Video summary

The truth about "fast" headphone drivers

Main summary

Key takeaways

Product Review

Product / topic reviewed

This video isn’t a review of a specific headphone/IEM model. Instead, it’s a deep-dive into the claim that some headphone/IEM “drivers are fast,” and whether common measurement tools—CSD (cumulative spectral decay) graphs and square-wave tests—really prove “driver speed.”

Main points & key claims (what the video argues)

  • “Fast vs slow” in audio is mostly about perception, typically described as:
    • Sharper transients
    • Clear separation/delineation between instruments (less “smearing”)
  • Some people claim sound feels fast because the driver itself is physically fast, and they cite measurement graphs as support.
  • The host argues that CSD graphs and square waves are largely dominated by frequency response, so they’re commonly misused as evidence of “driver speed.”

Measurement tools: CSD graphs (pros/cons)

What CSDs are supposed to show

  • CSD graphs add a time (decay) dimension to audio behavior, visualizing how quickly sound “settles.”
  • The intuitive idea: a “fast” driver would show faster decay, with less lingering resonance.

Why the video says CSDs are unreliable for “driver speed”

  • CSD ridges often line up with frequency-response peaks, meaning decay patterns may reflect acoustic resonances shaped by tuning rather than driver mechanics.
  • The time/frequency relationship is complex, complicating interpretation (the host argues people often oversimplify the math and perceptual mapping).
  • Frequency response strongly influences decay patterns, including effects from:
    • housing
    • pads
    • crossover (for IEMs)
    • EQ (explicitly demonstrated: the same IEM/driver shows different decay after EQ)
  • Windowing/visualization parameters change conclusions: even with the same underlying data, different CSD visualization settings can make treble resonances appear/disappear or make bass look “slower.”
  • Overall critique: using CSDs to infer “driver speed” is often like “post-talk rationalization” (compared to palm reading).

Measurement tools: square waves (pros/cons)

  • Square-wave testing is often framed as measuring how quickly a driver can start/stop.
  • The host cites UC Irvine professor Craig Stark to refute that interpretation.

Core explanation (square waves ≈ frequency response proxy)

  • A “square wave” is essentially many harmonics stacked together.
  • Because of this, square-wave measurement outcomes are mostly a proxy for frequency response.

Demonstration discussed in the video

  • Using a JVC FX7 in the coupler, the host shows that applying EQ changes the square-wave shape to look more square without changing the driver.
  • This implies the “driver speed” interpretation doesn’t hold up.

Conclusion

  • CSDs and square waves are dominated by frequency response, so they don’t directly prove driver speed.

Survey results (comparisons and numerical outcomes)

Audience survey: can people predict “fast” from CSD graphs?

  • The host ran a poll using three CSD plots (planar, dynamic, and balanced-armature).
  • Result: “Option B” won (the dynamic-driver option).
  • Explicit option mapping:
    • Option 1: Letshuoer S12 (planar)
    • Option 3: RSV Mark II (all-balanced armature; ~$500–$600)
    • Option 2 (winner): JVC FX7 Gummy (dynamic driver; $7)

Notable contradiction

  • In a previous “what sounds fast?” survey, respondents mostly cited planar and more expensive IEMs.
  • But in the CSD “looks fastest” prediction, participants selected a very cheap dynamic.
  • Host takeaway: there’s no clear evidence that the common CSD “speed” interpretation corresponds to perceived speed.

Broader correlations the host claims explain “fast/slow”

The host argues perceived speed tracks more strongly with sound signature than with driver type:

  • Bright / lean / neutral tunings are generally described as sounding faster
  • Warm / bassy tunings are generally described as sounding slower

He claims this can be predicted from frequency response graphs:

  • Example: JVC FX7 Gummy (mid-bass excess) → predicts slower
  • Example: Letshuoer S12 (more treble energy) → predicts faster

Pros (implicit strengths of the measurement discussion)

  • The video treats CSD and square-wave analysis as potentially useful diagnostic tools—just not reliable evidence of driver speed as commonly interpreted.
  • It encourages a more data-informed approach focused on prediction and validation.

Cons / criticisms (explicit negatives)

  • CSDs: misleading for “driver speed” because of:
    • frequency-response dependence
    • sensitivity to windowing/visualization
    • resonance peak overlap
  • Square waves: commonly misunderstood because they reflect harmonic/frequency response structure, not a direct driver “speed” metric.
  • Overall: the idea of “driver speed” as a meaningful, measurable property (as popularly used) is challenged.

User experience / real-world listening angle

  • The host emphasizes that the subjective sensation of speed is real—listeners consistently report it.
  • However, the sensation isn’t treated as literal mechanical driver speed.
  • Instead, it’s linked to frequency response (e.g., treble energy and tonal balance affecting perceived transient crispness).

Overall verdict / recommendation

  • Recommendation: Treat “driver speed” claims based on CSD graphs and square-wave distortion as unreliable unless the analysis properly accounts for frequency response effects and uses consistent measurement methodology.
  • To get “fast-sounding” results, focus more on tuning / frequency-response characteristics (bright/lean vs warm/bassy) rather than assuming driver type or CSD “fast decay” proves physical speed.

Unique points mentioned (consolidated list)

  1. “Fast vs slow” usually means sharp transients and instrument separation.
  2. Common belief: planars/electrostats are “fast,” others “slow.”
  3. CSD graphs are used as evidence for driver speed but are criticized.
  4. CSD ridges often correspond to frequency response peaks/resonances.
  5. The time dimension in CSD interpretation is more mathematically complex than it appears.
  6. EQ changes CSD decay patterns even with the same IEM/driver → suggests frequency response dominance.
  7. Windowing/visualization parameters can dramatically change CSD conclusions.
  8. CSD-based “driver speed” claims are likened to palm reading / post-rationalization.
  9. Square waves are misunderstood: they’re harmonic summations, thus largely reflect frequency response.
  10. Applying EQ changes square-wave appearance without changing the driver → supports FR dominance.
  11. Perceived speed is real but not literal driver speed; it’s tied to FR / sound signature.
  12. Survey #1: people think planars and expensive models sound fast.
  13. Survey #2 (CSD prediction): dynamic driver Option B won.
  14. CSD-predicted winner: JVC FX7 Gummy ($7), despite host’s impression that it doesn’t sound fast.
  15. Examples used: - Letshuoer S12 (planar): often described as fast; has more treble energy - JVC FX7 Gummy: mid-bass excess; predicted to sound slower - RSV Mark II: all-BA, expensive (~$500–$600) but did not win

Speaker-specific views

  • Mark Ryan (Super Review host):
    • main argument and data interpretation
    • hosts both surveys
    • performs/illustrates measurement reasoning
    • concludes CSD and square waves are dominated by frequency response
  • Craig Stark (UC Irvine, via discussion/disclaimer):
    • provides key correction that square-wave interpretation for “driver speed” is wrong
    • supports the “square wave is harmonic sum / frequency response proxy” explanation

Original video