High-Resolution Audio

[hello_hello]

High-Resolution Audio Demystified (https://www.youtube.com/watch?v=Z5S_DI99wd8)

This mightn't contain anything new for everyone, but I found it interesting enough to sit through the whole 55 minutes (YouTube video).

Mainly because he says stuff that makes sense, such as it's very, very hard to tell the difference between a CD and a decent 256k MP3, vinyl is dumb, or that most 96k/24bit audio today is a complete waste of time, while at the same time offering an argument that maybe there's a place for "real" high resolution audio.

Anyway.... I found it interesting and thought maybe someone else would too. ;)

[leonccyiu]

Hello

I understand that this topic results in heated debate on online forums and whatever discussions don't remain civil for long. I don't have much time this weekend, but I'll definitely watch through the entire video, but having read a similar article from Monty Montgomery from the Xiph foundation https://people.xiph.org/~xiphmont/demo/neil-young.html , I'd like to share some of my input having researched this thoroughly myself.

Since the video was published (2015) and the Xiph article was written which were backed up studies claiming that double blind tests among audiences showed no perceptible differences between CD quality and Hi-Res audio, an analysis of the data of the same studies by Dr Joshua Reiss of Queen Mary university has found that there is a small but statistically significant preference for Higher resolution audio, and this improved when the listeners were subject to some training. I have met him myself as Queen Mary university is my former university and it was mentioned that in some of the studies, the integrity of the data is questioned as the statistical chance of the correlations of the data occurring was incredibly small.
http://www.aes.org/e-lib/browse.cfm?elib=18296

What led me to research this was that I had previously heard the arguments about why Hi-Res audio is a waste of time, but didn't think much of it until one day when I was looking for am album from one of my favourite Japanese artists, and I couldn't find anywhere to legally or illegally obtain her music accept this one Japanese music store which only sold it in 24bit 96khz. The files were much bigger than I was used to, especially since there was no lossless compression but since I had plenty of hard drive space I thought I'd give it a go. Among playing it back, immediately I heard qualities to the sound that I had never heard before in any CD, the correct colour light indicator was on on my DAC indicating the correct sample rate was fed. I literally listened all the way through the album instead of skipping like I usually do. I thought to myself that there's no way the sound could similar be down to good mastering alone.

Later on when I obtained/purchased more albums/singles, some in 24/192 as well which sounded even better than 24/96, that's when I went back to the Xiph article to re-examine the evidence.

I'll quickly address the points made before expanding in further detail later, and provide some links from Bob Stuart co-founder of Meridian/MQA.

The first part concerns bit-depth, I believe 16 was chosen as it was a multiple of 8 which is how many bits there are in a byte, and also hence why the next step up is 24 bits. So each additional bit to a sample provides double the possible combinations and since decibels is a logarithmic measurement, so simply we can roughly say that each additional bit allows roughly an additional 6db to to quantized free of distortion from quantization error. For 16 bits, if you add 2 to 16*6, that gives you roughly 98db that can be quantized without distortion which is a huge volume since db is a logarithmic measurement. 24 bit would allow about 144-146db which exceeds the dynamic range that human hearing is capable of. This is of course if the sampling and reconstruction are perfect, which is isn't unlike what's claimed in the Xiph article and the video as well I assume, and I will come back to that later when talking about low pass filters. Another factor to consider is that ADC's and DAC's are mostly what are known as sigma-delta DAC's which use oversampling and noise shaping techniques during the sampling and reconstruction process. Ladder DAC's or R2R DAC's which directly convert from digital to analogue also introduce their own distortion problems. The 98db of dynamic range that can be quantized without distortion is really a theoretical maximum, and in practice you'll get quite a fair bit of distortion which is why studios today sample at 24bits (processing is down at 32/64 bit floating point, if 16bit was free of distortion, they could just convert it to 24bit for processing). There is dithering of course which adds noise to randomize the errors and allows an even greater dynamic range than the bit-depth would suggest, and when dithering from 24 to 16 bit, while the noise is inaudible ultrasonic noise, you're still not getting the full dynamic range in the original recording. When I compared 24bit 44.1khz to 16bit 44.1khz, the 24bit version sounds more dynamic, I'll try and provide some ABX tests later.

The second part is that it's argued that the Nyquist Shannon Sampling Theorem perfectly reproduces all frequencies below half the sampling rate. This is only in theory if you have an infinite delay to sampling and infinite delay to reconstruction which of course is impossible in real world conditions. You need an infinite delay to sampling because there doesn't exist a perfect brickwall filter that is required to prevent aliasing. All the information in between the samples is lost and interpolated by a formula involving fourier transforms, what is also not mentioned in the xiph article is that reconstruction also produces aliasing, and when filtered, this distorts the phase and amplitude of the signal and introduces other artifacts such as ringing. With a higher sample rate, you've got a much larger space to filter between the intended audio band and the aliases.
Please read https://en.wikipedia.org/wiki/Reconstruction_filter
This is an updated paper titled what Nyquist didn't say explaining why you can't just sample at just over twice the intended band and expected a reasonable amount of success.
http://www.wescottdesign.com/articles/Sampling/sampling.pdf

Since all the frequencies below 20khz are not being perfectly reproduced, the question then becomes, are the distortions from the low pass filters in sampling and reconstruction audible? There is an emerging amount of evidence that they are.
http://www.theabsolutesound.com/articles/meridians-master-quality-authenticated-the-interview/
A quote from the article
"We wanted to understand one of the paradoxes of digital audio, which was why 96kHz sounds better than 48, and 192 sounds a little better than 96, and 384 sounds a little better than 192, and 768 can actually sound better than 384! Each time we’ve got a doubling of data rate, but we’ve got the rapidly diminishing return of incremental sound quality.

That’s a puzzle because we can’t really hear frequencies above 20kHz. From the pure frequency-domain point of view, hi-res sampling doesn’t seem to make a lot of sense, which is why you see some skeptics who don’t actually think it through to the next stage."

http://bobtalks.co.uk/uncategorized/what-is-mqa/

"Recent hearing research provides support for the long-standing notion that the time-domain performance of anti-alias and reconstruction filters – most especially steep digital linear-phase filters – is responsible for perceptible degradation of sound quality. Recently, direct evidence for the audibility of low-pass filters used in digital audio has been published.

It has been known since at least 1946 that the Fourier time-frequency uncertainty inherent in conventional signal analysis can be ‘beaten’ by human listeners, and by a significant margin. Indeed, recent experimental studies have shown temporal discrimination at least 5 times higher."

http://www.head-fi.org/t/800264/watts-up/180#post_13160315

http://www.head-fi.org/content/type/61/id/1792671/width/500/height/1000/flags/LL
http://www.head-fi.org/content/type/61/id/1792672/width/500/height/1000/flags/LL

Illustration of a sinusoidal wave being incorrectly reproduced compared with the original.

That's all I have time for for now, but I hope you find this interesting and research this through yourself, I've become a Hi-Res convert and absolutely enjoy listening to 24/192 when I can find it and when it's priced reasonably, I use fairly cheap Iems as I wear glasses, and I use a fairly cheap DAC and my smartphone, but to me this makes a big difference rather than just splashing money on ever more expensive DAC's which have better reconstruction filters and headphones to listen to CD/lossy music.

http://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality

You might like to give this a go as well, I think you might do better than you think, from the comments, people can identify the lossless version 6/6 times.

[leonccyiu]

I'll also provide some content on my google drive such as the 24/96 samples that can from my phone, the LG life is good samples which I have resampled to 48khz using a high quality resampler dbpoweramp(SSRC) and dithered down to 16bits for you to compare.

[JarrettH]

The guy seemed pissed off

[hello_hello]

leonccyiu,
Thanks for all the info. There's a lot to digest. I haven't had a chance to look at your links yet but I will later.... probably tomorrow.

My problem.... I don't have a particularly great sound system for doing serious listening tests. I don't even own any top of the line quality headphones.... but there's a reason for that. I've spent my whole life working in the live audio scene (mainly clubs and smaller venues), so at home the last thing I'm interested in is more noise and therefore I've never spent much money on home equipment.

I'll admit I'm somewhat sceptical about a lot of claims when it comes to audio. There's been discussions here where people have made claims such as being able to tell without looking whether their car audio player is playing a CD or an MP3, which I find almost impossible to believe, and it's hard not to be sceptical when you read reviews such as the following one. If the HDD/SSD digital audio is stored on can make that much (or maybe even any) difference to the sound I'll video myself eating a hard drive and post it on YouTube.

Listening To Storage (http://www.enjoythemusic.com/hificritic/vol5_no3/listening_to_storage.htm)
"If the Kingston SSD stood apart from the disk drives for its mostly good yet quite alien character, drive four made itself known for entirely the wrong reasons. This Corsair drive (another SSD) conspicuously highlighted vocal sibilants, and had a hard, relentless quality that was impossible to miss."

This is one of my favourites.
AudioQuest Vodka Ethernet Cable and Diamond Ethernet Cable (http://www.audiostream.com/content/audioquest-vodka-ethernet-cable-and-diamond-ethernet-cable#XsKrJ5ik4iZSCA0u.97)
Not so much for the review, but for the comment the reviewer posted later. I hope to become enlightened enough to understand what it means one day.
"The only thing easier than hearing a difference is not hearing a difference."
Anyway......

I've tended to sit in the "CD quality is enough" camp, at least as a final delivery format, although maybe that's because I grew up listening to radio and LPs and cassette tapes and it's made me eternally grateful for even the humble 128k MP3, although I still tend to be very sceptical when it comes to 24/96 versions of old analogue recordings etc..... I really can't see how it'd sound better than CD quality, but for "real" high resolution audio I do have an open mind. I'm a little sceptical, but I do have an open mind, just the same. I recall reading about a study where brain activity was monitored while subjects listened to CD quality music, then high resolution music, and apparently the latter did make different parts of the brain light up. I'll try to find a link for that one later.

Anyway, I've never in my life heard "real" high resolution audio but I do wonder about the 24/96 deal if you're listening to audio on a sound system that's only capable of reproducing audio up to 20k, plus all the theory in the world is nice, but in the end the audio we listen to is still created by speakers that'd colour the sound way more than any other part of the process.... but I'll return tomorrow and follow your links.

Cheers.

[leonccyiu]

That sure is a lot to digest, please take your time.

May I ask what audio equipment you posess?

I agree, a lot of claims made in audio are questionable such as the materials used in cables or even premium digital cables. I did read about the Audioquest Ethernet cable, I think it's a bit ridiculous. I spoke to an Audioquest rep actually about premium hdmi cables and they said it was hard work defending their products.

I personally don't bother with all this where you're supposed to use an ssd instead of a hard drive, run your laptop off the battery etc to minimise interference and jitter.

Anyway back to the topic, higher sample rates isn't about reproducing frequencies beyond 20khz, human hearing declines sharply past 13khz and 20khz is already a generous limit. It's about the distortions caused by the low pass filters during the sampling and reconstruction process. They've improved a lot and the filtering is down digitally now, but the distortions they cause are still audible.

I personally think 24/192 is the biggest difference that can be made to sound quality providing you already have decent headphones/speakers and a reasonable DAC. While more expensive headphones may sound better, you're at a point of diminishing returns past a certain point.

Another reason that I doubt high frequencies play much if at all a part in why Hi-Res sounds better is that if that were the case, 24/192 shouldn't sound better than 24/96 and apparently 768khz sounds better than 384, I can't comment on that, but because of the way sound propagates through air, high frequency content loses energy much faster than lower frequencies and are pretty much non-existent at what those sample rates are capable of reproducing.

If you grew up listening to LP's then I guess you're older than me, I've hardly listened to LPs, only recently now that their making a come back. It's been said that Vinyl potentially surpasses all digital formats, I have Vinyl rips in 24/192 which sound absolutely stellar to me and are much better than the CD versions.

I was at a Vinyl demonstration from clearaudio last year with a vinyl enthusiast who insists Vinyl blows CD out the water, and pointed out to me the separation and spaciousness and naturalness of the sound. I'd have to agree with him, the CD sound flat and congested to me. Of course Vinyl as an analogue format degrades in quality with each playback and has all kinds of noise problems, and the quality really depends on your playback equipment.

I have 24/192 of old analogue recordings, whereas newer recordings which are heavily processed are only done at 24/44.1 or 24/96, despite the noise which maybe present in some old analogue master tapes, in many ways they really surpass modern 24/96 masters which have too much dynamic range compression (loudness wars) and are too processed.

[hello_hello]

I own an old Sony "all-in 1" sound system which sounds "okay" but I do 99% of my listening to stuff through a set of stereo PC speakers with a sub. They're Logitech THX approved speakers (I can't remember the model number). They sound quite reasonable but they're not overly expensive and obviously far from "audiophile" quality, but they're in the same room as my computer where I spend way too much of my time at home. I should invest in something better one day. I'll probably build a new PC quite soon & I might at least spoil myself with a decent pair of headphones while I'm at it.

I'll admit I'm very sceptical about vinyl sounding better than CD in any way. For starters the record spins at a constant speed which means the surface moves faster under the stylus near the outside than the middle, but I've never heard vinyl enthusiasts discuss a possible change in fidelity as the music progresses. The old 45 singles span at 45rpm for a reason though.
Then there's the pre-emphasis. I'm sure you're aware music on vinyl has much of the low frequencies reduced and they're boosted again on playback to keep the width of the grooves sensible. There's a type of CD pre-emphasis too but nobody uses it any more. CDs are at least "flat" in that respect. Sapphire or diamond stylus, or are they made with audiotanium tips or something exotic these days? Elliptical or spherical? Do players still use a stylus cartridge you insert one way for playing LPs and flip the stylus over to play singles, or was that my grandmother's record player?
Back in the turntable days wasn't they're a "direct drive" camp and a "belt driven" camp with both claiming one method of spinning the record sounded better than the other? I kind of remember that. Maybe both methods still sound better than CD. ;)
There's also the stylus moving through a groove thing which does my head in as far as stereo separation claims are concerned. The channels aren't completely independent. How does a stylus vibrating up and down as it vibrates left and right improve stereo separation?

I must not have bookmarked it, but I read an article about vinyl records and resolution I found quite interesting a while back, because vinyl being a physical medium it's made of stuff, and stuff is made of particles, and at some point particles give vinyl a finite resolution because the grooves can't be perfectly smooth (a bit like film grain in a way sets the resolution of film), and the article discussed the effective quantization error produced by a surface made of stuff and it being self dithering (I think) .... I remember it being an interesting read though. I'll try to find it again later.

