[Extract from John Powers, "The History of Presidential Audio Recordings and the Archival Issues Surrounding Their Use" (1996). Used with permission.]
There are approximately eight hours of recordings from the Roosevelt administration. These primarily consist of fourteen press conferences held in the Oval Office between August 23 and November 8, 1940. The machine was often left on after the press conferences ended, inadvertently recording meetings, office conversations, and room noise. Recording technology was still very primitive and much of the recordings, despite recent digital enhancement, is unintelligible. They are filled with static, loud echos, feedback, machine hiss, and other noises. 
Roosevelt made the decision to record his press conferences following an incident in January 1939. That month the New York Times printed Roosevelt's private comments to members of the Senate Military Affairs Committee. Not only was Roosevelt angered that his comments had been leaked to the press, but he was also furious because the press account was inaccurate. In a press conference three days later, Roosevelt called the quote attributed to him in the Times “a deliberate lie.” He directed his official stenographer, Henry Kannee, to find a way of ensuring that his comments were accurately recorded. 
Initially, Kannee used a rudimentary Dictaphone machine. Unfortunately, the machine did not work and Kannee continued his search for a machine that was satisfactory. In June of 1940, an RCA representative presented Kannee with a gift: an experimental model of the RCA Continuous-film Recording machine which inventor John R. Kiel developed. As tape recorders, per se, had not been invented, this prototype utilized both motion-picture sound technology and motion-picture film. Kiel's idea was to use a motion-picture machine “filling a reel of film with nothing but one sound track after another, side by side.”  The machine used a 35 millimeter film called "scribed acetate sound film." Sound was recorded on the film transversely as opposed to longitudinally. The result was that a substantial amount of conversation could be recorded on a very short piece of film.  The machine incorporated a voice-activated record mode. Roosevelt or Kannee could turn the machine on, but it would only record when a sound activated the system. It is for this reason that many office conversations, other meetings, and room noises are recorded. 
The RCA Continuous-film Recording Machine was large and bulky: it was over three feet tall and almost two feet wide. The Secret Service installed this machine directly under the Oval office, concealing it in a specially built chamber with a padlocked door so that White House staffers who used the room to store gifts would not be able to see the machine. Only Roosevelt, Stenographer Kannee, his successor Jack Romagna, inventor Kiel, the RCA representative, and the Secret Service agents who installed the machine, were aware of the recorder's existence. A single RCA microphone was installed in the lampshade on the President's desk, which explains why Roosevelt's voice is recognizable but other voices are not. Wires ran from the lampshade down the side of the desk. One set of wires threaded through a hole in the Oval Office floor and into the machine concealed in the room below. The second set of wires went through a hole in the President's desk to a control box located in his desk drawer. Roosevelt could activate the system by pushing a button on the control box. Likewise, Kannee, who had the key to the padlocked closet containing the machine, could activate the recording system by flipping a switch on the machine itself. 
There has been much speculation as to why Roosevelt recorded for only eleven weeks in the fall of 1940. As stated earlier, he was worried about being misquoted. By recording his dealings with the press, he could ensure an accurate record of what was said. According to his stenographers, Kannee and Romagna, Roosevelt did not like the idea of using the machine on a regular basis. However, Roosevelt was running for an unprecedented third term as President. Robert Butow, in his American Heritage article, states that he believes that this Presidential campaign played a key role in Roosevelt's decision to record. 
 Telephone interview with Julie Dumont, Archives Specialist, Roosevelt Library, May 12, 1995. See also: The Recorded Speeches and Utterances of Franklin D. Roosevelt, finding aid, Roosevelt Library.
 Public Papers of the President: Franklin D. Roosevelt, 1939: The War and Neutrality, Russell and Russell Company (New York: 1969) pp. 112-115.
 Robert Butow, “The FDR Tapes,” American Heritage, February/March 1982, pp.13-14.
 Robert Butow, “How FDR Got his Tape Recorder,” American Heritage, October/November 1982, pp.109-112.
 Ibid. p. 111. The article goes on to state that fifteen feet of this scribed acetate sound film could record over one hour of conversation. Butow speculates that this is the reason that there was only one reel of film used by FDR during the time he recorded his press conferences.
 Ibid. p. 111.
 “The FDR Tapes,” p.15.