It's difficult to read an audiophile guide as an analogue engineer
hackaday.com
hackaday.com
It's not a lack of courage. The subjectivity is the point. Copy/pasting a comment of mine from last week:
Whenever I think of audiophiles, I imagine a couple of older guys hanging out on a weekend. Because they're the kind of dudes who love maximizing and making numbers go up, it needs to have something to do with optimizing performance. Because they're somewhat wealthy and their success is a big part of their identity, it needs to be something expensive.
Then, because they're friends who want to feel useful to each other, it needs to be riddled with inscrutable lore. That way there's always some handy bit of research, some nuggets of wisdom that they can share with each other. That's what this is really about: feeling like they can bond by sharing expertise about some complex hard to optimize problem.
The actual problem itself is essentially arbitrary and subjective. In fact, it's better if it is arbitrary. Because any problem that can actually be solved means reaching an end to the ability to kibbitz and geek out. That's why bringing up objective metrics is such a faux pas in the audiophile world: it kills the game. And the whole point is to play the game together, forever.
eg their review of the Loxjie A30
https://www.audiosciencereview.com/forum/index.php?threads/l...
The site's owner has several hundred thousand dollars worth of measurement gear - Audio Precision analyzer, GRAS headphone rig, Klippel NFS rig, etc.
Isn’t this the point though? Does it actually require that much gear to get measurements that are accurate enough that no human can perceive the difference? With harmonic distortion for instance, we know what is inaudible and yet people keep lauding companies that push ever farther past it.
Does it actually require that much gear to get
measurements that are accurate enough that no
human can perceive the difference?
For speakers, yes, you need some of this sophisticated gear if you want to do objective (as opposed to purely subjective) reviews.
The Klippel NFS is an automated rig that measures a speaker's output from many different angles and uses some nifty math to subtract the room's influence, giving results more reliable than a physical anechoic chamber. That rig costs something like $100K; I forget.
Getting a 360 degree picture of a speaker's sound output is pretty important. The majority of sound we hear indoors is reflected. You are, essentially, hearing the sum of what your speaker is doing at every single angle all at once so if we're going to characterize a speaker's performance in an objective way we need something like that.
Of course, you don't need need a $100K rig. You can measure a speaker in your suburban back yard (or better, an open field) from a few different angles and get a pretty decent approximation of all of that. I've done it quite a few times! All you need is a $60 calibrated mic, a protracter, a tape measure, and some free software. And the willingness to annoy your neighbors with loud tone sweeps. Usually takes me a few hours to measure a speaker; most of the time spent is just dragging things around so I can do 20 minutes of measurements from 5 or 6 different angles.
However, if you're looking to do this with large numbers of speakers and have lots of money like Amir does, the Klippel certainly starts to look attractive.
For measuring DACs and amps, no, I guess you don't need a professionally calibrated AP analyzer. I think you can spend a couple hundred bucks and get pretty close. Maybe less.
For measuring headphones, yeah, I think you need something like the GRAS headphone measurement rigs that are $6K-$13K. Or maybe the $299 one from MiniDSP is just as good. I don't know!
The calibrated microphone is good for correcting some rough low frequency room reflections and frequency response unevenness and measuring spatial response, but not really measuring other distortion and noise. It's only typically -60 to -80 dB SNR. Even though distortion adds, it'll end up in the noise, and cheap ADC plus microphone preamplifier are typically not as high quality as DAC and speaker amplifier either adding its own cruft.
The headphone measurement couplers are more of an art than science for the moment. You can, in fact, make a decent-ish rig for a thousand or so, but MiniDSP one is not it. For plain equalization I'd trust your own ears with sine tones and some bandlimited noise instead. For measurements, it's just poor.
Klippel NFS is something I have personally seen exactly once, and I cannot attest the quality of its measurements (esp. room subtraction) personally. Almost nobody has it.
You also need a crane to hoist the speaker 20 meters into the air, or you have ground reflections to contend with. (You could bury the speaker so its front is flush with th ground for a hlaf-space measurement and do some math with that, of course.)
https://www.minidsp.com/applications/acoustic-measurements/l...
There are limits to that, admittedly. You're still going to get that bounce below a certain frequency (height dependent) so you'll still want to get it off of the ground as high as possible. At 2 meters you can get accurate measurements down to 90hz-ish IIRC. Actual in-room bass response is highly room dependent anyway so that's not totally the end of the world, but of course it's not ideal, and overall it goes to show you how impressive and valuable the Klippel is.
The audio industry is over a century old, and we've poured zillions of dollars into it. We know (on an objective, measurable level) what sounds good.
(Actually SET is a valid topology and can be done well. As can be a pure transformer amplifier, transformerless tube, BJT, FET, current mode in a variety of classes. Even open loop, funnily enough. They have different strengths and sometimes applications.)
The challenge to "mr. objectivist" is always the ABC-HR test. It's similar to ABX (or double blind randomized controlled trial) but you also assign comparative number rating at each pairwise comparison like in MUSHRA. It detects people who cannot rate things or detect a difference but lie about it too. Preferably with a known total garbage device as a low anchor or two to keep the rating scale honest. Needs a digitally controlled relay switch box for amplifiers, a bit of programming, a lot of amplifiers, potentially multiple loads and a lot of time with material that may expose flaws. The test has been used to tune psychoacoustic codecs with great successes. (It's easier with software only.)
