice or image may be reversible error even when identification itself is permissible).
In the end, the majority distinguishes this case from Evans by claiming that Agent Scovel was an eyewitness because he listened to the recordings and spoke briefly in-person with Johnson, whereas the detective in Evans had no personal interaction with the defendant. Although this is accurate, it is insufficient under the heightened Evans standard to make Agent Scovel an eyewitness or to give him the requisite prior special familiarity with Johnson’s voice to reliably identify it in the recordings.10 I would reverse for a new trial.
. The Florida Supreme Court decided Evans after Johnson’s tried.
. I concur with the majority that Johnson’s other claims of error lack merit.
.Expert testimony is also permitted by Evans, but is not an issue in this case.
. Because this case concerns lay opinion testimony, I need not distinguish Vilsaint v. State, 127 So.3d 647 (Fla. 4th DCA 2013), which the majority concedes is a case regarding authentication. See opinion at 648-49.
. The State could have allowed the jury to determine for itself if Johnson’s voice was heard on the incriminating recordings by playing the recording of the pretrial hearing at which Johnson testified, or by compelling him to give a voice exemplar, either before or at trial, in order to evaluate the physical properties of his voice. See Fla. R. Crim. P. 3.220(c)(1)(B); United States v. Dionisio, 410 U.S. 1, 93 S.Ct. 764, 35 L.Ed.2d 67 (1973).
. Likewise, I do not believe Agent Pederson was an eyewitness or had prior special familiarity with Howard’s voice to reliably identify it in the recordings.