Head-Fi.org › Forums › Equipment Forums › Sound Science › Why do OpAmps sound different?
New Posts  All Forums:Forum Nav:

Why do OpAmps sound different? - Page 3  

post #31 of 143
Quote:
Originally Posted by leeperry View Post

I was mostly referring to the hearsay of self-proclaimed experts you can read on internet boards...where ppl tend to depict real world experiences as hard "facts". So before quoting someone on his personal perception of audio equipment, hear it for yourself. I'm sure everyone here has already bought audio gear based on raving reviews, and ended up falling flat on his face(for me, it'd be the Pro750 and the Musiland 02US). We do all hear differently, and have different expectations too.


Well, he's mostly talking about measurements which are far more important than anecdotes or even your own subjective personal experience in an area as muddled and prone to cognitive bias as this.  These aren't headphones where differences are obvious enough to smack you in the face.  Much more care than sighted listening is required in order to draw meaningful conclusions.  Here's something related that I just posted in a different thread.

 

Feynman is full of awesome quotes, but this one is especially applicable.

 

"The first principle is that you must not fool yourself, and you are the easiest person to fool."

post #32 of 143
Quote:
Originally Posted by Willakan View Post

Simple solution: Buy no equipment based on the opinion of others, except when objective measurements for the type of product (eg headphones) of the correct nature and applicability are extremely hard to come by (in that case amalgamate as many opinions as humanly possible, weighting by reliability, and observe general trends...bit OTT but hey...)

I stopped reading "reviews" after I conclusively proved that a popular DAC was simultaneously worse and better than another popular DAC.



The only information from DAC or amp reviews not featuring measurements I take seriously are statements about concrete things like build quality, features, pot balance, and what not.

 

I can usually take a headphone review from even a rabid cable lover seriously because headphones sound different enough on their own that they generally don't have to subconsciously invent ones.

post #33 of 143

Originally Posted by Willakan View Post

I gave an example from someone who held the opinion that opamps had sound signatures, who is also well-respected due to his work in designing DIY audio stuff, in a effort to demonstrate that DC specs really, really are not applicable. I assumed you would find this more persuasive than a purely scientific proof.

 

 

You expect me to let someone else's opinion prevail over mine, even though it's a complete opposite of my very own real world experience? The person you just quoted is a DIY enthusiast, I've got a sound engineer diploma and have worked in a recording studio for over a decade.

 

Again, did you compare all those chips yourself? What did *YOU* hear?

 

Did you check the OPA627 datasheet? AC specs are much tighter in the higher rating chip...this always helps blind rollers. Why don't you give it a shot? I presume you to be interested in opamps considering that you're posting in this very thread. Cut/pasting other ppl's opinions and feedback about their own auditory perceptions won't learn your anything at the end of the day tbh.


Originally Posted by Willakan View Post

 

if someone claims to hear differences between two opamps which both measure far into the realm of inaudibility, I call BS

[..]

Also, interpreting TI saying that single-channel opamps offer lower crosstalk and general interaction between channels as an admission that dual opamps sound "coloured and plasticy" is a rather liberal interpretation, to say the least. Additionally, the reason Benchmark used those "inferior" opamps is that as implemented in their product, they felt that they did the job very well - this is reflected in the measurements.

 

Whilst you may view the opinion of people who have not "heard it for themselves" as invalid or of less value, other people who may not share your views on the validity of opinions, especially when the arguments used to justify the opinions in question do not require "hearing it for yourself", may disagree.


Oh indeed, you can have fantastically measuring gear that sounds terrible...or the opposite.

 

There's no excuse to use el-cheapo NE5532 and LM4562 in the $1K Benchmark DAC-1, at the end of the day ppl listen to this DAC with their own ears.

 

Here's the kind of feedback I would expect from this horrid combo: http://ravenda.wordpress.com/2009/08/14/audiogddac19/


Benchmark DAC1 ($995 ~ harsh, flat out boring)

 

I'm well aware of how 5532 and 4562 sound, and that's exactly what I would have to say about them together. Sure, they'll measure amazing...that's fantastic news if you're listening to your music through an oscilloscope \o/

 

To really measure opamps, you'll have to go a lot deeper than SNR/THD/IMD and so...but tbh, I've got nothing to prove...nor do I really care about how to measure the SQ difference between LME49710NA and LT1028ACN8.

 

Opamps are what colors the sound the most, they're really the link between the DAC chip output and your ears: http://forum.rightmark.org/topic.cgi?id=4:504-3


Everybody -in different locations, at different times, without knowing from each other- told the same story, that they found the differences between opamps more important than the differences in dac chips.

Edited by leeperry - 7/6/11 at 12:32pm
post #34 of 143

Originally Posted by maverickronin View Post

 

Well, he's mostly talking about measurements which are far more important than anecdotes or even your own subjective personal experience


OK, perfect! All the opamps measure the same in RMAA so that means that they all sound the same...because we all know that SNR and THD are the be-all/end-all of audio and that RMAA is a high level professional benchmarking tool. Case closed, well done biggrin.gif

 

If someday, you become curious enough to try stuff by yourself(the so called "real world experience")...you might end up being pretty surprised eek.gif


Edited by leeperry - 7/6/11 at 12:40pm
post #35 of 143

I forgot to add that the free version of RMAA goes through DirectSound(which isn't bit-perfect), so good luck w/ that.

post #36 of 143

I think you misunderstand. I am already of the opinion that any differences between the BP and AP version would be completely inaudible based on objective evidence. Tangent's opinion on the subject is not what I based this on. Regarding your incredulity that it would have any bearing on your opinion whatsoever, I was not aware and could not reasonably be expected to be aware of your background in audio.

Secondly, the measurements I referenced were not done with RMAA and are rather comprehensive.

The essential problem remains that my beliefs lead me to conclude that what you perceive to have heard is not of much import, whilst your beliefs lead you to conclude that the fact I have not "heard things for myself" similarly invalidates my argument. 

From my perspective, I'm afraid anecdotal evidence is not going to persuade me to hurl science outta the nearest window. I also respectfully disagree that the Benchmark is something that measures well that is somehow audibly deficient. The Benchmark provably, by objective criteria, does absolutely nothing that is within a thousand miles of audibility to the sound. This has been verified by multiple objective sources, not including Benchmark's own measurements (as a company that sells mastering equipment it is somewhat unlikely that these are inaccurate). If you think something sounds harsh/dull through the Benchmark, there are various possibilities:

 

1. The recording is harsh/dull.

2. You prefer a degree of colouration in your sound. Nothing wrong with that.

3. You are horribly biased towards the Benchmark based on what you have read about it and thus heard what you expected. It would be churlish not to give you the benefit of the doubt in this regard, however.

 

I will not hear a word against what is to all intents and purposes perfect. And no, I haven't heard itbiggrin.gif

 

If I want more base, or to recess the treble, or to compensate for some deficiency of a transducer, I equalise or apply other forms of digital signal processing, from an objectively neutral and well-measuring base.


Edited by Willakan - 7/6/11 at 2:18pm
post #37 of 143

...

Edit: taking back what I was going to say


Edited by Mad Max - 7/6/11 at 2:48pm
post #38 of 143

Originally Posted by Willakan View Post

 

The Benchmark provably, by objective criteria, does absolutely nothing that is within a thousand miles of audibility to the sound. This has been verified by multiple objective sources


So you're back to the idea that all the DAC's sound the same and couldn't be DBT'ed I presume? Again and again, which ones have you heard? I'm sorry to keep asking the same question, but it would be like discussing about pastel colors on a forum aimed at colorblind ppl.


Originally Posted by Mad Max View Post

...

 

Don't be shy, we're amongst friend biggrin.gif


Edited by leeperry - 7/6/11 at 2:45pm
post #39 of 143
Quote:
Originally Posted by leeperry View Post

OK, perfect! All the opamps measure the same in RMAA so that means that they all sound the same...because we all know that SNR and THD are the be-all/end-all of audio and that RMAA is a high level professional benchmarking tool. Case closed, well done biggrin.gif

 

If someday, you become curious enough to try stuff by yourself(the so called "real world experience")...you might end up being pretty surprised eek.gif


I'm talking about stuff that kicks RMAA to the curb like a $2 whore.  The kind of stuff that real, qualified engineers use to design audio circuits.

 

I've tried swapping opamps in stuff and I've never heard any plainly audible changes.  There are changes I think I might have just barely heard but with how long it takes to swap them its not even a big enough change for me to be sure its actually real.  I question my senses.  I know that they are fallible and I know how easily my perceptions can be influenced.

 

Also since you seem to think bit perfect is so important, I saw an article which showed that the entire process of feeding a digital signal to a Benchmark DAC1, converting it to analog, feeding the analog to a Benchmark ADC1, and turning it back to digital was bit perfect.  I'd link it again if I had it.

post #40 of 143

Originally Posted by maverickronin View Post

 

I'm talking about stuff that kicks RMAA to the curb like a $2 whore.  The kind of stuff that real, qualified engineers use to design audio circuits.


great! where can we find opamps measurements using those killer devices?


Originally Posted by maverickronin View Post

I've tried swapping opamps in stuff and I've never heard any plainly audible changes.  There are changes I think I might have just barely heard


Could you be more specific? Which device? Which opamps? using what amp/headphones? And what source? I've got a hard time imagining that an audiophile couldn't DBT a LT1028 from an AD797 or an OPA827. If he really can't, he should roll them once a day in order to get a good sense of the changes.
 

Originally Posted by maverickronin View Post

since you seem to think bit perfect is so important, I saw an article which showed that the entire process of feeding a digital signal to a Benchmark DAC1, converting it to analog, feeding the analog to a Benchmark ADC1, and turning it back to digital was bit perfect.  I'd link it again if I had it.


bit-perfect is important for measurements, coz getting your RMAA input skewed by DirectSound kinda kills the whole point of taking measurements in the first place...just like they say on your first link "bad testing is worse than no testing".

 

post #41 of 143
Quote:
Originally Posted by leeperry View Post


great! where can we find opamps measurements using those killer devices?


Could you be more specific? Which device? Which opamps? using what amp/headphones? And what source? I've got a hard time imagining that an audiophile couldn't DBT a LT1028 from an AD797 or an OPA827. If he really can't, he should roll them once a day in order to get a good sense of the changes.


Not many places, since they're pretty damned expensive.  I'm saying that's the kind of thing that companies that know their stuff like Benchmark and Anedio and who design for transparency use.  If they used cheaper opamps that's likely because more expensive ones didn't measure much or any better in that circuit.  Of course if you're aiming for "musciality" or something them measurements don't matter nearly as much

 

Nwavguy said he's working on an article about opamp rolling article though.  Might end up being a while though, he's mentioned a few other difficult and long sounded articles he wanted to do as well..  Recently he's been putting up an article every Wednesday.

 

Also I don't really feel like being much more specific since I can never keep all those damn numbers straight.  I'm not actually sure what opamps I have in my Maverick D1 right now.  My profile may no longer be accurate.  My XM6 has 2xAD8065 and 2xBUF634 but I didn't notice much of anything between the ADs and the cheaper TI opams that were stock.  Headphones are what's in my profile and list of former stuff.  I got/get the most use out of my SE530s, modded T50RPs, and HD650s until I got the T50RPs.

post #42 of 143

TBH, I don't really see the Maverick D1 as a good choice for opamp rolling...reasons?

