opamps THD+N/IMD/SNR measurements don't mean jack IRL, so let it go humm'kay?
Jan 1, 2012 at 12:02 PM Post #46 of 134
Quote:
I'm going to stick my nose in here, and probably where it doesn't belong, but...
 
Why not ask Dr. Kevin Gilmore, Electrical Engineer at Northwestern University 1-847-491-2962 
http://www.head-fi.org/t/197035/who-is-kevin-gilmore the author of the above quoted prose?
 
...instead of going after Lee Perry.  Lee has every right to argue IC as someone posting how ubiquitously wonderful is the OPA627 in XYZ amp.
 
For the record: I own a version of Dr. Gilmore's Dynalo and it is probably one of the "cleanest", "freshest", "purest" amps I own with lots of bottom end "SLAM", the "sweetest", "juiciest" mids and surgically "precise" highs  
bigsmile_face.gif


Because he says stuff like this:
 
"[size=small]The first course of research posits that much distortion in audio signals is actually noise, which may be too low to measure but is still audible."[/size]
 
This is in the context of a NE5532 graph which displays a high order harmonic at -135 dB. I'd love to see a blind test of that. leeperry is a wonderful example of confirmation bias. Whenever he quotes an engineer, it's best to look to another engineer for answers instead.
 
The article does go on to describe the input errors that can affect performance. All of which will create measurable artifacts and would appear in tests of different loads, like the ones Voldemort does.
 
 
Jan 1, 2012 at 12:17 PM Post #47 of 134
Quote:
Quote:
I'm going to stick my nose in here, and probably where it doesn't belong, but...
 
Why not ask Dr. Kevin Gilmore, Electrical Engineer at Northwestern University 1-847-491-2962 
http://www.head-fi.org/t/197035/who-is-kevin-gilmore the author of the above quoted prose?
 
...instead of going after Lee Perry.  Lee has every right to argue IC as someone posting how ubiquitously wonderful is the OPA627 in XYZ amp.
 
For the record: I own a version of Dr. Gilmore's Dynalo and it is probably one of the "cleanest", "freshest", "purest" amps I own with lots of bottom end "SLAM", the "sweetest", "juiciest" mids and surgically "precise" highs  
bigsmile_face.gif


Because he says stuff like this:
 
"[size=small]The first course of research posits that much distortion in audio signals is actually noise, which may be too low to measure but is still audible."[/size]
 
This is in the context of a NE5532 graph which displays a high order harmonic at -135 dB. I'd love to see a blind test of that. leeperry is a wonderful example of confirmation bias. Whenever he quotes an engineer, it's best to look to another engineer for answers instead.
 
The article does go on to describe the input errors that can affect performance. All of which will create measurable artifacts and would appear in tests of different loads, like the ones Voldemort does.
 


I'm not sure where you mean to go with that quote, KG follows this quote by explaining that the measured harmonic structure of the distortion may affect what you hear.
 
PS: That the graphs shows harmonics at -135 dB is irrelevant, there's no reference for the the dB scale, neither are the measuring conditions explicited, the graph is here as a general picture, it would be a mistake to mae something more of it (unless you read the source)
 
 
Jan 1, 2012 at 12:35 PM Post #48 of 134
I swapped out the stock opamps in my Xonar ST for some LME49720's, and what I heard was probably expectation bias. I'm sure I couldn't DBT the difference. Probably why I haven't bothered getting any other opamps.
 
Jan 1, 2012 at 12:44 PM Post #49 of 134


Quote:
I swapped out the stock opamps in my Xonar ST for some LME49720's, and what I heard was probably expectation bias. I'm sure I couldn't DBT the difference. Probably why I haven't bothered getting any other opamps.



Which is why I never changed the ones that are in my STX. I haven't seen any real proof for op amps making a difference when implemented properly. I trust Asus, with all the money they make, to have designed something that works to the best of its abilities. I'd trust them over some guy who DIYs on the weekend. 
 
Jan 1, 2012 at 12:56 PM Post #50 of 134
 
As a Professional Electrical Engineer with approx. 25 years experience let me just say that in the real world folks, this is how Engineers design stuff:
First they design it,
then they prototype it,
then they re-design,
then they prototype again,
etc.
then they release it for build and human consumption.
In the prototype phase it is necessary for the designer(s) to touch, smell, actually USE the thing they are designing, whether it be a car or a fighter jet or a piece of audio equipment. In the case of a fighter jet test pilots will fly it for the designer and tell the designer what to fix, optimize etc. Equipment very rarely works the way a designer want it to in the design and simulation phase. You gotta build it, use it and let others use it and maybe even break it.
 
Would anyone buy a car off a company that had never driven their equipment or had test drivers drive it, they had only done computer simulations and measurements?
No, the designers actually to use the product they are designing it and get someone else to use it.
Measurements aren't everything, but they certainly aren't nothing!
The truth is somewhere in the middle and don't let anyone tell you otherwise.
 
BTW, an Op Amp MUST use feedback to be used as an audio amplifier. A lot of feedback. Period.
 
Disclaimer: I only have 1 years experience actually designing analog signal processing equipment, it was not audio equipment but operated in a bandwidth of approx. 50-7,000 Hz, it was Sonar equipment. I've built audio stuff for my own personal use, and have built some custom stuff for a recording studio and a few musicians.
I have spent most of my career working in power conversion and distribution.
 
Sorry if I sound cranky..............my apologies if I do.
BTW, happy New Years folks!
biggrin.gif

 
Jan 1, 2012 at 1:02 PM Post #51 of 134
Amp measurements are test drives. They use real signals, they aren't just computer models.
 
Cars vs. amps isn't a good analogy. Cars vs. headphones would be. Measurements can determine a car's performance, same with a headphone, but not things like comfort which are subjective. There is no comfort when it comes to amps. It has one job, which is to pass an amplified signal on.
 
In fact, claiming subjective impressions of an amp trump measurements is very much like someone claiming their SUV gets good gas mileage because they like it.
 
Sure would be nice if someone proved that measurements aren't everything, rather than type up rhetoric and flawed analogies.
 
Jan 1, 2012 at 1:20 PM Post #52 of 134
Quote:
Amp measurements are test drives. They use real signals, they aren't just computer models.


Yep.  A large suite of tests like Voldermort does (including measurement and blind listening) is pretty much equivalent to taking a car out to the race track to get a lap time.
 
A dScope and a switchbox is probably cheaper than hiring The Stig too...
 
Jan 1, 2012 at 1:40 PM Post #53 of 134
 
am now too annoyed to care.

 
Oh, that's sad...you'll be sorely missed. But it's always the best that go first as they say.
 
Why are you quoting people off random forums?

 
Random? I can quote a truckload of ppl stating that those crappy 5532/4562/OP275 IC's sound plain bad...anyone with even a tiny interest in opamps knowns that really, it's only sad to read that you didn't get the memo. Most of these chips are called "low cost" by their manufacturers, did you really expect a sub-$1 IC chip to sound as good as a full blown discrete stage? And without having ever heard such a thing at that. Would you call this stubbornness or naivety?
 
I'm going to stick my nose in here, and probably where it doesn't belong, but...
 
Why not ask Dr. Kevin Gilmore, Electrical Engineer at Northwestern University 1-847-491-2962 
http://www.head-fi.org/t/197035/who-is-kevin-gilmore the author of the above quoted prose?
 
...instead of going after Lee Perry.  Lee has every right to argue IC as someone posting how ubiquitously wonderful is the OPA627 in XYZ amp.
 
For the record: I own a version of Dr. Gilmore's Dynalo and it is probably one of the "cleanest", "freshest", "purest" amps I own with lots of bottom end "SLAM", the "sweetest", "juiciest" mids and surgically "precise" highs  
bigsmile_face.gif


Thanks for the support! I was under the impression that only Team Naysay was bored enough to come onboard on the first of January, assuming that the others were too busy enjoying such a beautiful sunday with their family and all.
 
There are designers with irrational beliefs. I think you'll find that the majority of electrical engineers would contest Mr. Gilmore's statement as an absurdity, unless it has been taken out of context. Degree'd designers can still hold irrational beliefs and can manage, sometimes, to design good gear in spite of them.
 
The following link (recommended by NwAvGuy on his blog - thus regrettably allowing the chronically irrational to dismiss it out of hand) is a nice discussion of lots of irrefutable maths that conveniently kills every major NFB myth from a Mr. Bruno Putzeys.
 
http://www.linearaudio.net/userfiles/file/letters/Volume_1_BP.pdf
 


I would personally be dubious about anyone agreeing with that NWA guy.
 
So if I get this right, you personally know better than any design engineer that doesn't believe in the holy power of teh opamp? Am I getting this right? So they're all delusional, is that it? Joke aside, apart from a PhD in Audiophool Debunkism, what exact credentials would allow you to make such blatant statements again? Would you like us to call you "Doctor"? Or "Professor" maybe? Be very careful because madness starts when you're the only one who's right against the world. It can be treated though if you don't wait too long and accept your condition.
 
I swapped out the stock opamps in my Xonar ST for some LME49720's, and what I heard was probably expectation bias. I'm sure I couldn't DBT the difference. Probably why I haven't bothered getting any other opamps.


Well, 49720 is simply 4562 with a new name...did you expect a miracle of some sort? Its manufacturer gives them away for a reason ^^
 
Which is why I never changed the ones that are in my STX. I haven't seen any real proof for op amps making a difference when implemented properly. I trust Asus, with all the money they make, to have designed something that works to the best of its abilities. I'd trust them over some guy who DIYs on the weekend. 


Oh wow that's what I call taking risks baby, Living the life! Manufacturers make a 400% markup and their resellers another 200% on  top, would you really expect them to slap AD797's on the board when it sells for $150? Do the math.
 
Jan 1, 2012 at 1:54 PM Post #54 of 134
 
Originally Posted by Chris J /img/forum/go_quote.gif
 
Measurements aren't everything, but they certainly aren't nothing!


Poor measurements are a bad sign, but good ones don't mean jack...and are certainly not the warrant of anything. The latest Sabre DAC chips include built-in I/V opamps, and the manufacturers using those chips throw killer THD measurements as if they meant anything meaningful. This is the real con, and I rest my case.
 
Jan 1, 2012 at 1:56 PM Post #55 of 134


Quote:
Well, 49720 is simply 4562 with a new name...did you expect a miracle of some sort? Its manufacturer gives them away for a reason ^^



 
So which opamps would you suggest I get that will make a very noticeable difference?
 
Jan 1, 2012 at 2:07 PM Post #56 of 134

Quote:
Originally Posted by Head Injury /img/forum/go_quote.gif
 
Because he says stuff like this:
 
"[size=small]The first course of research posits that much distortion in audio signals is actually noise, which may be too low to measure but is still audible."[/size]
 
This is in the context of a NE5532 graph which displays a high order harmonic at -135 dB. I'd love to see a blind test of that. leeperry is a wonderful example of confirmation bias. Whenever he quotes an engineer, it's best to look to another engineer for answers instead.
 
The article does go on to describe the input errors that can affect performance. All of which will create measurable artifacts and would appear in tests of different loads, like the ones Voldemort does.
 


     Quote:
There are designers with irrational beliefs. I think you'll find that the majority of electrical engineers would contest Mr. Gilmore's statement as an absurdity, unless it has been taken out of context. Degree'd designers can still hold irrational beliefs and can manage, sometimes, to design good gear in spite of them.
 
