Yes — there are several measurable differences between even excellent modern Class D amplifiers and many Class A/AB amplifiers that
could plausibly contribute to subjective preferences, even when conventional metrics like THD+N and frequency response appear essentially perfect.
The key point is that “excellent conventional measurements” may not fully characterize all aspects of amplifier/speaker interaction or human perception.
Some measurable candidates include:
1. Distortion spectrum and harmonic structure
Two amplifiers may both have vanishingly low THD, yet produce
different distributions of distortion components.
For example:
- Some Class AB or Class A amplifiers produce predominantly:
- 2nd harmonic
- low-order even harmonics
- Some Class D amplifiers may exhibit:
- higher-order residuals
- more complex ultrasonic distortion products
- noise-shaped artifacts
Even if these are extremely low in amplitude, some listeners may be sensitive to:
- harmonic distribution
- intermodulation character
- time-varying distortion behavior
Many audiophiles subjectively prefer low-order even harmonics because they can slightly reinforce tonal richness or spatial impression.
2. Distortion vs frequency and load
A major historical weakness of Class D has been load dependency.
Modern Purifi and Hypex designs have dramatically reduced this, but measurable differences can still exist in:
- distortion under reactive loads
- phase shift into complex impedances
- HF linearity with real loudspeakers
A resistor dummy load does not fully simulate:
- crossover back-EMF
- reactive phase angles
- dynamic current demands
Some Class AB amplifiers maintain extremely linear behavior into awkward real-world speaker loads.
3. Ultrasonic noise and filtering behavior
Class D amplifiers inherently switch at high frequencies (typically hundreds of kHz).
Even when far beyond hearing range, residual ultrasonic energy can:
- interact with tweeters
- intermodulate in downstream electronics
- alter behavior under complex loads
Measurable aspects include:
- residual carrier leakage
- output filter response
- ultrasonic IMD products
Some researchers suspect that very small ultrasonic artifacts may subtly affect perceived treble texture or spatial cues.
This remains debated.
4. Feedback topology and transient behavior
Modern Class D often relies on:
- extremely high feedback
- sophisticated error correction
- switching reconstruction filters
Some audiophiles argue that:
- recovery from transients
- overload behavior
- dynamic settling
can sound different even when steady-state measurements are excellent.
Possible measurable areas:
- transient intermodulation distortion (TIM)
- multitone IMD
- settling behavior
- overload recovery
Certain Class A/AB designs may clip or recover more gradually.
5. Output impedance and damping behavior
Class D amplifiers often achieve:
- extremely high damping factor
- ultra-low output impedance
Counterintuitively, some listeners prefer the slightly higher output impedance of some Class A/AB designs because it can:
- subtly alter bass behavior
- slightly reshape speaker response
- affect crossover interaction
This is measurable.
A speaker’s frequency response can vary slightly depending on amplifier output impedance.
6. Noise spectrum differences
Two amplifiers may measure equally quiet in broadband noise terms, yet have different spectral distributions.
For example:
- Class D may exhibit ultrasonic noise shaping
- linear amplifiers may have smoother wideband noise spectra
Human hearing is extraordinarily sensitive to:
- low-level modulation
- correlated noise
- temporal masking effects
Conventional SINAD numbers may not fully capture perceptual significance.
7. Crossover behavior in Class AB
Ironically, one measurable imperfection of Class AB may contribute to listener preference.
A high-bias Class AB amplifier often has:
- tiny residual crossover artifacts
- low-order nonlinearities
These may create a subtly euphonic presentation.
This parallels why some listeners enjoy:
- vinyl coloration
- tube harmonics
- transformer saturation
Absolute accuracy and subjective preference are not always identical.
8. Power supply behavior
Many classic high-end Class AB amplifiers use:
- huge linear transformers
- massive capacitor banks
- enormous current reserves
This can produce measurable differences in:
- burst power delivery
- rail sag characteristics
- dynamic headroom
- clipping softness
Some listeners perceive this as:\
Modern switching supplies can equal or exceed this technically, but implementation matters.
9. Measurement limitations themselves
This is perhaps the most important issue.
Common published measurements emphasize:
- THD+N
- SINAD
- 1 kHz test tones
- steady-state behavior
But human hearing may be more sensitive to:
- multitone complexity
- time-domain behavior
- modulation effects
- speaker interaction
- spatial microcues
Research continues into whether current metrics fully correlate with perception.
Important perspective
The measurable differences between top modern Class D and excellent Class A/AB amplifiers are usually:
- extremely small
- often near or below established audibility thresholds
In controlled blind tests:
- many claimed differences disappear
- level matching becomes critical
- expectation bias can dominate
But that does
not necessarily mean all preferences are imaginary.
It may mean:
- Some subtle measurable factors are underappreciated.
- Some listeners are unusually sensitive to certain artifacts.
- Existing measurements do not perfectly map to perception.
- Some preferences are for pleasant coloration rather than strict accuracy.
Current consensus among many engineers
Designs from Purifi and advanced Hypex NCore modules are generally considered to have pushed Class D performance beyond the point where topology itself is likely the dominant audible factor.
At that level:
- speaker quality
- room acoustics
- mastering
- DSP/room correction
typically matter far more than amplifier class.
