Of three points that the defense has argued on appeal, we concentrate primarily on whether the setting for a voice identification by the complaining witness was so sugges*115live as to deprive the defendant of a fair trial. We decide that the voice identification procedure, although not without flaw, adhered sufficiently to the standards prescribed in Commonwealth v. Marini, 375 Mass. 510, 517 (1978), and was not impermissibly suggestive. We also consider two other points raised by the defendant, one relating to the denial of a continuance and the second to appellate inspection of mental health records. The defendant was convicted of aggravated rape.1 We affirm the convictions.
1. Facts. There was evidence at trial from which the jury could find the following. On March 28, 1994, the victim, a woman in her late forties, was raped and indecently assaulted in her own bed by an intruder. Some time past 3 a.m., the victim, who had been awakened by a telephone call, heard a door squeak and then saw a man standing beside her. There was some illumination from an adjoining industrial park, and a 100 watt bulb lit in a spare bedroom, to which the door was open, shed additional light. There was also a night light in the bedroom in which the victim slept, an additional light source that the intruder yanked from its socket. For the next two hours, the intruder repeatedly raped the victim vaginally, anally, through cunnilingus, and by compelled fellatio. During the attack, the man spoke continuously to the victim in a soft voice. The victim watched as the man left through a window, and when he was gone, she at once telephoned a friend who, in turn, called police.
To the police the victim gave a description of her attacker: he was white, with dark shoulder-length hair and a mustache. He smelled of cigarettes. He was wearing a dark flannel jacket with leather sleeves and something like “American Legion” sewn on the back. His underwear shorts were boxer type, white, and with a snap fly. He was shod in white sneakers, with a blue stripe and some red trim. His watch was of an electronic type; it had beeped during the attack. On the basis of the victim’s description, a police artist drew a composite picture. A lieutenant of the Franklin police, who saw the composite, thought the image resembled that of the defendant Paul DeMaria, and, on the day after the attack, presented the victim with a photo array that included a picture of DeMaria. She selected DeMaria from *116the array, although expressing concern that the person in the picture wore glasses, whereas her assailant had not.
Nine days after the incident, the police brought DeMaria and nine other men in for a lineup identification. By that time DeMaria had shaved off his mustache and cut his hair. For the lineup, the police planted above DeMaria’s upper lip a false mustache. When DeMaria stepped forward from the lineup, the victim visibly trembled.. She asked to hear his voice. In the lineup DeMaria wore a placard bearing the number four. The phrase “number four” was intoned several times by the police and the victim during this portion of the lineup.
At that point the victim was excused while the police and representatives from the district attorney’s office arranged for a voice identification. The victim was readmitted into the room and asked to stand with her back to the men in the lineup. Each man in the lineup was asked to read a passage from a Boston Globe sports column: “My goal is to play in the big leagues, with Boston or without them. I just want to play in the big leagues.” The participants in the lineup were asked to read out loud starting with participant number ten and working back to number one. So, for example, after number ten had spoken the passage a State trooper called out, “Number nine, come forward, please.” Ultimately number four was called forward and read the passage. The victim made a response recorded in the transcript of the identification proceeding as inaudible, but the response caused one of the two troopers directing the proceeding to say, “Just repeat, please.” Number four, who was the defendant, repeated, “I just want to play in the big leagues,” whereupon the victim, whom the stenographer described, “witness visibly trembling and crying,” said, “That’s him — that’s him. That’s his voice.” The victim then listened to numbers three, two, and one, asked to hear number four again, and confirmed her identification. She also asked to see number four walk once more and confirmed her visual identification.
2. The voice identification. The claim of prejudice derives from the consistent use of number four in relation to the defendant, both at the visual identification and the voice identification. The defendant argues that the victim, having first identified number four visually, albeit with a touch of hesitation, was programmed to react to number four when she was listening to the voices of the participants in the lineup. The consequence, as the defendant sees it, was that the victim was subjected to impermissible suggestion by number.