The following link (recommended by NwAvGuy on his blog - thus regrettably allowing the chronically irrational to dismiss it out of hand) is a nice discussion of lots of irrefutable maths that conveniently kills every major NFB myth from a Mr. Bruno Putzeys.
 
http://www.linearaudio.net/userfiles/file/letters/Volume_1_BP.pdf
 


So why don't you so called "engineers" go after and contest Dr. Gilmore, an obviously respected and vetted PhD at an Ivy League University, instead of us low life solder iron jockeys?
Setup a thread for "the majority of electrical engineers" and have at it -
popcorn.gif

 
 
Jan 1, 2012 at 2:27 PM Post #57 of 134


Quote:
 

Poor measurements are a bad sign, but good ones don't mean jack...and are certainly not the warrant of anything. The latest Sabre DAC chips include built-in I/V opamps, and the manufacturers using those chips throw killer THD measurements as if they meant anything meaningful. This is the real con, and I rest my case.


The worse crime however is for those deploying such uber-chips to 1) Use the chip manufacturers specs not actual measurements and  2) not measure the complete circuit which is often rather worse measuring - there is a manufacturer (popular round here) that does this - and achieves the feat of taking a 32 bit chip and creating an (approximately) 16 bit circuit - inspired engineering at it's best ! - granted analog stages incur more noise but even so losing 1/2 the bits is a tad careless
wink.gif

 
Jan 1, 2012 at 2:34 PM Post #58 of 134
Regardless of the designer, I would strongly contest any statement that:
 
1) Makes reference to unmeasurable degradation mechanisms.
2) Makes generalisations as to the minimal use of negative feedback providing better sound.
 
The statements there may have been taken out of context, I don't know. It's not very contentious to say there is no evidence behind them as they are presented here, but they may have been misrepresented.
 
 
@leeperry: Cheap-ish audio opamps are prettymuch perfect for many applications, such as in DACs and headphone amps. This is hardly a new idea, expressly minted by me to annoy you.
 
Jan 1, 2012 at 3:39 PM Post #59 of 134
leeperry,
 
Not in a million years will endless anecdotal accounts of how certain levels of feedback or specific IC opamps or any other sort of electrical difference sounds will as such be considered evidence.  Not even one iota. 
 
Regarding what one professor considers at a Big Ten (i.e. not Ivy League, for the record - not that it matters but we might as well get ALL of our facts right while we're at it) university thinks sounds better is more than a bit presumptuous considering that he happens to sell what he's hawking to you.  Not to mention on top of that that he is an electrical engineer, not a psycho-acoustics expert.  Not that we are, either, but guess what?  We're not trying to hawk an amplifier design from which we profit.  You know another thing about designing, creating, and selling products?  Marketing decisions trump engineering decisions as often as not.  There's a big market of gullible audiophools out there who eat up zero feedback amplifiers - if you know you can cater to them and profit, why not?
 
I'll bet you'll come back with a snip about how he wouldn't risk his reputation on such an assertion - I'll retort in advance with another reminder that he is an electrical engineer (and a very, very good one, yes), not a pscho-acoustics expert.  His claims as to subjective sound aren't directly related to his field of expertise and thus do not threaten his reputation as such - for example, I'm sure he could design a massive negative feedback design amplifier more than competently if he so desired.
 
But if it's all about an appeal to authority to you, the late Dr. W. Marshall Leach from Georgia Tech would like to share his low-TIM negative feedback amplifier design with you:
http://users.ece.gatech.edu/mleach/papers/lowtim/feb76feb77articles.pdf
 
The esteemed Nelson Pass also wrote a good white paper on feedback and distortion with a relatively neutral point of view (considering that he's designed numerous amplifiers of both types); it's an easy read for the non-EEs in the room:
http://www.passlabs.com/pdfs/articles/distortion_and_feedback.pdf
 
His conclusion on what would sound better? Left at an "I would like to think"...  He's got products to sell, yes, but he knows when he's beyond his area of expertise and his absolute knowledge.
 
Personally, I have no problem with either design, as long as it is properly implemented.  And that's the key.  But obviously that doesn't matter because if something anecdotally sounds bad then it is "evil", right?  :wink:
 
Jan 1, 2012 at 3:56 PM Post #60 of 134
Quote:
I would personally be dubious about anyone agreeing with that NWA guy.


