Appellants Lorenzo Allen, Frances Sylvester Lindsey and Roger Lee Harrell were convicted by a jury of armed bank robbery in violation of 18 U.S.C. § 2113(a), (d) and of conspiracy to commit that offense in violation of 18 U.S.C. § 371. The main issue on appeal concerns the propriety of testimony by a police officer and a parole officer identifying Allen and Harrell as the individuals appearing in bank surveillance photographs. Finding no error in the use of this testimony, or in other actions asserted as error by appellants, we affirm.
I.
On February 12, 1985, shortly before 10:00 a.m., three men robbed a Dominion Bank office in Portsmouth, Virginia. All three were dressed as construction workers and wore hardhats and goggles; two of the men had guns. Customers and employees of the bank were ordered into the vault, and consequently were unable to identify any of the robbers. At some point during the holdup, however, the men removed their goggles, allowing bank surveillance cameras a better view of their faces. They escaped with $33,650.
A bank customer outside the building during the robbery gave police the license number of the getaway car. A police officer spotted the car, gave chase, and finally apprehended the car’s only occupant, Michael Joseph Spicer. Spicer was acquitted in a separate trial. Fingerprints taken from the car were determined to be those of appellants Allen and Lindsey.
After the surveillance photographs were processed, Special Agent Cross of the FBI showed them to other law enforcement of*935ficials in an attempt to identify the bank robbers. Cross took three photographs to U.S. Probation and Parole Officer Michael McDonald, who immediately identified appellant Harrell, a federal parolee at the time. Agent Cross made no suggestion to McDonald as to the possible identity of the individuals in the photographs. Cross also showed the photographs to Detective William L. Garrison of the Norfolk Police Department’s Robbery Squad. Garrison later identified appellant Allen as one of the individuals in the photographs. Garrison had known Allen for fifteen years, though not in connection with any criminal activity.
Prior to trial, appellants Harrell and Allen sought to suppress identification testimony by McDonald and Garrison. The district court rejected these efforts, and held the identification testimony admissible so long as the occupation of the witnesses was not revealed by the prosecution. At trial, foundation for the testimony was established outside the presence of the jury, and the defense was given complete opportunity to cross-examine these witnesses both before the jury and outside its presence. In their testimony, McDonald and Garrison identified appellants Harrell and Allen, respectively, as the individuals in bank surveillance photographs. Testimony revealed to the jury neither the occupation of the witnesses nor any past criminal activity of appellants. Defense counsel on cross-examination attacked the ability of each witness to make the identification. Two other witnesses also identified Harrell as one of the men in the bank photographs.
II.
On appeal, Harrell and Allen renew their objection to the identification testimony of McDonald and Garrison. They argue first that this opinion testimony was neither “rationally based on the perception of the witness” nor “helpful to a clear understanding ... of a fact in issue,” as required by Fed.R.Evid. 701. Second, they argue that even if otherwise admissible, this testimony should have been excluded under Fed.R. Evid. 403 as unfairly prejudicial because defendants were limited in their cross-examination for fear of revealing their prior criminal activities. The admission of testimony over these objections, however, is left largely to the discretion of trial judges. United States v. Brainard, 745 F.2d 320, 323 (4th Cir.1984), cert. denied, — U.S. -, 105 S.Ct. 2320, 85 L.Ed.2d 839 (1985). United States v. Gould, 741 F.2d 45, 51, n. 6 (4th Cir.1984). We see no abuse of that discretion here. The testimony admitted by the district court was both rationally based and helpful, and did not unfairly prejudice defendants’ case.
A.
Fed.R.Evid. 701 governs the admission of lay opinion testimony. The rule provides that non-expert testimony “in the form of opinions or inferences is limited to those opinions or inferences which are (a) rationally based on the perception of the witness and (b) helpful to a clear understanding of his testimony or the determination of a fact in issue.” These conditions for admission were satisfied here.
Both witnesses gave opinions based upon “personal observation and recollection of concrete facts.” United States v. Jackson, 688 F.2d 1121, 1124 (7th Cir.1982), cert. denied, 460 U.S. 1043, 103 S.Ct. 1441, 75 L.Ed.2d 797 (1983). McDonald had served as Harrell’s parole officer since August, 1984, approximately six months before the bank robbery. He had seen Harrell on six or seven occasions, in meetings of ten to twenty minutes each. These monthly meetings lasted until Harrell was taken into custody. Thus, McDonald’s familiarity with Harrell provided him a rational basis for testimony that it was Harrell who appeared in the surveillance photographs.
Likewise, Officer Garrison had repeated contact with appellant Allen. He had known Allen for fifteen years, seeing him at least twelve times a year. Most of these encounters were on the street as Garrison walked his beat. He last saw Allen in November or December, 1984, two to three months before the bank robbery. Allen suggests that this gap in Garrison’s knowl*936edge leaves the witness without a rational basis to identify him as one of the individuals in the photographs. Though Garrison did not testify to Allen’s appearance on the day of the robbery, his fifteen years of acquaintance provided ample basis for his testimony that Allen was the man in the •photographs. Moreover, the opportunity for observation of the defendant usually goes to the weight of the testimony, not to its admissibility. United States v. Jackson, 688 F.2d at 1125. The discontinuity in the witness’ knowledge was emphasized on cross-examination and was for the jury to balance and assess.
Rule 701 retains the traditional concern for invasion of the role of the jury by requiring that lay opinion testimony be helpful to the jury. Thus, such testimony is improper when it involves only “meaningless assertions which amount to little more than choosing up sides.” Fed.R.Evid. 701, Advisory Committee Note. Where, however, “there is some basis for concluding that the witness is more likely to correctly identify the defendant from the photograph than is the jury,” United States v. Farnsworth, 729 F.2d 1158, 1160 (8th Cir. 1984), the testimony is helpful and should be admitted.
Appellants urge that the testimony here offered nothing that the jury could not “plainly see for itself.” Both defendants identified by the witnesses were present during the trial. Thus, appellants argue, the jury was in as good a position as the witnesses to draw conclusions about the identity of those in the photographs. Accordingly, the testimony was not helpful and should have been excluded.
We believe, however, that testimony by those who knew defendants over a period of time and in a variety of circumstances offers to the jury a perspective it could not acquire in its limited exposure to defendants. Human features develop in the mind’s eye over time. These witnesses had interacted with defendants in a way the jury could not, and in natural settings that gave them a greater appreciation of defendants’ normal appearance. Thus, their testimony provided the jury with the opinion of those whose exposure was not limited to three days in a sterile courtroom setting.
This fuller perspective is especially helpful where, as here, the photographs used for identification are less than clear. The picture identified as Harrell’s, for example, shows an individual with his jacket hood pulled over his head so that his hair, forehead and right eye are not visible. The two photographs identified as Allen’s are likewise incomplete reproductions of the man in the bank. One shows a profile of a man wearing a hardhat, rubbing his forehead, with his mouth open. The other is also little more than a blurred profile, with most of the left half of the individual’s face hidden. Given the limited glimpses afforded by these photographs, we conclude that the testimony of witnesses familiar with defendants was especially helpful to the jury’s crucial identification of the men inside the bank.
We disagree with appellants’ suggestion that identification testimony by lay witnesses should be considered helpful only when a defendant’s appearance has changed between the crime and trial and the witness can testify to defendant’s appearance at the time of the crime. Many courts have noted the helpfulness of lay opinion testimony in such circumstances. See, e.g., Farnsworth, 729 F.2d at 1160; United States v. Barrett, 703 F.2d 1076, 1085-86 (9th Cir.1983); United States v. Borrelli, 621 F.2d 1092, 1095 (10th Cir.), cert. denied, 449 U.S. 956, 101 S.Ct. 365, 66 L.Ed.2d 222 (1980). There will, of course, usually be some changes in defendant’s appearance before trial, however subtle. As the district court noted, “the dress of the individual at the time of the crime will generally not coincide with his appearance in court at the time of trial.” In any event, we see no reason to make evidence of changed appearance a prerequisite to the use of identification testimony. Accord, United States v. Jackson, 688 F.2d 1121, 1125-26 (7th Cir.1982). As explained above, this testimony was helpful to the *937jury, and that is all Rule 701 requires. Accordingly, we reject the assertion of appellants Allen and Harrell that the lay identification testimony was inadmissible under Rule 701.
B.
Appellants next assert that this testimony, even if admissible under Rule 701, should have been excluded as unfairly prejudicial under Rule 403. That rule provides that relevant evidence “may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice.” Appellants find in this testimony two types of unfair prejudice. First, they assert that the jury would easily conclude that the witnesses were law enforcement officials and attribute to defendants a history of criminal activity. Second, citing United States v. Calhoun, 544 F.2d 291 (6th Cir.1976), they argue that “the prosecution’s use of these particular witnesses made it strategically undesirable for the defense to attack their credibility fully or explore reasons for bias on cross-examination.” Farnsworth, 729 F.2d at 1160. We see no merit in either contention.
Nothing at trial gave the jury any reason to conclude that McDonald and Garrison were involved in law enforcement. The court specifically instructed the prosecution not to divulge the occupation of either witness before the jury, and neither direct nor cross-examination hinted at this information. We reject appellants’ speculation that, despite the restrictions on the testimony of these witnesses, “the jury obviously knew who and what they were.” To do otherwise would impose a per se prohibition on testimony by police and parole officers on the conjecture that all such personnel fit some readily discernible personality profile.
Appellants’ main objection to the testimony of Garrison and McDonald, however, arises not from the danger of jury divination of the witnesses’ occupations but from a perceived inability to probe these witnesses for bias. Such a probe, they claim, would inescapably divulge to the jury the reason for the asserted bias — the witness’ knowledge of defendants’ prior criminal activity. Defendants claim this limitation on cross-examination made the testimony unfairly prejudicial to their case. For this proposition, they rely on Calhoun, 544 F.2d 291, which found unacceptable the admission of identification testimony by a defendant’s parole officer. For the reasons below, however, we cannot accept the reasoning of Calhoun and find no unfair prejudice from the decision to admit the testimony. Accord, Farnsworth, 729 F.2d 1158.
