|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
70.26.29.245
In Reply to: RE: Measurements vs. Reviewers posted by hahax@verizon.net on August 18, 2015 at 20:38:52
Hi,
You are correct in saying that measurements in the right hands. But the right measurements, and enough of them.
The real key to my article isn't that measurements can't be used at all to predict. Rather, what it says is that you must have sufficient measurements to predict, and that, in my opinion, no magazine produces nearly enough measurements to predict anything with great accuracy, including us. You also have to have the skill to do so, which few have.
Over to Dr. Toole. I have tremendous respect for him. At Harman, they do the necessary number of measurements to accurate predict (full sound power and more), which is incredibly time-consuming. They also have skilled engineers able to properly interpret them. So they've gone the extra mile . . . and then some.
A few charts in a magazine, though, won't tell you a great deal. That's my point.
Doug Schneider
SoundStage!
Follow Ups:
I agree. I was just afraid that audiophiles who mostly believe listening always tops measurements, no matter how good and how many measurements, will take your article that way. In fact I'm sure nothing will change their minds especially when it involves their own systems.
I find that frequency measurements are pretty useful, especially at the low end where rooms have the most effect. And that a top end that rolls off a bit sounds more like live music and not artificially bright. I spent considerable time balancing the position of speakers, listening couch and a small forest of bass traps and ended up with results that sound and measure pretty flat in the bottom four octaves.
Two important criteria for me, however, are not well represented in measurements: coherency and image size:
I discovered as a teenager that I was particularly sensitive to hearing the different directivity patterns of multi-way speakers. Many popular 70s era JBL speakers sound to me like the equivalent of a fun house mirror with multiple ranges of sound having different distribution patterns into the room. Which likely explains why I'm a full range electrostatic freak. I highly value consistent directivity across the entire range.
There are lots of compact monitors that measure great, have pinpoint imaging, etc. that never sound live to me. More like looking through the wrong end of a telescope. Miniature versions of reality. Or that I'm listening to the concert from a considerable distance.
"not well represented in measurements: coherency and image size:"
Well, they aren't well represented by Stereo capture or reproduction either.
Stereo at best only manages localization along the horizontal axis. The data for the rest aren't even captured let alone accurately reproduced. So your pinnae are left just flapping in the breeze or squished against your head if you use headphones.
Look at the trinkets like crystals and tiny resonators that folks stick on the wall to at least have some specular sources off that axis. Apparently even bogus data on those axes beats having none in the quest for live-sounding sound...
Rick
Rick
Stereo at best only manages localization along the horizontal axis
I've listened to many speakers, mostly tall line sources that do create a lifelike image height.
"I've listened to many speakers, mostly tall line sources that do create a lifelike image height."
Me too, including my current ones. BUT I believe that that's due to the characteristics of our pinnae (and head-related transfer Fn. in general) along with the radiation patterns of our speakers rather than any sort of actual vertical positional data in the recording. There is simply no mechanism for the latter in normal stereo recordings. Now mind you that is NOT the case with binaural recordings experienced with headphones.
Try an experiment, generate some high frequency noise by rubbing your thumb and finger together (if you have dry skin like me, or rub a piece of paper if not) with your arm's dangling then sweep your arm along a forward arc to vertical while looking straight foreword. You will notice that the response peaks a little above your eyes (Leastwise mine does.) with near nulls when your arms are hanging down or overhead.
Since our brain believes our ears, it expects HF sound sources to be largely at our eye level since that's the response peak. A line source peaks at it's center as that's the only direction with symmetrical arrival times so that's where the sound "comes from". Soooo you hear HF coming from there regardless of it's direction at capture.
Regards, Rick
The most convicing theory I've heard of height perception from line sources is that the brain is timing the floor bounce. I'm not aware of experimental verification, though. In the case of point sources, the HRTF localizes in conjunction with head movement the sound to the height of the loudspeaker, overriding the recorded height cue, but in the case of a full-height line source the vertical image is in effect at infinity.
