New Test Reveals How Vintage CD Players Outperform Modern Models Thanks to One Forgotten Design Choice

The study shows how overlooked testing methods hid the true performance of both vintage and current models.

Lab tests have long been the final word on how CD players perform. Yet those tests rarely went beyond a few single-tone checks, leaving out the kinds of stresses that music actually creates.

Recent testing by NTTY filled in the gaps with a method that looked at clipping, distortion under load, and how digital filters behave with shaped dither. The results changed the story of how older non-oversampling and newer oversampling players really compare.

How the New Testing Methods Work

For years, CD players were judged with a narrow set of lab checks, usually a single 1 kHz tone and a simple distortion number.

Those tests showed if a player could “start the car,” but not how it handled the road. They rarely stressed intersample peaks, ignored how multiple notes interact, and often relied on analysis windows that hid real problems.

To dig deeper, NTTY built his own reference test CD and a new set of methods designed to show what earlier tools overlooked.

• Smarter signal design
• Music-like loads
• Capture and analysis choices
• Modern content context

Smarter signal design

The first big change was in the signals fed to the player.

The disc included signals you won’t find in the usual AES17 toolkit. Instead of the standard 1 kHz or 997 Hz test tone, NTTY used 999.91 Hz. That slight shift breaks up repeating patterns that would otherwise fool FFT plots, and it “self-dithers” without raising the noise floor.

He also added a set of intersample stress tones at 5,512.5 Hz, 7,350 Hz, and 11,025 Hz. These create internal peaks of +0.69, +1.25, and +3.0 dB, enough to reveal when oversampling filters or DACs run out of headroom.

If those peaks clip, fast transients like cymbals can sound brittle.

Music-like loads

Real music is never just one tone, so NTTY used 1/10-decade multitone signals, with dozens of frequencies at once. If the whole span stays clean, a player can “clear” all 16 bits of a CD in realistic conditions.

He paired this with intermodulation checks, not just SMPTE and CCIF but also the AES17 DFD IMD test. These reveal how low and high notes interact, something single-tone THD tests rarely show.

Capture and analysis choices

To keep results honest, all captures ran at 192 kHz so analysis extended to 96 kHz and ran different measurements and methods:

• Ran −6 dBFS dashboards alongside 0 dBFS sweeps, since many players behave differently just below full scale.
• Measured digital-out and analog-out together. So, if the digital feed matched the file but the analog trace showed problems, the fault was in the reconstruction stage, not the data.
• Checked clock accuracy with a 19,997 Hz tone. Deviations of a few ppm are far too small to hear but prove the method’s discipline.
• Noted transport and servo behavior (A.K.A. how players handled scratches, burns, and track jumps) since reliability matters as much as fidelity.

Modern content context

Finally, NTTY tested how dither and noise shaping affect what we see.

Rectangular dither, triangular dither, and shaped dither each behave differently. Shaped dither in particular pushes noise where hearing is less sensitive, and when the analysis is limited to 500 Hz–4.5 kHz, shaped files can show in-band performance up to 118 dB(A).

He also ran −12 dBFS linearity sweeps, which better expose how older R2R DACs and modern 1-bit chips behave with low-level content.

Measurement Results Across Vintage and Modern CD Players

With the new test set, the results told a story that spec sheets had missed for decades. Each test revealed strengths and weaknesses that showed up only under real-world conditions.

• Intersample-over results
• Multitone and IMD
• Step-level and harmonic behavior
• Linearity and dither
• Digital vs analog output
• Other performance notes

Intersample-over results

Stress tones at 5,512.5 Hz (+0.69 dB), 7,350 Hz (+1.25 dB), and 11,025 Hz (+3.0 dB) showed how players handle hidden digital peaks. Measurements were taken across 20 Hz–96 kHz with no weighting.

• Denon DCD-900NE (oversampling): collapsed to −19.1 dB THD+N at 11,025 Hz, showing hard clipping.
• Yamaha CD-1 (non-oversampling): managed −78.1 dB at 11,025 Hz but stumbled to −35.3 dB at 7,350 Hz.
• Teac VRDS-20: held up better, though not immune.
• Onkyo C-733: also showed strain at the highest peak.

