The acoustical experts who concluded that two gunmen were firing at President Kennedy when he was assassinated in 1963 have issued a stinging rebuttal of FBI criticisms of their work.

Rep. Louis Stokes (D-Ohio), who headed the now-defunct House Assassinations Committee that commissioned the acoustical work, put their replies into the Congressional Record recently with some crisp comment of his own about the FBI.

Stokes said it was now clear that "the FBI fundamentally misunderstood our scientific and evidential analysis.There was, in short, no basis for the FBI conclusion that our work was invalid.

The House committee's studies focused on a Dallas police recording of sounds apparently picked up on a motorcycle transmitter in Dallas' Dealey Plaza when Kennedy was shot. The scientists commissioned by the committee found that the tape was what it purported to be and said it showed a 95 percent probability that a second gunman had been firing at Kennedy from the so-called "grassy knoll" while three other shots were coming toward him from another direction.

The findings led the committee in late 1978 to the conclusion that the president "was probably assassinated as the result of a conspiracy." Last fall, however, in a special 22-page report submitted to the Justice Department, the FBI maintained that the acoustical findings on which the committee relied were based on faulty premises and "must be considered invalid."

The scientists for the House committee, James E. Barger of the Massachusetts firm of Bolt, Beranek & Newman, and Mark Weiss and Ernest Aschkenasy of New York, all signed the joint response.

". . . [We] disagree completely with the conclusions of the FBI," they said in a report to Stokes this spring. "We have studied the FBI's report and we find that the FBI failed to understand either the methods that we used or the nature of the problem that was posed to us.

"As a result, in their report the FBI asserts premises that are irrelevant, makes deductions from our report that are incorrect, and presents findings that are unsupported."

The main points of the debate concern the FBI's assertions that the House experts "did not scientifically prove that a gunshot was fired by a second gunman from the grassy knoll area of Dealey Plaza . . . [and] did not scienfitically prove that the Dictabelt recording of Channel 1 of the Dallas Police Department (DPD) radio system contains the sounds of gunshots."

But the FBI, the experts said, took other sounds on the recording, which did not originate in Dealey Plaza, in an effort to show that the noises the experts identified as gunshots might not have come from Dealey Plaza either.

Barger, Weiss and Aschkenasy said they were quite aware that the recording included sounds that had been picked up by other police microphones, in other locations, during the same time frame. But that in itself, they emphasized, proves nothing.

All the FBI did on this point, Barger and several colleagues from his firm added in a separate memo, was to define the problem. "Our method was actually to solve the problem." In short, they said:

"We determined where in Dealey Plaza the transmitter would have had to be if it were to have recorded the assassination gunfire sounds as they appeared on the DPD recording. It was found later by the HSCA [House Select Committee on Assassinations] that there was a motorcycle with a radio transmitter where we had found it must be. We are unaware of any contrary information contained in our results, and we believe that the FBI conclusion is unsupported."

The FBI review, made public last Dec. 1, relied heavily on the bureau's investigation of a November 1979 shootout between members of the Ku Klux Klan, the Nazi Party and the Communist Workers Party in a residential area of Greensboro, N.C. The FBI said one of the shots in that confrontation, which was videotaped by local TV crews, produced impulses that virtually matched those on the Dallas police tape attributed to a shot from the grassy knoll.

The House committee's scientists, however, said the FBI failed to supply the date needed to back up this statement and, in any case, apparently "misunderstood" the approach they took. What they did, they said, was not to establish certainty, but rather the "high probability" that the sounds they pinpointed on the Dallas police Dictabelt was picked up by a microphone that was moving through Dealey Plaza at the time of the assassination "and that the recording contains the sounds of gunfire."

Beyond that, the House experts said, their work also showed, "with high probability," that one group of impulses on the recording "corresponds to a gunshot sound fired on the 'grassy knoll' of Dealey Plaza."

In another memo that Stokes also had printed in the Record, the committee's former chief counsel, G. Robert Blakey, said the FBI report also ignored a wide range of circumstantial evidence, such as the testimony of 21 witnesses who said in 1963 that they heard a shot from the grassy knoll area.

"The most charitable reason that can be offered on why the FBI report misunderstood the scientific and analytic work of the select committee is that the bureau's technicians were inexperienced with the sophisticated statistical and acoustical procedures employed by the committee's scientists," Blakey added.

"A less charitable comment," he said, would be to note that the FBI apparently is still unwilling to admit that it failed to conduct an adequate investigation of the president's murder in the first place.