Audio Blind Testing - You Are Doing It Wrong! (Video)

[respice finem]

Yes, but, I'm afraid the manufacturers will not do this, at least not at a large scale.
We probably all remember the "ancient" German Hi-Fi norm DIN 45500. While it's almost laughable today, it wasn't made by (or for) deaf people, and it's relatively safe to assume, the majority of listeners will not hear the difference to anything better than this (at least when applied to the whole audio chain). But, leaving it at that and focusing on room treatment /correction would kick a large part of manufacturers out of the market, because they wouldn't be capable to compete with the "big boys" of room treatment, or room correction, as Dirac, Audyssey etc. But everybody can solder a DAC or amp together and boast with "better" specs. The other thing, cheap room treatment: I guess it will remain in the DIY domain, at least where I live (Germany), simply because of the cost of manual labour.

 

[ahofer]

Research into just audible thresholds are the province of academic psychoacoustics, and there is a literature on it. Even a wikipedia article.

And a thread…https://audiosciencereview.com/foru...-thresholds-of-amp-and-dac-measurements.5734/

 

[watchnerd]

Since headphone amps were brought up:

How problematic is it to take the headphone out of a headphone amp and feed it into an mic input ready ADC like the RME ADI-2 Pro?

My first thinking was that one should be able to use the mic inputs from an impedance POV, but then I realized I don't know what the reference voltage / input level would be.

 

[Newman]

So this discussion now goes to why Harman et al. have not focused on what appears to be a most relevant piece of research for consumer electronics: "A what point can we say changes in sound quality are rigorously inaudible?" The insight from this research would allow consumer electronics to focus on more important areas: features, reliability, UI design, surround immersion, DSP room correction, etc.

Let's say the answer to those questions is that any decent transistor circuit is so far below "rigorously inaudible" that the research you say is missing will really be research into "just how badly, stupidly, even deliberately, some say heroically, mis-designed does good budget circuitry have to be before it starts to affect home audio perception". Let's say it's true, because the existing understanding of human hearing thresholds truly is that far above routine specs of transistor electronic gear, and because there is no rigorous evidence that the simple logic is wrong, ie that we already know the answer by proxy.

I'll tell you what happens next: wine-taster-style audiophiles will rip into such research as methodologically flawed; amateur statisticians will insist the numbers don't add up; Michael Fremer will remind us how he has never heard digital electronics that doesn't make him headachy and sick in a few minutes; everyone who owns, or wants to own, or builds and sells expensive electronics -- especially the bad stuff that measures above the thresholds -- will pooh-pooh the whole business of science ever being able to tell us anything; and people on audio forums will carry on about what specific experiments they would like to see that isn't available, even though the above hypothetical research was only done to answer their last barrage of requests for 'missing' research to be done....and we will end up exactly where we are today.

The challenge is to let go of denial, instead of erecting ever more hurdles for others to cross before we will listen.

cheers

 

[JJB70]

Audio electronics were a "solved problem" years ago. DACs are a commodity. For the electronic part of the chain audible transparency is available for peanuts. The amplifiers and DACs in wireless speakers and headphones are not usually anywhere near state of the art and if offered as standalone items would probably be derided yet how many people have any complaints about the audible performance? I think the true basis of objective audio evaluation in terms of sound quality is what is audible, not just in terms of whether it is possible to discern differences but whether such differences are noticeable or affect normal listening. For audio electronics that means pretty much everything is fine providing specs like power output are appropriate.
So in some ways much of the audio hobby is obsolete. This is not just a provocative bit of devils advocacy, I think the mainstream market has got it right in recognizing that the way forward is wireless speakers and headphones, audio gear other than wireless and sound bars, which was once very much part of the mainstream, is now a niche.
However, it's not just about logic and numbers. People still appreciate industrial design, premium build quality, good through life support, UI, feature set etc and buy stuff which probably doesn't sound any better but which gives them a lot of pleasure. Then there is upgradeitis which again is easy to criticize but how many can honestly say that they have never bought something that they didn't need but just wanted?

 

[Newman]

The only trouble with 16 wireless speakers is that they need 16 power cords...

 

[Bruce Morgen]

The only trouble with 16 wireless speakers is that they need 16 power cords...