It's next to impossible to run one of these unless you're super rich and/or an audio shop.
If you have not done that, you are a number chaser not an objectivist. A Pythagorean.
Amplifiers are often enough roughly audibly different based on their topology, especially when driving weird loads, even if our minimalistic measurements show they supposedly should not be.
And on the other hand, people have made amplifiers that pass the number checks but produce sound with an obvious annoying signature. (And the fun part is, these are either chip amps or sold as random audiophile equipment to people who are supposed to have ears and know what high fidelity means.)
We don't know what to measure and how, besides a few total fail checks.
And this is why people still use tube amplifiers
with design from 70s, just with modern high quality
power supplies and EMF practices, right?
In a nutshell, "what sounds good" to the most people most of the time is accurate signal reproduction, plus some tweaking of the output to conform to a house curve like the Harman target response.
There is some room for personal preference, of course. And of course some kinds of distortion can sound pleasing such as even order harmonic distortion.
If you have not done that, you are a number
chaser not an objectivist. A Pythagorean.
It's like saying you can't believe the moon exists unless you've been there. I've read the controlled research by Toole et al, and while he could certainly be lying, I haven't seen a credible refutation. Haven't ever seen an electron with my own eyes either but there's some pretty compelling research proving they exist.
We don't know what to measure and how, besides a few total fail checks.
Let me turn this around. How do you feel it's possible that we're 100+ years and many generations of audio engineers into this multi-trillion dollar industry and we still don't know what sounds good? According to you we literally haven't even figured out how to measure this stuff? What a laugh.
That would be an absolutely remarkable state of affairs.
How about photographs and video displays? Same thing? We're just spitballin' it there, too? Got a bunch of artistes in the TV factory just sort of tweaking each set till it looks good enough?
Our "distortion metrics" are barely validated garbage that has not been corrected since 60s. (Including THX standard which uses them, a pretty funny joke. It does better at room and speakers.) Hence, they are only useful as a check not a rating.
We do know what sounds good, obviously it is an application of a bass boost knob. /s
What's under the question is what sounds accurate (or accurate enough) and high fidelity. And the metrics we have typically do not answer it in the positive, only in a weak negative. People designing amplifiers typically do not chase these numbers too hard either, since that ends subjectively audibly poorly. We do not happen to have perfect amplifiers, it turns out.
The metrics are not psychoacoustic for one, nor have been correctly validated. Better ones have been proposed but not applied. There are some definitive statements like "distortion 80 dB lower than signal cannot be heard in headphones" but there's no validation of it. (Plus it seems suspicious compared to what we know about psychoacoustic masking and absolute thresholds of hearing. Plus time domain behavior might be audible too in some cases, we just don't have an audio metric for that.)
Some people have taken to plotting harmonic distortion graphs for multiple orders separately already, but it is way rare and still hard to interpret and it still does not differentiate audibly different (and differently annoying) kinds of distortion.
https://www.google.com/search?q=audibility+site%3Aaes.org
https://www.google.com/search?q=preference+site%3Aaes.org
This discussion is over.
If I missed a paper, point it out in a less lazy way. The last one that did something was in 1984 by Edward Cherry with a funky tunable distortion generating ampliifer. Even that was not a panel evaluation.
There are more or less two benchmark quality amplifiers out there (I know of, and happen to be Benchmark) done by chasing the numbers, oscilloscope traces (because metrics are insufficient), a lot of modern error correction maths (90s era) in addition to validation by ear. I gave up trying to decode their feedforward plus feedback topology in detail, it's hard and not the usual example. These cost serious money. Trades some power efficiency too. One could use one of these to evaluate any and new metrics more easily, but nobody does that.
Basically, the story of every Gentoo userbase.
The problem is, for the rest of us, who just want to buy decent audio equipment and don't have lots of disposable wealth, their hobby fills the space with so much noise. Try searching for reviews of a pair of decent bookshelf speakers. Even add the word "budget" to your search. You'll get all these WhatHiFi style mags who are also playing this inscrutable status game trying to convince you that you'll need to spend thousands of dollars on speakers, even budget speakers. And they're experts so of course most people believe them.
The reality is that once you go over a few hundred dollars for small home speakers you'll get massively diminishing returns. I would say basically zero return in fact between 500 and 10,000 dollars. You should spend money on furnishings to make the space sound better instead.
The problem is, for the rest of us, who just want to
buy decent audio equipment and don't have lots of
disposable wealth, their hobby fills the space with
so much noise
While not perfect, and while they don't provide as much objective data as I'd like, I have always sound Wirecutter's coverage of audio equipment to be of a very high caliber. No snake oil.
On the hobbyist side of things Erin's Audio Corner and Audio Science Review have some great data driven reviews.
I would say basically zero return in fact
between 500 and 10,000 dollars
If so, then yeah, I think I'm almost on board with you. My main system is basically a pair of < $500 JBL studio monitors crossed over to a couple of subs.
If not, then bass and total output are definitely going to be lacking with that $500 system. Most $500 small, commercial speakers are going to struggle to hit 40hz with any kind of authority.
Another thing that a lot of speakers get wrong is off-axis performance. Most of the sound we hear is reflected so it helps if speakers are doing the right thing in as wide of a cone as possible. Paying more tends to get you better performance in this regard but it's a rather weak correlation.
