A sound generating device of an electronic keyboard instrument, which eliminates overlap between frequency bands to avoid sound wave interference to stabilize acoustic characteristic. Woofers are disposed at a bottom part of a speaker box and directed downward, and squawkers are disposed at an upper part of the speaker box and directed upward. In a DSP, waveform data selected from waveform data groups are input into a MIX, and a waveform signal containing high and low frequency band components is produced from a signal output from the MIX. In a distributor, the waveform signal is converted into an analog signal which is then separated into a musical tone signal only containing the high frequency band component and a musical tone signal only containing the low frequency band component, and these musical tone signals between which there is no frequency band overlap are supplied to the tweeter and the woofer.