The $2 or so worth of opamps in the amp he designed sound pretty good to me...
 
Seriously though, you need a better argument than, "I hear a difference when I know what I'm listening to".  You appear to be influenced by price and the appearance of complexity.  You keep bringing up price and say it can't be good because its cheap.  Since you're posting this on the internet you have proof right in front of you that what used to be expensive and not very good can become cheap and wonderful as technology advances.  By that string of logic my overclocked i7 must be inferior to the ENIAC regardless of measured performance or the user's experience.  You may ask why the 5532 is still in use after all these years if technology advances but there's a simple answer.  It was already plenty good enough for many applications when it was created and human hearing hasn't got any better since then.  You also seem to be claiming that price directly correlates to quality.  In this case, you're missing the very large economies of scales involved.  If you magically inverted the demand for discrete and opamp based amplification the prices would pretty much switch too.  Since many semiconductors have low marginal costs to produce they are often cheap simply because they are popular, not because they are low quality.
 
You posts also seem to imply that an opamp is is some kind of simple stopgap solution while discrete designs are better because just they're more intricate.  They only look that way if you treat them as some sort of "black box" that operates by magic.  Once again its like comparing the ENIAC to my i7.  A silver square about 1 inch on a side with 1,366 gold pins on the bottom doesn't look like much compared to 17,468 tubes, 7,200 crystal diodes, 1,500 relays, 70,000 resistors, 10,000 capacitors and around 5 million hand-soldered joints until upon closer inspection you discover that the humble looking silver square contains 731 million transistors.  Of course opamps aren't as complicated as CPUs but they do condense circuitry into a single small chip and hide the complexity from the end user.
 
None of this is to say that discrete designs are bad.  In the context of headphones and line level outputs the usual reason to prefer opamps is that they can get you the same performance for less money.  What opamps don't do as well as discrete designs is output large amounts of current.  That's not a very big deal for headphones use since only a select few require that sort of juice.
 
By now you're probably complaining about all sorts of assumptions I've made, especially the one about equivalent performance between opamp and discrete.  "Screw the numbers," you say.  "Use your ears," you say.  Fine.  Maybe measurements are useless.  Go ahead and use your ears, but only use your ears.  Demonstrate that when you have no clues except for the sound that you can tell the difference between two opamps that should measure beyond commonly accepted thresholds of audibility.  If the differences are so night and day that people who can't discern them should give up music then you'll have no problem.  If you back off to say that the differences can't be remembered after a small switching delay then how did you notice them in the first place?  If you're going to say something about blind listening tests being inherently flawed then devise a new test or just give up.
 
Even if blind tests somehow made it impossible to hear such differences even when the differences really were audible in normal circumstances that wouldn't give anyone else a reason to believe you.  The reason is simple.  You would have created an unfalsifiable hypothesis.  No objective test could count as evidence against it.  To an outside observer who hasn't shared your personal experience reality would appear identical whether or not it was true.  It is therefore indistinguishable from an infinite number of fantasies and no one else has any reason to believe it themselves.  You can go on believing it if you want to but you would have no business affirming its truth or telling others to believe it.
 
Please take a look at it from our perspective.  We've never measured these things, we've never heard these things, and the evidence from many other DBTs suggests that no one else has heard them either.  In addition to that we have psychological evidence that shows how people can perceive differences which aren't there and spontaneously and unintentionally create false memories.  All you have is personal testimony which is notoriously unreliable and among the weakest form of evidence.  We all have to base our acceptance of such anecdotes on what we already know about the world.  Given my current understanding of reality you're only a few steps away from claiming to have psychic powers or having been abducted by aliens.  That assessment of your claims will stand until or unless I come across new evidence.
 
I'm not saying you didn't experience what you claim and I'm not saying you're crazy.  I just have a different explanation for your experience.
 

Users who are viewing this thread

Back
Top