Sound Quality Vs. Measurements - Page 303 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 3rd March 2012, 02:48 PM   #3021
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
Very well said, T.
  Reply With Quote
Old 3rd March 2012, 02:54 PM   #3022
Previously: Kuei Yang Wang
 
Join Date: Nov 2002
Hi,

Quote:
Originally Posted by dvv View Post
So far, it appears we can't rely on classic distortion measurements because the one with higher THD levels may well sound better nevertheless, assuming that the THD levels on both are below say 0.07% PEAK.
Then it would seem your favourite HK Amplifiers all fail the requirements, as do your treasured Audax Speakers.

So it would seem that even more than 0.07% THD do not cause problems.

Ciao T
  Reply With Quote
Old 3rd March 2012, 03:17 PM   #3023
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
"Maybe you made the perfect Amp?"

Not hardly. It's the 120 I was simulating. LTSpice was showing harmonic distortion only 80dB down right at the output of the voltage source. Model is a perfect source; no internal impedance. So, less than perfect generation is what I was seeing. I bet they have their reasons. I have 16 bit 48K generators unless anyone can lead me to a Freeware generator that does better than that. Although I don't expect to rely on simulation, it is part of learning what can be done and where to look. Same with the noise simulation. I see them as a place to get a hint.

I am still convinced we can measure the differences. It is likely to be a dynamic problem , masking problem, and preferences all rolled up so there is not a number that will always be "better". A set of differences, not the traditional bench tests. I think that is established. What set is the question.

In my limited understanding, BJT's are a bit more linear. JFETS have the advantages you mentioned, plus the ability for the nifty trick John came up with on his input design. Quite clever as symmetry is the goal, not an absolute. Wonder what the schematic of my "new" C5 Nak looks like. I guess it is all FET as it is a Pass design. Seems to work quite well.

I only need good enough to convince my wife to stay in the room when I am playing Harry James or Copland.
  Reply With Quote
Old 3rd March 2012, 03:38 PM   #3024
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by john curl View Post
Be careful, bcarso. Don't think that hard work is exactly the equivalent to sonic successfulness.
No, don't worry, I don't (would that life were so simple!). On his and other meter-reader terms, his designs are successful.

What is interesting, perhaps, is how DS alternates between, in essence, declaring that he's wrapped things up and the rest of us don't need to bother (for example his "blameless" adjective freely applied, almost always to his own stuff [although he granted Stochino kudos for the latter's non-slewing amps]), and then deciding to share some new thing of his that represents, in his opinion, an advance.

Brad
  Reply With Quote
Old 3rd March 2012, 03:58 PM   #3025
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by KSTR View Post
AFAIK, not yet. I really hope they make it, as soon as other researchers will have taken efforts to back up their hypothesis (or disconfirm it, worst case).

-------:------

Looks like they get ignored here, too :-/

Maybe I should quote the very first sentence of chapter 8, Conclusions (emphasis mine):


Personally this statement struck me like a flash of lightning, with regard to its consequences....
Agree, that was stunning to me too. And their logic and experimental support of this seems unassailable, although I need to devote serious study to the piece and not just the enthusiastic but necessarily brief perusal. I'm going to do some poking around among some of my friends (one whose PhD is specifically in psychoacoustics) to get their read, if they can be persuaded to devote the required attention.

Unfortunately, as many know from bitter experience or by observing it, many academics or former academics are as lazy and stubborn as the "rest of us", and a consequence of this is to resist genuinely new information and theories, unless recommended by a respected colleague. As well remember that in academia, to be found wrong is considered to be shameful and career-limiting, despite how many times one has gotten it "right".

Having said that, there's plenty of "crank" stuff out there and it often is sent to academics. The UCLA Astronomy reading room used to get such all the time, as well as other professors there, particularly those with something approaching popular notoriety (George O. Abell for example, who wrote a widely-adopted textbook and cataloged clusters of galaxies and planetary nebulae, thus has an ongoing presence. They often misspelled his name, naturally, a "bad sign" for openers). When I see this stuff (I collected it when people tossed it out for a while) it seems as if it's easy to recognize, and yet saying that is a bit along the lines of the famous statement about not being able to define "pornography", but being sure one knows it when one sees it.

Although this material is not, by any means, as earth-shattering as Einstein's early work, there have been some recent discussions about whether or not, in today's climate of referreed journals, Einstein would have ever gotten published.


Brad

Last edited by bcarso; 3rd March 2012 at 03:59 PM. Reason: spelling
  Reply With Quote
Old 3rd March 2012, 04:17 PM   #3026
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by ThorstenL View Post
[snip]
Plus I would again IN THE STRONGEST POSSIBLE TERMS suggest reading Samuel Groner's commentary on Self's Amplifier book.

Ciao T
I confess I was somewhat apprehensive when Samuel was working on that paper, which began as a review as it were for Linear Audio, but wound up on his website, as I didn't know how well DS would respond to criticism. But we were pleasantly surprised that he was supportive, even appreciative. This is another reason why I mentioned "mellowing".

Brad
  Reply With Quote
Old 3rd March 2012, 05:06 PM   #3027
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by tvrgeek View Post
"Maybe you made the perfect Amp?"

Not hardly. It's the 120 I was simulating. LTSpice was showing harmonic distortion only 80dB down right at the output of the voltage source. Model is a perfect source; no internal impedance. So, less than perfect generation is what I was seeing. [snip]

In my limited understanding, BJT's are a bit more linear. JFETS have the advantages you mentioned, plus the ability for the nifty trick John came up with on his input design. Quite clever as symmetry is the goal, not an absolute. Wonder what the schematic of my "new" C5 Nak looks like. I guess it is all FET as it is a Pass design. Seems to work quite well.

