The images come from other movies(often pornographic films). He simply cuts pieces of film from the reels that he wants and splices them in where he wants them (usually between reels of the main film, where it is easiest because he doesn't have to cut the reels of the main film). It isn't explained any more than using a doorknob on the way from one room to another is explained. He has access to the material and the tools needed because he works as a projectionist.
There is some noticeable sound, as he is only adding a few frames, too few to notice a break, consciously, but enough to register subconsciously. At about 29 or 58 Frames per Second (FPS), 6 frames would be a blink. He splices on a frame with a sound channel, almost literally cutting and pasting it in between the regular film, played by the projection system in sync. The film has video and audio on the same cellophane reel, side by side, read by different parts of the projection system. That's how movie projectors work.