Anything beyond that is brand and greed with minimal returns, unless it's a truly exotic design which may not necessarily be better. (Ribbon speakers, electrostatic, isodynamic...)
Definitely. OP was talking about bookshelves. Personally I'd definitely want a sub to go with any bookshelf speaker.
>Most of the sound we hear is reflected so it helps if speakers are doing the right thing in as wide of a cone as possible. Paying more tends to get you better performance in this regard but it's a rather weak correlation.
There's differing opinions/preferences on that, and what you might prefer will also depend on the room size.
What experts generally do agree on is that the dropoff should be smooth with no peaks or resonances off-axis.
The classic examples I'm aware of are Revel vs Kef.
Revel aims for wider dispersion than Kef but other than that, they're pretty similar in terms of specs, both are very linear on-axis with a smooth dropoff.
The number of subwoofers I own has become a running joke between my brother and I, so I don't need to be sold on their value but they definitely add bulk, cost and complexity so I never want newcomers thinking they absolutely need them to have fun in this hobby.
I enjoy a lot of bass-heavy music so I do feel I need one, but if I only had the budget for average speakers and a sub, or great speakers without a sub, I'd go for the great speakers without a sub.
I'd probably go for good speakers with a cheap sub rather than great speakers without a sub though.
This kind of dichotomy often occurs throughout tech product categories: sometimes you want the consumer thing that has optimized its defaults to mid-market taste, other times you want an unbiased, configurable professional tool. Rarely do you want a hybrid of both.
I found out about Audio Pro through some WhatHiFi-esque publication, checked WhatHiFi's reviews and those were raving about Audio Pro's speakers even though they are pretty inexpensive relative to the whole audiophile environment.
Colored speakers will color anything you play through them.
Possibly that means that some genres or particular albums could sound better, but the flipside of that is that others will sound worse.
That's not what I want, since you never know how a particular album might be mixed. I want to have the widest range of "sounding good".
Maybe if I wanted to have multiple speaker setups in the living room for different genres, but otherwise...
If you're not aiming for accuracy, then what are you aiming for?
One way to think of monitors is that they don't bias or color the sound. That's technically true, I suppose.
But another way to think of monitors is that they're designed to be the superset of all possible listening experiences. If there is any flaw in your sound that could show up on some set of speakers, you should be able to perceive that flaw on your monitors too, so that you can fix it.
Of course, if a mix sounds good on some particular sound system, it's nice if they sound good on the monitors too. But that's less of the point. Monitors are intended to provide actionable feedback for audio production decisions. They're trying to shine a bright light on every dark corner of your mix so that nothing is hidden from you that might show up for a listener.
From that perspective, monitors are sort of the worst speakers for casual listening: They're designed to emphasize as many production flaws as possible.
It's also important to think about how the music you're listening to was mixed and mastered. The people who made production choices for the music you listen to are targeting a range of listeners with mostly average speaker setups. If your speakers are consumer-oriented and middle of the road, then they're more likely to be in the sweet spot for how the album was mixed.
For example, if the mix engineer knows that most speakers overly boost the lows to make things more hyped, they might tame those lows in the mix to get a more balanced result. Now if you listen to that mix on "unbiased" monitors, the lows will be weak.
It's sort of a Keyne's beauty contest: speaker purchasers are trying to pick speakers that make recordings sound good, while engineers are trying to make recordings sound good on the speakers they picked. There isn't really a fixed ground truth. An accurate reproduction of a record deliberately engineered for biased speakers won't necessarily sound good.
That's one way to think of it.
I like to think of it as "revealing every production detail". I want to hear the skill of both the musician and the production.
I'm personally very heavily biased toward well-produced music though.
The alternative is masking details, which is not what I want.
I could get that with cheap speakers in a crappy room, or in the car with the engine running.
Great mixing with a great performance on revealing speakers is an amazing experience.
>they might tame those lows in the mix to get a more balanced result
So you buy speakers with extra lows, and then when you change music to something mixed with neutral bass, it's overpowering?
I'd rather have neutral, calibrated speakers and have most music be balanced and then I can EQ the badly mixed music if I want.
It's a moving target, but I'd much rather have a zeroed rifle. Yes I know the analogy falls apart when the target changes distance but it's the first thing I thought of.
I really truly want to objectively know whether the £X,000 I might commit to an "upgrade" is actually an upgrade, even if "objectively" in fact means just "subjectively by a clear majority of a large ish sample".
Obviously any serious purchase needs a personal listening audition first, but I want to know I'm not wasting time by even shortlisting certain candidates.
Otherwise it's - where to even start from literally hundreds of amps and speakers?
That said, it's ok to blow your money on whatever someone wants. We don't need to judge others for doing what they want, it comes off as jealousy.
I agree that you "shouldn't yuck someone else's yum" -- but I don't think it comes off as jealousy. Accusations of jealousy always come off (to me, anyway), as a ridiculous position taken by people who feel defensive but have no defense.
Oh, man, this is so true. I recently bought some audio gear. I wasn't going for "cheap" as much as "good value" -- but it was extremely difficult to figure out what's good and what's not, mostly because the hobbyist crowd (where I would expect to find out the ground truth) generates so much noise.
In the end, I got a decent recommendation here on HN and was very happy with it.
100%.
But the good news is that speakers have never been better or cheaper, and there's more good measurements of speakers out there than ever. For $500 you can get better speakers than you ever could, if you ignore the audiofool noise.
