Oh boy indeed. This is the full
“time-domain smear” audiophile script with citations taped on afterward.
He’s making a few distinct claims, and nearly all of them fall apart.
The big errors
1. He assumes the conclusion
He hears a difference, then declares the cause to be:
- brickwall filtering
- temporal smear
- pre-ringing
- lack of “stop”
- loss of black space
- elevated noise floor
None of that is established. It’s just
post-hoc storytelling.
2. He treats standard reconstruction as a flaw
A DAC reconstruction filter is not “adding energy that wasn’t in the recording” in the way he implies. It is reconstructing the
band-limited analog waveform represented by the samples.
That waveform is the target.
The idea that a competent DAC “can’t stop a note” because of linear-phase filtering is just nonsense. A transparent DAC does not add an audible temporal tail.
3. He confuses impulse response plots with audible errors
Yes, a linear-phase filter has pre/post ringing in its impulse response.
No, that does
not mean:
- smeared imaging
- woolly bass
- higher audible noise floor
- reduced instrument separation
- loss of microdynamics
Those are giant leaps.
4. “Higher noise floor”
This one is especially funny.
If he means actual noise floor, it is measured.
If he means “perceived haze,” then he is no longer talking about noise floor.
You don’t get to rebrand “I feel less blackness” as an objective noise-floor increase.
5. “10 µs timing cues” abuse
He is dragging in localization literature and acting as if that means a DAC reconstruction filter with normal audibility-safe behavior destroys imaging.
That does not follow.
Interaural timing sensitivity:
- is about binaural localization cues,
- not “my DAC’s generic FIR filter ruins music.”
Completely different context.
6. “128-tap generic filters cannot resolve…”
This is more technobabble than argument.
A DAC filter does not need to “resolve 10 µs timing cues” the way he imagines. It needs to correctly reconstruct the band-limited signal. If it does that transparently in the audible band, the job is done.
7. “No-filter mode sounded better”
That is the funniest part.
So the DAC with:
- more HF droop
- more imaging
- less correct reconstruction
sounds “better” to him.
Fine. That’s preference.
But that directly undermines the claim that he’s reporting pure fidelity rather than liking coloration.
The practical reality
If the Loxjie really had:
- less low end
- higher noise floor
- harsh highs
- less separation
then one of these should be true:
- it measures that way,
- there was a level mismatch,
- some settings differed,
- expectation bias / uncontrolled switching dominated.
The last two are by far the most likely.
What’s really going on
This is the classic pattern:
- prefers an older colored device
- dislikes a transparent one
- interprets neutrality as sterility
- invents a time-domain theory afterward
That’s not engineering. That’s narrative repair.