Multitone and IMD

When tested with a 1/10-decade multitone, well-kept vintage units achieved about 110 dB of clean span. That means they can “clear” the full 16-bit content of a CD under music-like loads.

Intermodulation checks also showed that some modern designs, which looked fine on single-tone THD struggled when bass and treble hit together.

Step-level and harmonic behavior

Simple dashboards often praised modern 1-bit DACs, but stepped-level sweeps told a more complicated story.

At −20 dBFS, an older R2R ladder DAC produced about 30 dB more second-harmonic distortion than its 1-bit rival. Yet at higher levels the differences were smaller, and the real surprise came with dither.

Linearity and dither

Without dither, R2R players lost linearity at very low levels. With rectangular dither, they improved, but noise spread across the band.

With shaped dither, they held linearity down to −110 dBFS and showed in-band performance close to 20-bit.

The famous 118.5 dB(A) figure wasn’t from an analog player at all but from a shaped-dither WAV analyzed over 500 Hz–4.5 kHz.

It proved what the method could reveal, not a blanket claim about all CD hardware.

Digital vs analog output

Another important check compared the digital stream to the analog output. The digital feed matched the file, proving it was bit-perfect. The distortion and clipping seen in analog traces came from reconstruction filters and DAC stages, not from the data on the disc.

Other performance notes

Crosstalk tests showed good separation, with vintage units often exceeding −100 dB even at 10 kHz, giving a stable stereo image.

Bandwidth plots highlighted trade-offs: oversampling filters kept the top octave flat and blocked ultrasonic images, while NOS designs rolled off earlier and passed more out-of-band energy.

What the Results Mean for Listeners

Lab plots are useful, but the question most people care about is simple: what do these differences mean when you press play? NTTY’s expanded tests show how design choices in CD players can shape what you actually hear and how a player behaves day to day.

• What you might notice in the sound
• Analog out or transport?
• Why content matters
• Filter trade-offs in plain terms
• Beyond sound: reliability

What you might notice in the sound

When a player clips on intersample peaks, sharp sounds like cymbals or strong vocal sibilants can take on a hard edge. Players that keep headroom in check avoid this and sound cleaner on hot, modern masters.

Multitone and IMD tests point to another area: separation. Units with low distortion under complex loads keep bass from muddying treble and maintain a stable stereo image, which listeners describe as clearer space around instruments.

Shaped-dithered content can also reveal quieter ambience, giving recordings a “blacker” background and more subtle detail between notes.

Analog out or transport?

Not everyone uses a CD player the same way. If you listen from the analog outputs, the reconstruction filter and DAC design inside the player directly affect the sound.

If you use the digital output into an external DAC, the story changes. NTTY confirmed the digital stream can be bit-perfect; in that setup, audible differences mostly come from the DAC, not the transport.

Why content matters

The impact also depends on what’s on the disc. Older CDs and vinyl transfers rarely push levels high enough to trigger intersample clipping, so oversampling filters mostly matter for their tonal balance at the top end.

Modern pop and EDM, mastered loud, are more likely to hit those hidden peaks, which is where players with real headroom, or non-oversampling designs, can sound less strained.

Classical and jazz recordings, especially those mastered with shaped dither, make low-level linearity more relevant; that’s when vintage resistor-ladder DACs sometimes surprise listeners with natural ambience and micro-detail.

Filter trade-offs in plain terms

Oversampling filters were designed to fight the natural roll-off of DACs and block ultrasonic images. That helps keep the top octave flat and suppresses out-of-band junk.

The downside is possible ringing and clipping if a disc pushes above 0 dBFS between samples.

Non-oversampling players avoid those traps but give up some high-frequency extension and can leak ultrasonic images. Which trade-off you prefer depends on your ears and your music.

Beyond sound: reliability

Sound quality isn’t the only factor. NTTY also ran drive and servo checks with scratched and burned discs, and with sudden jumps in track pitch and velocity. These tests highlight which players hold on to playback under rougher conditions.

For anyone who still uses discs daily, that resilience can matter as much as clean graphs.