Also the electronics are much better as well. Room calibration makes a huge difference, Dirac and Audyssey are great.
>zero return in fact between 500 and 10,000 dollars
For me, the tipping point is around $2k for speakers.
I aim for measurable accuracy when I'm looking for speakers. I mean, if you don't care about accuracy, then what are you buying?
There are measurable (and audible) differences up until $2k. After that in most cases it's aesthetics, questionable "sound signatures", and ridiculously overpriced and useless stands.
I agree the room should be treated, no matter what speakers you're using.
The less nice version is that it's all marketing and for-profit woo: Buy these gold cables with 20% nitrogen and your MIDIs will sound crisper! Buy these fiber optic cables made with Genuine Himalayan Sea Salt and your bits will be better-formed! Buy my optoelectric isolator or you might as well be hooking your speakers to your turntable with a rusty wire coat hanger!
"Riddled with inscrutable lore" is a prime diagnostic feature of all this; it recurs in pseudoscience of all kinds, and the audiophile experience is pseudoscience-adjacent at least whenever it makes claims about reality. It's motte-and-bailey, to some extent: This sounds "Better", where "Better" is either completely subjective, or the result of objectively untrue and blatantly impossible claims they make about the gear they're selling.
If everyone knows it's just pretend, sure. But false advertising and lying are morally wrong.
In general, yes, I agree with you that false advertising and lying are wrong. I'm also a pretty rational, skeptical person who is "anti-woo".
At the same time... in the past couple of years I've become much more open-minded about the positive psychological value of just choosing to believe some shit in the absence of evidence. It's good to know that you've chosen that belief for non-evidentiary reasons. Maybe put it in a different mental bucket.
But I no longer believe that the "things I believe but can't justify" bucket should be empty. The world is huge and scary and uncertain and awful things can happen without any notice or any ability to control it. I could get a traumatic brain injury. My wife could realize she no longer loves me. My kid could get hit by a car.
I have no evidence to justify a belief that those things won't happen. But I need some level of faith that things are gonna be OK (even if they might not) so that I can function.
If someone is selling kelp and claiming it cures cancer, that's a deep evil because it prevents someone from seeking an actual effective cure to a real material problem.
But if someone has some disposable money and wants to spend it buying the subjective illusory experience of better sound and the identity of a person who is really into audio, I won't judge a company who wants to satisfy that demand.
We all just need some fucking comfort sometimes, you know? And if a $400 HDMI cable gives it to someone, I sure as hell won't judge them for it.
It was a review of shampoo. Basically lumped all those high end shampoos and the common brands you find in the grocery store together.
It basically said to me -- shampoos really don't matter, many get your hair clean and shiny.
EDIT: hmmm... I think it might be September 2000 issue of consumer reports magazine. Can't seem to read it on archive.org.
A bunch of the "shampoos" are combined with conditioner which means they're neither a good shampoo nor a conditioner. (Because those two are contradictory. Shampoo is to remove stuff from hair and skin, conditioner is to deposit.)
It also happens that quite a few shampoos have a lot of aggressive alcohol or phenolic preservatives. Certain people are sensitive to these (most are not).
Some "men" shampoos also have wrong pH, way too acidic, causing skin and hair problems in the long run. Some even with pH of < 3. (Properly should be between 3.5 and 5.5)
And then there is the difference in strength of surfactant (SLS etc.) but that should affect just the dose.
So yes, they're definitely all the same. /s
The actual grift there is using weird oils, fruit additives which do nothing or just make the shampoo worse.
Not that I buy into that conclusion myself.
I buy shampoo (and toothpaste) without SLS.
'pH of dog shampoo'> The actual problem itself is essentially arbitrary and subjective. In fact, it's better if it is arbitrary. Because any problem that can actually be solved means reaching an end to the ability to kibbitz and geek out. That's why bringing up objective metrics is such a faux pas in the audiophile world: it kills the game. And the whole point is to play the game together, forever.
Oh, so it's just like sports or politics. Makes perfect sense in retrospect.
I only read using crystal aligned fonts. I recommend you do too. Available from bitstream, only $30,000 per letter. Ligatures not included.
There are folks (and associated communities) out there with engineering backgrounds and sophisticated industry-standard measurement gear: AudioPrecision analyzers, GRAD headphone measurement rigs, Klippel NFS scanners.
https://www.erinsaudiocorner.com/
https://www.audiosciencereview.com/forum/index.php?reviews/
I realize it's fun to dunk on all of the "crazies" who are tweaking their sound with crystals or whatever, but they are the lunatic fringe.
Ultimately it's not too different from a lot of industries. Of course, folks hawking their products make some ludicrous claims at times. Just like car ads promise you a happy life with your family and dog if you buy a Subaru or Ford or whatever, there are some wild claims about audio equipment at times. However, there's also a very solid bedrock of actual engineering going on that can be understood and appreciated. And in my experience that's how most people do this hobby.
Although, part of me I guess stupidly would like to reclaim "audiophile" as well.
That said, I still have a set of Buchardt A500s that measure anechoically flat down to 25hz. I paid $4500 for these. I also have a set of Shure KSE1200 that I tune to Harman with PEQ using my lifetime subscription to Roon. These were $1650 as an open box.
Aspirational ASR setups using high end studio monitors can push $10k. I believe the owner of ASR has a system exceeding $30k.
None of these purchases are within the realm of normal, they just have backing by measurements of being pretty great… things that matter in the service of verifiably good sound.
It would be wonderful to snatch that term back. The objectivism in the audio space seems to be taking a bigger mindshare as the years roll on.
Assuming you're using digital sources... roughly $500 for a decent 2.0 setup, $1000 for a nice 2.1 setup, and $1500 for a nice 2.2 setup. Less if you have the ability and knowhow to shop the used market, although the used audio market for speakers and subwoofers is kind of a pain because shipping costs effectively limit you to local pickups.
At that point you're comparing pretty favorably to a lot of things. Like a single generation of gaming on a specific console. And audio systems don't really become obsolete. They'll sound good for decades. Until capacitors leak and need to be replaced, same as any electronics.
The $10K system will be a cut above (assuming you've chosen well - it's certainly possible to spend $10K and wind up with something mediocre) but the vague price range I mention above it IMHO where you start hitting diminishing returns.
https://www.nytimes.com/wirecutter/reviews/best-bookshelf-sp...
https://www.nytimes.com/wirecutter/reviews/best-computer-spe...
I'll also be a bit heretical here and suggest that EQ is important. Either in software, or with something like: https://www.nytimes.com/wirecutter/reviews/best-computer-spe...
However... if you're happy with your current speakers, perhaps there's no reason to upgrade! A lot of the choices I've linked to above will come with an increase in size and/or complexity. And I think those Edifiers are pretty well regarded in general.
EQ is absolutely important for the vast majority of systems. But use proper PEQ that has been derived from measurements and don't just wing it by ear. Just as important in most environments is using some sort of room correction (and preferably room treatments).
Go here, click on speakers, sort by price, and follow the preference score (generally around 5+ is an excellent speaker on the low end), then read specific reviews for more detail about their strengths/weaknesses:
https://www.audiosciencereview.com/forum/index.php?pages/Rev...
There are a handful of brands, while not cheap, that provide excellent value as up/down their lines they are nearly infallible, such as Genelec, Revel, and Neumann. There are also some that are usually excellent such as Kef.
The $10K systems I reference really are about just top of the heap and provide excellent value based on raw performance. One may not need that level of perfection and power, but they're objectively that amazing.
Things like the Genelec 8361A: https://www.audiosciencereview.com/forum/index.php?threads/g...
Or the Dutch and Dutch 8C: https://www.audiosciencereview.com/forum/index.php?threads/d...
That said, agreed you can put together nice systems for not much. There are systems for just a few hundred dollars that perform well, at least for nearfield/small room situations.
Audiophile: crackpot with a $2000 bundle of boutique digital audio player and headphone amp hooked up to some head cans worth more than my monthly income riding the subway.
Another example is distortion. You've got guys obsessing over 120db SINAD vs. 110db when CD quality audio only has 96dB of dynamic range, the average room has a noise floor of 30-40dB, and the dynamic range of most recordings is like 30dB at most. I think ASR tends to have a fairly realistic view there; some guys just like chasing "engineering excellence" even if it's not audible and as long as they're realistic that's fine.
(There's at least some small basic for reality there: distortion is cumulative along the audio chain, so if you want X dB of SINAD you need > X dB SINAD at each step of the way)
But, I haven't seen folks obsessing about jitter too much in general. I believe that you've encountered it though!
For DACs, I think SINAD (for those unfamiliar, this is equivalent to THD+N) really does tell something very close to the whole story. Call it 95%, 99%, 99.9%, whatever you will. The remaining N% is stuff we probably don't yet have reliable measurements for. I'm not saying it's magic and unmeasurable stuff, just that we don't have a solid set of measurements to capture it.
For headphone amps, or DAC+amp combos, there are definitely things not well represented by SINAD. I think most of this is related to power delivery (which he does measure) and I think the rest is probably related to slew rate or something thereabouts. But that's just a wild-ass psuedo scientific guess on my part.
Second partially true (as validated by intersubjective tests) guess is that the harmonic structure matters. It is a major component of the coloration, and funnily enough low non-tapering one is discernible from low tapering one despite what SINAD/THD graphs suggest. Amount is discernible too, especially of second and third harmonic, and total of higher harmonics, but we have not established the thresholds. And for lower harmonics, mostly at lower volumes unless it's a lot. Intermodulating injected ultrasound medium to high level garbage (as produced by some poorly implemented power supplies in bad PSRR amps) is sometimes detectable and tests tend to miss it. It also sounds different from just bad SINAD/IMD. Crossover distortion also sounds pretty bad especially at low volumes.
AES researchers (Earl Geddes, Lidia Lee, 2003) tried for a better perceptually weighted harmonic and intermodulation distortion metric in the past. It went nowhere because industry always wins. Even that is only a part of the answer.
And if the amplifier does outright clip, either in current or in voltage, that sounds just rough.
You arent going to hear intermodulation distortion at -110dBfs regardless of how you weight it.
I want zero distortion or as close to it, not euphonic distortion unless I put it in there via equalization or effects or whatever.
Power delivery is a secondary parameter for qualitative analysis. Although for end user it may be actually the main parameter and SINAD is a secondary.
Amir refused to retract or correct the review.
Similarstory with his review of a broken Anthem AVR.