 

-The maker is using some old entry level DAC chip, and he's so ashamed of it that he's burning the DAC chip package w/ acid: http://cdn.head-fi.org/3/31/316e09e3_007.JPG

 

-It's been speculated that the DAC chip is CS4334(circa 1998): http://www.head-fi.org/forum/thread/446764/maverick-audio-dac-amp/720#post_6333834

 

-This chip does support 24bit, but barely reaches 16bit SNR(96dB): http://cirrus.com/en/pubs/proDatasheet/CS4334-35-38-39_F5.pdf

 

And that's the DAC chip SNR, which will eventually get worse when getting through the DAC board..also this DAC is using a tube at the last stage....so a noisy entry level DAC chip + a tube stage + a cheapo volume pot are all deal breakers for proper opamp rolling tbh.

 

Well, rolling an OPA2134 for an OPA2604 would be pretty much impossible to DBT...if you don't recall what chips you tried, we can safely assume that you haven't tried the top of the crop when it comes to audio opamps. Yet, you seem so convinced that all opamps will always sound the same in any given situation.


Edited by leeperry - 7/6/11 at 4:38pm
post #43 of 143
Quote:
Originally Posted by D. Lundberg View Post



 

The crosstalk measurment only tells you how strongly the channels are separated. If the channel separation is poor they will start to "bleed" into each other, and if it gets really poor you'll be listening to mono instead of stereo (or you may get stereo widening).

In this case the crosstalk is nowhere near where it starts to be audible, but I assume the measurments are for the line outputs and not the headphone output?

The headphone output will probably tell a completely different story (at least if it's poorly designed) when it comes to crosstalk, frequency response and distortion.

 

http://en.wikipedia.org/wiki/Crosstalk_measurement

 


The crosstalk measurement (yes, I know what crosstalk is) matches what I heard and I know the design wasn't intended to perform in a completely linear manner. So unless there is another measurement which gives a similar V-shaped frequency response matching what I heard, then, as far as I'm concerned, the effect of it is audible.

 



Quote:
Originally Posted by Willakan View Post

 

@Currawong - If you look at the SG-Acoustics measurements, you'll see that the real reason for the strange sound of the Moon discrete opamp is absolutely atrocious distortion measurements - from "high to very high" in the words of the writer. This distortion, aside from being gigantic, also varies with frequency. -70db vs -80db of crosstalk is extremely unlikely to be the factor causing the sound signature you commented on IMO.

 

Doesn't exactly fill you with faith regarding AudioGd's engineering prowess, looking at their opamp measurements. A Burson opamp (unsure if the same model as used in their recent products) is also tested - the author proclaims surprise after noting it appears to be identical to the AudioGD Earth discrete opamp, which boasts poor linearity performance "with every respect" and similar distortion as the other AudioGd opamp.

 

We aren't talking about how good or bad these things are. The question was: Why do OPAMPs sound different? That question is, essentially: Why do amplification circuits sound different? I provided an answer, with measurements as an example. Aside from that, the owner of Audio-gd, as well as other people I have consulted when I want to understand what is going on in audio gear, are highly experienced engineers. Unless you have similar qualifications, I think I'd take their explanations over yours. 

 

It's clear a number of people in this forum have the attitude one the one hand that something has to be proven to be true and everything else is placebo. When it is proven, but goes against whatever argument they have, it's "inaudible" or something else. I see little to no interest from people here in learning anything, especially not genuine science (discovering the truth), not even learning how audio equipment works, just an interest in bashing beliefs over other people's heads.  

 

post #44 of 143
Quote:
Originally Posted by Currawong View Post

 

It's clear a number of people in this forum have the attitude one the one hand that something has to be proven to be true and everything else is placebo. When it is proven, but goes against whatever argument they have, it's "inaudible" or something else. I see little to no interest from people here in learning anything, especially not genuine science (discovering the truth), not even learning how audio equipment works, just an interest in bashing beliefs over other people's heads.  


x2

 

Scientific inquiry is more a journey than a final destination.  For every 20-30 posts in the 'science' forum that isn't a rehash of the same continuous argument I might hope to gain a singular insight.  Back to stealth mode.

