A method and apparatus for encoding input signals, such as digital data, by so-called high-efficiency encoding, in which pre-echo and post-echo are suppressed. A gain control position decision circuit detects an attack portion and a release portion of an audio signal entering an input terminal. An acoustic model application circuit finds a masking level based on a psychoacoustic model of the input signal. A gain control decision circuit determines the gain control value adaptively selected in accordance with the masking level. A gain control circuit controls the gain of the audio signal entering the input terminal in meeting with the gain control value.