I'll concede a CD and vinyl pressing of the same master (analogue or digital) could potentially sound quite different, and I suspect some people will prefer the vinyl version for whatever reason, and I'd argue there's nothing wrong with that, but the day I hear an original master recording, the CD and the vinyl versions compared. and it's proven the vinyl pressing reproduces the audio (relative to the master) more accurately than the CD, I'll film myself eating an LP followed by a 45 for dessert and upload it to YouTube. ;)

I can remember when CDs were first released, being far more expensive than LPs initially, generally only audiophiles bought them. It's a funny old world.

I was going to have a look at your links shortly but you made me type all that and now I need to go back to the real world for a while, but I'll return later and try again. :)

[JoeyMonco]

I was at a Vinyl demonstration from clearaudio last year with a vinyl enthusiast who insists Vinyl blows CD out the water, and pointed out to me the separation and spaciousness and naturalness of the sound. I'd have to agree with him, the CD sound flat and congested to me. Of course Vinyl as an analogue format degrades in quality with each playback and has all kinds of noise problems, and the quality really depends on your playback equipment.

Was the vinyl and the CD pressed from the exact same master mix? If not, then the comparison is heavily flawed. It's like when people compare Dolby Digital and DTS and trying to proclaim one is superior to the other when both tracks are from completely different mixed masters. You can't actually make a comparison that way.

I have 24/192 of old analogue recordings, whereas newer recordings which are heavily processed are only done at 24/44.1 or 24/96, despite the noise which maybe present in some old analogue master tapes, in many ways they really surpass modern 24/96 masters which have too much dynamic range compression (loudness wars) and are too processed.

But nothing about CDs inherently means that the master has to be DRCed. You're conflating DRC with the inherent quality of a CD. A properly-mastered CD is going to reproduce the source master far closer than any vinyl will from the same master and can reproduce all the "warmness" that comes from tape hiss, pops and crackle just as well as any vinyl. It's just that the master mix needs to not be heavily digitally processed to do so. Plenty of early CDs had wonderful masters with wide dynamic range.

[johnmeyer]

I'm old enough that the first records I played were 78 rpm. I've seen -- and heard -- it all. Some of my earliest memories are visiting various audio showrooms with my father and listening to various equipment. Also, when you purchased a record, they had listening rooms where they would let you "audition" the record before you purchased it.

Fast forward to 1970, and I'm buying my first stereo system with my own money. Even for a first-time buyer like me, I was able to go to a standard retail store, and listen to various mashups of equipment: pick the turntable, receiver, and speakers you wanted, and they would hook them up for you so you could listen to what that mix of equipment produced. Then, if you wanted to try different speakers with that turntable and receiver, you could cycle through each speaker. Same with the turntable and receiver.

Before I ever went to the store, I read all sorts of magazines. Back then, there were at least half a dozen magazines that reviewed each piece of equipment. These weren't reviews like you find on some of the tech sites today that are done by someone with lots of attitude but no real credentials, but instead were written by people who had spent decades doing this sort of thing. The reviews were often 10-20 pages and included both a listening test with real ears, as well as some pretty advanced testing with high-end audio test equipment.

Then, as I got older, I was able to go to the CES show in Las Vegas, towards the tail end of the audio era (1984). I not only went to all the booths on the main floor, all of which were playing the sonically amazing "Thriller" album that had just been released on the new CD medium, but I also went to dozens of suites set up in hotel rooms where the low-volume, high-priced esoteric equipment was being featured. Want to spend $10,000 (1984 dollars) on an amplifier? They had it.

I was blessed with really good hearing, and what I remember from all of this was that there was a HUGE difference between equipment: the source, the amplifier(s) (pre-amps were usually separate, for high-end setups), and the speakers.

The most important item -- by far -- both then, and now, is the speaker. If you listen to music on any of these "psycho-acoustic" systems that use small speakers and a subwoofer, you are not getting real fidelity. Those setups use some neat tricks pioneered by Amar Bose, but your brain is "making up" much of what is lost by not having something that can move a lot of air. Also, unless you have a subwoofer with feedback (like the original Velodyne subwoofers), you are not getting much more than "thumping" out of that box. It really doesn't matter what your source might be: vinyl, CD, Hi-Def audio, high sampling rate, etc., you will never get anything approaching the fidelity that is possible with a truly great set of speakers. Words like "transparency" were often used to describe what you'd get with fabulous ($10,000 and up) speakers. I heard them with my own ears, often in comparison with other speakers, and the differences were amazing.

As a result, I have gigantic speakers for my home theater, and also still use my dad's old Altec 605C speakers (1950s manufacturing) in my living room. Neither of these are the "ultimate" (I don't have $10,000 to spend on speakers), but they are a major purchase that last a lifetime (the Altec 605 still fetches $3,000 because it still sounds great).

So, my take on all of this is that the place to focus your attention is the speakers, not the source.

[feisty2]

Anyway back to the topic, higher sample rates isn't about reproducing frequencies beyond 20khz, human hearing declines sharply past 13khz and 20khz is already a generous limit. It's about the distortions caused by the low pass filters during the sampling and reconstruction process. They've improved a lot and the filtering is down digitally now, but the distortions they cause are still audible.


aliasing and ringing?
I mainly learned some stuff about video (not audio) processing but I think audio filters follow the same rule here..
I also have no intuitive impression about what aliasing and ring "feel" like in the audio world..
things are much easier to observe in the video world, aliasing = the feeling that edges are going jaggy, ringing = the feeling that there're ripples/halos/other annoying things like mosquito noise around the edge.
anyways, higher sampling rate is not necessarily the only solution here, I don't know about audio, but at least in the image/video world, there're neural net based resampling filters that could avoid any of these artifacts.

[Groucho2004]

things are much easier to observe in the video worldNot really, you just need an oscilloscope and audio spectrum analyzer to see and measure these things.

[manolito]

Not true, there are things in the human brain which do not obey these measurements... :p

It is not much different for video and audio. For video many people agree that you need a certain amount of grain to make it look good. Take away the grain and people will complain about the "plastic" or "soap opera" look.

For audio every audio engineer who already worked in the good old analog days knows that a certain amount of tape hiss makes a recording sound more transparent. Remember the religious war between valve and transistor fans? The harmonics introduced by valve amplifiers are certainly artifacts, but they make high pitched strings sound better.

And the same about vinyl vs. CD. There is no doubt that the first commercial CD players sounded harsh due to the steep slope of the low pass filters (I know, I still own the first Sony CD player model in mint condition). But this changed rapidly, the manufacturers indroduced oversampling, the DACs got much better.

I remember a controversial test report in some US High End Audio magazine where they actually praised a portable CD player (I think it was an Aiwa) for its natural sound. Later it turned out that this playeer had a noticable rolloff of the high frequencies which obviously made the filter ringing less obvious.

And anyone remembers the frequency response plots of the ultra expensive high end moving coil cartridges? They all had a raise in the frequency response starting at around 16 kHz caused by some stylus resonance. And this made them sound more transparent.

My take on all this is that it is not so much about the method (vinyl vs. CD, valve vs. transistor). It is much more about how much care is taken to perfect the method which is used.


Cheers
manolito

[johnmeyer]

I still own the first Sony CD player model in mint condition). Me too! CDP-302. Bought in 1985. Very expensive (around $500, I think). It was a huge splurge, but I bought it because the sales guy played some scratched discs in several other units, and this was the only one that would play them without problems.

I've had to replace the door drive belt. Also, one year ago, I finally replaced the batteries in the remote. I've never before had batteries last more than a decade. These lasted thirty years. Unbelievable. (And they didn't leak ... the voltage just declined until the remote didn't work).

