not everyone has the same ear so the answer depended on person. in fact, there are quite some people who can't figure out the difference between 128k mp3 and lossless WAV.
I also thought so in the beginning, but after reading several more level-headed takes like the one
linked above and understanding a bit more on human hearing as well as the science behind
digital audio, I've quickly come around. I also thought that maybe *I* had a "golden ear"; sad truth is I probably don't.
This will be a long a rant, sorry reader!, so feel free to skip it:
<rant>
To get the easy out of the way, Re 128k mp3 and lossless WAV, I believe that brain adaptation has a large effect to play here. People accustomed to 128kbps MP3 for all their lives played from poor sources on poor gear will simply have a brain not used (or not caring) about subtle notes or details in a piece. If you reverse it, I suspect that people accustomed to lossless WAV from good gear for all their lives would immediately be shocked by the poor sonic qualities and
averaged blandness of a 128k mp3. Their brains would simply be accustomed to all the detail retrieval in lossless files. Many today fit in the 128k mp3 generation... OTOH, those who had a taste of unabridged playback in the golden age of analogue sound (the vinyl era) will often hark about how poor today's audio reproduction is (although they've probably never had access to good digital recordings and decent DACs and gear.)
As for the more contentious part, let's start from the facts.
Adult humans do NOT hear above 20 kHz. And this is a generous upper limit; from what I understand, beginning with late 20s there is a progressive roll-off in the upper hearing limit and many young adults would be lucky to hear anything above 15 kHz. Most people with access to the wallet required for audiophile tastes---and with sufficient interest and knowledge to get into such debates---will be in their 20s and above. But let's go incredibly generous, and allow for hearing at frequencies of 25 or even 30 kHz. Given the possible intermodulation effects at high frequencies, some allow that content useful for human hearing may be present in
sampling rates (or
speeds) of
about 50 kHz to 60 kHz, i.e. hearing rates of 25-30 kHz. We're scraping the top of the hearing barrel here...
The other important point is that the Nyquist theory, which is at the heart of digital audio reproduction, dictates that to fully retrieve frequencies up to a given hearing rate (e.g. 20 kHz) you need twice the sampling rate (e.g. 40 kHz). According to this white paper (
Sampling Theory For Digital Audio By Dan Lavry):
"The great value offered by Nyquist’s theorem is the realization that we have ALL the information with 100% of the detail, and no distortions, without the burden of “extra fast” sampling.
“Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth. What is the audio bandwidth? Research shows that musical instruments may produce energy above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not reach 40KHz."
This is one of the reasons for the choice of 44.1 kHz sampling rates for redbook CDs (i.e. retrieved frequencies of up to ~22 kHz), and 48 kHz for much of professional audio/video gear (i.e. frequencies up to 24 kHz).
This right here is already hi-res.
Some will point out that in other digital domains (e.g. networking, imaging) sampling rates of 3-4 times are used instead of 2 times. While I'm not qualified to settle this argument, I will point out that 96 kHz sampling rates allows fully retrieving 48 kHz frequencies at 2x, 32 kHz at 3x and 24 kHz at 4x.
At this point (
96 kHz) we've exhausted most of the theoretical arguments for quicker sampling speeds, and for human applications very little audio quality will be left on the table. If you're really, absolutely paranoid of missing out on some sounds that you might be able to hear, I can see why you would choose to play back digital audio at 192 kHz sampling rates to retrieve frequencies of 48 kHz assuming the generous 4x multiplier. (Although skeptics will go as far as saying that
24/192 Music Downloads are Very Silly Indeed.)
But going above 192 kHz is unadulterated overkill. So why 384 kHz? To reproduce 96 kHz frequencies at 4x (or 192 kHz ones at the more conventional Nyquist rule of thumb of 2x)? If your intended audience is audiophile dolphins or bats, then yeah, sure. But for humans, that's over the board: no adult human will ever hear these frequencies, however you torture them. And where do we stop? Should we start recording and reproducing digital audio at 100 MHz? I can see why 384 kHz sampling rates to infinity and above will appeal to marketeers and poorly informed consumers, but there has to be a limit to the insanity. I'm wondering which wonder DAC will break ground into supporting 768 kHz next... And why?
The theory is very clear, and the math is simple and damning. Bottom line from all this rant is that once we reach support for 96 kHz (and 192 kHz, if we're incredibly generous about the limits of humans and technology), the actual DAC engineering and implementation at 44.1/48 and 88.2/96 kHz will matter much more for audio fidelity than support for exotic sampling rates like 384 kHz. And support for exotic sampling rates may negatively affect fidelity at lower sampling rates.
As for the forgotten 16 bit vs 24 bit vs 32 bit, in digital audio this will not increase fidelity in any way. Technically speaking this serves merely to define the noise level: more bits, lower noise threshold. As far as bits per sample are concerned, the quality of the master recording probably matters more than support for 32 bit samples. I don't have any evidence on this, but I suspect that 32 bit samples won't overcome poor recording quality. And if you're not hearing the noise threshold, then even 8 bit samples at 96 kHz would serve just fine for hi-res reproduction.
</rant>
Lmao. I would never put 384/32 files on my computer which has a 1TB hard-drive, let alone on the 128gb micro-SD card in my X3ii.
but to me, they sound the same as Redbook CD, 44/16. The one hope I have is that maybe, if record producers think we care enough to pay $25 for an album instead of $9.99, maybe, just maybe they'll do a better job of recording, mixing, mastering and converting/dithering to 44/16. The word length matters more to me than sampling rate, because I hate hearing a noise floor (noise floors ruin the whole audio illusion to me). Of course, you need an amp with a low noise floor too. That's one big plus of the X5ii. I've read on other forums how "all amps have a noise floor". (At least that's the quote you usually hear from manufacturers of amps that have very audible noise floors!) I'm sure the X5ii has a noise floor too, but I can't hear it at all. Not even with my Sure SE846. Hopefully the X7 will be the same in this regard.