And I agree with you. I remember the ELAC speaker review you're referring to. In fact, a friend of mine had a friendly business relationship with the speaker designer. He spent (wasted) a lot of time doing damage control thanks to that review. And while the speaker designer remained professional and polite in public, I am told that he had some much spicier remarks in private as one would suspect. So I think Amir was quite wrong there.
But I don't think some isolated bad calls mean that ASR is not science-driven. By and large, he gets it right and his contributions have been tremendous. We're talking about a handful of controversial moments... versus literally thousands of quality objective reviews and a real wealth of data.
You can dislike Amir's methods, but you can't deny that he has transformed the audiophile space by just creating a website for his life passion and sharing his thoughts.
I feel a lot better buying hardware that has been thoroughly tested by ASR. I've even sent Amir some gear to be measured, he's the best of the bunch.
I recall an urban legend about an audio engineer who added a "richness" control to a piece of studio equipment that did absolutely nothing but allowed some rich person to feel like they engineered an album.
Over a long enough distance (~30 m), it's better to use ADC <-HDMI-> DAC usually than send line-level signals due to losses involved unless one wishes to spend zillions on the most perfectly engineered and widely-incompatible shielded coaxial cable imaginable.
Another is denial of Nyquist–Shannon. If you can't hear over 8 kHz, then you don't need 256-bit @ 96 kHz sampling. It's more likely a badly-chosen lossy compression codec (or it's parameters), a nonuniform DAC, or some other source of analog signal loss is causing loss of fidelity by the time it reaches ears.
Further is the belief that lossless codecs cannot be replaced compressed lossless or sufficient parameters of some lossy formats.
Obsession with vacuum tubes and/or vinyl. I get the ceremony of vinyl, but it's not going to improve over perfect digital reproduction.
Can't most folks hear frequencies higher than 8 kHz? I don't disagree with your conclusions since I don't care much about audio quality, but I feel like I'm missing something here.
20hz-20khz is the human range, more or less. For most adults it's more like 15khz.
The 44.1khz sampling rate of CD audio was chosen because it can represent frequencies up to 22.05hz (44.1khz / 2) which safely covers the frequency range of human hearing with a bit of headroom for good measure.
Edit: thanks for all these great responses!
If you mean "does a 44khz only output a square (or triangle) wave at 22kHz, the answer is "no."
If you have two alternating values at 44kHz, the output will be a perfect (to the limits of noise &c.) sine wave, since the DAC is band limited.
If this is still confusing to you, I recommend watching https://www.youtube.com/watch?v=FG9jemV1T7I
You can only capture all the information if you sample at twice the rate of the highest frequency of interest.
If it would, it wouldn't be a sine wave anymore, and you could decompose it into its component of phase and amplitude shifted sine waves, which will occupy all kinds of places in the frequency domain, many of them higher than your initial 22 kHz.
That's in fact what we mean when we say "a 22 kHz signal": An ensemble of all of these overlapping waves represented as a single signal. If you're sampling that with an ADC, you're not only sampling the the 22 kHz wave, but you're also sampling it – and that, but only that, can in fact be described with a single bit per ADC readout.
Looking at it from an information theoretical point of view, the Kolgomorov complexity of "a sine wave of 22 kHz and amplitude x" is pretty minimal – just from that sentence, you can perfectly recreate that signal with no sampling whatsoever, and one ADC readout tells you its amplitude and phase. It can carry a bandwidth of 0 Hz.
This reaches the edge of my knowledge, but I think this is basically because the discrete values are 'slewed' (aka integrated aka low-pass filtered) into a continuous signal by the DAC.
See: https://en.wikipedia.org/wiki/Sampling_(signal_processing)
a_i sin(2pitf)
where each a_i is the amount of a given frequency present. If every wave sample you have goes [1,-1,1,-1,1,-1...] then 100% of your wave is accounted for as a sin wave at the sampling rate. There is zero additional frequencies present with unknown amounts. If your wave samples went like [0,1,0,-1,0,1,0,-1...] the only frequency of sound present in the signal is the one at half the sampling rate. The whole wave is accounted for.
On a more physical level, its ok if your sample only sees +1/-1 every other sample, because when that electrical signal turns into pushing a magnet the magnet has to accelerate smoothly in accordance to Newtons laws. In other words, flipping the voltage between +-1 as a 40khz square wave becomes a perfect 20khz sine wave when it reaches the speaker.
The reason for 96kHz or even 192kHz sampling is that your filter can have a much gentler roll-off.
And the studio can master the recording from a high sample rate recording using an enormous FIR filter, because latency and computing power are mostly irrelevant.
High sample rate equipment just allows you to listen to studio intermediate data without having to have it downsampled first.
But there are problems. For example, if there is any non-linearity in the equipment this can easily cause inaudible tones in the ultrasonic range to intermodulate with each other to produce tones that are audible.
Nothing is simple.
Or no?
Play with it yourself - record various notes, on various instruments, do the FFT to break it into the frequency domain, and you'll see quite a few harmonics over the base note. Strip those out, and it sounds muddy.
For those old enough, an analog phone line was typically limited to 300-3300Hz, give or take. It's more than enough to pass clearly understandable human voice, but neither it is the same as being next to someone.
There are also a lot of the high frequencies present in the sort of transients that cymbals and snare drums and such put out - and I'll suggest that most audio compression algorithms mangle them quite badly until you get up into the lossless or "may as well have used lossless" bitrates.