[Groucho2004]

Not true, there are things in the human brain which do not obey these measurements... :p
I was referring to aliasing and ringing which can easily measured.

[hello_hello]

Also, one year ago, I finally replaced the batteries in the remote. I've never before had batteries last more than a decade. These lasted thirty years. Unbelievable. (And they didn't leak ... the voltage just declined until the remote didn't work).

30 years! &$% me!
The Samsung TV remote here had the batteries replaced at four years but barely needs them after five because two thirds of the buttons don't work any more. :(
It's never been sat on or mistreated or even used heavily. They just stopped.

[Motenai Yoda]

@leonccyiu then you can watch even the "monty" videos on xiph.org where he can reach -120dB on 16bit using real consumer equipment, and also show how the 1kHz armoinc sinusoid is exactly reproduced.

Also you said you found 24bit/96kHz tracks sound better than 16bit/44.1kHz, but it isn't a blind test as you know witch track was playing.
I'm even a lot curious about how much this "small but statistically significant preference for Higher resolution audio" is.

[manolito]

Me too! CDP-302. Bought in 1985. Very expensive (around $500, I think).

Mine is the CDP 101, bought in 1983 for 2.300 DM (German Mark). At this time only this Sony and a Philips model were available. The Philips was a little cheaper, the error correction was better, but the mechanical quality was a lot more solid for the Sony. It probably is a collector's item already...

[johnmeyer]

Wow, that IS an old unit. Truly 1st generation.

[StainlessS]

OT, In 2000, my Casio Scientific calculator battery finally gave out @ ~13 Years of age (real old fella, common lithium button cell, about 20mm diameter, supposed to have lifespan of max 10 year I think). Had been using it 30 mins earlier, and auto shut off after about 60 secs non use.
Then on switch on half hour later, Kapowy !!!, exploded and blew the battery cover off, nasty dense black smoke and gaping mouth exposing black nasty stuff.
Calculator still works fine now (@ about 30 Y.O.) but a few more batteries been used since then (maybe 3 or 4).

EDIT: Kapowy !!!, guess who said that, no pun intended. :) [EDIT: OK, was probably Robin who said it, but nearly him].

[Ghitulescu]

Not true, there are things in the human brain which do not obey these measurements... :p

It is not much different for video and audio. For video many people agree that you need a certain amount of grain to make it look good. Take away the grain and people will complain about the "plastic" or "soap opera" look.

For audio every audio engineer who already worked in the good old analog days knows that a certain amount of tape hiss makes a recording sound more transparent. Remember the religious war between valve and transistor fans? The harmonics introduced by valve amplifiers are certainly artifacts, but they make high pitched strings sound better.

And the same about vinyl vs. CD. There is no doubt that the first commercial CD players sounded harsh due to the steep slope of the low pass filters (I know, I still own the first Sony CD player model in mint condition). But this changed rapidly, the manufacturers indroduced oversampling, the DACs got much better.

I remember a controversial test report in some US High End Audio magazine where they actually praised a portable CD player (I think it was an Aiwa) for its natural sound. Later it turned out that this playeer had a noticable rolloff of the high frequencies which obviously made the filter ringing less obvious.

And anyone remembers the frequency response plots of the ultra expensive high end moving coil cartridges? They all had a raise in the frequency response starting at around 16 kHz caused by some stylus resonance. And this made them sound more transparent.

My take on all this is that it is not so much about the method (vinyl vs. CD, valve vs. transistor). It is much more about how much care is taken to perfect the method which is used.


Cheers
manolito
An almost perfect study, carried out with acoustic instrument artisans, studio gear and studio engineers, revealed that if the tools are properly used there is no difference between lamps and transistors. However the ICs were visibly worse, which may be inferred from the fact that they may not be 100% balanced (transistors may be hand-picked to match, so do resistors and capacitors).
In amateur world, where "louder is better", all of them distort, yet the tubes sound a bit better, because they fill in the even harmonics (inaudible per se but giving volume).

As for the HiRes, a statistical study revealed again that there is no statistically difference.
http://www.aes.org/e-lib/browse.cfm?elib=14195
However, one should be rather critical in understanding what it was said, given that statistics is largely unknown and most of those knowing statistics will abuse it :). The study does not say there is no difference and that nobody hear it/them, just that most people do not hear it/them.
By far, the most striking difference lies in the fact (it is a fact) that the CDs are badly mastered, whereas the HiRes, as a niche product, is more carefully mastered (otherwise they can't show any differences and therefore why would one buy it at 3x the price???).

And yes, the speakers are the most important link in the audio chain, so important that ranks 2 and 3 are empty :)

Now to the punctual issue of "transparency".
Like with "warmness" of analogue music, this "transparency" is nothing than the acoustical image given by the surplus of trebles. One can simply use an equaliser and raise the 8kHz or 16kHz, on a good source otherwise the tape hiss will interfere.
Linked to this there is also another common myth amongst vinylisers, that the LP reaches 50kHz or any beloved number, but above the CD. Yes, indeed, the spectrum does indicate signals above 20kHz, yet these are resonances and aliases, in a word noise. It would rather be hard to be music, because all lathes do employ a digital delay (a sort of TBC to reuse video terms) that has the features of a CDR recorder (44k1/16) - so the vinyl is limited by the CD.
Now, because most CDs are over-emphased (lots of trebles), people discovered that taping them, in other words using the limiting properties of the tape, makes them more listenable. The same could do an equaliser, but these vanished in the '90ies because CDs did not need them, the LPs equally, only the tape needed them to raise the trebles and bass, lost during recording.

[hello_hello]

As for the HiRes, a statistical study revealed again that there is no statistically difference.
http://www.aes.org/e-lib/browse.cfm?elib=14195

The main subject of the lecture in the video I linked to was what constitutes "high resolution" audio, and he believes if you limit the bandwidth to standard 20-20k it's not "high resolution" no matter what the bitdepth or sampling rate.
He even mentions the study you linked to and states that although the so called "high resolution" samples used were 96/24 or whatever, they were all 20hz-20k so he says himself it's hardly surprising nobody could tell the difference between 44/16 and 96/24 because that's not his definition of "high resolution".

He admits he can't prove you can tell the difference between CD quality audio and his definition of high resolution audio, which isn't frequency limited to 20k, and I'm still sceptical myself, and he even admits if "real" high resolution does make a difference to the listening experience it'll be a very niche market, but the subject of the video I linked to isn't about high sampling rates and bitdepth as such.

His argument.... and I find it a little hard hard to disagree, is we're at the stage where audio can be captured accurately up to 40k and beyond (and reproduced), so why not do it? The analogy was something like, we're heading towards 4k or 8k being the standard for video in a world where 1080p is usually more than adequate for the average living room, so why not do the same with audio, just in case? I tend to think it'd generally be a waste of time, but I can't prove that either.

A different study (the one you referred to is referenced in several places) but it's the one leonccyiu linked to in post #2. You can download the PDF. I've only just started reading it but here's what seems to be the initial highlight.
http://www.aes.org/e-lib/browse.cfm?elib=18296
4 CONCLUSIONS
4.1 Implications for Practice
The meta-analysis herein was focused on discrimination studies concerning high resolution audio. Overall, there was a small but statistically significant ability to discriminate between standard quality audio (44.1 or 48 kHz, 16 bit) and high resolution audio (beyond standard quality). When subjects were trained, the ability to discriminate was far more significant.

[Ghitulescu]

I shortly checked the Reiss article - it brings nothing new.
WHy HiRes should be better is however discussed in several focused articles, like
http://www.aes.org/e-lib/browse.cfm?elib=12372
http://www.aes.org/tmpFiles/elib/20170207/18046.pdf (free)
http://www.aes.org/tmpFiles/elib/20170207/15398.pdf
http://www.aes.org/e-lib/browse.cfm?elib=16632

In the end, nobody can prove anything, since the standard appears to move from one target to another.

In fact, most studies actually say that most people do not hear and do not care, while specialists, like wine tasters, can discriminate under certain conditions certain things (variations of one parameter while others were kept constant etc etc).

[JoeyMonco]

The main subject of the lecture in the video I linked to was what constitutes "high resolution" audio, and he believes if you limit the bandwidth to standard 20-20k it's not "high resolution" no matter what the bitdepth or sampling rate.
He even mentions the study you linked to and states that although the so called "high resolution" samples used were 96/24 or whatever, they were all 20hz-20k so he says himself it's hardly surprising nobody could tell the difference between 44/16 and 96/24 because that's not his definition of "high resolution".

That sounds like the same ridiculous argument that the guy who claims to have invented e-mail uses. He sets up criteria (http://www.inventorofemail.com/definition_of_email.asp) that nothing can meet such that he can always claim that nothing can be "email" but his own software thus to complete the circular argument he is the sole creator of email.

He admits he can't prove you can tell the difference between CD quality audio and his definition of high resolution audio, which isn't frequency limited to 20k, and I'm still sceptical myself, and he even admits if "real" high resolution does make a difference to the listening experience it'll be a very niche market, but the subject of the video I linked to isn't about high sampling rates and bitdepth as such.

So then right there you can dismiss his argument out of hand. A hypothesis that can't be proven is not a useful hypothesis. Sounds like classic case of begging the question.

His argument.... and I find it a little hard hard to disagree, is we're at the stage where audio can be captured accurately up to 40k and beyond (and reproduced), so why not do it? The analogy was something like, we're heading towards 4k or 8k being the standard for video in a world where 1080p is usually more than adequate for the average living room, so why not do the same with audio, just in case? I tend to think it'd generally be a waste of time, but I can't prove that either.

The basics of human hearing points to it being a waste of time unless your goal is to make better audio quality for your dog to listen to. Just in the United States alone something like 15% of the population suffers from noise-related hearing loss which involves most of them not being able to distinguish sounds above 15khz (and many times a frequency even lower). Pretty much no one is going to hear the difference other than some audiophile placebo effect.

[Ghitulescu]

Pretty much no one is going to hear the difference other than some audiophile placebo effect.

This is the argument I always use - the audiophile are generally deaf, because by the time a normal person reached the amount of spare money needed to buy one of those expensive toys, he'd lost his hearing accuracy.
To be more specific :) it's like those people hanging with supermodels 1/3rd their age...

[JoeyMonco]

This is the argument I always use - the audiophile are generally deaf, because by the time a normal person reached the amount of spare money needed to buy one of those expensive toys, he'd lost his hearing accuracy.
To be more specific :) it's like those people hanging with supermodels 1/3rd their age...

Or even if they aren't old, their system is likely to cause noise-induced hearing loss so in effect they're basically robbing themselves of hearing even the frequencies that are realistic.

[johnmeyer]

Have any of you actually sat in a proper listening room, and done A/B comparisons with truly high-end equipment? The differences are really not that subtle (although I will admit that I was in my 30s when I auditioned all this super-high-end equipment).

[hello_hello]

That sounds like the same ridiculous argument that the guy who claims to have invented e-mail uses. He sets up criteria (http://www.inventorofemail.com/definition_of_email.asp) that nothing can meet such that he can always claim that nothing can be "email" but his own software thus to complete the circular argument he is the sole creator of email.

That's probably the worst analogy I've ever read.

So then right there you can dismiss his argument out of hand. A hypothesis that can't be proven is not a useful hypothesis. Sounds like classic case of begging the question.

I can't prove it'll ever be possible to travel between the stars in a practical way, therefore the idea should be dismissed out of hand.
He said he'd really like to do a proper study himself but it's expensive. He didn't say it can't be proven, but he admitted it hasn't been. He came across as being genuinely honest about it to me.
The guy spent most of the lecture refuting lots of what he sees as audio snake oil, so I don't think he's a nutter. He's a knowledgable guy who appears to be keeping an open mind when it comes to the effect of reproducing frequencies above 20k on the listener. He thinks, because he's recorded and listened to a fair bit of it, that somehow it makes music less fatiguing to listen to, but admits he doesn't even know why.

The basics of human hearing points to it being a waste of time unless your goal is to make better audio quality for your dog to listen to. Just in the United States alone something like 15% of the population suffers from noise-related hearing loss which involves most of them not being able to distinguish sounds above 15khz (and many times a frequency even lower). Pretty much no one is going to hear the difference other than some audiophile placebo effect.

I'm sceptical myself, but he claims somehow the brain might know when ultra-high frequencies are present, and there appears to be a small amount of evidence to suggest it's possibly true. It could very well be nonsense, but I'm not dismissing it completely out of hand simply because you haven't proved it's not possible.

I shortly checked the Reiss article - it brings nothing new.
WHy HiRes should be better is however discussed in several focused articles, like
http://www.aes.org/e-lib/browse.cfm?elib=12372
http://www.aes.org/tmpFiles/elib/20170207/18046.pdf (free)
http://www.aes.org/tmpFiles/elib/20170207/15398.pdf
http://www.aes.org/e-lib/browse.cfm?elib=16632

In the end, nobody can prove anything, since the standard appears to move from one target to another.

You're just offering generalisations as fact instead of offering facts, while posting a link to a pdf that says you're wrong. Okay....

4. DISCUSSION
Findings from the listening tests suggest that expert listeners can detect differences between musical excerpts presented at 88.2 kHz and 44.1 kHz. Moreover, the qualitative analysis of verbal descriptors indicates that these differences were perceived in terms of spatial reproduction, high frequency content, timbre and precision.

I'm not saying I believe it yet. I haven't even read the whole pdf.

In fact, most studies actually say that most people do not hear and do not care, while specialists, like wine tasters, can discriminate under certain conditions certain things (variations of one parameter while others were kept constant etc etc).

When you've made up your mind.....

[JoeyMonco]

Have any of you actually sat in a proper listening room, and done A/B comparisons with truly high-end equipment? The differences are really not that subtle (although I will admit that I was in my 30s when I auditioned all this super-high-end equipment).

Yes, but is high-end equipment going to allow me to hear frequencies only audible to dogs?

[JoeyMonco]

That's probably the worst analogy I've ever read.

Seems pretty analogous to me. The guy simply sets up criteria that he knows nothing can meet so he can then simply try to dismiss any evidence he doesn't like because then he can say "it's not true high-res audio". No different than the guy who claims things can only be e-mail only if it meets some ad-hoc criteria.

I can't prove it'll ever be possible to travel between the stars in a practical way, therefore the idea should be dismissed out of hand.

If your method of interstellar travel is based on an unproveable hypothesis then yes it should.

He said he'd really like to do a proper study himself, but that it's expensive. He didn't say it can't be proven, but he admitted that it hasn't been. He came across as being genuinely honest about it to me.
The guy spent most of the lecture refuting lots of audio snake oil, so I don't think he's a nutter. He's a knowledgable guy who appears to be keeping an open mind when it comes to the effect on the listener of reproducing frequencies above 20k.

