A programmed data processor receives input voltage measurements for a number of speaker drivers, wherein each of the voltage measurements may be a sensed or estimated sequence of time-domain samples of a respective speaker driver input voltage that is over a different time frame. The processor obtains a sensed shared current, being a measure of current in a single power supply rail that is feeding power to each of a number of audio amplifiers, while the audio amplifiers are driving the speaker drivers in accordance with a number of audio channel signals, respectively. The processor computes an estimate of electrical input impedance for each of the speaker drivers using the sensed shared current and the input voltage measurements. Other embodiments are also described and claimed.