After a jury trial in the Superior Court, the defendant was convicted of armed robbery while masked, assault and battery, and assault by means of a dangerous weapon. He appealed from these convictions, arguing that (1) the judge abused her discretion by admitting in evidence five photographs of the defendant selected by the victim at an out-of-court identification procedure and (2) the judge erred in failing to give the defendant’s requested jury instruction that there is no correlation between a witness’s confidence in his or her identification and the accuracy of that identification. We transferred the case from the Appeals Court on our own motion, and we affirm.
*5901. Background. The jury could have found the following facts. At approximately 5:15 a.m. on July 18, 2002, a man entered a variety store in Lowell armed with a large kitchen knife and wearing a hat and a cloth mask over his face, with only his eyes visible. He shouted to the victim, the store clerk, several times, “Give me the money,” and stood next to the victim while he opened two cash registers. After taking $350 from the registers, he ordered the victim to “get down” on the floor and punched him in the face. He then took three packages of Newport cigarettes off the counter. As the perpetrator was leaving, a customer entered the store. Brandishing the knife, the perpetrator ordered him to the floor, but the customer ran out the door, and the perpetrator ran out in the opposite direction. When the police arrived shortly after the robbery, the victim described the perpetrator and told police that he was “Spanish.” The robbery also was recorded by the store’s video cameras.
On July 24, 2002, the victim informed a Lowell detective that, during the robbery, he had recognized the perpetrator’s voice as belonging to a frequent customer, although he did not know the man’s name. This customer had visited the store almost every day for six months to buy Newport cigarettes, and the victim had spoken with him briefly on each occasion. The victim also had seen this customer multiple times at a nearby supermarket; in fact, ten days before the robbery, they had an argument at that supermarket about the price of stamps at the variety store. The victim explained that he did not immediately disclose this information to police because he feared retaliation by the defendant. When the victim went to the Lowell police station the next day, the detective entered information the victim provided about the defendant’s appearance (age, ethnicity, and hair style) into the department’s photographic imaging system, which generated a computerized display of 1,344 photographs of men fitting this general description. While the detective waited outside the room, the victim reviewed these images and selected five different photographs of the defendant, Hector Cruz, identifying him as the customer who committed the robbery. On August 2, 2002, Lowell police arrested the defendant, who was later indicted on charges of armed robbery while masked, assault and battery, and assault by means of a dangerous weapon.
*591At trial, the victim testified to the facts described above; he also stated that he had no doubt that the defendant was the perpetrator. The defendant challenged the victim’s identification and presented an alibi defense.1 The defendant presented a motion in limine to exclude the five photographs, but the judge denied this motion, over the defendant’s objection, finding the photographs relevant to the issue of identification. The judge also observed that the photographs lacked any indicia of mug shots, and she instructed the jury to draw no negative inferences from the fact that the police possessed photographs of the defendant. The defendant requested a jury instruction stating that there is no correlation between a witness’s certainty in his or her identification and the accuracy of that identification. The judge declined to give this instruction, noting the defendant’s objection, but she instructed the jury that a witness may be honest but mistaken regarding an identification.2 The jury convicted the defendant of all charges.
2. Standard of review. Because the defendant preserved both issues on appeal with timely objections, we review the proceedings below for prejudicial error. This requires a two-part analysis: (1) was there error; and (2) if so, was that error prejudicial. An error is not prejudicial if it “did not influence the jury, or had but very slight effect”; however, if we cannot find “with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error,” then it is prejudicial. Commonwealth v. Flebotte, 417 Mass. 348, 353 (1994), quoting Commonwealth v. Peruzzi, 15 Mass. App. Ct. 437, 445 (1983).
*5923. Photographic evidence. The defendant maintains that the judge abused her discretion in admitting the five photographs in evidence. He argues that the Commonwealth had no need to introduce these photographs because the victim did not use them to identify the perpetrator; instead, he maintains, the victim recognized the perpetrator’s voice as belonging to a frequent customer, then selected photographs of that customer, who may or may not have been the person who committed the robbery. The Commonwealth contends that the photographs were necessary to prove its case because the victim’s ability consistently to identify the defendant out of more than 1,300 photographs demonstrated his familiarity with the defendant and his corresponding ability to recognize the defendant’s voice.
