|More info here.|
|"Sunrise Sentosa" - March 2019, Singapore|
As discussed before, I do believe (at least here in North America), Stereophile has been the most reasonable when it comes to portraying the audiophile hobby... By "reasonable", I mean that it at least has the courage to include objective measurement results with product reviews, though not always and not with the most contentious products - ever seen a magical audio cable, one of the many ridiculous Synergistic "Research" products, or even MQA fully reviewed or measured?!
The measurement results are often very telling of the actual performance of devices out there. A good example recently is the review of the Naim ND5 XS 2 streamer which while I'm sure sounds very competent, from a performance perspective is lacking compared to many modern devices when judged at the highest technical levels. Have a peek at the jitter performance through the SPDIF TosLink input and the loss of low-level precision with the undithered 16-bit LSB signal. These limitations have already been overcome without great difficulty in devices much less expensive than the Naim's asking price of US$3500. Clearly, the asking price of something in itself is a subjective estimation of what a company believes the market will bear and can be divorced from the quality of its utilitarian high-fidelity playback function.
As expressed previously, we need to remember that these days, objective measurements can be so precise and audio electronics have achieved such a level of performance that if a device "measures bad and sounds good", maybe the reviewer's hearing just isn't good enough to tell a difference. After all these generations of digital equipment, should we not be entertaining the "possibility" that sonic differences have already surpassed human hearing abilities (the acuity of which sadly degrades as we age)? A reviewer these days can probably say essentially any combination of positive comments or minor negatives about a reputable US$300-$30,000 DAC/streamer and it would still be just fine because audible differences might already be at best very minor, if not insignificant.
|"Eternally Mid-20th Century" (except for the cars!), March 2019, Lukang, Taiwan|
Presumably it is with some sadness that Atkinson "sighed" at the feedback he received from his "longtime" copyeditor Mr. Lehnert about soundstaging. It made little sense thus inspiring the need to talk about the basic principle of what we audiophiles call "soundstage" - the perception of sonic localization, and the testing methodology as it applies to hi-fi. This article seems to address the frustration of Mr. Atkinson having to make a decision as Editor between two subjective perspectives (the submitted text from Mr. Rubinson and Mr. Lehnert's review). Ultimately, describing his adjudication of the matter which not surprisingly includes discussion of a controlled "standard" signal like the mono pink noise to check for the ability of a hi-fi system to allow faithful reproduction of a synthetic signal that is intended to be "staged" right at the center between stereo speakers when heard.
Atkinson is right. "Soundstage", as perceived by the mind, is a secondary phenomenon of the technology feeding the auditory system with a reasonable facsimile of a recording through 2 channels. It's a result of the placement of "sound objects" be they voices, instruments, noises as captured by the microphone in whatever configuration, processed by the audio engineer in the studio, and then laid down in the 2-channel carrier whether as physical media or virtual files. Whether the final recording sounds like it has a wide soundstage, allows pin-point placement, sounds diffuse, or is presented as a "wall of sound" was to a large extent decided in the process of the audio recording and production long before you and I got to hear it on our systems.
|"Gardens, Supertrees, and the S.E. Asian Trade Flotilla" from 650 feet, March 2019, Singapore|
While the mind enjoys music and reconstructs the "soundstage" of the performance, all the "machine" is doing is reproducing the source it was fed. If the recording contains old-skool ping-pong effects, can the DAC/(pre)amp/speakers reproduce this with high channel separation? Is the equipment (especially speakers) "fast" enough and have accurate time-domain performance (ideally time-aligned drivers) to extract those subtle differences in correct phase? Is the resolution good enough to accurately reproduce the nuances as well as the dynamic bursts accurately from the two channels? These (and more) can be each assessed objectively and in isolation. Together, when done well enough, the intent is that the mind has a much easier time extracting all that wonderfully recorded "soundstaging", tonality, rhythm, texture, dynamics, etc. We as audiophiles can appreciate this level of achieved fidelity, thus the "value" and joy of owning a good high-fidelity system; willing to part with significant disposable income to obtain these products (hopefully rarely needing to take out loans!).
