|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
142.79.101.139
In Reply to: RE: Yes- Measurements and listening go hand in hand posted by Ross on November 21, 2020 at 17:33:26
Listening can tell us what we hear, and measurements can tell us why we hear what we hear.
why some like THD or SINAD have little to no correlation with what we hear.
Follow Ups:
> > > > why some like THD or SINAD have little to no correlation with what we hear.
Very weak arguments They have a lot to do with what you hear. Raise them up high enough, you'll hear them. You'll be able to correlate them and all that stuff.
The thing is, good engineering has brought the levels down so they don't matter -- but only because the engineering has got them that way. Lower them anymore and it really doesn't matter -- but that doesn't mean they don't matter at all.
Doug Schneider
SoundStage!
They have a lot to do with what you hear. Raise them up high enough, you'll hear them.If you're referring to Electrophonic brown goods gear, sure. Which is what I started with in 1970 as a kid. ;)
The tests performed by the "deceitful sorcerer" , however, have little to do with what you actually hear with high end gear.
Lower them anymore and it really doesn't matter.
Once again, we agree.
What my experience tells me, however, is those who pursue extreme levels on paper compromise overall sound quality in the real world. Switchers and op amp based devices are not yet ready for prime time in the highest quality systems.
Edits: 12/05/20
I think that Stereophile includes measurements more as a way for them to differentiate themselves from their competition than because such measurements are often truly significant. Gives the magazine one more thing to print and thus occupy space.
No not really ...!
Stereophile started doing measurements as away to keep up as it was Standard for the industry back then , this was way before opinionated mags became the norm like today ..
Objective and subjective reviews were standard , Magazine of choice for me was "Audio" with Bascom King ..
With amplfiers you can basically tell everything from a complete set of bench testing , Stereophile while extensive is not complete or consistent in application , some protection of the innocent is necessary ...
:)
Regards
Nt
E-Stat said:
"why some like THD or SINAD have little to no correlation with what we hear."True, while the human perceptual system has a wide tolerance to THD and SINAD, let's not get too extreme and suggest "little to no correlation". As per the blind test results using purposely distorted audio linked below, the overall results based on 55 respondents who felt they could hear a difference, there was a preference for distortion less than 0.3% (-50dB) being "better" sounding. While DACs and most good amplifiers can achieve better than this, many speakers would not be able to claim this level of low distortion (across the whole audible spectrum).
While it's not necessarily the end-all-and-be-all of technical measurements, it's good to keep in mind that low THD/high SINAD does correlate with low nonlinear distortion and in general I think it's nice to aim for <0.1% (-60dB) as a reflection of a "transparent" source device which will likely be well received as being good sounding. Obviously formal blind testing recommended to better understand/verify this.
Obviously this is a generalization. As with most human characteristics distributed over a "normal" curve, some audiophiles might actually prefer more distortion, while others demand even lower distortion.
-------
Archimago's Musings: A 'more objective' audiophile blog.
Edits: 11/26/20 11/26/20
To believe that artificially added distortion bears any resemblance to the far more complex reality of reproducing music.
Enjoy your parlor games. :)
E-Stat said:
"To believe that artificially added distortion bears any resemblance to the far more complex reality of reproducing music.
Enjoy your parlor games. :)"
Not sure if this is a fair assessment, man... Kinda rude in fact. Maybe you can provide some decent intellectual discussion?
Surely you must think at some high level, like say 3% or 5% THD actually is audible, right? If so, then there must be some kind of threshold whereby audiophiles do care about high-enough-fidelity playback?
-------
Archimago's Musings : A 'more objective' audiophile blog.
you never listen to gear. Just look at numbers on paper.
To each his own I guess. :)
nt
evaluate audio gear by listening to high quality recordings?I find that approach provides far more insight to understanding how a DUT fares reproducing live, unamplified music than simplistic numerically based tests. It's easy to score well on THD based tests-just dial in boatloads of corrective feedback! Which to these ears results in compromised real world performance.
Years ago, there was a jerk here (since banned) who referenced a distortion perception test he referred to as "audiophile repellent". He posted his results and assumed no one would accept his challenge. I bettered his ability using the headphone output of a basic laptop and modest Shute IEMs. The THD threshold was in the several percent range IIRC.
Boring recording. Boring test. Been there, done that. Wasted time.
edit: I located that post here . -36 db relates to 2.5%.
Edits: 11/29/20 11/29/20 11/30/20
...on why you believe the described experiment and results are a "parlor game" (deception)?
Such a test proves what it proves but bears little resemblance to actual qualitative factors in music reproduction. Harkens back to the irrelevancy of the THD wars in the 70s. What may look great on paper reduced to a single metric tells you very little - unless your favorite tunes are uncorrelated sine waves!
I have an Ncore amplifier that is highly rated in terms of SINAD but fails horribly at high frequency reproduction and imaging. It has a curious lack of center fill.
"Dare, even today, to extol the virtues of an amplifier as having really low
distortion and some know-it-all will stand up and say "you know measurements
don't say it all; remember the 80's when we were flooded with amps that had 0.00001% distortion and sounded all screechy", bystanders nodding vigorously"
Might be worth reading via the link below.
As well, you might want to ask yourself this: If a product can't perform well on "simple" sine-wave tests, how is going to perform well under supposedly more complex conditions?
Doug Schneider
SoundStage!
Thats some pretty low distortion numbers by 80's standard .. :)
Might be worth reading via the link below.I am quite familiar with Putzeys and his work - along with using one of his Ncore amplifiers for the garage system. Arguably, his designs are the best switchers available today. While it has a wonderful midrange, I find the top end is compromised as compared with the VTLs. But it works great with New Advents and having 300 watts/channel on hand provides nice headroom. Lots of bang for buck but have no interest with it driving the stats in the main system.
As well, you might want to ask yourself this: If a product can't perform well on "simple" sine-wave tests, how is going to perform well under supposedly more complex conditions?
It's not a case of inherently linear design not performing well on sine wave tests. Rather - how many zeroes to the right of the decimal point actually improve the listening experience? Nelson Pass has demonstrated that such an assertion has reality bass-ackwards . Designs using boatloads of cascaded feedback do downright poorly on complex conditions. Here is a visual from the referenced article:
As for me, I'll pass (pun intended!) on introducing complex distortion products for which the ear is more sensitive. Some feedback can be good, but lots is not necessarily better when the objective is determined by what you hear - as opposed to what you see on paper.
Which is likely why I am not a fan of op amp based designs as they have complex topologies and rely upon large amounts (> 40db) of corrective feedback. Have you ever replaced such with discrete FETs and compared the differences? I found them very enlightening with a Music Hall DAC where I replaced three TI chips almost three years ago with Burson FETs.
My home theater/MC system uses Emotiva and Oppo products with TL072 and 5532 op amps, respectively. The result is neutral, but opaque sounding as compared with the Audio Research combination used upstairs which are zero feedback, class A designs using matched JFETs and triodes in the case of the SP20.
Edits: 12/02/20
You might enjoy this book. Yes, it's technical but not loads of math. Dr. Kolinummi has sections describing in engineering terms why steady state distortion measurement is hardly a complete examination of audio performance.
As Nelson observes, it differs from other analyses of distortion by looking at open loop linearity - an area in which switchers and op amps are weak.
They require heroic amounts of correction to behave properly.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: