State v. DELATORRE-VARGAS

*504 SERCOMBE, P. J. Defendant was convicted of two counts of first-degree robbery, ORS 164.415, one count of first-degree burglary, ORS 164.225, one count of aggravated first-degree theft, ORS 164.057, two counts of coercion, ORS 163.275, one count of unlawful use of a weapon, ORS 166.220, two counts of menacing, ORS 163.190, and one count of attempted interference with making a report, ORS 161.405(2)(e); ORS 165.572. On appeal, he contends that the trial court erred in denying his pretrial motion to suppress out-of-court photographic and voice identifications made by two witnesses. He argues that the identifications were irreparably tainted by suggestive police procedures. We conclude that the identifications were obtained using suggestive procedures and that the state failed to demonstrate that the identifications were independently reliable. See State v. Classen, 285 Or 221, 232, 590 P2d 1198 (1979) (setting forth the test for determining the admissibility of pretrial identifications). Accordingly, we reverse. We are bound by the trial court’s findings of historical fact where there is evidence in the record to support them. State v. Najibi, 150 Or App 194, 198, 945 P2d 1093 (1997), rev den, 326 Or 464 (1998). However, we are not bound by its legal conclusions drawn from those facts. Id.] State v. Ray, 157 Or App 601, 604, 971 P2d 490 (1998). In making that legal determination, we limit our review to the evidence before the court at the time that it decided the motion. 1 Najibi, 150 Or App at 199 n 4. On December 19, 2006, two masked men, one armed with a handgun, entered the back door of the Willamina Pharmacy shortly after closing time. Two pharmacists, Kotaich and Bowman, were present. The intruders forced the pharmacists into a back office and bound their hands with zip ties. The intruders then demanded certain narcotics, which they stole. Before leaving, they also stole some cash. The robbery lasted about eight minutes. *505 Kotaich was able to give a rough physical description of the intruders to the police. He reported that one of the perpetrators was about 5'6", had a “heavier build,” and was dressed in faded blue jeans and a plain black sweatshirt. Kotaich reported that the other perpetrator, who was armed and did all the talking during the robbery, was 5'8" or 5'9", weighed about 180 pounds, and was wearing darker clothing. Kotaich was unable to give any further description of the perpetrators’ appearance because they were wearing makeshift masks — knit wool “beanies” pulled down over their faces and necks with only eyeholes cut out. Kotaich did not get a good look at the taller individual. Kotaich did not think that either perpetrator was a customer of the pharmacy because he did not recognize the taller man’s voice. However, Kotaich reported that “the bigger guy had a very distinct voice that was unusual and a higher pitch” and opined that he would “remember that voice forever.” On February 7,2007, Detective Steele of the Yamhill County Sheriffs Office contacted Kotaich to make a voice identification. 2 Steele used a recording of defendant being interviewed by police about robberies in a different county. Steele testified that he had selected a “benign” portion of the recording to play for Kotaich, but he could not recall what portion of the recording he had played or what was being said on the recording. He made no attempt to put together other voice exemplars to play for Kotaich. However, before playing the tape, Steele cautioned Kotaich that he “may or may not recognize this voice, and the fact that [Steele] was playing [it] for him he should not read anything into.” According to Steele, after just a few seconds of listening to the recording, Kotaich said that he was “certain” that the voice on the recording was that of the taller perpetrator. Sometime after the robbery, Steele had learned that another witness — Hoyer, a customer of the pharmacy — had seen someone outside the store about an hour before the robbery. Steele did not obtain a description of the individual from Hoyer before meeting with her. Instead, on February 10, he brought five black and white photographs to Hoyer’s *506 home. At least three of the photographs were of suspects, including defendant. Steele made no attempt to prepare a photo array of similar-looking individuals: four of the individuals were apparently Caucasian and one was apparently Hispanic; only one individual (the Hispanic man) displayed visible facial hair; one individual had short blond hair, one had short brown hair, and three had dark brown or black hair of widely varying lengths; and the individuals’ facial structures differed. 3 Before showing the photographs to Hoyer, Steele told her, “You may recognize somebody in this photo throwdown, you may not. Just because I’m showing you these photographs does not mean that the person I’m interested in is here. So, look at them very carefully, and if you do choose one that you say you saw, be very, very sure.” Steele presented the photographs to Hoyer and, after about seven seconds, she pointed to defendant’s picture (the Hispanic man with facial hair) and said, “He resembled this one.” Steele then asked her for a description of the clothing of the man she saw outside the pharmacy. Hoyer indicated that the man was wearing a dark-colored hooded sweatshirt and a dark-colored stocking cap. She had never seen him before the day of the robbery. Defendant moved before trial to suppress both the photographic and voice identifications on the ground that they had been procured using suggestive procedures and were therefore inadmissible. At the pretrial hearing on defendant’s