Going back to the question of clowns and jokers. Do they still exist in audiophilia? (Consider that a rhetorical question! :-)
|"Salt Fields", March 2019, Jingzaijiao, Taiwan|
Over the years on discussion groups and forums, it's interesting that the words "clown" and "joker" have been used to describe some in the audiophile press. Yes, there is inherent humor in watching neurotic grown men obsess over their toys and I certainly don't think audiophiles need to take themselves too seriously. However, there comes a point though where the ongoing presumably unintentional, foolish beliefs or conclusions expressed by some writers draw contempt and ridicule. This is unbecoming of the "professional" regardless of what can be considered "fun". The unintended effect is that when repeatedly unchecked, eventually the magazines and "audiophile hobby" end up as the butt of jokes, and audiophiles are seen as nothing more than fiscally unwise, poorly educated, faith-based, superficial and gullible consumers.
|"Distance", March 2019, Jiufen, Taiwan|
In all that time, I think Stereophile (and Mr. Atkinson) could have taken a stronger leadership role to advance audiophile education, addressed mythical audio products, and made it clear when some manufacturers are obviously selling snake oil. Not only would this clean up the poor actors in this hobby, but also would empower magazine staff in the process to write reality-based opinions and provide reasonable editorial feedback. Indeed, even further promote the importance of objective evaluation as a way to minimize biases.
Some in the Industry might not be happy with a no-nonsense publication willing to call out those of poor repute. Certain companies of certain products will not advertise in such a respected publication, but the best ones will, especially those that provide true technological progress, worthy new features, and can bring value to the consumer. In turn, the consumer will be steered to purchase products worthy of their hard-earned dollars. IMO, this is the only way to build a firm foundation for an Industry based on engineered devices with an honest media aligned with consumer interests.
|"Singapore Flyer", March 2019, Singapore|
Atkinson: Do you see any signs of future vitality in high-end audio?
Holt: Vitality? Don't make me laugh. Audio as a hobby is dying, largely by its own hand. As far as the real world is concerned, high-end audio lost its credibility during the 1980s, when it flatly refused to submit to the kind of basic honesty controls (double-blind testing, for example) that had legitimized every other serious scientific endeavor since Pascal. [This refusal] is a source of endless derisive amusement among rational people and of perpetual embarrassment for me, because I am associated by so many people with the mess my disciples made of spreading my gospel. For the record: I never, ever claimed that measurements don't matter. What I said (and very often, at that) was, they don't always tell the whole story. Not quite the same thing.Looks like Mr. Holt expressed the embarrassment of having clowns and jokers all around and associated with him. Compared to the late 2000's, I suspect if JGH were alive today, he'd probably recognize that thankfully there seems to be more rational audiophiles online these days in the late 2010's.
I wonder what Mr. Atkinson thinks about that response above. It would certainly be fascinating if in the rear-view mirror, and with some distance from running the business side of Stereophile, his assessment could end up not much different from the late JGH in the years ahead.
|"Gao Mei Sunset", March 2019, Taiwan|
Mono Pink Noise
The file was created in 24/96, in total 1.5 minutes long. I've made the sound fade in and out over 20 seconds at the beginning and at the end to make sure there's stability of the focal point at varying amplitudes. I've also included an inverse polarity version where instead of the narrow inter-speaker imaging, one should hear unfocused diffuse sound from your speakers. Peak RMS amplitude around -15dBFS which isn't too loud. Over the years, I have used a variant of this pink noise track along with jitter samples and MP3 recordings at various quality (128kbps, 192kbps, 320kbps) when auditioning sound systems. These kinds of tracks can certainly help show off technical insufficiencies within seconds even if pristine "audiophile-quality" recordings still sound excellent.
I hope you're enjoying the music, art, creativity, design, and culture all around us. Wishing you fantastic results with your high-fidelity system; which in itself holds great beauty at the intersection of science and art... Being "in the middle" is often the best place to be.
great article ...thanks.....ReplyDelete
Since I'm in the comment section of the article you cite, my views on the subject there should be clear, but sometimes more is better ;-). So...
"if a device "measures bad and sounds good", maybe the reviewer's hearing just isn't good enough to tell a difference"
I have to diverge a bit from you here. One cannot presume that measure "bad" = sound "worse". Only controlled listening tests can determine subjectively whether a "bad" or "good" measurement is A) Audible and B) Preferred. I know you are aware certain types/amounts of "distortion" can be perceived as more pleasing, to some, under certain conditions.
There is a certain forum founded by a believer who has always advocated that better measured "performance" is automatically better, regardless of audibility. Same person regards the word "audible" as Kryptonite. I applaud your use of listening tests as the ultimate arbiter. They are!
Regarding stereophonic "accuracy", that is a bit of an oxymoron. Stereo itself is an inaccurate construct http://www.aes.org/e-lib/browse.cfm?elib=9136
Electronic "accuracy" is possible through the signal chain, but once transduced 3 dimensionally by loudspeakers into soundwaves/a soundfield (pardon the pun), "accuracy" becomes a subjective interpretation, not any absolute.
I design loudspeakers. I make no pretense of "accuracy" in the 3d/perceptual sense. "Pleasingly realistic semblance of audio memory"? Perhaps ;-).
"I wonder what Mr. Atkinson thinks about that response above"
Since I've been quoting it for nearly a decade:
Happy April to you also
Hi AJ, thanks for the response.Delete
Yeah, I certainly do not believe that there is an automatic assumption that bad measurements = worse sound. It of course all depends on the final outcome of the sonic chain with the speakers & room being the most important arbiters to the overall experience.
Much of the time here, I'm measuring DACs and the like so for the most part, I think that highly accurate reproduction of the intended signal at least doesn't result in "bad sound". More like I'd like to place emphasis on the inverse: good measurements = good sound. One could subjectively of course make it sound potentially "better" by adding euphonic distortion, etc. Which is why I'd like to emphasize that audiophiles should try for "transparency" first then if they choose, feel free to explore euphonic options like DSP or NOS or vinyl or tubes with distortion... But just don't assume that everyone will like these options or that they're necessarily "better".
As for the term "accuracy", you're right, it's tough when by definition, "stereo" itself can never be an exact "copy" of the sound. For synthetic music and studio mixes, there was never an actual performance in the first place. I think at best, the only accuracy we can hope for is that the system reproduces the source accurately. Not some idealized concept of "does this sound natural?". Some recordings will sound "natural". Some recordings like say the Edgar Moreau I discussed recently:
has the intention of sounding natural but the production itself is suboptimal for this desire.
Thanks for the link!
"I think at best, the only accuracy we can hope for is that the system reproduces the source accurately."Delete
At the risk of sounding pedantic...
Yes, the **electronic** system should pass the "source" without audible degradation. All "accuracy" ends at the speaker terminals, after which transduction to soundwaves occur. There is no reference beginning nor end soundfield to determine "accuracy".
Not yet anyway ;-).
At best, one could attempt a "perceptual" reconstruction of some original (as linked). Even there, subjective verification might prove difficult.
Never mind very few people would desire any such thing. Being somewhat immersed in audiophile culture, JG Holts ruminations are apparent. Folks like whatever sounds good to them, whatever that is. Preference rules. I'm fine with that.
Remember though that even from the perspective of technical accuracy, there are parameters we can measure on the analogue end as well. Speaker measurements can certainly show devices with flatter frequency responses. Step response can show time-aligned drivers. There are of course the "spinorama" measurements for off-axis performance... Then we can also look at measurements from the "sweet spot" and DSP room correction to aim for improvements for the primary listener.
Yeah, ultimately whatever sounds good to any audiophile is fine with me as well... That is just the position one has to take in respect for freedom. Doesn't mean we should ignore or not be educated about what constitutes "high fidelity" of course!
Again, as a speaker designer/AES member, I'm acutely aware of what we can measure...and what that represents, including spinoramas. On/off axis determine perceived spectral tone and timbre. Key word there is perceived.Delete
There is no perceptual "reference" to compare in room measurements to vs "the source".
Please don't get me started on DSP "room" correction ;-). It's speaker correction, based on some room interaction...and subjective taste. The room is unchanged.
Use care when trying to define "fidelity", unless a reference exists.
As far as stereo "soundstage" goes, forget it ;-).
Good write up Arch, and I like the pics of Taiwan.ReplyDelete
As far as stereo playback goes, I'm a bit like you. I aim for transparency. We never can quite reach it, if only due to room acoustics and as AJ Soundfield suggests, speakers will always be a limitation. However, if true Hi Fidelity is what one wants then transparency must be the goal.
Like you say, if one prefers a bit of colour it is best to start with a transparent set up and go from there, whether it is old fashioned tubes and vinyl or modern parametric equalisers and DSPs.
I'm partial to a bit of colour, particularly around the low to mid bass region, but even so I usually resist the temptation to boost the lower registers. The main reason is that the greater the transparency and lack of colour in the playback chain, the closer you hear what the artist/producer/mastering engineer intended. You really do hear a striking difference between albums or even with same albums that may have come from different source tapes or masterings.
While excellently produced recordings will always sound their best on a transparent set up (I'm using the term best in an the objective), it can be a curse on poorly produced recordings as every flaw is easily heard. In that latter case, a bit of added colour or less transparency can take the edge of a bad recording.
Transparency is a double-edged sword. It lets us hear into the recording warts and all without potentially pleasant rose-tinted "colored" glasses.
In fact, because of the above, I have various room-correction DSP settings that I choose to use depending on mood. Some of them based on my favourite speaker frequency responses for example (typically to accentuate the mids for that "warm" sound), and I will use the "accurate" setting when I'm up for a night of "hi-fi" listening :-).
For me, the great thing about learning about this stuff and doing things like room correction DSP is to achieve the power to modify sound beyond just buying more good. We already have very powerful tools to achieve this thanks to the power of computer audio.
A good dialogue. Thanks again Arch, for an Objective perspective. Your advise has saved my lots of $$ and I have avoided Upgradits.ReplyDelete
Good to hear that the information can be useful. As a hobby there are always opportunities to find things to put money into of course! The important thing is to be aware of what likely has value for those $$$ rather than stuff that are IMO just a waste of time...
"Some in the Industry might not be happy with a no-nonsense publication willing to call out those of poor repute."ReplyDelete
It once existed. It was called The Audio Critic. RIP, Peter Aczel.
Yeah, those old issues were excellent.Delete
In its day, was The Audio Critic ever "successful" in the sense of gaining much readership Steven?
One very sad thing for music lovers (and audiophile) is, while the electronics behind our audio systems has gotten to a very high quality level, recording techniques are getting worse and worse with the loudness war obsession, much too close miking techniques and low-end monitoring bookshelves speakers for the final mix. It is incredible to oberve the sound quality differences (from excellent to eurk...) when you run a playlist that is played, let's remind it, from the same audio system.ReplyDelete
So while we now have extremely performant cars, we rarely have high octane gas to run them full speed..!!
Well said DColby.Delete
Modern music and the recording technique for most pop/rock albums remains at a poor level overall even though at least most are not as awful as some in the early 2000's.
Maybe it's a conspiracy... They're purposely releasing crappy dynamic compressed music on digital (CD and hi-res) so people purchase the less compressed vinyl which cannot be replicated. :-(
Dunno whether there's "method to the madness" or just plain silly "loud albums sell!".
I am Robert from Hungary, Europe. You have a great blog, it's a joy to read!
I tend to prefer your attitude about objective measurements. It's a shame that most of today's blogs and magazines are all about subjective sound quality of hi-fi components. I think subjective reports of design philosophies are important but today's marketing-oriented subjective reports are totally useless.
I have a comment for the objectivity-subjectivity topic. If I understand well your philosophy is that high fidelity is all about precise reproduction of the recorded signal. Objective measurement helps deciding between components. I think there is one problem with objective measurements: how do you decide which measurement is important? Decision is easy when a component is clearly better in every way.
But in real world this is rarely the case. Every high fidelity system is full of tradeoffs.
I am a big fan of full range speakers for example. There are big drawbacks of full range systems. There is problem with doppler. Frequency response is rarely ideal. But phase response is usually outstanding. These are nearly ideal point source also. Multi-way systems are better in frequency response but I've never heard any multi-way system which is capable reproducing of the holographic imaging of a good full range speaker. Dog barking on the famous Amused to death album is a good test track. On my full range speaker it is precisely on the left side and totally lively. On multi-way speakers it is also on the left side but strongly blurred. This is not unique example. So which is better? Multi-way or full range? Which is better in reproducing the recorded signal? I think one can not decide based on measurements.
Other example is digital filters. You made great measurements about this. Frequency and noise are outstanding with "sharp" filters. But impulse response is usually poor in this case. Noise and frequency response is poor with "NOS" filters but impulse response is outstanding in this case. Noise and also deviations in impulse response is inaudible based on tests so which is more important? Which is better in reproducing the original signal? I think decision is not easy. I think objective measurement helps of course but only that is not enough even if our goal is precise reproduction.
Thank you for the comment. Very true, there are many objective measurements we can take and out of all of them, not all need to be significant contributors to actual fidelity when we listen. For any one of us, there are of course limits to our hearing and also training in how we interpret what is heard as well.
When it comes to DACs and digital gear in general, I think we pretty well have most things down pat. In fact, I would suggest that if the noise level, frequency response, IMD sweep, THD, and crosstalk results look similar between 2 DACs, I think it would be difficult to hear a difference at all (measured with a good modern ADC like the RME ADI-2). This is of course assuming that the listening level is the same.
As you can see, over the years I have supplemented the basic measurements with impulse responses, and jitter results. I do not believe the impulse response itself is that important (unless it starts affecting frequency response like with the Pono) nor do I have strong beliefs in audibility of jitter as expressed here:
You bring an important point though about speakers. That is where I think the complexity comes together and we need to spend more time considering. The variables are so much more nuanced in that interaction between the "speaker-amplifier complex".
Assuming the amplifier matches well with the speaker, for sure, full range sounds different from multiway from panels from omnidirectionals from whatever-the-heck Bose 901's are :-).
This is why I certainly appreciate the work Harman has done and what John Atkinson does in Stereophile with speaker measurements. Those measurements are alot of work and there is much data needed to grasp a more complete picture of these devices. Even with all that data, there is no guarantee that the speaker-room interactions will result in "better" sound overall.
My sense then is that while the definition of hi-fi based on objective parameters can be relatively easily applied to digital gear, the effect/non-effect of cables, preamps, even amplifiers... Speakers are going to be a different matter that will be highly idiosyncratic even if we can generally capture some objective results like anechoic frequency response, sensitivity, step response (remember, not all coaxial drive speakers end up time coherent such as the Tannoy Churchill back in the day), etc...
BTW, when it comes to speakers I do like what Mitchco did recently in one of his articles comparing the KEF LS50 vs. massive JBLs:
High quality binaural recordings might be an interesting way to help us understand relative differences. Already DSP room correction is a powerful tool in the search for good sound...
Yup, watch'd the recent Youtube video, from Harmon:Delete
I am a 40 year owner of Klipshorns, and the realism of High Dynamic speakers for Live reproduction has been my Gold Standard. Owned a few pair of JBL's of the era before I got the Khorns. Poor rez music and recording stick out like a sore thumb on these speaker types. My bud with a full 800 series B&W/Bryston with three DB1's has a amazing sound, but it makes almost all music Low rez and above sound good, but not realistic/live?? I am a big Fan of Binaual 5.1, AIX Mark Waldrep BlueRays. They have become my Gold Standards.
Thank you for the answer! Sorry for the late answer!
What I wrote about DACs was just an example about that every component in high fidelity have trade-offs. Sometimes the tradeoffs are below hearing threshold fortunately. :-)
I actually tend to agree with you about that good measured DACs sound same. Although hi-fi forums are full of subjective reports about sound of different DACs. I'm usually sceptical with that. Yes I read your measurements about jitter and other digital artifacts. Jitter examples was very interesting.
I've read Mitchco's article about KEF LS50. It was very interesting, thanks for suggestion! It's very interesting to see how room affects the response of the KEF.
"The variables are so much more nuanced in that interaction between the "speaker-amplifier complex"."
Yes but I would add one more component: room-speaker-amplifier complex. I think objective measurements are difficult in speakers partly because performance of speakers is highly affected by the room. You can measure speaker in a measurement environment but performance in a real room will be different. And of course it depends on the speaker placement also.
In the last 2 months I listened to many hifi gear at local shops and even at local manufacturers. I just wanted to know the philosophies. Now I tend to prefer active speaker systems. I actually listened to a KEF LS50W at a local dealer's shop with my own CDs but that did't convince me.
I also appreciate John Atkinson's measurements at Stereophile! It is a highly valuable source not just for speakers but other components too. I appreciate their CD tests with Pierre Verany CD in particular. It clearly reveals that price has nothing to do with objective quality in high fidelity. There is a minimal amount of cost if we want acceptable quality. But above that there are very expensive transports with mediocre tracking and error correcting capabilities and reasonably priced ones with quite good tracking and error-correcting capabilities. I think it is true for all hifi gear.
Binaural recordings are good but I think this technology is incapable to give us information about the soundstege of the stereo system. (Which is also an illusion as you quoted above.) It is good for capture tonal differences though.