If you lop off everything above 8khz, things sound "muffled", like you've put a thin blanket over the speaker.
No need to believe me, try it yourself in Audacity or any music playing app with EQ controls.
A plucked string in FLAC vs something like 128 or 160kbit MP3 sounds radically different. The same is true of the initial hit of a cymbal - they're just wrong when compressed.
We all know that JPEG is a really good format, but like all lossy formats it struggles with sharp pixel-perfect text. More generally, lossy data compression struggles with those sorts of razor-sharp details.
I don't know why some fight tooth and nail against this in the audio domain.
In practice, yeah, MP3 is usually good enough. An isolated plucked string or cymbal crash is not something you come across too often. And that's why MP3 fares so well in listening tests against uncompressed audio.
But if you really want a full fidelity experience it's not the ultimate choice, any more than a JPEG of the Sistine Chapel is a 1:1 substitute for the real thing.
It really depends on what you're listening to. If you're, say, a fan of high speed bluegrass, it's an awful lot of isolated plucked strings. :)
Yeah I was definitely simplifying in order to keep things under 50,000 words hahaha
The high-speed bluegrass I enjoy has so many string events, so close to each other (in time), that almost none of them are isolated.
There are however some folks who have much higher top ends, 24k to 32k and maybe more. hard to test. They hear bats talking.
There's also things like hyperacusis which add an extra dimension of sensitivity to things others can dismiss completely like dynamic range and even phase.
https://en.wikipedia.org/wiki/Hyperacusis https://en.wikipedia.org/wiki/Hyperacusis
As a millennial, I'm always slightly amused (and maybe a little horrified) to see my peers spend endless time and more than a little money hunting down vinyl releases of the music of our generation. I also listen to a lot of music from earlier generations and have noted that older collectors were and are thrilled to see that music get well-transferred and mastered releases on a digitally perfect and durable medium and generally consider those to be the definitive release.
https://dr.loudness-war.info/?artist=Taylor+Swift&album=Red - one fairly recent album I found in which the vinyl version seems to have quite a bit more dynamic range than the digital or CD versions.
There's also the benefit of vinyl in that a single sale will typically send quite a bit of money actually to the band - vs streaming sending fractions of a penny, if that, per play. One of the best ways to actually support a band you like is to buy the LP, even if you don't have a turntable.
Perhaps the mastering engineer did that intentionally? Surely the vinyl dynamic range doesn't exceed 96db.
> Ugh, this frustrates me. You are just saying that the _mix_ put on vinyl is better than the mix put on CD. That same mix put on CD would be even better due to higher fidelity of the format. Vinyl is a strictly worse format... unfortunately it's also an easier source of better mixes.
Demand better mixes, not vinyl.
I just have no idea how to go about doing that, short of demonstrating that people are willing to pay good money for good mixes on vinyl, and perhaps a few people getting the idea that a good mix on CD might sell too. But this, historically, has been the "SACD" and related versions of an album, which seem to mostly not exist anymore.
Another is denial of Nyquist–Shannon
Over a long enough distance (~30 m), it's better to use
ADC <-HDMI-> DAC usually than send line-level signals
due to losses involved unless one wishes to spend zillions
on the most perfectly engineered and widely-incompatible
shielded coaxial cable imaginable.
https://www.aviom.com/blog/balanced-vs-unbalanced/
Further is the belief that lossless codecs cannot
be replaced compressed lossless or sufficient parameters
of some lossy formats.
Don't get me wrong. High bitrate lossy codecs are almost always good enough for nearly everything. It's what I listen to most of the time.
I get the ceremony of vinyl, but it's not going to
improve over perfect digital reproduction.
The one kink here is that sometimes vinyl releases are mastered better than their digital counterparts; many releases these days have that "brick wall" loudness wars mastering applied that smashes out all of the dynamic range. So sometimes there are vinyl releases that actually do sound better (though it has nothing to do with the format itself)
I'm a bit fuzzy as to what you're arguing here, but I think most people would agree that, at some bitrate, you're going to be effectively lossless with various compressed codes. That bitrate, though, is high enough that you're no longer saving that much over something like FLAC. Disk is cheap enough these days.
I just tossed Dark Side of the Moon into a test. It's all 16 bit, 44.1kHz sample rate (CD quality).
wav: 434MB flac: 249MB mp3 (-V0): 80MB
A factor of 3 savings for MP3 vs lossless, with no generation losses, seems worth doing for audio.
Tubes, the point isn't the "perfect reproduction" over digital, because a fairly cheap digital (relatively speaking) manages acoustically transparent reproduction (as long as the headphones or speakers can keep up). People just like how tubes sound. It's also a bit of a rejection of the digital, consumer tech, "can't build it, can't repair it, don't really own it" world we live in. I've got a Bottlehead kit half assembled on my workbench right now...
Vinyl, similarly, I think is a rejection of that which has become background noise and digital and data-logged. Also, in terms of sound quality, be careful not to compare "The technical capabilities of a CD" with "The technical capabilities of vinyl." Compare albums, as delivered - https://dr.loudness-war.info/ is a useful resource here.
