After trial by a jury defendant stands convicted as the masked robber in a holdup at the Pierrepont Hotel in Brooklyn. His identification was based largely on the recognition of his voice by a single witness. The principal issue on this appeal is whether it was error for the trial court, during the witness’ cross-examination, to have refused to allow an in-court experiment by which the defendant sought to test the witness’ ability to identify another voice. By its *703affirmance, the Appellate Division held it was not. On that score our view is consonant with its decision.
The witness, Judith Skolnick, an assistnt to the hotel manager, was unshakeable in her testimony. She described in detail how she was accosted outside the manager’s office by a tall, slender man with bushy sideburns, who, wearing a stocking mask over his face and, apparently aware of the fact that she had a key to the door, ordered her at gunpoint to "Open the door or I’ll kill you”. She thereupon admitted the robber and a female companion to the office, where he commanded her to lie down on the floor .while the manager was forced to open the safe. The robber seemed familiar with the office layout and with the location of Ms. Skolnick’s purse, which he pilfered. For the next 20 or 25 minutes, as she lay on the office floor, she was continually exposed to his voice as he talked incessantly, "babbling” in ordinary conversational tones in Spanish to his companion and in English to the victims. Having resigned herself to the situation once it became evident that the intruders were not intent on physical harm, Ms. Skolnick’s tension lessened and her aural observations became increasingly alert. At one point, he asked "Do you have security here? You need some security here” or words to that effect. Before the robber and his companion departed, he tied his two victims up and politely offered to call the front desk to effect their release once he was clear of the hotel. As soon as they were gone, Ms. Skolnick told the manager that she was "[one] hundred and five percent sure” the robber was Agilio Acevedo, a security guard employed at the hotel.
Ms. Skolnick based her identification primarily on Acevedo’s voice. He had, she later testified at trial, a "slight speech impediment” which was especially apparent in the unique way he had of pronouncing the word "security”. His use of that word during the robbery had "confirmed” her recognition of his voice. According to her further testimony, Acevedo, a polite, friendly and loquacious young man, had then been working out of her office over a period of almost two years up to the very time of the robbery, in a series of assignments, the last of which was the one as security guard. During the entire period, Ms. Skolnick had had literally "hundreds” of conversations with him and thus, by the time of the crime, had come to have everyday familiarity with the sound of his voice.
Given Ms. Skolnick’s unusually extensive and extended opportunity to come to know the defendant’s speech pattern *704as it existed prior to the robbery and her full opportunity to hear him during its commission, it is not surprising that defendant did not attempt to challenge the admissibility of her voice identification testimony (see People v Dinan, 15 AD2d 786, affd 11 NY2d 350; People v Abelson, 309 NY 643). Instead, defense counsel, during her cross-examination, sought leave to subject her to the following test of her ability to recognize voices: After the witness had been blindfolded she was to hear the voice of defendant’s brother, John Acevedo, uttering the words, "open the door or I will kill you” and "you need security here”. John Acevedo had worked for the hotel as a part-time painter for some six months; at the time of the trial, he had done no work there for nearly two years, and Ms. Skolnick had not heard him speak during all that intervening period. The trial court refused to permit this in-court experiment.
Testimonial evidence, by which witnesses, whether in narrative or question and answer form, relate events which they perceived at earlier points in time, is held in no higher regard than is demonstrative evidence from which Judges and juries can themselves form perceptions through the direct use of their own senses (4 Wigmore, Evidence [3d ed], §§ 1150-1160; 21 NY Jur, Evidence, § 355, p 487). Indeed, both courts and commentators have noted, with respect to demonstrative evidence, that, when validly and carefully used, there is no class of evidence so convincing and satisfactory to a court or a jury (Jones, Evidence . [3d ed.], § 653; see United States v Skinner, 425 F2d 552; cf. Clark v Brooklyn Hgts. R. R. Co., 177 NY 359; Note, Real Evidence: Use and Abuse, 14 Brooklyn L Rev 261).
However, though tests and demonstrations in the courtroom are not lightly to be rejected when they would play a positive and helpful role in the ascertainment of truth, courts must be alert to the danger that, when ill-designed or not properly relevant to the point at issue, instead of being helpful they may serve but to mislead, confuse, divert or otherwise prejudice the purposes of the trial (see People v Fiori, 123 App Div 174; Kratche v New York Cent. R. R. Co., 228 App Div 820). When there is such a threat, the trial court itself must decide in the exercise of a sound discretion based on the nature of the proffered proof and the context in which it is offered, whether the value of the evidence outweighs its potential for prejudice (see Uss v Town of Oyster Bay, 37 NY2d 639, 641; People v Buchanan, 145 NY 1; McMahon v Brooklyn & N. Y. *705Ferry Co., 10 App Div 376; 21 NY Jur, Evidence, § 376, and cases cited therein).
The voice identification test proposed by the defendant in the case before us patently would not have duplicated the circumstances which surrounded Ms. Skolnick’s voice identification of the defendant (see Thomas v Central Greyhound Lines, 6 AD2d 649; Washington v Long Is. R. R. Co., 13 AD2d 710; cf. Starr v People, 28 Col 184). Her testimony was based on her recognition of the defendant’s voice on the day of the robbery itself when she still had immediate familiarity with it; the test of her ability to recognize the voice of defendant’s brother John was to take place two years after she had last heard it. The number of her prior verbal interactions with the defendant was proved to have been in the hundreds; there was no proof of the number of the concededly fewer interchanges she had had with John during her far more limited acquaintance with him. Her identification of the defendant was based not only on her excellent prior knowledge of his voice but also on 20 to 25 minutes of conversation during the robbery itself; her identification of the brother was to depend on the sound of two short sentences alone. Her identification of the robber had been made with her eyes open after observing an individual whose build, height, bushy sideburns, manner and mannerisms came through to her despite his mask and coincided with those of a person whose voice she recognized; her identification of the brother was to be made in a personality vacuum created by her proposed blindfolding.
Even putting aside all these essential differences, it was not her ability to identify the defendant, the real question involved, which was to be tested; the test was to be of her ability to recognize the voice of another. But her identification had not been of a previously unknown or little known voice; it was not sought to be tested with a subject with whose voice the witness had a comparable familiarity at the time she testified. And counsel made clear that his offer of proof did not include participation by the defendant in the demonstration.
Finally, insofar as the speech defect Ms. Skolnick described in defendant’s enunciation of the word "security” is concerned, that was well litigated at trial. In the course of Ms. Skolnick’s cross-examination, defense counsel had appellant stand up and repeat the words "You need security” so that the jury was able to make its own judgment as to whether there was in fact any peculiarity to it. Also, during the *706defendant’s and his brother’s testimony, each used the word "security”, the defendant at least four times. Summations by both sides thereafter rang with debate over whether that word had or had not been pronounced with an "elongated s” sound. In short, it is hard to see how it can be said that there was not ample demonstrative evidence on this point.
Under all these circumstances, the Trial Judge most certainly did not abuse his discretion in refusing to permit the proffered test.
There is here yet another matter requiring our consideration. Defendant was convicted on multiple counts, two of robbery in the first degree, two of robbery in the second degree, one of grand larceny in the second degree and one of grand larceny in the third degree. Though all his sentences were to be served concurrently, he urges that certain of the counts on which he was convicted should have been dismissed on the ground that they were lesser included offenses (CPL 300.30, subd 3), a doctrine rooted in the proscription against duplicate punishment for the same act (for extended discussion see People v Savarese, 1 Misc 2d 305).
The specific contention is that the two counts of robbery in the second degree and the count of grand larceny in the third degree were lesser included offenses to the two counts of robbery in the first degree. Section 160.10 of the Penal Law defines a robbery in the second degree as one committed either while "aided by another person actually present” (subd 1) or while displaying "what appears to be” a gun (subd 2), thus authorizing separate counts if both a gun and an accomplice are used. In the prosecutor’s discretion, a robbery committed while displaying what appears to be a gun may also be charged as robbery in the first degree (Penal Law, § 160.15). Our courts have read these two statutes as countenancing the coexistence of the two counts, noting that the proof of "aid” and "actual presence” of the associate under section 160.10 are not identical with any element that must be proved under section 160.15 (People v Williams, 47 AD2d 262). As the Williams case points out, CPL 1.20 (subd 37) indicates that one crime is only included within another when it is impossible to commit the crime in the greater degree without also committing "by the same conduct” a crime in the lesser degree. Whenever the putative lesser crime requires proof of an element not required by the greater one, as is the case here, there is no inclusion.
*707Although the defendant has not asked us to dismiss the count of grand larceny in the second degree, we pause to note that the analysis above applies to it as well. While the District Attorney, both at the Appellate Division and in this court, "conceded” that grand larceny in the second degree was a lesser included offense on these facts, it is not (Penal Law, § 155.35) the section under which this defendant was indicted, requires proof that the items taken exceed $1,500 in value; no proof of value is required for robbery in any of its degrees. Consequently, it is possible to commit robbery without also committing larceny in the second degree and the latter is not included in the former (People v Williams, supra, at p 265).
The defendant does, however, challenge his conviction for larceny in the third degree on the same grounds. This count stemmed from his theft of Ms. Skolnick’s rings and was based on subdivision 5 of section 155.30 of the Penal Law, which defines grand larceny in the third degree as applicable when property, "regardless of its nature and value, is taken from the person” of the victim. It is impossible to commit robbery without also committing grand larceny under this definition, and the third degree larceny count is, consequently, a lesser included offense in these circumstances (People v Grier, 37 NY2d 847).
Accordingly, the order of the Appellate Division should be modified by reversing so much of it as affirmed the conviction of the defendant for grand larceny in the third degree and it should otherwise be affirmed.
Chief Judge Breitel and Judges Jasen, Gabrielli, Jones, Wachtler and Cooke concur.
Order modified in accordance with the opinion herein and, as so modified, affirmed.