Listening trial summary: AGD Audion Mk. III vs. NAD M23 (v. 1)
I wanted to share a summary of the latest round of ongoing Class D evaluations in my system.
Previous sessions compared the NAD M23 against a DIY Hypex NC500 (matched OPA2107 op-amps) and my Pass XA-25. Those trials established that Class D can compete in my room and gave me a solid baseline on the M23’s character. The AGD Audion Mk. III enters as the next step up — GaN-based Class D at a higher price point.
Methodology
Three consecutive evening sessions, February 24–26, 2026. Amplifiers swapped within each session on the same material. Volume calibrated with a 250 Hz test tone, C-weighted, at the listening position. Sessions ranged from 82.5 dB baseline up to a deliberate high-volume stress test at 91.3 dB on the final night.
System
Holo Audio Spring DAC → solid-state preamp (Burson-based) → amplifier under test → JBL 15″ woofer / Beyma AMT tweeter on Townshend platforms. Treated room with diffusion, absorption, and bass trapping. Audience power conditioning.
Test Material
The playlist was varied and included well recorded material. Mix of Qobuz streaming and local files including an SACD rip at 24/176.
Tracks included:
- Orchestral (Tchaikovsky 5th/Honeck, Sleeping Beauty/Järvi, Daphnis et Chloé/Pappano, Mahler 6th/Rattle, Ives Overture & March 1776, Wagner Siegfried-Idyll/Janowski, Elgar Cello Concerto/du Pré),
- Piano (Gould var. 1 from Goldberg Variations, Colom, piece by Chopin),
- Jazz (Bill Evans Waltz for Debby, live),
- Vocal (Ella and Louis "Cheek to Cheek," Regula Mühlemann, Mozart, Vorrei Spiegarvi, Oh Dio,),
- Ambient (Eno, "Deep Blue Day")
- Rock (Steely Dan "Green Book").
Findings
The AGD was the better amplifier; sometimes it wasn’t close, at other times it was more comparable than I expected. The details help break this down.
Imaging. In general, the AGD’s imaging is more definite, with a faster, more dramatic presentation of dynamic events — particularly the way an orchestra executes complex passages, or the way a soloist stands out from the orchestra’s backing.
N.B. There were many places where the NAD held its own, so these remarks are highlighting differences which showed up a significant number of times, and repeatedly on certain recordings.
Bass was excellent on both: tight, fast, creating a real sense of musical involvement. The bass on both amps anchored performances and generated toe-tapping, rather than just filling in the bottom end.
Tonality. Here, the AGD was clearly better, exemplified on tracks with well-recorded voice and piano. Gould’s piano sounded richer, more natural, more percussive – the piano was "there." On the NAD, the piano’s notes sound thinner – almost but not quite like a fortepiano rather than a giant Grand. In other words, the AGD's "midrange character" is a bit better; NAD in comparison is slightly thinner.
The "shouty" or "cringey" factor. This is about tonality, but I’ll list it separately. The AGD never produced the "shouty" quality associated with a Class D amp. The M23 sometimes did, especially on louder tests. The diagnostic moments for this difference were Steely Dan’s "Green Book" (a closely miked snare hit), on Eno’s "Deep Blue Day" (the synthesizers’ swells became metallic and bright with the NAD), and on Muhlemann’s loudest soprano belt-outs on the Mozart piece. At these moments, the M23 felt like "oof, that’s loud, turn it down." Bright, metallic, and fatiguing. This did NOT happen with the AGD.
Instrument separation. The AGD had better instrument separation in symphonic passages, and a more "coordinated" sound where instruments felt like part of the same performance. N.B. Here, though, the NAD often was comparable.
Dynamic range for both is excellent. They scale convincingly.
Background noise floor — both are dead quiet.
Summing up: Key Differentiators
- Imaging & soundstage: AGD often more definite. Sometimes had a wider stage.
- Macrodynamics: AGD faster and more dramatic, often. Crisper, rhythmically, with fast or sharp dynamic changes.
- Bass: Both excellent. Instrument tonality on AGD’s bass was better but bass tightness or quantity was good on both.
- Tonality: AGD more organic, natural, especially on voice and piano; no harshness on aggressive material. NAD can sound shouty on transients and turns metallic at high SPL.
- Instrument separation: AGD resolves complex passages (Ives, Ravel) with clarity. NAD competitive on simpler material but loses resolution in dense orchestration.
- Cohesiveness: AGD sounds coordinated — instruments belong to the same performance. NAD often competes but sometimes less integrated.
- Behavior at high SPL: AGD scales cleanly with volume. NAD hits a tonal ceiling — brightness and fatigue that represents a real-world performance limit.
Outstanding questions and Practical takeaways
Questions I was left with:
- How does the NAD M23 v. 2 sound? Would it compete more effectively with the AGD Audion III than the M23 v. 1?
- How do less expensive AGDs sound – such as the Tempo?
- How would Atmasphere’s Class D amp sound, compared to these? Would it also avoid the shouty or glare of the M23 at higher volumes or on some recordings?
- How would other similar competitors sound – the Bel Canto, etc. There are many out there.
Prices and value:
The compared units were not perfect matches: these two amplifiers are at very distinct price points; the NAD v. 1 can probably be found for $2500 to $3000; the v. 2 is around $5k new or maybe $4.5k, used. The AGD Audion III monoblocks are closer to $8k new, $6k used.
If I was sufficiently impressed by the M23’s v. 1 performance (versus the AGD), I would probably seek out a used NAD c298 because it has the same Eigentakt v. 1 module, same company, and while probably a less beefy design, my guess is that for the price, used (approx. $1650 to $1900), one could get mostly there and save a lot of money. I also expect the M23 v. 1 price to keep dropping with the v. 2 now being promoted.
Onwards and upwards!