Real-world conventional amplifier. A real-world amplifier, by design, has a limited amount of available current. As the techs deploy the system, Spk A is connected and the amplifier adjusted for the target SPL. When Spk B is connected, the SPL from Spk A drops by 2 dB. When Spks C and D are connected, the SPL from Spk A (and B) drops by 3 dB. This behavior is dynamic and program-dependent, with the amplifier essentially behaving as a compressor.
Alternately the amplifier just goes into distortion with the increased load. In either case, the amplifier’s behavior is now non-linear. The designer must still deal with the line loss from the cables.

“Constant power” amplifier. A constant power amplifier produces the same audio power into any load impedance. As the techs deploy the system, Spk A is connected and the amplifier adjusted for the target SPL. When Spk B is connected, the SPL from Spk A drops by 3 dB. When Spks C and D are connected, the SPL from Spk A drops by 6 dB. This behavior is dynamic and program-dependent, with the amplifier essentially behaving as a compressor. Even though the SPL drops (dynamically) the amplifier can handle the load and doesn’t overheat. Again, the designer must still deal with the line loss from the cables.

Transformer-distribution system. The ideal scenario (Amplifier A) can be realized using a transformer-distribution system. Each loudspeaker is connected through a 250-watt autoformer. (An autoformer is essentially a step-down coil that is smaller and lighter than a transformer, and they’re often used in lieu of transformers in high power distribution schemes like the one described in this article.)

Not only would this type of system perform like the ideal scenario, but the line loss due to wire resistance would be reduced since the autoformer makes the 8-ohm load look like 20 ohms to the amplifier. Figure 3 shows a screen capture from the “High-Z Calc” of the CAF Viewer for a single load on the amplifier. Note that the required amplifier size would be 1 kW if all four loudspeakers are connected using AWG 13 wire with a home run back to the amplifier for each loudspeaker. Once can simply increase the Qty field to 4 to account for this.

“Self-powered” loudspeakers. Another approach is using internally powered loudspeakers. The required “daisy chaining” is now done at line level ahead of the amplifiers. The impedances are much higher here (measured in kΩ or kiloOhms) and line loss is completely mitigated. This approach requires AC power cables in addition to a twisted pair for the signal. In some cases, the loudspeakers can be AC-powered along the route, reduced the required wire gauge for the extension cords.

Subwoofer array. Since we’ve come this far, there’s one more scenario that should be considered. Instead of a parade route, let’s stack all four loudspeakers and make a subwoofer array. The difference is that this sums the acoustic outputs of the loudspeakers, whereas before they were isolated. Nothing will change with regard to the amplifier loading or behavior, but we can expect an increase in SPL along with a shaping of the radiation pattern, depending on how the subs are stacked.

Due to the long wavelengths at these frequencies the acoustic summing will approach 6 dB per doubling of boxes, or +12 dB overall. Since the ideal case rarely occurs, let’s call it +10 dB. This means Amp A and the self-powered subs will yield +12 dB. Amp B will yield +7 dB. Amp C will yield +4 dB. Yes, in all cases you get “more” but this serves to obfuscate the fact that the amplifiers are tanking under the load.

In conclusion. All of these approaches can “work.” It’s the job of the sound system designer to understand the strengths and weaknesses of each, and then select one that makes sense for the application.