There are test signals you can use to test height , it's been sometime now but this was sold on a CD some decades ago, you have to control the reflections and this a Linesource Ribbon with wings does pretty well , you can hear up and across pretty easy.
A Large Panel like an ESL does not work well in this test for obvious reasons.
It seems to me it wold be easy to test with a reverb unit, just set a very short delay. Maybe I'll try it myself if I get a chance and report back.
Years ago, after reading an article somewhere, I tried listening to white noise on headphones while sweeping the notch filter on a distortion analyzer. Sure enough, the sound moved up and down, a consequence of the HRTF. Bet that technique would work with a line source, too -- it doesn't seem to be as idiosyncratic as some HRTF effects.
The LEDR test was developed by EASI , helps To have good vertical response ...
Thanks, fun test. Tried it on my computer speakers (old Monsoons) and the height didn't work very well, but the depth did. Then tried it with earbuds and the height worked well but the depth ended up pulling the sound *behind* my head. Will try it on the big line sources next, but probably not for another week since I have to assemble my new HTPC first.
Have to disagree i do get different recording heights from ( linesource) mine, never experienced such with a point source speaker . A pity Mags wont review linesource speakers anymore, maggies being the exception ..
Regards
Edits: 08/21/15
BUT I believe that that's due to the characteristics of our pinnae (and head-related transfer Fn. in general) along with the radiation patterns of our speakers rather than any sort of actual vertical positional data in the recording.Perhaps you read more into my comments than I intended. Some speakers provide a lifelike image size while mini monitors don't. Even mono images can sound real to me if not miniaturized. :)
I don't find any metric to quantify that (important to me) criteria.
edit: Regarding the number of channels used, I run Polk mini monitors in the HT system. It's Oppo 103 can play MC SACD recordings and I've got a couple of Telarc examples. While they sound nice and spacious, the MC thing cannot compensate for the vertical image compression as far as fooling me that the sound source is live.
Edits: 08/21/15
"Stereo at best only manages localization along the horizontal axis"
Are you saying that stereo can't do depth?
Alan
"Are you saying that stereo can't do depth?"
Damn, that's an interesting question! I don't know if it helps that aspect or not but it shouldn't hurt.
My sense of it is that depth (real depth, not reverb) is mostly conveyed by the direct/echo ratio and the loudness. I suppose that stereo could provide an additional insight to the distance if you knew the mike spacing and source size, say just two Mics and a grand piano.
What's the right answer?
Rick
Listening always, always tops measurements, 100% of the time, no exceptions. That statement is tautological if one's hobby is listening to music.
It doesn't mean measurements aren't useful. I use measurements all the time in shortlisting gear to audition, and to help set up my system, and to treat my listening room. By the final arbiter is always listening.
What if measurements tell you a product has very poor fidelity and you like it? That's perfectly fine for you as an individual but has nothing to do with the concept of reproduction or of reviewing for fidelity rather than sounds good to me.
And yes, since I can't do measurements I mostly rely on my ears to choose equipment. But I do look to printed matter, either on paper or on the web for as much input, especially measurements to help me.
Hypothetically, if I auditioned these W-B Square One speakers and liked them, and then I saw the measurements, I would want to go back for another round of listening with different source material chosen to be more sensitive to the upper midrange response in particular. I would try to listen for artifacts from that 1 KHz on-axis resonance and see if I could detect a general midrange coloration resulting from that "smile" shaped response from 100 Hz to 1 KHz. Maybe it would change my listening impressions, maybe it wouldn't. A lot depends on the in-room response, so I would also want to measure that. For example, the variations in the on-axis response above 1 KHz look like they will be offset by a reverse pattern above and below the horizontal plane, so I'll bet the power response is pretty flat above 1 KHz. It really is a shame that no in-room measurements were made for this speaker. I would particularly like to see a measurement taken in John Marks' room and compared to his reference speakers.
My position is that measurements are tools, not arbiters of sound quality. Regardless of what the measurements say, the only way to really know whether a component is high fidelity or poor fidelity is to listen to it.