Yup. That's why my main L+R speakers and sub are active, but my ambience/surround speakers are passive (relative) antiques: I want to be able to control mains power to the whole system from a single, conveniently located power strip.

 

[2M2B]

Ah, reminds me of how Hydrogenaudio got mad when I couldn't tell 170kbps Musepack on their samples & my samples. But could tell AAC/Vorbis/MP3 with my own samples even at 320kbps, And also admitted 160kbps Opus was just as transparent beyond one rare sample needing 400kbps.

 

[Spocko]

I'll tell you what happens next: wine-taster-style audiophiles will rip into such research as methodologically flawed...

cheers

ROFL, this line really says it all. I keep forgetting about the infallibility of the self-appointed audio sommelier - as in any religion, it comes down to unshakeable faith. Like an atheist walking into the Vatican with an affidavit signed by Mother Teresa and expecting the Pope to actually listen.

 

[krabapple]

Ah, reminds me of how Hydrogenaudio got mad when I couldn't tell 170kbps Musepack on their samples & my samples. But could tell AAC/Vorbis/MP3 with my own samples even at 320kbps, And also admitted 160kbps Opus was just as transparent beyond one rare sample needing 400kbps.

Why would 'Hydrogenaudio' (an entire forum) get 'mad' about that?

Very respected posters there have reported ABX data showing they could hear a difference between lossless and 320kbps, on some samples. Nor does 'Hydrogenaudio' claim any lossy codec to be always transparent.

 

[krabapple]

I'll tell you what happens next: wine-taster-style audiophiles will rip into such research as methodologically flawed; amateur statisticians will insist the numbers don't add up; Michael Fremer will remind us how he has never heard digital electronics that doesn't make him headachy and sick in a few minutes; everyone who owns, or wants to own, or builds and sells expensive electronics -- especially the bad stuff that measures above the thresholds -- will pooh-pooh the whole business of science ever being able to tell us anything; and people on audio forums will carry on about what specific experiments they would like to see that isn't available, even though the above hypothetical research was only done to answer their last barrage of requests for 'missing' research to be done....and we will end up exactly where we are today.

The challenge is to let go of denial, instead of erecting ever more hurdles for others to cross before we will listen.

cheers

Bravo! Perfectly expressed by someone who's obviously a veteran observer of the cranks, creeps, and grandiosely ignorant who swarm this hobby.

 

[2M2B]

Why would 'Hydrogenaudio' (an entire forum) get 'mad' about that?

Very respected posters there have reported ABX data showing they could hear a difference between lossless and 320kbps, on some samples. Nor does 'Hydrogenaudio' claim any lossy codec to be always transparent.

You ask me, I came back with 5 users throwing personal attacks at me and then shortly after got banned. When In a thread about what codec do you use my main two were 170kbps MPC & 256kbps MP3, because I could tell AAC/Vorbis & even Opus at 320kbps on my hard samples. They seem to be very touchy about Musepack being immune to artifacts at 170 ~ 256kbps, I've noticed that too with 256kbps MP3 if set to 48KHz.



Here a censored screen grab's I got before I was banned.

 

[krabapple]

If you're the one who wrote the 'Quote' -- it doesn't contain any ABX test results. Extraordinary test results would usually also have a link/description of the sample and settings somewhere along the line so others could try to replicate. Did the full text or some prior post of yours have these? By itself the quoted text would certainly be a TOS8 violation.

And are claims like 'worse by a long shot' from just your ABX results, or from large-scale data?

Also seems like there was already history by that point?

 

[2M2B]

If you're the one who wrote the 'Quote' -- it doesn't contain any ABX test results. Extraordinary test results would usually also have a link/description of the sample and settings somewhere along the line so others could try to replicate. Did the full text or some prior post of yours have these? By itself the quoted text would certainly be a TOS8 violation.

And are claims like 'worse by a long shot' from just your ABX results, or from large-scale data?

Also seems like there was already history by that point?

Jog on dude It was a thread about what codec you use daily not a public blind test. I hope your joking that is anyway to act like when someone has a opinion since threatening me Is not going to get me pump out a ABX log since I noticed you lot ignored my Vorbis vs MP3 sample I posted as a thread over there.