Keeping an open mind based on what? We simply cannot hear anything above that and numerous people can't even hear the full range of frequencies anyway due to either age-related or noise-related hearing loss. Once he has a well-defined, proveable hypothesis that explains the method that would allows us to do this then it can be entertained.

I'm sceptical myself, but he claims somehow the brain might know when ultra-high frequencies are present, and there appears to be a small amount of evidence to suggest it's possibly true. It could very well be nonsense, but I'm not dismissing it completely out of hand simply because you haven't proved it's not possible.

Unproveable hypotheses should always be dismissed out of hand. That's the realm of pseudo-science.

[hello_hello]

Seems pretty analogous to me. The guy simply sets up criteria that he knows nothing can meet so he can then simply try to dismiss any evidence he doesn't like because then he can say "it's not true high-res audio". No different than the guy who claims things can only be e-mail only if it meets some ad-hoc criteria.

FFS. Watch the video before you comment on it. The guy records what he calls "high resolution" audio himself (audio that's not frequency limited to 20k) and has the ability to reproduce it. He even sells it. It's not some vague, forever moving goal post.
My current TV can only reproduce a maximum resolution of 1080p, but I've heard rumours there's 4k content available.

He might be wrong. If there is a discernable difference, maybe it's the higher sample rate and bitdepth that's responsible after all, and the additional high frequencies have nothing to do with it. I don't know. I'll happily admit it.

If your method of interstellar travel is based on an unproveable hypothesis then yes it should.

I don't have a method for interstellar travel, but that still doesn't mean it won't one day be possible.

Unproveable hypotheses should always be dismissed out of hand. That's the realm of pseudo-science.

There's a difference between "unproven" and "unprovable" you've not appeared to grasp.
Anyway, I'm not going to turn the thread into a pointless argument about it, because I've said repeatedly I'm sceptical, just not completely dismissive of the idea, and you're not offering a single piece of evidence to prove your claims are correct either.

[johnmeyer]

One thing to point out about the quality of audio reproduction is that frequency response is only one measure of a system's ability to accurate reproduce various sounds. For instance, systms having similar frequency response may reproduce transients (like those produced by cymbals) quite differently. You can read about this in various links given by this Google search:

Transient Response Google Results (https://www.google.com/search?q=audio+transients+frequency+response&ie=utf-8&oe=utf-8)

The reason why frequency response doesn't tell the whole story is that it is based on Fourier transforms which, in turn, are based on the assumption that you are measuring a continuous signal (i.e., it is the same for an infinite amount of time in the past and the future). For most purposes, this assumption works, because the difference between a waveform that is continuous and one that is changing all the time isn't much. However, when you have an impulse (an infinitely short, infinitely steep waveform), Fourier analysis breaks down. And, it is not just the math that blows up, but so does its ability to properly give you the correct answer about how a system will perform.

The fact that frequency response is not sufficient helps explain why, even though most people can't hear much in the frequencies above 10,000 Hz, they can still hear differences between systems that are flat through 20,000 Hz.


And, you need to look at dozens of other specs besides this one: harmonic distortion; intermodulation distortion; clipping; and many more.

Bottom line: you have to do A/B comparisons. The proof is in the listening, and it is not a placebo effect. In the old days, the audio magazines did lots of double-blind studies involving multiple people. These were really great and, for instance, were able to show that there was no audible difference between lamp wire and "Monster Cable" despite all the silly claims made by the ridiculously expensive cable manufacturer.

[Ghitulescu]

Transients are like their name already said, high frequency spectra - the higher the bandwidth the higher the probability that these transients are reproduced.

The differences in sound are usually given in the deviations from the linear curve (or whatever curve) of one appliance vs another - it's similar to playing with en Equaliser with plenty of "slices". However, the mechanical reproducers, the loudspeakers and headphones, do differ from linearity in an extreme manner (one can see the graphs) - which poses the real question - when someone master a piece of music using a pair of near field monitors, would the sound be perceived the same with another speaker? The answer is no, but luckily most speakers have similar shapes of the curves.

Also many fantasy concepts that abound in the high end literature can easily be traced down to a single parameter, the phase difference (between channels and between individual frequencies within the same channel).

I did not want to contradict you, just to explain that some unmeasurable things are actually very measurable.

[johnmeyer]

Transients are like their name already said, high frequency spectra - the higher the bandwidth the higher the probability that these transients are reproduced.

<snip>

Also many fantasy concepts that abound in the high end literature can easily be traced down to a single parameter, the phase difference (between channels and between individual frequencies within the same channel).

I did not want to contradict you, just to explain that some unmeasurable things are actually very measurable [emphasis added by johnmeyer].I don't think you entirely understood what I posted. I was in no way arguing that you couldn't measure characteristics that contribute to better reproduction; instead I was arguing that frequency response alone was not sufficient.

Also, you failed to understand what I said about the relationship between frequency response and transients: there is certainly some relationship between frequency response and transients, but that as the transients get more abrupt (an "impulse" is the mathematical extreme example), the less likely the frequency response measurements will properly characterize the system's ability to reproduce those transients. That is because frequency response depends on the Fourier math which requires that the waveform be infinite in the time domain, which is never the case with any waveform but, with transients, is so far from being the case that this measurement no longer has any useful meaning.

Finally, the whole point of my finishing my short summary of other measurements that are needed with "and many more" is that there are dozens of others that I didn't list (hum, noise, linearity, ... the list goes on).

So, to repeat: in no way was I arguing or stating that one cannot measure or characterize all or part of an audio reproduction system; I only argued that frequency response alone is not sufficient.

[Ghitulescu]

An amplifier having a linear bandwidth of 1Hz to 100 GHz would certainly reproduce any "audio" transients. 100Khz bandwidth is something very common today, unlike in the '50ies, it helps a lot reproducing transients downto 10ns.
I know how Fourier applies to "music".
However, any recoding method is somehow limited in bandwidth. It is my belief that whole digital freed the audiophiles (but mostly the regular people) from a lot of shortcomings, it simply committed suicide - digital cleaned the sound up to the point where, after getting used to the high quality, people started to feel things that they otherwise never ever had the possibility to notice. And those people liked to go back to eg vinyl or tape, because these low quality (IMHO) sources mask a lot of these issues. One simply cannot hear the "broken glass effect of the brickwall digital filter" because it is simply dunked 20dB below in the tape hiss :) or the resonance of the arm-cartridge masks it.

Just in case I wasn't clear - a transient is an impulse that may be described by a frequency equivalent, taking into considerations a lot of other factors, like cycle time, amplitude/fall/raise time etc. It may even be a compound sound.

I repeat, it is not my intention to argue with anyone. My goal is to demystify the HighEnd claims. Most of the "sonic differences" can be traced down to irregular (non linear) spectrum, whereas most concepts like the famous soundstage may be tracked down to phase errors, which confuse the brain on the location of the sounds, in particular when they also disagree with the sound levels :).

[johnmeyer]

<snip>

I repeat, it is not my intention to argue with anyone. My goal is to demystify the HighEnd claims. Most of the "sonic differences" can be traced down to irregular (non linear) spectrum, whereas most concepts like the famous soundstage may be tracked down to phase errors, which confuse the brain on the location of the sounds, in particular when they also disagree with the sound levels :).Here is some support for your point of view:

Testing audiophile claims and myths (http://www.head-fi.org/t/486598/testing-audiophile-claims-and-myths)

He provides dozens and dozens of links to "ABX" studies, many of which show pretty conclusively that most audiophile equipment does not provide enough of an advantage over cheaper equipment to be discernible. [For those not familiar with ABX, the listener compares two hidden products ("A" and "B") and says which he prefers, but then has to listen once more to an unknown product "X" and say whether it is "A" or "B"]. Almost half of the links are to tests of Monster Cable, or similar interconnects. Those tests have been around since the 1970s and obviously show that Monster Cable provides no advantage.