Usually, the systems that sound the most real to me are built with components that measure reasonably well, but may not have the most impressive looking measurements in their class.
The world seems to be full of components that ace a typical suite of bench tests but don't sound real and are not high fidelity. This observation seems to be most common with digital electronics, and common with amplifiers, but sometimes with speakers too. It's very common for audiophiles (and some speaker designers) to focus almost exclusively on the near field, on-axis, pseudo-anechoic response. Which is unfortunate since that particular measurement doesn't tell you much about how a speaker will sound in the far field in a typical room.
Not so Dave , no designer worthy would focus on any one set of measurements , practically a full day of measurements is required to get a solid picture , of course with enough experience with a particular design you can then focus on key results because of past knowledge on hand ...
Regards
There seems to be plenty of designers who are shooting for flat on-axis nearfield pseudo-anechoic response as their most basic design goal, and who optimize their design based primarily on that one metric.
Listening and measuring is not mutually exclusive , its actually the only way to do a proper review of anything, ideally i would like to see a second subjective opinion on a product but of course this would require more time and expense and deadlines being what they are maynot be possible.40 yrs of hearing what others think sound great and scratching ones head have me heading to the measurements before reading any reviews , only those whom i have a personal interactive audio relationship with would i put any weight into their opinions and only because you get to know each others biases and Idiosyncrasies over time .
Alot can be gleaned from proper measurements at the very least is amplifier compatibility and possible room interactions ..
Regards
Edits: 08/19/15
I find the arguments that are so strongly against measurements a little odd.
Likewise, I also find those who think there are companies/designers out there that measure, but don't listen, just as odd.
I've been to many audio companies around the world, and the best ones measure and listen -- over and over again. In fact we just produced a SoundStage! InSight episode on another, Monitor Audio. See link below.
Doug Schneider
SoundStage!
Hi, I knew the anti measurement group would chime in on this one. My favorite amps and ones I use are little vacuum tube Heathkits that have far from state of the art distortion measurements. I don't really know when my ears might detect distortions but if you read John Atkinsons findings on the Vinnie Rossi integrated amp the measurements don't matter crowd might cringe. The amp is very far off the stated stats, going in the wrong direction. I'm far from a recording engineer so sometimes the stats are the only thing that I can use as a barometer to shop other than just ones review. That's why my stupid idea of internal debate within the same publication. It hasn't been done yet so why not a try...might make some good reading as well as a good Hifi shoppers tool and that's what the magizines all about. Isn't it?.......sincerely Mark Korda
It depends on the stats you're using. I was given a good explanation why an amp with variable feedback sounded best with no feedback(and thus higher harmonic distortion) and the reason was measureable. You need to choose the right measurements if the device actually sounds good(difficult to prove subjectively) and seems to measure poorly.
I find the arguments against listening odd also. And I can sort of blame a good friend, Gordon Holt, for it. Though I believe it was a misunderstanding of his beliefs. He wasn't against measuring, just against measuring that didn't correlate to sound. We've learned a lot since Stereophile came out in the early 60s and we often can measure to describe performance.
It's always amazed me how much audiophiles are often against measurements. I t appears that as a group they are of very high IQ and are also very well educated. But when it comes to audio they forget all this and revert to acts of faith.
Trusting your hearing is not an act of faith. If you don't trust your hearing how could you ever buy anything? By measurements. Good luck
Alan
Preferably by both hearing and measurements so each side can moderate the other. Too bad we don't get enough measurements. We do get a lot of words which are often even harder to judge than measurements.
I'll just say that with all the companies I've been to, no matter how many measurements get done, I've never found one where listening doesn't happen. It's all part of the process -- and usually an iterative thing.
Doug Schneider
SoundStage!
Of course listening is still needed to see if the measurements actually correlate. I just know there are measurements that do correlate. And I know Floyd Toole believes that also.
Hi,
And Dr. Toole is definitely correct. At Harman, they do measurements that allow them full sound power, as well as directivity at pretty much ever angle. As I said in my article, companies such like that one do volumes of measurements that publications don't.
Doug Schneider
SoundStage!
Doug,
That's the way it's supposed to be. The manufacturers are the ones who do the measurements. Not us. I expect no less from Harman. That's their job. Not our job.
Actually that is part of the reviewer's job, to verify a manufacturer's technical claims. Your excuse is can only mean to me that you have neither the equipment nor the skill in all likelihood to make those measurements. Instead you just give your opinion. So you review an amplifier that claims 500 watts per channel, you review it and tell us how great it sounds, and then when I buy it I find out it only puts out 200 watts per channel. Is that your idea of a review? Well it's not mine.
John Atkinson provides speaker measurements without ever telling us how much distortion a woofer puts out. All that supposed deep bass might just be doubling but he and we'd never know it until we hear it for ourselves. I've yet to see one of your reviews that was critical of any equipment. Everything you publish seems to me like an extension of the manufacturer's advertising. Well they're good for a laugh if nothing else.
Without measurements to back it up a review is a practically useless anecdote.
As you can tell from this long discussion, I'm on the side of measurements. Unfortunately most reviewers don't have access to these tools. But I do believe a review can still be useful especially from a reviewer who has written often. You can get a sense of his biases over a long period and use this to know what he's hearing. And I do believe there have been a few really good subjective reviewers such as Gordon Holt and recently Roy Gregory who you can glean tons of info from.
Something like Dayton Omnimic is easy to use and can be purchased on Amazon for $300. I think every audiophile should be equipped to make in-room acoustic measurements and should use them to help system/room setup and optimization. But I know that few do.
If I were reviewing speakers, I would want to make measurements to help with placement and to get the most out of them.
It's not enough. As Doug Schneider has written in this discussion it needs lots of measurements and someone who knows how to interpret them. Otherwise designing or criticizing speakers would be a snap.
You said most reviewers don't have access to measurement tools, which is nonsense. For acoustic measurements, any reviewer has access to suitable tools.
You don't need an anechoic chamber. A chamber is surely nice, but reviewers like John Atkinson and Martin Colloms are able to make useful measurements without one. And more importantly, most speaker designers manage to do OK without one as well.
I'm not even asking for reviewers to possess the engineering background necessary to interpret a full suite of measurements. If all reviewers would just use basic in-room FR measurements to optimize speaker positioning, and then include an in-room measurement with comparison to their reference speakers, we would be a lot better off.
In his speaker reviews, John Atkinson often includes a spatially averaged in-room response compared to another reference. This one plot is more useful than all his other loudspeaker measurements combined, because it correlates best with his subjective impressions. Also, by comparing and contrasting the measurements of various speakers in the same room, you can separate the speakers' signature from the room's signature. I don't know why every reviewer doesn't do this.
I'd actually like measurements we usually don't see and could be hard to do. In one case I'd like to see the transfer function of each driver and its crossover, that is the frequency response of an isolated driver/crossover combo.
Why so hard on opinion only reviewers, a tier one reviewer opinion trumps all , no need to listen just buy ..
:)
I do find some purely subjective reviews to be useful as long as they include comparisons with other equipment of similar design & price or with references I'm familiar with. When that's lacking as well, it's hard to call it a review.
I agree with you.
Doug Schneider
SoundStage!
Hi Peter,
I guess it depends what you figure your job is.
As a review, measurements are valid. As I explained in my article, they might not tell you everything about the way a product sounds, but they do tell you certain aspects about the product, and they can help you to tell a product has been competently designed or not. I don't take magazines too seriously that don't do them.
Doug Schneider
SoundStage!
Have to agree ....
In an interview Peter Walker once was asked what the QUAD listening room looks like.
He answered that they do not have one and everything is designed by measurements alone although his engineers were allowed to take stuff home if they really wanted to listen to it.
Things may have changed since he passed away and the company was sold to China.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: