Ulysses Biggins appeals from his conviction for possessing and distributing heroin in violation of 21 U.S.C. § 841(a)(1). After a brief jury trial, Biggins was given concurrent three year sentences. Two issues are raised on appeal. First, Biggins argues that the prosecution established an insufficient foundation for the introduction into evidence of a tape recording and a re-recording of a conversation incriminating him. Second, Biggins argues that it was error for the trial court to admit evidence of an offense with which the appellant was not charged. We affirm.

I. Facts

Clarence Lydes was a confidential informant for the Drug Enforcement Administration. Lydes met Biggins at a bar and discussed with him the possibility of obtaining narcotics. Appellant told Lydes to contact him through Bertha Coudgo. On May 19,1975, Lydes went to Coudgo’s apartment[*66] to purchase narcotics from the appellant. The appellant told Lydes to return on the following day, when appellant would deliver the heroin. Appellant did not appear the next day. On May 30, 1975, Lydes and DEA agent Audis Wells returned to Coudgo’s apartment. The appellant there agreed to sell one ounce of heroin for $1,600 at Lydes’s apartment later that day.

Lydes’s apartment was monitored by means of electronic surveillance. From his vantage point in an apartment across the street, Agent John Anderson of the Palm Beach County Sheriff’s Department monitored and recorded the conversation that ensued within Lydes’s apartment. During the course of that conversation, the appellant sold Wells one ounce of heroin for $1,500. Shortly thereafter, Wells obtained laboratory confirmation that the substance he had purchased was heroin.

At the appellant’s trial, the government introduced the original tape recording of the conversation in Lydes’s apartment and a re-recording of the original tape, ostensibly filtered for noise.

II. Establishing a Foundation for the Admission of Sound Recordings

Admitting sound recordings into evidence at a criminal trial presents discrete dangers to which courts have been justly sensitive. In the oft-cited case of United States v. McKeever, 169 F.Supp. 426 (S.D.N.Y.1958), rev’d on other grounds, 271 F.2d 669 (2d Cir. 1959), the court established seven criteria as conditions precedent to admitting sound recordings. The Eighth Circuit has recently adopted that test in the context of electronic monitoring. United States v. McMillan, 508 F.2d 101, 104 (8th Cir. 1974), cert. denied, 421 U.S. 916, 95 S.Ct. 1577, 43 L.Ed.2d 782 (1975). ^[1]

Although we neither adopt nor reject that test as a whole, we think that certain of its requirements may justifiably be imposed on the party seeking to introduce sound recording evidence. The court properly admits a sound recording into evidence only when the party introducing it carries its burden of going forward with foundation evidence demonstrating that the recording as played is an accurate reproduction of relevant sounds previously audited by a witness. As a general rule, at least in the context of a criminal trial, this requires the prosecution to go forward with respect to the competency of the operator, the fidelity of the recording equipment, the absence of material deletions, additions, or alterations in the relevant portions of the recording, and the identification of the relevant speakers.

This burden properly falls to the government because it has access to such information in a way the criminal defendant does not. A defendant will often hear the tape recording for the first time in court. More so than photographs or other demonstrative evidence, sound recordings are susceptible to alterations that may be impossible to detect. It is therefore important that the defendant be alerted regarding any possible uncertainties or distortions in the recording before it is introduced as evidence against him.

Nevertheless, the trial judge has broad discretion in determining whether to allow a recording to be played before the jury. ^[2] The standards for foundation evi[*67] dence we adopt serve the paramount purpose of ensuring the accuracy of the recording. Strict compliance with the government’s particularized burden is the preferred method of proceeding. If the trial judge independently determines that the recording accurately reproduces the auditory evidence, however, his discretion to admit thq evidence is not to be sacrificed to a formalistic adherence to the standard we establish. If there is independent evidence of the accuracy of the tape recordings admitted at trial, we shall be extremely reluctant to disturb the trial court’s decision even though at the time that decision was made the government had not carried its particularized burden of going forward.

In the case at bar the appellant objects that the government’s foundation for the tape recordings was deficient because it failed to satisfy the McMillan test in several ways. First, appellant argues that the government failed to prove the competency of the operator. Second, the appellant contends that the government failed to establish the accuracy of the recording. Third, the appellant claims that the government’s witness failed to identify a voice on the tape as that of the appellant.

The only foundation for the admission into evidence of the original or the filtered recording is found in the testimony of John Anderson, who monitored the conversation in Lydes’s apartment. Anderson was not explicitly shown to be a competent operator of electronic monitoring equipment. We know neither that he was trained in the use of this equipment nor even whether he had ever used it before. Anderson averred only that he was “on electronic surveillance” for DEA. On the other hand, Anderson evinced some familiarity with the techniques of electronic surveillance, and it would be a reasonable inference that. he was competent to use the monitor and tape recorder. But a greater uncertainty exists. Some person whom Anderson did not know and about whose competence there is no evidence both re-recorded and filtered the original tape of the conversation. These are not sophisticated operations, to be sure, but an incompetent operator could alter the conversation, perhaps without being aware that he had done so. Nevertheless, these defects in the government’s foundation evidence are inconsequential under the circumstances of the case at bar. Because there was testimony that the re-recording was an accurate transcription of the original, the competency of the second operator is less important.

With respect to the most critical issue, the recording’s accuracy, the government’s foundation evidence is deficient in ways that, under other circumstances, might be fatal. Anderson testified that the re-recording was a “duplicate” of the original, but he did nothing to verify that the original was a faithful recording of the conversation that took place in Lydes’s apartment. This stands in sharp contrast to the care with which the authenticity of the sound recording was established in McMillan. In that case, the agent testified that he replayed the tape of a conversation between defendant and informant in the informant’s presence to verify the accuracy of the recording. See United States v. McMillan, supra, 508 F.2d at 105; see also States v. Onori, 535 F.2d 938, 946 (5th Cir. 1976).

Nevertheless, in the circumstances of this case there was sufficient evidence that the tape recording was authentic outside the foundation testimony established by Anderson. Lydes and Wells testified regarding the conversation in Lydes’s apartment. The tape recording portrayed that conversation precisely as they described it. The record thus reveals a correspondence between the agents’ accounts of the conversation and the version evidenced by the sound recording sufficiently close to lessen the importance of explicit testimony that the recording itself was accurate.

Similarly, neither Anderson nor anyone else testified that the tape has not been altered. Although no one affirmatively suggested that the tape had been altered,[*68] the party seeking to introduce such evidence has the burden of going forward. Here again, however, the government’s proffer is saved by the close correspondence between Lydes’s and Wells’s testimony regarding the conversation in Lydes’s apartment and the sound recording of that conversation.

Finally the appellant asserts that the government failed to identify Biggins’s voice on the tape. This contention is merit-less. Lydes, who participated in the conversation with defendant Biggins and thus was familiar with Biggins’s voice, clearly identified the defendant’s voice on the tape. Fed.R.Evid. 901(b)(5) makes clear that the witness’s familiarity with the voice sought to be identified, whether the familiarity developed before or after the time of the recording, is sufficient to ensure reliable voice identification. In this circuit, we have agreed that such familiarity is sufficient. See United States v. Ladd, 527 F.2d 1341 (5th Cir. 1976).

In sum, the district judge acted correctly in admitting the sound recordings into evidence in the case at bar.

III. Other Crimes

Biggins claims that the court erred when Bertha Coudgo was allowed to testify over appellant’s objection that she received cocaine from the appellant. The appellant’s objection to this use of other crimes evidence is meritless. The court promptly admonished the jury that Biggins was charged only with a heroin offense, not with a cocaine offense. The defendant relied on an entrapment defense. Consequently this evidence of a prior similar offense was relevant to show predisposition and intent. See Fed.R.Evid. 404(b). Finally, well before Coudgo’s testimony regarding Biggins’s other offense, the appellant’s counsel had himself elicited the information that appellant had previously given Lydes drugs for the latter’s personal use.

The judgment of the district court is

AFFIRMED.