Voice processing method and device, equipment and storage medium

The embodiment of the invention provides a voice processing method and device, equipment and a storage medium. The method comprises the steps that the duration and the sound signal amplitude of each word in a wake-up word voice signal are acquired; judging whether the sound signal amplitude of the word is consistent with the sound signal amplitude of a target word in a preset registered voiceprint, and if not, zooming the sound signal amplitude of the word so as to enable the sound signal amplitude of the word to be consistent with the sound signal amplitude of the target word; the target word is a word which is the same as the word in the preset registered voiceprint; and judging whether the duration of the word is consistent with the duration of the target word, and if not, zooming the duration of the word to enable the duration of the word to be consistent with the duration of the target word, thereby obtaining a zoomed wake-up word voice signal. In this way, the problems that the speed of the input wake-u