*117Concern has been expressed in reported cases about the danger of unfair prejudice to the suspect that inheres in voice identification. Commonwealth v. Marini, 375 Mass, at 516-517, and authorities there cited.2 That opinion cautioned that (1) a witness who has a basis for making an identification by sight ought not to be asked to make a voice identification unless the witness asks to hear the voice; (2) a one-on-one audition ought to be avoided, i.e., there should be an audition lineup; (3) the witness ought not to be viewing the suspects as she listens to them; (4) the words spoken by the suspects ought not be the same as those heard by the witness at the scene; and (5) the voice recognition test ought to be conducted close to the time of the crime. Id. at 517. The Marini decision, at 517, acknowledges that there will be occasions where one or more of those criteria may be overlooked, as in an immediate postcrime showup identification. Significantly, the identification procedure in this case adhered to all of the precautionary criteria set out in Marini. Compare Commonwealth v. Gauthier, 21 Mass. App. Ct. 585, 587-588 (1986). Did calling out the suspect in the two phases of identification by “number four” inject a fatally suggestive mistake?
In answering that question, what the victim said at the suppression hearing and what the motion judge found as fact take on great importance. The prosecutor asked the victim:
“Now, . . . had you burned in your mind any particular numbers of any participants when you returned to the [lineup] area?”
The witness answered, “No.” She said in answer to the next question that she did not know whether the exact same people were in the voice lineup as the face lineup. She reacted and made her identification on the basis of what she heard. The victim’s testimony warranted the motion judge’s finding — she after all had the advantage of observing the witness — that, “I credit the testimony of the victim that she focused on the voices, and not on the numbers being called.” We do not disturb a motion judge’s subsidiary finding of fact if it is supported by the evidence. Commonwealth v. Mahnke, 368 Mass. 662, 689 (1975), cert, denied, 425 U.S. 959 (1976). Commonwealth v. *118Gutierrez, 26 Mass. App. Ct. 42, 45 (1988). In Commonwealth v. Melvin, 399 Mass. 201, 207 (1987), for example, an identifying witness had seen an intruder leap from a third-floor balcony and land on his shoulder. There was an element of suggestiveness, therefore, in a picture of the defendant with his arm in a sling. The trial judge in the Melvin case accepted the testimony of the witness that he had a vivid memory of the assailant and was identifying the “individual, not the sling.” Ibid.
Obviously, it would have been better had the authorities not called the defendant by the same number in both the visual and the voice lineups. The question, however, is not whether that was suggestive — we may assume it was — but whether it was so suggestive as to have deprived the defendant of due process of law. See Commonwealth v. Dickerson, 372 Mass. 783, 789 (1977). We think not when looking at the total picture, the parts of which included the preceding visual identification of the defendant, the strong emphasis by the witness on voice quality, and her disavowal of reaction to a number label. Compare Commonwealth v. Miles, 420 Mass. 67, 78-80 & n.15 (1995).3 It was open to the defense to argue to the jury that the suggestiveness that had infiltrated the voice identification rendered it unreliable and, indeed, defense counsel did just that. See Commonwealth v. Melvin, 399 Mass, at 208.
3. Denial of continuance. On the first day of the trial both the defense and prosecution learned for the first time that the victim had tested positive for chlamydia, a sexually transmitted disease. Defense counsel proposed to introduce evidence of that fact (through hospital records) and argue to the jury that, because the disease is highly contagious and has a two- to five-week incubation period, the absence from the prosecution’s case of evidence that DeMaria had contracted the disease tended to show he was not the rapist. On the fourth day of trial, the defense brought out the fact of the positive chlamydia test while cross-examining the victim’s treating physician. The prosecution responded with an expert who testified that there was a likelihood that the test was a false positive. Some time during *119the morning session of the previous trial day, i.e., the third, the Commonwealth’s expert, a physician from the Department of Public Health, had been subjected to a voir dire examination. At that time, defense counsel moved successfully for funds to consult with an expert. The trial judge also adjourned the trial at about 1 p.m. that day so that defense counsel might work on obtaining responsive medical testimony.
The next morning, a Friday, defense counsel reported that she had identified and reached an expert but that he was on vacation. Counsel asked for a continuance. The judge denied it, giving defense counsel the weekend to reel in her designated expert or any expert. On Monday, the expert was still not available and defense counsel moved for a three-day continuance to the following Thursday. The judge denied the motion.
Whether a motion for a continuance should be allowed lies within the sound discretion of the trial judge. Commonwealth v. Watkins, 375 Mass. 472, 490 (1978). Hunnewell v. Hunnewell, 15 Mass. App. Ct. 358, 363 (1983). When the ground for a continuance involves late disclosure of evidence without any showing of bad faith or carelessness by the prosecution, “a defendant is required to show material prejudice from the disclosure before a new trial can be considered.” Commonwealth v. Hamilton, 426 Mass. 67, 70 (1997). It is not a minor matter to suspend a jury trial for three days. There is no material prejudice here. The chlamydia information was less than dynamite. The best the defense could do with it was to pound the absence of evidence that DeMaria had caught chlamydia and defense counsel did just that rather adroitly when she made her closing argument to the jury.4 The trial judge did not abuse her discretion in refusing a continuance.