Funny how you're asking me to dump proof but when anyone asks you lot to show ABX logs of Musepack VBR method artifacting by leaving music/sounds at 88 ~ 105kbps it nothing but radio silence?.

Here more screenshots of how nasty It got there, the thread where I was hounded the moment you lot tried pretend you didn't ignore me the moment I did link proof like with Vorbis vs MP3.

Here the thread where I linked samples that Vorbis/MP3/AAC failed on but MPC at 175kbps was transparent on.

 

[krabapple]

Mmmnope, if you make claims like that there, doesn't matter what the topic is. Posting that you like something is fine. Extraordinary claims about what you heard require evidence. TOS8 applies there. Seems you just weren't clear on the HA concept...and still aren't.

Your idea that there is some collective 'lot' there that reads every post you make is also amusing.

 

[2M2B]

Did you just ignore the bottom link me showing samples for what I was talking about or you just gonna keep calling people trolls for no reason?.

Edit: And once again a HA user runs off the moment they're called out. lol

 

[SIY]

ROFL, this line really says it all. I keep forgetting about the infallibility of the self-appointed audio sommelier - as in any religion, it comes down to unshakeable faith. Like an atheist walking into the Vatican with an affidavit signed by Mother Teresa and expecting the Pope to actually listen.

At least actual sommeliers pass blind tasting exams as part of their certification.

 

[Killingbeans]

Funny how you're asking me to dump proof but when anyone asks you lot to show ABX logs of Musepack VBR method artifacting by leaving music/sounds at 88 ~ 105kbps it nothing but radio silence?.

Dude, you need to grow some thicker skin. I don't see anything "nasty" in any of your links.

I have no idea what Musepack is, and I don't really care. But rule #1: You make the claim -> you cough up the evidence. It's nothing personal, just a common practicality to minimize wasted time and confusion. Verification comes later.

 

[krabapple]

Did you just ignore the bottom link me showing samples for what I was talking about or you just gonna keep calling people trolls for no reason?.

Edit: And once again a HA user runs off the moment they're called out. lol

Some of us have lives IRL, thus there are times when ASR is not onscreen. There are also things called time zones, look into them.

But sure, let's waste more time on you.

The first link you posted here in evidence of your dismissive treatment by HA was to imgur screencaps of a HA thread called Vorbis better than opus? - which is not "a thread about what codec you use daily." The Vorbis thread was started by someone who made a dubious claim about Vorbis's audible inferiority based on spectrograms, but quickly acknowledges his error.

Later on in that thread a certain 'ani_Jackel3' makes a sweeping, unsupported and clearly TOS8-violating claim that "overall Vorbis is worse than AAC/MP3 by long shot not even transparent on most samples at 256kbps while AAC/MP3 are to be blunt."

That's the first quote in your first screencap.

ani_Jackel3 was called out on that (including by shadowking, who hoped you'd show evidence), and never got back to that thread with ABX logs and samples to back it up. But ani_Jackel3 does not appear to be banned at HA. His profile shows him is 'Full Member". His post (from July 2021) is still up on the thread. There's no posts in the HA Recycle Bin from that thread/time.

As for the 2 subsequent links you posted here ..the first shows someone (you?) being chided for "expecting [HA readers] to be able to quote you by heart" rather than simply linking to your evidence first. (hmmm I sense a pattern here...)

And then finally a link to ani_Jackel3's posted samples (no ABX logs)...from Sept 2020....that claim to demonstrate Vorbis 160kbps failing on two samples vs lossless..on a thread about Musepack.

Then in August of this year 2021, a certain poster named' 2M2B' appears on that thread, to answer a question directed at ...ani_Jackel3. He's not banned either.

Feel free to keep semi-coherently complaining about your supposedly raw treatment there. But don't ask anyone to care much at this point.

 

[Spocko]

And a thread…https://audiosciencereview.com/foru...-thresholds-of-amp-and-dac-measurements.5734/

But let's take this thread in the context of the caveat by @amirm

"Just a note: an exhaustive analysis and proof simply does not exist for any of this. As such, take them as guidelines which are better than having none."

And with those guidelines in mind, you're pretty safe getting anything that's green or better. If you're noise floor is above average, might as well get orange rated gear because what's the point?