When comparing audio quality, subtle differences in sound may be difficult to detect where unmistakable differences are too obvious to ignore. As an example, the radio in my car can receive both analog and digital FM. The radio antenna easily unscrews, and to prevent theft, I removed the antenna. With the antenna off, I noticed that the FM stations were being received analog (verified by the occasional sound of multipath). I drove the car without an antenna for five years. A few weeks ago I decided to put the antenna back on. The difference in sound was dramatic. Listening to dense audio passages sounded more like a bunch of noise, and less dense audio had a very annoying midband sound. I removed the antenna and threw it away.

I was involved with one of the first digital FM transmitter installations. Digital FM uses audio codecs with lower bitrates to fit more stations into a single spectrum. Digital broadcasting was created to allow more channels on a single frequency to increase revenue, not for high-quality signals. It was no surprise that the analog signal sounds better than the digital signal.

Class A amplifiers are primarily the audio circuits that I work with. I have not tested, worked with, or listened to any Class D audio amplifiers. I have avoided Class D amplifiers because of a previous experience in radio broadcasting with PDM modulation that was introduced in 1975. As the transmitters aged, the sound quality degraded, probably due to component value changes with age. The system in 1975 used 70 KHz switching. Current PDM systems use much higher switching frequencies. With the higher frequencies, PDM circuit boards are subject to effects common with RF circuits. PDM circuits are prone to anomalies that can degrade performance.

Class D Mechanisms:
The incoming audio signal is converted into a series of very fast high-frequency on-off pulses. The pulses are all the same amplitude, but their width determines how long they stay on and changes to match the original audio signal. The amplifier's power transistors act like switches, rapidly turning fully on or fully off in sync with the pulses. On the output a special filter, usually made of coils and capacitors, smooths out the rapid pulses. The high-frequency parts are removed, leaving behind the original audio waveform. Class D amplifiers are much more efficient than Class A, capable of delivering a power output of 200 watts a channel or more while drawing less utility power.

Reading about the process of the input signal of a Class D amplifier being converted to high-frequency pulses and an output stage operating as on-off switches, I have to wonder if the high-frequency switching and simulated switched analog output are generating a potpourri of harmonics, and are any of those harmonics leaking through the output filter?

Apparently there are problems related to Class D amplifiers. In particular, the high switching frequency and circuit board layout. Switching frequencies above 100 KHz will be more susceptible to circuit design and circuit board layout flaws. Such high frequencies require engineering similar to RF transmitter design with careful attention to circuit layout. You can read Rod Elliott's page on the subject. Rod has done extensive testing of PDM topology.

As with any amplifier, components are subject to age deterioration. In my example above, where adding or removing an antenna made a noticeable difference, artifacts creeping into Class D sound may not be noticeable at first. If artifacts are obvious, you cannot simply replace capacitors or other parts that may be specially rated. On the other hand, with all that power at your disposal, you may have blown out your ears with periods of 100 dB SPL levels and can't hear artifacts anyway. Probably a worst-case scenario would be failed output switching that applies DC directly to the speakers, damaging them. Any electronic circuit is subject to failure, including speaker protection.

Class A Mechanisms:
Class A amplifiers provide linearity and low signal distortion, are simple and stable, and, for the DIY builder, are easy to work with. They provide very good signal reproduction as the operating point is halfway between cut-off and saturation. They offer a broad and flat frequency response.

About four years ago I became interested in low-power audio based on the notion that one should not have ear-damaging power in an audio system. That is, of course, assuming one values keeping their hearing in prime condition for years to come. For low-power systems, Class A amplifiers are a good match. At lower power levels, relatively speaking, the low efficiency of Class A is not a big issue.