4. Appellate review of the victim’s mental health records. There were two in camera examinations of the victim’s mental health records. First a judge of the Superior Court, several months before trial, examined the records under the defense’s *120Bishop5 motion and found that the “medical records of the complaining witness . . . are all privileged and of very doubtful relevancy.” At trial, the defense made a particularized motion for an in camera review of the victim’s medical records, see Commonwealth v. Bishop, 416 Mass. 169, 184 (1993), to look for evidence of bad eyesight, motive to fabricate, or any information concerning her chlamydia that may have been of an exculpatory nature. The trial judge allowed the motion and examined the records, which she described as records of a psychologist and a psychiatrist with whom the victim had. consulted during a period running from January, 1990, to August, 1991.
The judge, after examining those records, reported that “there are no entries and there is no record of any relevance to the subject matters expressed in your Bishop submission.” This occurred on the fourth day of trial, by which time the judge would have acquired an understanding of what the defense case was about and a more than ordinary sense of what was relevant to that case. The defense now asks us to make our own determination whether anything relevant to the defense and exculpatory in nature lodges in those records, even though a judge’s determination of what is relevant is subject to a deferential abuse of discretion standard. Commonwealth v. Pare, 43 Mass. App. Ct. 566, 572 (1997), S.C., All Mass. 427 (1998).
As to the three issues that the defense had raised, it is most implausible that records of mental health consultations conducted three and four years before the incident in question would have anything to disclose that would be useful. Mental health therapists are unlikely to have dwelled on the acuity of the victim’s eyesight, her chlamydia, which appears to have come to light three or four years later, or reasons for bias against a defendant whom she had not then met. We do not read Commonwealth v. Fuller, 423 Mass. 216 (1996), or Commonwealth v. Pare, supra, as automatically requiring appellate reexamination of the privileged records on a de novo basis because a defendant requests it. The Fuller opinion, at 224, dealing with records protected by an absolute privilege under G. L. c. 233, § 20J, rejects “virtually automatic in camera inspection” of private and sensitive materials of the kind here involved and requires, at 226, “a good faith, specific, and reasonable basis for believing that the records will contain exculpatory evidence *121which is relevant and material to the issue of the defendant’s guilt.” An in camera review, while less intrusive than public disclosure or disclosure to a defendant’s attorney, “is nonetheless a substantial invasion of the privacy of a complaining witness.” Commonwealth v. Fuller, supra at 225. We resist subjecting the victim’s conditionally privileged records (G. L. c. 233, § 20B) to inspection by three additional sets of prying eyes without some plausible hypothesis of relevance — as there certainly was in the Pare case. There the records in question went directly to the core issue to be resolved, the reliability of the juvenile witness reporting events that had occurred some years before and in circumstances where he may have had motives to fabricate. Commonwealth v. Pare, 427 Mass, at 431. There were also questions about whether the young complaining witness had been subjected to suggestive questioning. There is no such inherent relevance in the material the defendant wants reviewed in this case.
Judgments affirmed.
The defendant was also convicted of indecent assault and battery on a person over fourteen years of age and breaking and entering in the nighttime to commit a felony.
As to the use of visual analysis of spectrograms, commonly called “voiceprints,” see Commonwealth v. Lykus, 367 Mass. 191 (1975).
It is of at least parenthetical interest that the police, when they searched the defendant’s apartment — pursuant to a warrant — found items that the victim had described to them after the attack, namely, a leather-sleeved flannel jacket with lettering on the back, snap-fly boxer shorts, white sneakers with a blue stripe, cigarettes, and an electronic watch. At trial, the victim identified each item of clothing as similar to what her attacker wore.
Defense counsel said to the jury: “Then, they brought in some expert from Harvard [sic] to tell us all that maybe [the victim] had chlamydia, but maybe the test wasn’t very good. I suggest to you, folks, that if this man had chlamydia, that test would be touted as the greatest invention since penicillin. If this man had chlamydia, I suggest you would have an 8" x 10" glossy right here of the organism showing it. I suggest to you folks that he doesn’t have it. How does she explain that?”
Commonwealth v. Bishop, 416 Mass. 169 (1993).