In many cases, the CD mastering of an album is a brick wall compressed bit of rubbish, "optimized" for listening on earbuds on a subway. With both vinyl and the various "high def audio" formats, you're far more likely to find a master that's actually designed for a competent living room system - with the dynamics to match.
https://dr.loudness-war.info/?artist=Boston&album=Boston - look at the difference in dynamic range between some of the older vinyl pressings of the album and some of the newer CD pressings. One may very well find that the vinyl sounds "better" - because it's got dynamic range. The problem is that the capabilities of the CD aren't being used, because someone was told, "Compress it and make it louder."
This is something I ask every audiophile I ever meet. Why not use digital audio at all times and convert to analog as late as possible in the audio pipeline instead lf buying ten thousand dollar cables? No one ever gave me a good answer...
> If you can't hear over 8 kHz, then you don't need 256-bit @ 96 kHz sampling.
> Further is the belief that lossless codecs cannot be replaced compressed lossless or sufficient parameters of some lossy formats.
The added precision is good for archival purposes and audio production though. It's easy to reduce quality later if necessary but the reverse is not true. Artist performances are unique, singular points in time. I think they should be captured with as much precision and quality as possible.
> Obsession with vacuum tubes and/or vinyl. I get the ceremony of vinyl, but it's not going to improve over perfect digital reproduction.
Isn't this due to the fact that older vinyls were mastered differently compared to modern releases? Nothing stops a digital audio recording from sounding just as good but the loudness war in the music industry resulted in masters with less dynamic range. Older music certainly has a different quality to my ears.
If they want to keep the signal pure analog, that is enough of a reason.
Why would they want that? It's possible to perfectly reconstruct the original signal. The sampling theorem states that analog and digital audio are equivalent given a high enough sampling rate.
> If a function x(t) contains no frequencies higher than B hertz, then it can be completely determined from its ordinates at a sequence of points spaced less than 1/2B seconds apart.
I don't see the point of analog signals anywhere in the pipeline other than in a DAC whose output drives the actual speakers.
The honest answer is that absolutely nothing will provide improvement like what you've just experienced. Tripling your headphone budget gets you a number of beautiful, well-engineered cans with sound signatures tailored to your favorite sorts of music, but with improvements on the order of "that midrange has more detail, nice" than "holy crap!"
Woo-selling companies target anyone still convinced at that point there is some next, higher level of audio bliss they can reach with the perfect combination of kit. Cables made from exotic materials, fancy power conditioners, rocks and stickers have glowing "reviews" and marketing pseudoscience about sonic transparency and skin effect. They do nothing and give the audiophile industry a bad name, but they'll be with us until suckers stop paying for them.
Audio magazines were dependent on advertisers, of course, and many of those advertisers were the sorts of companies selling bullcrap gear like magic speaker cables and other such snake oil.
So you had the "journalists" in these magazines writing all this ridiculous crap singing the praises of magic speaker cables or whatever.
The cycle continues today, in a way, with some Youtubers equally dependent on the income from those sorts of sources.
The hobby as a whole has, THANKFULLY, largely moved away from the snake oil though. Lots of solid objectivists out there now. Snake oil salesmen can sway a magazine or an individual with money, but not an entire community. Lots more truth these days.
I also have a subwoofer box project. I modeled the response of my unit with a physics program WinISD. It incorporates properties of the subwoofer and the box to predict sound pressure levels.
WinISD also predicts air velocity from the port, which if too high will produce fart noises. Farting is solved by making the port cross sectional area larger, and making the port channel longer to maintain the resonant frequency of the port.
I have spent about $300 on supplies for the box. I’ll get a membership at a woodworking club to finish building it ($200/year, storage, tools, training). My 4.3cubic foot, 20-21Hz tuned sub box will enhance the response of the lowest audible 10-15Hz compared to the same speaker in a smaller box. It’ll fit nicely in my sedan trunk too.
With respect to measuring response, an O(n) sine sweep can measure basic response. I’ve found many speakers can’t handle combination of bass and midrange, but can handle either or somewhat fine. I’m not sure how to measure this aspect. Maybe diff the original track from measured response somehow.
I've got a real sense of the cello, and a feeling for the audience (little coughing!) and the size of the ISGM's room. Some breathing noises (Professor Carr is putting the work in) but the music is flowing.
In terms of a transfer function this chain must be far from perfect but it sounds ok if a little 'chocolate'.
I think it is all a bit subjective this stuff.
You use signal transformers to isolate inputs / outputs, like these:
https://cinemag.biz/mic_input/mic_input.php
https://cinemag.biz/line_input/line_input.php
They’re directly in the signal path of a preamp, and when overdriven, change the sound.
You can quickly make a reasonable guess at what the amplifier will do (or fail to) even before you build it with a bunch of EE knowledge and a few datasheets.
Take Class D amplifiers, for instance. They're complete and
total trash until you get into the $1000+ market
Here's one with state of the art performance for $800: https://www.audiosciencereview.com/forum/index.php?threads/t...
Lots of ones with totally acceptable performance for $100-$200 from Topping, SMSL, etc. By "totally acceptable" I mean they measure the same as affordable Class AB amps in the same price range. I've got loads of both in my house and have spent many hundreds of hours with each.
One unfortunate reality for Class D amps is that admittedly, the power ratings tend to be wildly inflated. If something is advertised as "100W into 8ohms" you can bet that it's more like 40-50W before distortion rises massively in a hockey stick curve like graph. However if the buyer understands this, they still often deliver very competitive value. After all, any amp sounds bad when pushed into distortion.
Class D, $72 and measurably better than a LOT of amplifiers out there.