motion, Steele was the only person to testify. In addition to the facts recounted above, Steele testified about the procedures normally used in pretrial identifications. He explained that the normal procedure for a photo throwdown involves assembling several photographs of “like individuals” *507 and showing those to the witness. He admitted that there were no pressing circumstances that prevented him from assembling a traditional photo throwdown in this case. Steele also explained that he was not aware of any protocol for voice identifications and admitted that there was no reason why he could not have provided more than one voice exemplar to the witness in this case. In addition, defendant read a statement at the hearing in order to provide a voice exemplar. Steele testified, and the trial court found, that defendant did not have a high-pitched voice. 4 Nonetheless, the trial court determined that the voice identification had not been obtained using “impermissibly suggestive” procedures and that, in any event, the circumstances suggested the identification was reliable. The court also concluded that the photographic throwdown was not suggestive. Consequently, it denied defendant’s motion to suppress. At trial, both witnesses recounted their pretrial identification of defendant, and defendant was ultimately convicted. On appeal, defendant renews his argument that both identifications were the product of suggestive procedures and, therefore, should be excluded under the principles announced in Classen. 5 With regard to the voice identification, defendant argues that it was suggestive to play a “single recording * * * of a single voice” for the witness, particularly where the recording was an interrogation by police about robberies in another county. Defendant further argues that the state did not meet its burden to establish that the identification was reliable notwithstanding the suggestive procedure. With regard to the photo throwdown, defendant points out that the detective, by his own admission, deviated from the *508 normal procedure by failing to select photographs of similar-looking individuals. That failure, according to defendant, improperly led the witness to identify defendant. Furthermore, defendant argues, the circumstances surrounding the identification do not support a conclusion that it was reliable despite the suggestive procedures. The state responds that the detective’s warnings prior to both identifications were sufficient to inoculate the identifications against any suggestive effects of the police procedures. As to the voice identification, the state also argues that the procedure was not suggestive merely because a single voice was played or because the subject matter of the recording was a police interrogation. Moreover, according to the state, using multiple voice exemplars would have increased the likelihood of suggestiveness. In addition, the state contends, the content of the recording was not influential because the detective used a “benign” portion of the police interview. In any event, the state argues that the identification bore sufficient indicia of reliability to be admissible. As to the photo throwdown, the state argues only that the procedure was not suggestive; it concedes that, if the procedure was suggestive, the identification was not reliable. Alternatively, it argues that any error with regard to the photo identification was harmless. For the following reasons, we conclude that the identifications should have been suppressed. Under Classen, the admissibility of challenged identification evidence hinges on a two-step inquiry: “First, the court must determine whether the process leading to the offered identification was suggestive or needlessly departed from procedures prescribed to avoid such suggestiveness. If so, then the prosecution must satisfy the court that ‘the proffered identification has a source independent of the suggestive confrontation’ or photographic display or that other aspects of the identification at the time it was made substantially exclude the risk that it resulted from the suggestive procedure.” 285 Or at 232 (footnote and citation omitted). Factors relevant to the second determination “include the opportunity that the witness had at the time to get a clear view of the persons involved in the crime and the *509 attention he or she gave to their identifying features, the timing and completeness of the description given by the witness after the event, the certainty expressed by the witness in that description and in making the subsequent identification, and, of course, the lapse of time between the original observation and the subsequent identification. * * * Obviously other facts may also be important, such as the age and sensory acuity of the witness, or a special occupational concern with people’s appearance or physical features, or the frequency of his or her contacts with individuals sharing the general characteristics of the person identified. Listing these and other relevant inquiries must not distract attention from the ultimate issue whether an identification made in a suggestive procedure has nevertheless been demonstrated to be reliable despite that suggestiveness.” Id. at 232-33 (footnote and citations omitted). We begin with the voice identification and conclude that the process used by the detective was suggestive. 6 As noted, Steele contacted Kotaich to make a voice identification, cautioned him that he should not feel compelled to make an identification, and then played “benign” portions of a recording taken from a police interrogation of defendant regarding robberies in another county. Playing a recording of a single voice for a witness, much like displaying a single photograph, is a nearly quintessential example of a suggestive procedure. See, e.g., State v. James, 240 Or App 324, 327, 245 P3d 705, rev allowed, 350 Or 532 (2011) (“Clearly, under the test set forth in Classen, a procedure in which a witness is shown only a pair of joint suspects and asked to identify them is an unduly suggestive procedure.”); State v. Rector/ Tremaine, 82 Or App 466, 477, 729 P2d 1 (1986), rev den, 302 Or 614 (1987) (“An identification is unduly suggestive if it unfairly singles out or points to a defendant as the suspect to be identified for a known crime or for a known reason.”); Sanchell v. Parratt, 530 F2d 286, 294 (8th Cir 1976) (“[SJhowing only a single suspect to the witness is the most suggestive and, therefore, the most objectionable method of *510 pretrial identification.” (Citation and internal quotation marks omitted.)). That process creates an unavoidable inference that the individual in the recording is a focal suspect of the police, and may lead the witness to feel that he or she should identify the individual’s voice. In that context, a preemptive warning to the witness does little to diminish the powerful inference that the circumstances otherwise suggest. See State v. Lee, 56 Or App 147, 153, 641 P2d 589 (1982) (“[T]he fact that the witness * * * was not told that the suspect was among the samples is not determinative on the issue of suggestiveness.”). We conclude that the voice identification procedure in this case was suggestive. See id. at 152-53 (voice identification procedure was suggestive where, although six recordings were played for the witnesses, the recording of the defendant’s voice was unscripted and contained background noise and therefore sounded more natural than the others); State v. Davie, 56 Or App 507, 513, 642 P2d 680, rev den, 293 Or 146 (1982) (in-person voice identification of single suspect was suggestive). The question remains whether the state demonstrated that the voice identification was reliable notwithstanding the suggestive procedure. Here, the witness was exposed to the perpetrator’s voice on a single occasion for eight minutes while the witness was under great stress. Cf. Ray, 157 Or App at 605 (identification was reliable where, among other things, the witness had encountered the defendant several times over a period of three years); Lee, 56 Or App at 153 (identification was reliable where the witness had “at least two face-to-face encounters with the suspicious man” and, during the second encounter, “took special note of his appearance”). Nonetheless, the witness expressed confidence at the time of the crime that he would be able to remember the perpetrator’s voice, and he expressed certainty in his identification of defendant’s voice at the time the recording was played for him. 7 The certainty of the witness’s *511 identification, however, is undercut somewhat by the fact that the witness’s original description — that defendant’s voice was “distinct” and “unusual” and had a “higher pitch”— was inaccurate, that is, it ultimately did not match defendant. Cf. Najibi, 150 Or App at 199 (identification was reliable where, among other things, the witness gave a detailed description of perpetrator’s features, “all of which were consistent with [the] defendant”). Finally, the identification occurred nearly two months after the crime. Cf. James, 240 Or App at 328 (where identification was made five hours after the crime, “[t]hat relatively short amount of time does not weigh significantly against the admissibility of the identifications”); Najibi, 150 Or App at 200 (noting that witnesses’ “memories may not have been fresh” where one month had passed between the crime and the identification, but also noting that “we have upheld identifications that occurred much longer after the crime”). Several of those facts are, in effect, neutral — that is, they do not add significantly to the reliability of the identification or could be argued either in favor of or against admissibility. The fact that the witness was in the presence of the perpetrator for eight minutes, although potentially significant in the context of a visual identification, is less persuasive in the context of an auditory identification. The state did not introduce evidence at the hearing to show how much talking the perpetrator did, the manner in which the perpetrator was speaking, or how carefully the witness paid attention to his voice. 8 Nor was there any evidence about the witness’s sensory acuity to voices. Similarly, the stress of a robbery at gunpoint could be cited as a factor both in support of and against the reliability of the identification. See State v. Lawson, 239 Or App 363, 381, 244 P3d 860 (2010), rev allowed, 350 Or 532 (2011) (so noting). Likewise, there is no bright-line rule regarding the lapse of time between the crime and the identification. Nonetheless, after two months’ time, the witness’s memory may not have been fresh. *512 What we are left with is the witness’s certitude, on the one hand, and his inaccurate description, on the other. Under those circumstances, there is little evidence that “negate[s] any substantial risk that th[e] identification was stimulated by the suggestive procedure.” Classen, 285 Or at 236. In other words, the state did not carry its burden to show that the identification was independently reliable. Thus, that identification should have been suppressed. We also conclude that the photographic throwdown was suggestive. The officer made no attempt to put together photos of similar-looking individuals, and, indeed, the individuals pictured in the throwdown