The only areas unexplored on cross-examination were the occupations and possible biases of these witnesses, because, as noted, the information might give the jury reason to know of defendants’ criminal activities, activities they understandably wanted to keep from the jury. Limitation of cross-examination was thus the result of a tactical choice by defendants similar to those frequently faced at trial. Nothing in the Rules of Evidence or any other source is intended to relieve criminal defendants from difficult strategic decisions. The decision of a criminal defendant to take the stand on his own behalf, for example, may result in revelation of prejudicial material a defendant would like to suppress. Courts, however, have felt no obligation to make the choice of a defendant to testify risk-free. See, e.g., Harris v. New York, 401 U.S. 222, 91 S.Ct. 643, 28 L.Ed.2d 1 (1971). Defendants chose to limit cross-examination, and we see no reason to insulate them from the natural consequences of that choice.*
Courts should, however, employ procedures that give a defendant the information necessary to make an informed dera*938sion as to the scope of cross-examination. See, e.g., Farnsworth, 729 F.2d at 1161; United States v. Pavon, 561 F.2d 799, 802-03 (9th Cir.1977). Here, the district court handled the matter in a commendable fashion. It allowed the defense to object outside the presence of the jury. Foundation for the testimony and complete cross-examination also took place with the jury absent and allowed the court to consider fully the value and possible prejudice from the use of this testimony before admitting it. This procedure gave the defense the opportunity to explore the alleged biases of the witnesses and weigh their responses when considering the scope of cross-examination at trial. The prosecution was instructed not to reveal the occupation of the witnesses before the jury, and the defense was free to cross-examine the witnesses as it chose. This procedure ensured that the decision to discuss the potentially prejudicial material was left entirely to the informed judgment of the defense.
Appellants nevertheless assert that the prejudice arising from this tactical decision so overwhelms the probative value of the testimony that we should relieve them of the choice by directing the per se exclusion of such testimony under Rule 403. This approach was adopted by the Sixth Circuit in Calhoun, but has been followed by no other circuit. See, e.g., Farnsworth, 729 F.2d 1158; United States v. Butcher, 557 F.2d 666 (9th Cir.1977). We reject Calhoun because “we disagree that the constraints on cross-examination in this situation are so extreme that the admission of a parole officer’s identification is per se an abuse of discretion.” Farnsworth, 729 F.2d at 1161. The examination of the witnesses outside the jury’s presence, for example, revealed little, if any, evidence that they were motivated by a bias against defendants. Thus, not only were appellants presented with an acceptable strategic choice, but also faced, at most, minimal prejudice from the choice they made. Given the probative value of the identifications, we decline to say the district court abused its discretion by admitting the testimony.
III.
Defendant Lindsey asserts that a statement he made to FBI agents identifying himself as one of the men pictured in the bank photographs was illegally obtained and should have been excluded. This contention is meritless. According to factual findings of the district court, three agents approached Lindsey’s car after one observed that he resembled an individual appearing in the bank photographs. One of the agents identified himself and asked Lindsey to step out of the car and onto the curb, out of heavy traffic. The agent showed Lindsey the picture, and Lindsey agreed with the agent’s suggestion that it looked like Lindsey.
We agree with the district court that this encounter constituted an acceptable stop under Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968). The agents here were well within the bounds of their authority, as articulated by the Supreme Court in United States v. Hensley, 469 U.S. 221, 105 S.Ct. 675, 681, 83 L.Ed.2d 604 (1985): “[I]f police have a reasonable suspicion, grounded in specific and articulable facts, that a person they encounter was involved in ... a completed felony, then a Terry stop may be made to investigate that suspicion.”
We also agree that Lindsey’s statement regarding the photograph was admissible. Lindsey was not in custody when he made that statement and accordingly was not protected by the requirements of Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966). See Davis v. Allsbrooks, 778 F.2d 168, 170 (4th Cir. 1985). We further find no merit in defendant Lindsey’s challenge to his subsequent confession to police after he voluntarily agreed to go to the police station and waived his Miranda rights.
IV.
Finally, appellants complain -that the prosecutor’s use of peremptory challenges to exclude blacks from the jury violated *939their Sixth Amendment rights. Evidence shows that the government’s six peremptory challenges were used to strike blacks who were members of the venire. Three of the twelve jurors who heard defendants’ case, however, were black. We recently rejected a similar challenge to the government’s use of peremptory challenges, United States v. Grandison, 780 F.2d 425, 430-31 (4th Cir.1985), and see no reason to reconsider that position here.
V.
We have considered the various contentions raised by appellants and find them to be without merit. The judgment of the district court is accordingly
AFFIRMED.
We find the claim of possible prejudice from full cross-examination particularly weak with respect to officer Garrison. Outside the presence of the jury, cross-examination of Garrison revealed that he had not known Allen in connection with any criminal activity, but had simply become familiar with Allen as he walked his beat in Norfolk. Thus, further cross-examination of Garrison would not have revealed that Allen ever engaged in any criminal activity.