I only need good enough to convince my wife to stay in the room when I am playing Harry James or Copland.
Having the simulator show the generator with distortion only 80 dB down is pretty strange and disturbing, but probably obvious if one knew the details of the software.

Something I've done when Circuitmaker refuses to run fourier analysis is to lay down a series of notch filters (bootstrapped twin-t) at the fundamental, and sometimes as well at low-order harmonics, and observe their outputs (and sometimes merely adding these to the schematic allows fourier to run!).

Broad-brush, without a long list of proper references, and neglecting lots of other effects:

BJTs: relatively high transconductance for a given operating bias current and collector-emitter voltage. Relatively high output impedance at the collector even at rather low collector-emitter voltages. High gm means that local feedback will reduce the distortion of a single common-emitter stage without as large a resistor (compared again to other devices). But that distortion is mostly due to the exponential function that properly describes this transconductance, and which is associated with all manner of distortion. One of the more successful (measurements now folks) strategies is to "predistort" the signal at the base with another nonlinear device, usually another transistor often "diode-connected"; Barrie Gilbert has made something of an industry out of various clever approaches in this connection, even to the point of the now-eponymous Gilbert multiplier.

JFET or DMOS: Often described as square-law devices. Modern low-noise parts depart from the simplest models when the details are considered (see, for a recent brief discussion Scott Wurcer's mic preamp piece in Linear Audio Vol. 1; for more detailed discussions Ed Oxner's book is excellent). Real devices also have some contact resistance, which means that even a common-source stage with no external resistance in the source will have a little local feedback. Were these parts truly square-law alone, and had very high drain impedance in saturation, in principle they would generate second harmonic distortion and sum-and-difference-frequency IM distortion.

FET transconductance for a typical bias current will be lower than for a bipolar (roughly speaking, similar device area yada yada). As a result the strategy of linearization via local feedback will be less effective than for bipolars.

Tubes (!). Well, to the extent a triode conforms to the first-order Langmuir-Child theory, the product of transconductance and plate resistance, "mu", is constant over a useful range of operating currents/voltages. Loading on the plate spoils this, but not too quickly. With ideal L-C triodes and no plate load, triodes are distortionless (but of course we have to have some sort of load).

For all three device types, operation in a long-tailed pair (Brit influence excuse me) will, with perfectly matched devices driven differentially and deploying outputs differentially, cancels even-order distortion. So will "push-pull" operation, except that P and N type devices are never all that well-matched, and tubes that use electrons-only require transformers. Boyk and Sussman seemed quite surprised and delighted that push-pull JFETs were IM-distortionless, but their IM distortion paper uses ideal devices for their modelling.

Brad
  Reply With Quote
Old 3rd March 2012, 05:33 PM   #3028
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Wife's critical ear testing results are in:
She still prefers the RB 951 to the HCA 1200ii playing my Paradigm studio 20's. Even well over 3dB louder, the trumpet blair on track one of The King James Version (Sheffield) is "smoother." When I played my Julian Bream, I thought the nylon strings kept the metallic edge away too. In highly subjective terms, When I compared the DH 120; RB 951 and 1500, I thought the 1500 to be , "clinically open". A level of detail that I thought maybe was not natural. So for our preference, with my current speakers and current CD's, we prefer the Rotel.

Now, on my woofers, the 1200 is VASTLY better than the DH-220 I was running. Slam dunk, Tex Avery eyeball pop, amp going on e-bay.

If I were to guess, the more limited bandwidth of the Rotel's may be masking bad things the tweeters are capable of doing. However, I tried a passive LP filter on the crossover and it did not help. We will repeat the tests again in the distant future after I build some Revealtor or Excel based speakers.

Denon S-Audio CD player, Nak CA-5 preamp for those wondering. No need to make judgement on fractions of a percent matching. Whatever this is, it is a simple good/bad and is not very sensitive to level. Level was enthusiastic and above speech, but not loud.

If someone would like to donate , oh say a Aragon, Krell, or Levinson, I would gladly see if our preference really is in favor of a deficiency, or just in favor of some design choices over others.
  Reply With Quote
Old 3rd March 2012, 05:39 PM   #3029
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Brad, thanks for the generalizations. Not being a experienced designer, just trying to understand what's going on, most helpful for me.

Of course, I am still sitting here wondering what the differences are that sets off my wife's sensitivity.
  Reply With Quote
Old 3rd March 2012, 05:52 PM   #3030
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
Quote:
…Whatever this is, it is a simple good/bad and is not very sensitive to level. Level was enthusiastic and above speech, but not loud.
the psychoacoustic testing literature is very firm on the level matching requirement - at several dB differences Loudness curves change perceived frequency balance, nonlinear hearing effects

Katz implies he can use these perceptual effects to set set levels "by ear" between recording venue, mastering studio monitors to better than ~ 1dB - but he does measure SPL in both for "ground truth" for this claim

but Clark's ABX curves shows that 0.1 dB level matching is required for level alone to have "no discernable" statistically significant perceptual effect

obviously many "clearly audible" differences don't require such strict level matching to identify - but you can't rule out a priori that level differences below 1 dB as being "the real" cause for any unqualified claim of "audible difference"

Last edited by jcx; 3rd March 2012 at 05:58 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Quality Control differences = variations in sound quality? KT Class D 3 4th June 2014 12:02 AM
Sound Card for Measurements Marik Solid State 2 2nd January 2012 08:59 PM
Sound Card Recommendations (For Audio Measurements) dchisholm Equipment & Tools 5 16th July 2011 09:40 AM
How to protect sound card during amp measurements? okapi Everything Else 13 2nd September 2008 03:06 PM
Sound cards - test and measurements jackinnj Everything Else 2 5th July 2003 03:02 PM


New To Site? Need Help?

All times are GMT. The time now is 05:04 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2