The amplitudes of seismic waves have always been a foremost concern of the seismologist to which considerable ingenuity was devoted. In the 1920s the problem was to magnify the ground motion sufficiently for detection. This was done at first by simple levers that moved mechanical pens. But at the start of exploration seismology, this had already been superseded by optical levers, photographic recording, and (soon after) electromechanical transduction followed by amplification. From the 1930s to about the early ’60s, devices of increasing complexity were introduced to compress the large amplitude difference between the first arrivals and the weakest reflections of interest to the limited dynamic range of the recording medium: first the paper record, then magnetic storage media, and finally the digital magnetic tape. This period can be identified with techniques known as automatic gain control (AGC). Soon after the introduction of digital recording techniques, the emphasis shifted: with intermediate digital storage, the limit to the dynamic range was no longer controlled by the properties of the storage medium. Now everything that passed through the acquisition unit could, in principle, be stored on magnetic disk or tape. At that time the aim became to record the ground motion as faithfully as possible. There were several technical developments on the way to achieve “true amplitudes” that, in turn, made exploration concepts like bright spots, seismic stratigraphy, and amplitude‐versus‐offset evaluation possible. However, the most significant innovation was what became known as floating‐point amplifier. It dominated seismic acquisition for about 25 years. Floating‐point representation of seismic signals allowed storage of the entire dynamic range in relatively economic words of about 18 bits. During the last decade, the quest for ever‐greater resolution—and the availability of mass‐produced components for hi‐fi audio equipment—led to the introduction of the sigma‐delta (Σ-δ) converter. With this device, the full range of the seismic signal (or rather the geophone output) is recorded in binary fixed‐point formats with 24 bits. With this development, the full seismic signal can be stored without distortion or loss of resolution.