This diagram explains how audio is encoded in IRENE images. The images produced from optical scan images are depth images, where gray-scale value in the image correspond to height of the surface, white is highest and black is deepest. This image shows that if one follows a groove around the cylinder the bottom of the groove undulates up and down, in and out of the cylinder surface. This undulation was engraved into the cylinder surface by the recording stylus and induces the motion in a playback stylus that would produce audio. The speed of the stylus directly relates to sound and so to produce audio all that is necessary is to find the data along the groove bottom, which is done using a processing algorithm, and take the derivative along the bottom.
This method is meant to illustrate how data collection works for a cylinder. During acquisition, individual measurements of the height at 180 points along a line are taken in quick succession, at slightly different positions, as the cylinder rotates, and then stitched together in analysis software.