 

ph34r.gif

 

post #45 of 143
Quote:
Originally Posted by leeperry View Post

TBH, I don't really see the Maverick D1 as a good choice for opamp rolling...reasons?

 

-The maker is using some old entry level DAC chip, and he's so ashamed of it that he's burning the DAC chip package w/ acid: http://cdn.head-fi.org/3/31/316e09e3_007.JPG

 

-It's been speculated that the DAC chip is CS4334(circa 1998): http://www.head-fi.org/forum/thread/446764/maverick-audio-dac-amp/720#post_6333834

 

-This chip does support 24bit, but barely reaches 16bit SNR(96dB): http://cirrus.com/en/pubs/proDatasheet/CS4334-35-38-39_F5.pdf

 

And that's the DAC chip SNR, which will eventually get worse when getting through the DAC board..also this DAC is using a tube at the last stage....so a noisy entry level DAC chip + a tube stage + a cheapo volume pot are all deal breakers for proper opamp rolling tbh.

 

Well, rolling an OPA2134 for an OPA2604 would be pretty much impossible to DBT...if you don't recall what chips you tried, we can safely assume that you haven't tried the top of the crop when it comes to audio opamps. Yet, you seem so convinced that all opamps will always sound the same in any given situation.

 


I mostly use its own head amp and I have changed the chips in both the sockets with 2-3 others that were recommended in the D1 thread.  The XM6 has a much fancier DAC but I'm not really sure I hear much of a difference either way.

 

Where did I say I was "certain" that they always "sound the same in any given situation"?  The situation matters a lot.  There are plenty of situations where blindly switching opamps could change the sound but from a technical point of view as long as the circuit isn't already atrocious then the only change you're likely to hear would be a degradation of the sound caused by new opamp becoming unstable, ringing, oscillating, or something in that particular circuit because it wan't designed for it.  If it ends up working fine in same circuit and is a decent op amp then any differences should be inaudible.

 

If opamp A and opamp B are both stable in circuit X and both opamps are already spec'ed past the point of audibility then how could you hear a difference between them?

 

Giving you the benefit of the doubt, there's something else to keep in mind before you dismiss a device because the opamps in it didn't work well with whatever you've tried them in before.  Its obvious that the parts can be different but no one is measuring or listening to a single part.  All the parts are combined into a whole, and designed to function together.  If you took a super car and a sports bike that had the same 0-60 and quarter mile times and switched the engines would you expect the same results?  That's essentially what you might end up doing when you roll opamps.  If that happens, have you proved that opamps sound different?  Technically you have, but you miss the spirit of the question which boils down to "Does using different but still good opamps improve the final sound of a well engineered device?"

 

That has to be what you're asking because the final result is the only thing that really matters.  As end users, this is like arguing over engines and steering geometries when we should be looking at lap times.  Are all those parts different?  Of course they are, but you don't win a race with just an engine and you don't make an amp with just an opamp or two and those differences only matter in the context of the rest of the device.  Also, if you're going after something other than transparency then you can relax "improves" to "changes".

 

If I was going to sum up my current position I'd say they can sound different but it probablydoesn't matter because modern ones measure so well and that sort of thing is best (and usually) designed around to make the differences a moot point in the finished product.

 

Of course this is hardly bullet proof logic.  As active devices, this is an order of magnitude of order of magnitudes more plausible than say, cable BS.  If I thought it was that improbable, I'd never have spent my own money trying it out.  I'm just combining my personal experiences with what I know of the theory.  AFIK, there's no real hard data on this sort of thing so I'm open to new evidence.  I'm looking forward to nwavguy's article where there should be some real data.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
This thread is locked  
Head-Fi.org › Forums › Equipment Forums › Sound Science › Why do OpAmps sound different?