Testing audiophile claims and myths

[Blaithin]

I agree too. One of the best recordings I've ever heard was an LP of Fiedler's Gaeitie Parisienne that was recorded in 1952. I think it was the second Living Stereo recording made by RCA. The stuff that matters are the judgements made by the musicians and engineers, not incremental increases in data rates. I think sometimes audiophiles spend more time listening to the formats of their music than they do the music itself.

I agree (again hahaha). That lesson was a hard-earned one for me...while I was never wholly on any side of the debate, I have always leaned more towards the more objective side. It is hard...even nigh impossible, however, to convince people who are more entrenched in the subjective rhetoric with factual arguments. It is like an [American] conservative vs liberal and vice-versa...no amount of facts will convince those who hold their beliefs with fervent religiosity. You cannot convince them...you can only shake them up and hope they start thinking more critically about their current direction.

Which is why my method is to rock the boat where there's common ground...to kind of poke around and see whether they're more of a hard-line audiophile or a music lover. Most people will start questioning themselves once you kind of ask what they're willing to sacrifice for sonic perfection because sonic perfection will eventually demand of you to abandon music you love for sounds you [supposedly should] love.

It is not too different from Abraham being asked to kill Isaac for God...though in a kinder interpretation, God is not so cruel as to test Abraham that far...Sound, however, is not so kind a god. Kind of like my friend...who woke up one day to the amazing system and recordings he's had since forever, but found he loved none...(he kind of broke down); Found that he was satisfying something that did not satisfy him. He's since dialed himself back and we're both firmly in the same direction music lovers first...audiophiles no longer.

P.S. I like to inject religious metaphors when speaking about audiophilia because the online audiophile community has always inspired a certain modicum of zealotry that I find funny. What with the miraculous "experiences" from pseudo-scientific stuff, and the fervent gesticulating about each upgrade as if experiencing God's grace, or being possessed by the Holy Spirit. (Plus, the Judeo-Christian mythos is easier to help re-frame my arguments to a pre-dominantly Western crowd)

 

[Blaithin]

I wonder why it is that whenever I ask the info on killer tracks, it always seems to be a really obscure foreign CD that costs a lot of money to order from overseas. I just did that for another person's killer track. I'm not eager to do it again. Why don't domestic CDs at reasonable prices ever have killer tracks? I bet there it has something to do with expectation bias.

There was a whole thread dedicated to the lossy-lossless thing over on ASR (been trying to find it)...I think there was a Tracey Chapman (known for Fast Car) track considered to be a killer track posted there...something about her voice having so much natural distortion that it works against mp3 compression pretty badly.

As for Amir's general sentiment for hearing differences in ABX testing (direct quote from a totally different thread).

It is hard for me too despite my training.

I should also note that training is not limited to hearing impairments but optimal way to take the test. Ability to loop small segments for example is very important as compression artifacts vary from moment to moment. You can have all the training in the world but if all you do is play one clip, then play the other, you may very well miss the differences.

He makes training for ABX'ing sound like Asian education hahaha. Training to take successfully take a test was never appealing to me, and it still isn't

Edit: Found it. Don't know if you can post ASR links here so I just put the thread title instead. Look for "Anybody Out There Who Hears a Difference Between 320 kbps MP3 and Red Book CD? What Differences Do You Hear?" thread title on ASR. The first few tracks linked via youtube are Suzanne Vega - Tom's Diner (Original version), and Tracey Chapman - Fast Car (Presumably not the 2015 re-master of the song).

 

[bigshot]

I think certain people who claim to have exceptional hearing are just serving their own egos because they want to be seen as "authorities". Certain people cheat on listening tests. And I called certain people out on it once. Certain people refused to answer my questions. Certain people need their own forum so they don't get questioned in a way that they don't want to be questioned. I don't believe what certain people claim, even if they now admit what I called them out on and they refused to admit before.

The way to judge the quality of music is to listen to music the same way you normally listen to it. Use your own equipment in your own living room playing a song. Sit on the couch and listen very carefully. Listen to the song a few times if you want. Listen to it a hundred times. Can't hear a difference? Then it doesn't matter.

 

[Blaithin]

I think certain people who claim to have exceptional hearing are just serving their own egos because they want to be seen as "authorities". Certain people cheat on listening tests. And I called certain people out on it once. Certain people refused to answer my questions. Certain people need their own forum so they don't get questioned in a way that they don't want to be questioned. I don't believe what certain people claim, even if they now admit what I called them out on and they refused to admit before.

The way to judge the quality of music is to listen to music the same way you normally listen to it. Use your own equipment in your own living room playing a song. Sit on the couch and listen very carefully. Listen to the song a few times if you want. Listen to it a hundred times. Can't hear a difference? Then it doesn't matter.

While I personally don't care too much about the detectability of lossy compression in music. I still put some trust in the other members and regulars of ASR even if we assume that Amir has his own agenda for doing what he does. I don't feel like you're giving enough credit to those other members for standing up to what ASR is all about when Amir, himself, fails to hold up that standard.

Amir gets dogpiled on a lot, and most often by his forum's most notable members, no less (a lot more than you'd think). This is especially when there's a lapse in the rigor that their particular forum demands. It's happened before, and it's happened far more times than you'd assume just in the last 2 months from his headphone reviews. If you bother to read past the first few pages, it's almost always the older members who end up keeping him in check, and "fighting" him over perceived problems in his methodology. If they verify what he says then I can, at least, be sure there's some merit in it.

You also have to kind of give them credit for not shutting down unwanted discussion. Their moderators aren't trigger happy, and many of the most unfavourable points raised against Amir by relative outsiders (Resolve, Antdroid, Crinacle, Chocomel etc. etc.) are not deleted, but instead are left out in the open, and heavily reinforced by regulars when they have good points to add to the discussion/debate [and they usually do] even if it puts Amir in a bad spot (take the whole MQA thing where his membership really roasted him for his views). That's more freedom of speech than most other audio forums I've ever participated in, and that, in itself, is admirable.

What is shut down quickly in ASR have pretty much been forum mud-slinging threads, and the general climate is incredibly civil even if you don't agree with their perspective. I've spoken with some members and they're not really the hardline objectivists they're broadly generalised as. That label is better reserved for Hydrogen Audio. I've posted quite a few subjective impressions in there, and, as long as I make sure I'm not claiming my experience as fact then they're cool with it, and even welcome it as long as its not off-topic.

ASR is not really an Amir cult so much as it is a mechanism for Amir to 1.) deliver his measurements and 2.) have those measurements challenged with the additional perks of forums in general. The guy gets into hissy fits when challenged, and uses questionable debate tactics. But at least he eventually takes the feedback so he's not completely ossified unlike the founders of other major bastions of audio.

All I want to say is, "Credit given where credit is due."

 

[bigshot]

You can believe in whatever you want. I'm sure there are nice people there. I don't disagree with that. But a certain person who runs the site is a fraud. That is my belief based on my past history with the certain person. It's all in the archive of posts here in Sound Science. I don't bother to follow the hijinx of people I know to be disingenuous. Other people can do that if they want. I don't care. I don't have to listen to people who "prevaricate".

 

[Blaithin]

You can believe in whatever you want. I'm sure there are nice people there. But a certain person who runs the site is a fraud. That is my belief based on my past history with the certain person. It's all in the archive of posts here in Sound Science.

I see. Well, I'm not here to convince you to change your mind about said person. If that is your wish then I won't pursue any more discussion in this particular direction.

 

[bigshot]

No problem. I like you. You're quoting me and telling me about someone I have no respect for. I'm struggling to be as polite as I can and still explain the situation. Does that make sense?

 

[Vamp898]

I don't know if exceptional hearing is related to that. My girlfriend for example does have (due to heavy medication) issues with hearing. Especially with high frequencies (pretty much everything above 8KHz). She can still hear them if they are loud enough, but often when we stroll, i can hear cicadas while she can't.

On of her most Favorite bands is fleet foxes and she have no issues to know which is the 128kbps AAC and what is the FLAC (not 320 AAC though)

So i think exceptional hearing isn't directly linked on how good your brain does the job in separating instruments, concentrating on certain aspects of a song and so on.

> I wonder why it is that whenever I ask the info on killer tracks, it always seems to be a really obscure foreign CD that costs a lot of money to order from overseas

The song i mentioned costs 262 Yen (roughly 2€) in Hi-Res Audio. Digital. Welcome to 2021, no need to order CDs from oversea.

Beside that, Japan does have the biggest music market in the world, so its absolutely no suprise that a lot of non-mainstream music comes from there.

Also the band i mentioned フレデリック (Frederic) does have tens of millions of clicks on youtube, they even are on spotify. Its not that obscure.

Another Song is from amazarashii called スピードと摩擦 (every weeb knows that song, sadly) which can easily picked from the AAC version.

I think its not related to obscure foreign CDs, i think its more that Japanese tend to produce and mix totally different than western people do.

In the west, most is recorded with tubes because "It sounds better (c)(r)(tm)" and so any micro detail is destroyed anyway. Then we tend to mix either extremely warm or extremely sterile.

That doesn't mean that every japanese song is easier to pick apart, but japanese tend to love complex songs with tons of micro sounds and effects and everything.

In some song from さユり there are sound effects in the 12-14KHz region in the background (who are preserved in the AAC btw, i can't pick heir Songs apart from the AAC to FLAC), there she is singing an the drums and guitar and bass plays. Or the 2016 Version of the Album ソルファ from Asian Kung Fu Generation (another very famous band because, afair, some of their songs where used in the Anime Naruto which i haven't watched so can't say for sure).

Even BABYMETAL Songs can (depending on the song) easily be picked apart and you can't say that BABYMETAL is a obscure oversea band.

Just because you live in your western music bubble doesn't mean others are wrong.

Japanese songs tend to be more detail focused and, yes, sometimes overproduced while western songs, in general (not all) are recorded either more warm or mixed sterile dead (especially every genre that has core in its name).

But there are Bands like Alcest from France (especially the first 3 albums) where you can also pretty easy hear the difference from the 320 AAC to the FLAC. It just depends totally on what you listen to.

I think Ecailles de lune and Souvenirs d'un autre monde are one of the prominent examples of western music where it is pretty easy to hear the difference between the AAC and the FLAC, even with the WH-1000XM3 (and that says a lot).

 

[bigshot]

I live in the good old US of A! We have a postal system that is inexpensive and no tariffs, but when I order things from Japan or Europe, the postal shipping is incredibly expensive. Not because of the US side of the mail, it's because of the international mail. I ordered one CD. I'm expecting that the first title mentioned had the most obvious artifacting. I'm not going to order more until I get this one and find out whether it is a killer track or not. I want 16/44.1 I don't want to convert twice to get it to AAC. That just means arguments about dithering and yada yada yada.

I'm not interested in anyone's musical taste. I'm interested in killer samples. If the killer sample doesn't sound bad, I'm not going to listen any further. I've seen the same kind of smoke thrown up in the past. Put forward your best argument and stand behind it.

 

[Vamp898]

I buy music digital/online at Mora (From Sony: https://mora.jp/) or レコチョク (From Rakuten: https://recochoku.jp/)

But スピードと摩擦 is available at Qobuz: https://www.qobuz.com/de-de/album/speed-to-masatsu-amazarashi/uxk60bua5v2sa

I can pick the AAC from the FLAC Version of this song every single time.

I just noticed a difference in Coward from the Interstellar Soundtrack, the part that starts at 07:55. at least i think i can hear a difference in the background with the WM1Z and the IER-M9, but did not do any blind testing.

 

[Blaithin]

No problem. I like you. You're quoting me and telling me about someone I have no respect for. I'm struggling to be as polite as I can and still explain the situation. Does that make sense?

Makes enough sense. I got enough of the undercurrent to know more or less how much. I chose to respect your wishes because it would've been counterproductive, off-topic, and wholly unnecessary to this thread. Anyways, this will be the last post about this just to get this whole thing back on track

 

[bigshot]

i think i can hear a difference in the background with the WM1Z and the IER-M9, but did not do any blind testing.

I'm not at all interested in anecdotal impressions. I am looking for proof. I ordered Aoega COAV. That supposedly has a clear killer track on it with a specific time where it artifacts. Either that is the case or it's BS. Don't give me weeaboo nonsense about Japanese music being more "detailed". Music is music. Artifacting has nothing to do with detail or complexity. It's just a sound that the codec wasn't designed to deal with at that data rate. I'm not going to order a dozen CDs just because someone tells me to. Blackwolf tells me Aoega COAV should prove it to me. I know him. He is a regular here. I trust that. I'm not ordering any more right now. If you want to make suggestions, rip me something lossless 16/44.1 and put it up for me to download and I'll check it. Better yet edit it to the 20 second section of the track where it artifacts. Then I can’t miss it.

 

[Vamp898]

I'm not at all interested in anecdotal impressions. I am looking for proof. I ordered Aoega COAV. That supposedly has a clear killer track on it with a specific time where it artifacts. Either that is the case or it's BS. Don't give me weeaboo nonsense about Japanese music being more "detailed". Music is music. Artifacting has nothing to do with detail or complexity. It's just a sound that the codec wasn't designed to deal with at that data rate. I'm not going to order a dozen CDs just because someone tells me to. Blackwolf tells me Aoega COAV should prove it to me. I know him. He is a regular here. I trust that. I'm not ordering any more right now. If you want to make suggestions, rip me something lossless 16/44.1 and put it up for me to download and I'll check it.

> Artifacting has nothing to do with detail or complexity

Huh? Thats a suprise. Maybe you should tell that the FFMPEG developers, they think different. But what do AAC Encoder developers know about AAC, right?

But it depends on what Version and Encoder of AAC we're talking about, but i mean the general AAC-LC that is available at every streaming service.

Not all AAC Encoders are the same. We first need to define, what AAC Encoder do we use in which versions and with which profile.

Harald Popp, leader of the department that develops the codecs at Frauenhofer IIS says that even though AAC is statistically transparent, it is not. If you're trained you can hear differences in an controlled setup.

That is why they increased the Maximum Bit-Rate to 576kbit/s with xHE-AAC and claim better sound quality at 320kbit/s compared to AAC-LC.

So even the developers of AAC say, that there is a difference and if you're trained, it can be heard and due to this, they are further improving the Codec.

And as said in my initial post, i am no longer able to pick the AAC Version apart from the FLAC when i use the most current AAC at 320kbit/s VBR (which goes up to 576kbit/s). With AAC-LC (320kbit/s CBR) i am.

And its dependend on the setup too. Its easier for me to pick them apart with the IER-M9 (much easier) than with the MDR-Z1R, so even the headphones play a big role in all of this too.

There are just too many factors to be considered to make a general statement which is why i said "With 320kbit/s CBR AAC-LC, i can hear a difference. With 320kbit/s VBR xHE-AAC, i can't"

Also this is the reason why blind tests made by Frauenhofer don't play a role for me too. What Player did they use? What Songs did they play? What Headphones did they use?

In the end, they make money with AAC and want to sell it, they have an interest in telling that it is, according to their statistic, transparent.

> rip me something lossless 16/44.1 and put it up for me to download and I'll check it.

All my music is digital and it would be illegal anyway.

The Track costs 2,99€ at Qobuz in Hi-Res, i'll pay you the 2,99€ and you can buy the Song with my money, is that a deal?

 

[bigshot]

Complexity and detail in music has nothing to do with artifacting. Randomness and sounds that the codec just wasn't designed to deal with does. One of the hardest things to encode, especially at lower data rates, is massed audience applause. White noise too. That isn't at all complex or detailed. It's sustained broadband randomness. But who listens to white noise? Nursery rhymes encode just as easily as Shostakovich. Musical complexity is irrelevant. Codecs are designed to encode all kinds of music.

When someone tells me they have a killer track that they're sure won't encode properly... but they haven't bothered to check it in a blind listening test, it doesn't instill confidence in me. I'm looking for a killer track that will artifact at AAC 256. I haven't found one yet. If I'm going to have to lay out money to buy the track, I would prefer it if the person recommending the killer track has actually checked it to make sure. I've been stung many times by people who claim that lossy is not transparent, and then once it looks like I'm actually going to check what they claim, they start backpedaling and shifting the goal posts (i.e.: you need trained ears / your equipment sucks / you must be deaf / you need to live with the sound for a few months to hear the difference.) I've got no time for that nonsense. And I don't want a hundred different suggestions of tracks that *might* be what I'm looking for. I want something you're sure of. And I don't care how fabulous the music is. Odds are any music that artifacts that badly is going to be pretty difficult to listen to.

Everyone knows that music can artifact with older codecs and less than ideal data rates. Personally, I think that AAC and higher data rates combined with VBR has completely eliminated the problem. If you have evidence that it hasn't, I'm all ears. This thread is called "Testing Audiophile Claims and Myths". That is exactly what I'm doing.

By the way, I am an archivist at a 501(c)(3) non-profit digital archive. I have fair use protection for experiments like this. If you want to send me a one minute clip of a killer track, it isn't illegal.

 

[Vamp898]

Complexity and detail in music has nothing to do with artifacting. Randomness and sounds that the codec just wasn't designed to deal with does. [...] I'm looking for a killer track that will artifact at AAC 256

256kbps CBR AAC-LC or 256kbps VBR xHE-AAC? They use different methods and fail with different things.

Are we talking about the Codec failing and causing artifacts, or failing to keeping information that is able to be perceived (those are different things).

Also to put this in relevance, you have no influence on what Codec in what Version Streaming sites use. Your statement is not generally true. AAC 320kpbs (depending on the settings an the method used), can be statistically transparent. That doesn't mean the AAC 320kbps you download from, for example, mora is transparent to you when you listen to it with your player and your headphones.

AAC-LC does fail with increased complexity. I don't know why you disagree with Frauenhofer (the inventor of AAC) and FFMPEG (who developed their own AAC Codec), the Source Code is OpenSource, you can just look how it works..

With complexity i don't mean how creative a song is, i think thats clear. An complex Piano piece is less likely to fail than just an average boring rock song.

With complexity i, and i think thats obvious by the examples i mentioned, mean the amount of sounds and overlapping going on. The more instruments that play, the better they are recorded to contain more detail and micro detail, the more they overlap and use effects like hall and reverb, the more sounds that go on, the more likely is the codec to fail.

It droppes informations that are assumed to be not perceived by the listener, but this highly depends on lot of factors (not just the listener). As said, i can pick apart songs with the IER-M9, i can not with the MDR-Z1R. So there are lots of factors that can cause if an sound is being able to be perceived.

Not to mention that there are Equalizers that increase the volume of sounds otherwise not being able to be perceived too.

AAC works on an algorithm that drops informations based on assumptions if it is able to be perceived by the listener _and_ at the same times, can artifact doing so. If the attack on a ride changes, there is no artifacting going on.

So if you have some insider information we're all not aware of that proofs, Frauenhofer (again, the inventor and main developer of AAC), FFMPEG and Wikipedia (who explains how AAC works) wrong _please_ share it.

Frauenhofer says, that there can be a difference and due to that, they increased the bit-rate with xHE-AAC and improved the codec further by using different technologies. They even planned to release an lossless codec a few years ago (i remember it was supposed to be called HD-AAC or something like that) but dropped that in the favor of xHE-AAC (and most likely because there are already enough lossless codecs on the market).

By the way, I am an archivist at a 501(c)(3) non-profit digital archive. I have fair use protection for experiments like this. If you want to send me a one minute clip of a killer track, it isn't illegal.

According to the law of which country? The US? We're not living in the same country and beside that, i would need to cut together pieces without knowing what your setup enables to hear what differences. You would need the whole song.

I offered you to pay for the song, i am not going to open Audacity and cut samples together.

Also, and this is a real world test that makes sense, you would have to compare the Hi-Res Audio Version Qobuz offers you and compare it to the AAC Version qobuz offers you. We're talking about real world here, not what could be possible if the world would be different.

Maybe AAC is fine and Spotify/Apple/Qobuz/Mora/レコチョク are messing things up when converting, that could be the case, but then, again, this is real world and if those services are messing things up when converting to AAC, you don't want to listen to the AAC Version.

As said, i can't pick them apart with the MDR-Z1R, maybe you use that headphone and don't hear a difference too. Sometimes there are new things i discover in songs after having listened to them like 100 times over months or even years.

There are even Instruments i can't hear with the MDR-Z1R but i can with the IER-M9 (like the triangle that plays in the background at some parts). Or in other words, when you know where and when it is, you are able to perceive it with the MDR-Z1R, but there are too many distracting things going on that without knowing its there, i would never ever have noticed that and so never would have noticed a difference from an AAC to the FLAC

If there is such a big difference going from one headphone to another, instruments that appear and disappear, there is absolutely no way for you to know if i am able to perceive it based on your own experiments.

Artifacts are easy to pick up because they are defects in the encoding. Changes in the ringing of a snare or the attack of a ride is something completely different.