While the end conclusion of this blogger is that many of the audiophile claims are unsupported, there are many flaws in both the tests linked to, as well as in his conclusions. Without going into all of them, the biggest flaw is that all the tests he cites are comparing audiophile equipment with really good standard consumer electronics. In other words, it is like comparing a Ferrari to a Lexus sports car where, in the end, the Ferrari cost 10x more because it looks really cool (which is his conclusion about why the high end stuff is still worth it). If instead you compare a Ferrari to a Chevy (or a Yugo), a lot more differences would emerge, and the differences would be substantial, important, and easily discerned.

It isn't worth spending any more time on this: despite the flaws in the ABX tests he links to, most modern amplifiers are amazingly good, and the differences that used to be commonplace back in the 1960s have pretty much vanished now that we have computerized design that has been fabricated into integrated circuits where that design includes all the lessons from the science and experience of previous generations of designers, and which can be included in any amplifier for a few pennies.

[Ghitulescu]

While the end conclusion of this blogger is that many of the audiophile claims are unsupported, there are many flaws in both the tests linked to, as well as in his conclusions. Without going into all of them, the biggest flaw is that all the tests he cites are comparing audiophile equipment with really good standard consumer electronics. In other words, it is like comparing a Ferrari to a Lexus sports car where, in the end, the Ferrari cost 10x more because it looks really cool (which is his conclusion about why the high end stuff is still worth it). If instead you compare a Ferrari to a Chevy (or a Yugo), a lot more differences would emerge, and the differences would be substantial, important, and easily discerned.
I do not consider them flawed, as one of the points is to show that quality does not depend on the money invested. They may show that above a price tag, spending money is a waste.
One point shown is and should be that many high end manufacturers and in particular sellers are dishonest.
One field that was somehow spared from this aggressive marketing was the pro field. Either the pro shops have fallen to "consumer level" (which some did actually) or the professionals started to be less pro in knowledge (also some did).

[hello_hello]

johnmeyer,
Just so I'm not getting what you said wrong, are you saying hat most ABX tests are flawed because they use top of the line audiophile gear and not more typical consumer stuff, or is it the tests is comparing that sed different types of equipment, making his comparisons flawed? It's no big deal, I'm just not sure what you meant.

I guess when you're performing listening tests the equipment used should be as a high quality as possible, or colour the sound as little as possible etc, but you reminded me of a Hi-Fi system comparison I tried many years ago. I owned an inexpensive "all-in one" system, although back then the manufacturers still tried to pretend they weren't "all in one" by making the amp and cassette decl sections etc appear to be individual components with the front panels design, even though they weren't.

I'll stop rambling..... the guy I shared a house with at the time owned a fairly expensive system. Too long ago to remember any details, but I recall it cost him lots of thousands back in the early 80s. One day for fun I unplugged his speakers and plugged them into my several hundred dollar system (unplugging my speakers) and it surprised me how little difference there was to the sound when the same speakers were used. Except for the volume. His amp was capable of producing hideous amounts of volume. Mine would start to distort by the time it was loud enough to hear in the next room, although there's a chance that was a speaker ohm-age thing that I was ignorant of at the time.

[johnmeyer]

johnmeyer,
Just so I'm not getting what you said wrong, are you saying hat most ABX tests are flawed because they use top of the line audiophile gear and not more typical consumer stuff, or is it the tests is comparing that sed different types of equipment, making his comparisons flawed? It's no big deal, I'm just not sure what you meant.I tried to be clear by providing a car analogy. The Lexus is still a "consumer" level car, but is at the very high end of that classification, whereas the Chevy or Yugo is at the lowest end. When comparing the Ferrari to the Chevy, there will be a lot of obvious differences whereas it gets harder to spot those differences when you compare to the Lexus.

The same thing is true of stereo equipment. In the ABX tests in that link I provided, they compared the "Ferrari" equipment with equipment that was at the very high end of what you would have found in places like Pacific Stereo (that was a normal, consumer grade, stereo store, back when stereo was a big thing). If they had instead used some of the middle-of-the-road equipment, then they would easily have found differences.

This is important because many claims attempt to say that there is no difference between any audio equipment. That simply is not true. Therefore the question is: where do the differences start to appear? And, are they always there, but just harder to find?

I can't tell you how many times, when restoring media, I think I have finally devised some restoration technique that really works when, after staring at the film or photo for hours, I suddenly spot something that I had been staring at all the time, but had overlooked. Once I know what to look for, I realize that there were huge differences there all along, but I just wasn't "tuned in" to know what to look for. The same thing is true in comparing audio equipment, tasting wine, restoring media, etc.

Finally, as some of those tests pointed out, the audio used for the tests makes a HUGE difference in the outcome. If you use high-compressed audio, you won't be able to hear most differences, especially if the audio contains distortion (I am listening right now to Jimi Hendrix' "Purple Haze," and it would be a horrible track to use for audio comparisons).

One of the most convincing demonstrations of the superiority of an audio component was when I auditioned the original Veoloydne subwoofer at Century Stereo in San Jose. They used Jennifer Warnes "Way Down Deep (https://www.youtube.com/watch?v=aY28-5cnbeQ)." It is the polar opposite of most subwoofer tests, since it contains no thunderous, crashing thuds, but instead requires that the subwoofer "articulate" the deep bass notes, something cheap subwoofers cannot do.

The differences were obvious, and the superiority of the very expensive ($2,000) Velodyne were obvious. It is the only time in my life I have ever purchased a piece of "audiophile" equipment.

If instead I had auditioned the subwoofer by playing the movie "Independence Day" I might have said the two subwoofers were equivalent because one thud sounds just like another.

So in addition to the point I already made, the audio used makes a huge difference.

[Ghitulescu]

I tried to be clear by providing a car analogy. The Lexus is still a "consumer" level car, but is at the very high end of that classification, whereas the Chevy or Yugo is at the lowest end. When comparing the Ferrari to the Chevy, there will be a lot of obvious differences whereas it gets harder to spot those differences when you compare to the Lexus.

The same thing is true of stereo equipment. In the ABX tests in that link I provided, they compared the "Ferrari" equipment with equipment that was at the very high end of what you would have found in places like Pacific Stereo (that was a normal, consumer grade, stereo store, back when stereo was a big thing). If they had instead used some of the middle-of-the-road equipment, then they would easily have found differences.
True.
This is important because many claims attempt to say that there is no difference between any audio equipment. That simply is not true. Therefore the question is: where do the differences start to appear? And, are they always there, but just harder to find?
Unfortunately, a lot of people fall right in the opposite direction, and give reasons for the ABX-haters to contest an absolute legitimate and scientifically proven means, probably the only one in a subjective world.
..... especially if the audio contains distortion (I am listening right now to Jimi Hendrix' "Purple Haze," and it would be a horrible track to use for audio comparisons).
A story goes saying that once they sent a tape of Jimmy to Japan for a pressing. The Japs returned the tape saying that probably it was defective or the shipping caused to it, for it was too distorted.
If instead I had auditioned the subwoofer by playing the movie "Independence Day" I might have said the two subwoofers were equivalent because one thud sounds just like another.