bore little resemblance to each other. Defendant was the only individual who had short dark hair, had any discernible facial hair, and appeared to be Hispanic. “For a display to be impermissibly suggestive it must somehow lead the witness to identify a person on some basis other than the witnesses] memory. A display could be impermissibly suggestive if, of all the persons shown, only one or two persons looked at all like the defendant.” State v. Maher, 72 Or App 543, 546, 696 P2d 573, rev den, 299 Or 314 (1985). Here, if the witness was looking for a Hispanic man with facial hair, she had only defendant’s picture from which to choose. That throwdown was suggestive. See, e.g., Classen, 285 Or at 236 (photo throwdown was suggestive, in part because “only one of the pictures showed a man with facial hair beyond a mustache”); Lee, 56 Or App at 152 (photo throwdown was suggestive where the “defendant’s was the only picture of a white man with short, dark curly hair”); cf. State v. Mackey, 86 Or App 691, 695, 740 P2d 231, rev den, 304 Or 279 (1987) (photo throwdown was not suggestive where it showed “six white men of similar age with substantially similar hair color, style and length, facial structure and facial hair, wearing similar clothing”). As noted, the state does not contend that the photographic identification was reliable despite the suggestive procedure. The relevant facts are these: the witness saw the individual outside the pharmacy, apparently in passing, and had no reason to pay particular attention to his appearance; the witness had never seen the individual before the day of the robbery; the witness did not provide any description, let *513 alone a detailed one, of the individual before the photo throw-down; almost two months had passed between the original observation and the photo identification; and the witness did not express certainty in her description, saying only that defendant “resembled” the individual she had seen. In light of those facts, the state’s concession is well taken. Finally, the state argues that the failure to suppress the photographic identification was harmless in light of the witness’s equivocal trial testimony and the other evidence introduced at trial. However, the state advances no argument that the failure to suppress both out-of-court identifications was harmless. An error is harmless where there is little likelihood that it affected the verdict. State v. Davis, 336 Or 19, 32, 77 P3d 1111 (2003). Having reviewed the record in this case, we readily conclude that the errors in admitting the identification evidence were not harmless. See Lawson, 239 Or App at 385 (noting that “eyewitness testimony is often believable and can wield considerable influence over jury decisions”). Consequently, the case must be remanded for a new trial. Reversed and remanded. Footnotes [1] At trial, several significant facts were revealed that supplemented or contradicted the testimony at the pretrial hearing. Although the parties have cited to some of those facts, we do not consider them in reaching our decision. [2] Steele also asked Bowman to make a voice identification, but Bowman could not identify the voice. [3] The trial court found that none of the individuals pictured in the throwdown appeared to he “Hispanic in nature.” Having reviewed the pictures ourselves, we reject that finding as unsupported by the record. Although we recognize that a certain amount of subjectivity is involved in identifying a person’s ethnicity, one of the individuals in the photographs — defendant—appears quite clearly to he Hispanic. The other individuals appear to be Caucasian, although one of them could at least arguably be Hispanic. On this point, we also note that Steele testified that defendant had a “medium” skin tone and that the other individuals in the photographs had a light skin tone; Steele later described defendant as being a “fair-skinned Hispanic” man. In any event, none of the pictured individuals bears significant similarities to each other. [4] The trial court also listened to a portion of the recording of defendant. [5] Classen announced, as a matter of state evidence law, that pretrial identifications must hear a threshold level of reliability in order to be admissible. 285 Or at 226, 232; State v. Johanesen, 319 Or 128, 130, 133-34, 873 P2d 1065 (1994). That rule is derived from, but independent of, the federal due process guarantee. See Johanesen, 319 Or at 133-34 (so noting); cf. Perry v. New Hampshire,_US _, 132 S Ct 716, 181 L Ed 2d 694 (2012) (discussing the federal standard). Here, defendant contends that the identifications should have been suppressed both under Classen and as a matter of federal due process. Because we decide this case under the test articulated in Classen, we do not reach the federal due process question. [6] We have previously applied the Classen test to voice identifications in the same manner as photographic throwdowns or lineups. State v. Lee, 56 Or App 147, 152 n 7, 641 P2d 589 (1982); State v. Davie, 56 Or App 507, 516 n 4, 642 P2d 680, rev den, 293 Or 146 (1982). [7] We recognize that the old “notion that a witness’s certainty in his or her identification of a person as a perpetrator also reflected the witness’s accuracy has been ‘flatly contradicted by well-respected and essentially unchallenged empirical studies.’ ” State v. Lawson, 239 Or App 363, 385, 244 P3d 860 (2010), rev allowed, 350 Or 532 (2011) (citation omitted). Nonetheless, our cases have continued to consider this factor in weighing the reliability of identification evidence. [8] Steele testified only that the relevant perpetrator made a few commands during the robbery.