Decisions about the admissibility of photographic evidence are “left to the discretion of the trial judge, and we will overturn the judge’s decision only where a defendant is able to bear the heavy burden of demonstrating an abuse of that discretion.” Commonwealth v. Waters, 399 Mass. 708, 715 (1987). The admission of photographs that form the basis of an out-of-court identification is “relatively routine.” Commonwealth v. Picher, 46 Mass. App. Ct. 409, 416 (1999), quoting Commonwealth v. Austin, 421 Mass. 357, 364 (1995). Police photographs used in out-of-court identification procedures may be admitted if the following criteria are satisfied: (1) the prosecution must show some need for their introduction; (2) the photographs should be offered in a form that does not imply a prior criminal record; and (3) the manner of their introduction should not call attention to their source. Commonwealth v. McAfee, 430 Mass. 483, 493 (1999). Commonwealth v. Rodriguez, 378 Mass. 296, 309 (1979). “In circumstances where (as here) the defense claims misidentification and alibi, the Commonwealth’s need for the photographs is apparent. . . . The photographs helped to buttress the reliability of the identification[s].” Commonwealth v. Picher, supra at 416. See Commonwealth v. Weaver, 395 Mass. 307, 309 (1985) (no error in admitting photograph where “assailant’s identity was a live issue at trial”); Commonwealth v. Gee, 36 Mass App. Ct. 154, 158 (1994) (“Where the sole issue at trial was the identification of the defendant, there can be little doubt as to the Commonwealth’s need to use the photographs”). *593Identification was a central issue in this case: the strongest evidence that the defendant was the person who committed the robbery was the victim’s identification of him as the perpetrator. Because the victim testified that he recognized the defendant based primarily on his voice, a key issue before the jury was whether the victim knew the defendant well enough to identify his voice. See Commonwealth v. Mezzanotti, 26 Mass. App. Ct. 522, 527 (1988) (“a voice identification may be considered by a jury as long as the witness expresses some basic familiarity with the voice he or she claims to identify”). The victim’s selection of five photographs of the defendant, with different clothing, hair styles, and facial hair, was relevant to the issue of familiarity: the victim’s ability to identify the defendant’s photographs could indicate how well he knew the defendant, which would be relevant to how often he had heard the defendant speak and, consequently, how familiar he was with the defendant’s voice.
The defendant contends that the Commonwealth had no need to introduce the photographs in this case because he “effectively conceded” that the victim’s identification of him as a customer was accurate. See Commonwealth v. Smith, 21 Mass. App. Ct. 619, 622 (1986), S.C., 400 Mass. 1002 (1987). We disagree. The defendant did not concede that the victim properly identified him as the perpetrator, or even that he was the same customer identified by the victim. The defendant testified that he did not go into the store frequently, that he did not speak with the victim often, and that there was no confrontation in the supermarket before the robbery, all of which is contrary to the victim’s testimony that he saw the defendant and spoke to him every day for six months and argued with him less than two weeks before the robbery. Furthermore, defense counsel, on cross-examination, challenged the victim’s ability to see and hear the perpetrator during the robbery and argued in his closing that the victim did not interact with defendant enough to be familiar with his voice. The Commonwealth bore the burden of persuading the jury of the accuracy of the victim’s identification beyond a reasonable doubt, see, e.g, Commonwealth v. Cuffie, 414 Mass. 632, 640 (1993) (Appendix); in determining whether the Commonwealth had met this burden, the jury *594properly could consider the victim’s ability to identify different photographs of the defendant, which could counter the defense argument that the victim did not know the defendant well enough to recognize his voice. See Commonwealth v. McAfee, supra at 493-494 (photographs selected by witness from array properly admitted where defendant argued that witness may not have seen perpetrator well enough to make reliable identification).
The defendant’s contention that the photographs were unduly prejudicial because they conveyed his criminal history to the jury is without merit. Whether the probative value of relevant evidence is outweighed by its prejudicial effect is within the judge’s discretion. Commonwealth v. Dunn, 407 Mass. 798, 807 (1990). Here, any prejudice to the defendant from the admission of the photographs was minimal. As the judge observed, the photographs had been sanitized and bore no indication that they were mugshots; no identifying marks or height charts were visible. See Commonwealth v. Waters, supra at 715 (photograph of defendant from shoulders up and wearing street clothes not immediately recognizable as mugshot); Commonwealth v. Blaney, 387 Mass. 628, 638 (1982) (prosecutors and judges must use reasonable means to avoid calling jury’s attention to source of photographs used to identify defendant); Commonwealth v. Smith, 29 Mass. App. Ct. 449, 452 (1990) (photographs properly sanitized where “no numbers, scales, signs or other signs of police origin were visible”). The prosecutor did not refer to the pictures as “mugshots” or connect them to the defendant’s prior criminal activity in any way. Commonwealth v. Blaney, supra; Commonwealth v. Smith, supra at 453. Moreover, the judge gave the jury a cautionary instruction on the source of the photographs, even more extensive than a similar instruction approved in Commonwealth v. Blaney, supra at 636-637 & n.7, to minimize any negative inferences regarding the defendant’s criminal history.3 We presume that the jury understood and followed that instruction. Commonwealth v. Jackson, 384 Mass. 572, 579 (1981). Given the relevance of the photographs to the identification of the defendant and the *595minimal risk of undue prejudice, the judge did not abuse her discretion in admitting the photographs of the defendant.
4. Requested jury instruction. The defendant also challenges the judge’s refusal to instruct the jury that the confidence of an identifying witness does not correlate to the accuracy of the identification.4 He cites several studies supporting his position and notes that this court acknowledged the weak correlation between confidence and accuracy in Commonwealth v. Santoli, 424 Mass. 837, 845-846 (1997). The defendant thus urges us to conclude that a cautionary jury instruction is required to reflect the “current science” regarding identification, particularly in cases involving voice recognition, which he suggests is less rehable than eyewitness identification.
The judge, however, did not err in declining to give the requested instruction. Although we have held that judges should no longer invite juries to consider the strength of a witness’s identification in the standard identification instruction, we did not hold that witness assessments of confidence could not be elicited or considered. See Commonwealth v. Santoli, supra at 846. In fact, the Santoli case held that witnesses could continue to testify about their confidence in identifications, and counsel could challenge such statements. See id. (significance of witness’s confidence “should be left to cross-examination and to argument of counsel”). See also Commonwealth v. Watkins, 63 Mass. App. Ct. 69, 74 (2005) (“Supreme Judicial Court. . . has not precluded witness testimony regarding certainty, or prohibited counsel from probing the subject or arguing about *596it”); Commonwealth v. Cowans, 52 Mass. App. Ct. 811, 814 (2001) (“We decline to extend Santoli to prohibit inquiry on direct examination about a witness’s level of confidence in an identification”); Commonwealth v. Drane, 47 Mass. App. Ct. 913, 914 (1999) (Santoli holding that judges should not instruct juries to consider witness confidence in identification “does not mean that counsel are not to be permitted to probe that subject or argue about it”). Here, defense counsel cross-examined the victim at length about his identification, and he argued in his closing that the victim could not be sure of the perpetrator’s identity. It was then for the jury to determine the weight to give to the identification, including any statements of certainty or uncertainty.5 Commonwealth v. Cowans, supra (“The relevancy, as well as the weight, of a witness’s level of confidence in an identification will vary from case to case, depending on a multitude of factors. The significance that should be given to the level of confidence is . . . for the jury to determine, not the court”).
The jury instructions on identification given by the judge, substantially those approved in Commonwealth v. Rodriguez, 378 Mass. 296, 310-311 (1979) (Appendix), as modified in Commonwealth v. Cuffie, supra at 640-641 (Appendix), and Commonwealth v. Santoli, supra at 845, were adequate. The factors provided for the jury’s consideration in assessing eyewitness testimony — including the capacity and opportunity of the witness for observation, the timing of the identification, conditions at the scene, past contact between the witness and the defendant, and any failed or inconsistent identifications — are applicable to assessing “earwitness” testimony as well.6 Although no specific instructions were provided on voice *597identification, language taken from the judge’s identification instructions, set out in the margin below, made clear to jurors that a witness may use different senses to make an identification, and that similar criteria should be applied to assess the reliability of that identification.7 Furthermore, by instructing the jury about the possibility of an honest but mistaken identification, Commonwealth v. Pressley, 390 Mass. 617, 619-620 (1983), the judge sufficiently conveyed that the victim, despite his certainty, could have identified the defendant erroneously as the perpetrator. “The judge is not required to grant a particular instruction so long as the charge, as a whole, adequately covers the issue.” Commonwealth v. Daye, 411 Mass. 719, 739 (1992), quoting Commonwealth v. Anderson, 396 Mass. 306, 316 (1985). It may be within the judge’s discretion to give a charge similar to the one requested by the defendant, but it was not required here. Cf. Commonwealth v. Hyatt, 419 Mass. 815, 819 (1995) (judge may, but need not, give jury instruction about cross-racial identification).
This court has recognized that “based on a trial record or on the published results of studies, or both, some new principle concerning the process of eyewitness identification may become sufficiently reliable ... to justify formulating a jury instruction.” Commonwealth v. Hyatt, supra at 818. However, this is not the appropriate case for us to consider whether such an instruction is warranted regarding the relationship between the confidence of an identifying witness and the accuracy of that witness’s identification. The defendant cites a number of studies, but he did not seek to call an expert witness on voice identification or the relationship between witness confidence and accuracy, although he was free to do so. See id. (expert testimony about eyewitness identification admissible at discre*598tian of judge); Commonwealth v. Francis, 390 Mass. 89, 98-101 (1983) (same). Consequently, there was no hearing or testimony regarding the reliability of these studies or their general acceptance in the scientific community, and there is no basis on which this court may now assess their reliability. See Commonwealth v. Lanigan, 419 Mass. 15, 24-26 (1994). Moreover, after conducting our own review of the research cited by the defendant, it is not at all clear that there is “no proven relationship” between confidence and accuracy. Although these studies indicate that the correlation between confidence and accuracy may be weak or limited, there does not appear to be a consensus within the psychological community that “no” relationship exists. Compare Solan, Hearing Voices: Speaker Identification in Court, 54 Hastings L.J. 373, 412 (2003) (witness’s “level of confidence correlates only slightly with the likelihood of accuracy”), and Yarmey, Earwitness Speaker Identification, 1 Psychol., Pub. Pol’y, & Law 792, 803 (1995) (“accuracy-confidence relationships generally have been found to be statistically nonsignificant, or when significant to be low-positive or low-negative”), with Yarmey et al., Commonsense Beliefs and the Identification of Familiar Voices, 15 Applied Cognitive Psychol. 283, 297 (2001) (study found correlation between confidence and accuracy in identification of different voices by the same subjects, which “suggests that some weight should be given by legal professionals to earwitnesses’ statements of confidence and accuracy for more familiar speakers than less familiar ones”); Wells, Eyewitness Identification Procedures: Recommendations for Lineups and Photo-spreads, 22 Law & Hum. Behav. 603, 622 (1998) (Wells) (“the most exhaustive review to date . . . suggests that witnesses who are highly confident in their identifications are somewhat more likely to be correct as compared to witnesses who display little confidence”); Sporer, Choosing, Confidence, and Accuracy: A Meta-Analysis of the Confidence-Accuracy Relation in Eyewitness Identification Studies, 118 Psychol. Bull. 315, 322 (1995) (“when limited to witnesses who make positive identification under laboratory conditions, confidence appears to be a somewhat stronger predictor of accuracy”). The wording of the defendant’s requested instruction thus does not appear *599accurately to reflect the current state of the scientific research, and the judge was under no obligation to correct or modify the requested instruction. United States v. Oreto, 37 F.3d 739, 749 (1st Cir. 1994), cert. denied, 513 U.S. 1177 (1995) (trial judge “is not required to edit a proposed instruction to delete the bad and preserve the good”). See Commonwealth v. Adams, 34 Mass. App. Ct. 516, 519 (1993), quoting K.B. Smith, Criminal Practice and Procedure § 1846 (2d ed. Supp. 1992) (requested jury instruction must be “substantially correct” for failure to give that instruction to constitute reversible error). We therefore conclude that the judge did not err in declining to give the instruction.
Finally, we note that the defendant cites several studies and books that have found that jurors are strongly influenced by eyewitness statements of confidence, perhaps more than any other factor, while other research suggests that confidence is not a reliable indicator of identification accuracy. See, e.g., E.F. Loftus & J.M. Doyle, Eyewitness Testimony § 1-3, at 3 (3d ed. 1997) (“eyewitness testimony is likely to be believed by jurors, especially when it is offered with a high level of confidence, even though the confidence of that witness and the accuracy of the eyewitness may not be related”); Wells, supra at 620 (there is “consistent evidence” that “the confidence that an eyewitness expresses in his or her identification during testimony is the most powerful single determinant of whether or not observers of that testimony will believe that the eyewitness made an accurate identification”); B. Cutler & S. Penrod, Mistaken Identification: The Eyewitness, Psychology, and the Law 186 (1995) (“jurors tend to undervalue viewing conditions that are known to predict identification accuracy and instead base their decisions in part on . . . witness confidence,” which is often a “poor predictor[] of identification accuracy”). Although the defendant made a timely objection to the Commonwealth’s inquiry about the victim’s certainty regarding his identification, he waived this issue by not raising it on appeal. See, e.g., Commonwealth v. Cundriff, 382 Mass. 137, 150 n.22 (1980), cert. denied, 451 U.S. 973 (1981), citing Mass. R. A. P. 16 (a) (4), as amended, 367 Mass. 921 (1975). We do not imply by today’s decision that we are *600disinclined to consider these issues in some future case on an appropriate and fully developed record.
5. Conclusion. For the reasons discussed above, we conclude that (1) the judge did not abuse her discretion in admitting the photographs of the defendant in evidence, and (2) the judge did not err in declining to give the requested jury instruction about confidence and accuracy in eyewitness identifications.
Judgments affirmed.
The defendant testified that he was asleep at a motel in Tewksbury at the time of the robbery.
The judge gave the following instruction regarding honest but mistaken identification: “In assessing the testimony in which a defendant was identified as a perpetrator of a crime, you must consider the possibility of a good faith error by the identifying witness; that is, in addition to assessing the credibility of the witness, you must also consider whether the witness is honestly mistaken in his identification of the defendant as the perpetrator of the crime. Even if you find that the witness is sincere and honest in his or her belief that a defendant committed the offense, you must still return a verdict of not guilty unless you are convinced beyond a reasonable doubt that the identification is accurate. Again, the burden is on the Commonwealth to prove that the identification, however honest, is correct.”
The instruction given by the judge matched almost verbatim the instruction requested by the defendant about the source of the photographs.
The defendant’s requested instruction read: “In weighing the testimony of an eyewitness, including the testimony of a witness who testified as to his observations and recollections, you should consider that there is no proven relationship between a witness’ [sic] confidence in his identification and the accuracy of the witness’ identification. An eyewitness’ certainty about an identification may have many sources and does not necessarily bear on the correctness of the identification. Indeed, experts have concluded that an eyewitness’ representation of confidence is among the least significant factors in predicting the accuracy of the identification. The correctness of the identification is for you, the jury, to decide.” (Citations omitted.) This language appears to be taken from an instruction proposed in E.F. Loftus & J.M. Doyle, Eyewitness Testimony § 12-13, at 343 (2d ed. 1992). In his reply brief, the defendant modified his position somewhat and argued that the judge could have instructed that the correlation between, the accuracy and the certainty of the identification is slight or limited, rather than nonexistent.
One commentator noted that it is “important to distinguish the type of situation that tends to lead to some confidence/accuracy relationship from the type of situation that leads to little or no confidence/accuracy relationship.” E.F. Loftus & J.M. Doyle, Eyewitness Testimony § 3-12, at 67 (3d ed. 1997). This suggests that the relationship between confidence and accuracy may vary depending on the circumstances surrounding an identification, and that confidence may be an appropriate issue for the jury to consider in some cases.
In fact, the factors provided for assessing the reliability of eyewitness identifications are similar to those suggested by the defendant in a proposed modified instruction on the reliability of “earwitness” identification. These include the witness’s opportunity to hear the speaker, conditions at the scene, *597the witness’s familiarity with the speaker’s voice, and any failed identification attempts. This proposed “earwitness” identification instruction does not appear in the record below: it seems that the defendant proposed it for the first time before this court.
The judge instructed the jury: “In general, a witness bases any identification he or she makes on his or her perception through the use of his or her senses. Usually the witness identifies an offender by the sense of sight. However, this is not necessarily so and he may use other senses to make an identification.”