So in addition to the point I already made, the audio used makes a huge difference.
Indeed. However, the audiophile still entrenched itself in the stereo field, not in surround, which appears to be exclusively the field of videophiles (yes, there are DVD-A and SACD and BD-A, but their amounts are negligible). Pretty much like their forefathers that criticised the "ping-pong" stereo :)
And indeed, the video industry favourise small tweeters coupled with a subwoofer, for exactly these are the sounds they have in movies. This is why they favourised the 5.1 scheme and not the old "quadro" (which would have been 4.0).
The games industry also benefit from 5.1 (or any x.1 scheme) for the exactly the same reasons.

It is a delight to hear surround (eg from a SACD/DVDA/BDA) with 4 or 6 fullrange loudspeakers and maybe a subwoofer.

[johnmeyer]

It is a delight to hear surround (eg from a SACD/DVDA/BDA) with 4 or 6 fullrange loudspeakers and maybe a subwoofer.Yup, that is what I have. No small speakers here: every one of the five main speakers (I have a 5.1, not a 7.1 setup) are full-range, with tweeter, woofer, & crossover.

[Midzuki]

..........

It is a delight to hear surround (eg from a SACD/DVDA/BDA) with 4 or 6 fullrange loudspeakers and maybe a subwoofer.

Yes.... HUGE and HEAVY wooden enclosures, each one containing a tweeter, a midranger and a woofer.

Old-School :devil: FTW! :cool:

[manolito]

When it comes to speakers I think nobody beats me in the "Old School" department... :p

My speakers are original Klipschorn corner horns bought in 1979. Before those I did experiment with quadro setups. But after experiencing the Klipschs I never went back, and even if you cannot compare the old quadro stuff to current Home Theater setups I never felt the need to try multichannel setups.

When I got the Klipschs there were no CD players, the vinyl mastering engineers had no clue about phase accuracy, so even Mr. Paul W. Klipsch recommended a mono center speaker to remedy this hole in the center of the sound stage. But when CDs became standard and recording and mixing techniques became much better, it always amazed me how well the Klipschs scaled to these improved sources. With an accurate stereo source there is absolutely no hole in the middle any more, no need for a center speaker (or a surround system).

And some thoughts about listening tests:
This is an extremey complex topic. To make out subtle differences between two sources it not only takes first rate amps and speakers. The room plays a big role, the listeners need trained ears, the loudness between the sources must be carefully calibrated using test tones (the louder source will always sound better), and the source content is essential. Like it has already been stated, for a Jimi Hendrix track which has already been recorded with a lot of distortion you will hardly be able to hear the difference between normal resolution and high resolution audio.


Cheers
manolito

[johnmeyer]

When it comes to speakers I think nobody beats me in the "Old School" department... :pWell ... maybe so, maybe not :)

In addition to having Klipsch speakers for the three front speakers in my media room, I have a separate living room which has two full-sized (very big) bass reflex enclosures with an Altec 605 in each one. These "sound of the theater" speakers were manufactured in the 1950s and 1960s, When someone looks up "old school speakers" in the dictionary, it says, "see Altec 605."

[Ghitulescu]

When I got the Klipschs there were no CD players, the vinyl mastering engineers had no clue about phase accuracy, so even Mr. Paul W. Klipsch recommended a mono center speaker to remedy this hole in the center of the sound stage. But when CDs became standard and recording and mixing techniques became much better, it always amazed me how well the Klipschs scaled to these improved sources. With an accurate stereo source there is absolutely no hole in the middle any more, no need for a center speaker (or a surround system).

In fact, for a long time, the live stereo events were captured according to the old adagio: 2 speakers therefore 2 mics. This created a "hole" in the middle. This was not immediately evident, for the customer media (not the pro tapes) was "fuzzy" enough to cover this hole more or less. Plus that it was so nice a stereo effect (remember, it was after stereo managed to get past the claim of the monoists that it only helps manufactures to cash twice for the same merchandise - something similar to biamping today), in particular with headphones. Many stereo albums were recorded voice to the left, instruments to the right (id est, the channels were completely separated from each other), until people at the mixers learned how to recreate the stereo (which means spatial not two :) ) image correctly. So, for large sources, like choirs, people started to use 3 mics (or more) and to give attention to the "hole". The CD or better said the digital recording in studios, due to its lack of self-noise and extremely low distortions, also helped the engineers to pinpoint and gradually eliminate all deficiencies of the old technology of capturing the sound - since we are in a video forum I say that a similar thing happened here, when HDTV was introduced, when the rumour circulated that many presenters (anchormen and in particular women) complained about HDTV to be too sharp that all the wrinkles are now seen :) :) :) this happened for real in audio studios....

[Gser]

All the high tech analyzes aside I can really hear the difference in the noise floor between 44/16 and 96/24 which is why I'm always happy to search for hi-res versions. It's really gives a sense of space, depth and clarity while as cd's sound dull and fuzzy in comparison.

[Ghitulescu]

All the high tech analyzes aside I can really hear the difference in the noise floor between 44/16 and 96/24 which is why I'm always happy to search for hi-res versions. It's really gives a sense of space, depth and clarity while as cd's sound dull and fuzzy in comparison.

I did this myself - took a HiRes source, in this case a stereo source, Santana, and downconverted it with foobar to CD. The sound was very good, yet in direct comparison something was missing from the CD.

[Ghitulescu]

http://oi49.tinypic.com/2yv29es.jpg

[manolito]

I did this myself - took a HiRes source, in this case a stereo source, Santana, and downconverted it with foobar to CD. The sound was very good, yet in direct comparison something was missing from the CD.

The downconversion might have flawed the comparison. Resampling is never lossless, and bitdepth conversion also has problems.

For a really accurate test you need to make a high quality stereo recording (critical content like a grand piano), mics need to be ultra expensive (e.g. Neumann). After the mics you need a first rate mic preamp. Then you split the preamp output into 2 AD converters, one converts to 24/96 (or even higher), the other one to 16/44.1. These AD converters need to have the same ultra high quality, preferably they should come from the same manufacturer.

After you spent so much dough and time and energy for the comparison I might be more willing to believe your results...


Cheers
manolito

[Ghitulescu]

The downconversion might have flawed the comparison. Resampling is never lossless, and bitdepth conversion also has problems.

For a really accurate test you need to make a high quality stereo recording (critical content like a grand piano), mics need to be ultra expensive (e.g. Neumann). After the mics you need a first rate mic preamp. Then you split the preamp output into 2 AD converters, one converts to 24/96 (or even higher), the other one to 16/44.1. These AD converters need to have the same ultra high quality, preferably they should come from the same manufacturer.

After you spent so much dough and time and energy for the comparison I might be more willing to believe your results...


Cheers
manolito
Two different ADCs are never the same. Also the playback on different machines are never the same.
This, the downconversion - faulted or not-, was to me the closest comparison I could think of, as everything, except the internal architecture of the DAC - influencing the way the signal is interpreted-, was identical.

PS: I know that downconversion can be carried out in several ways, as each variant has advantages and disadvantages.

PPS: I just said an opinion based on the aforementioned facts and tests. I did not impose a meaning, except that I believe there is such a difference.

But the biggest problem is not this, downconversion or not, but the fact (this is a fact) that the music is "faulted" at source, in other words, the studios provide us music for which even the CD standard is way too generous :) :) The question of HiRes for this music is rather philosophic, as one can have the originally distortions better reproduced :) :)