In the context we're talking here (the difference between a good and not so good opamp), we're talking the difference between perhaps -135 and -110 dB for THD.
-135 dB THD @ 1 KHz is very good (OPA1612). Classic opamps such as the NE5532 were around -110 dB. I'd be very surprised if anyone could hear -110 dB THD.
Every sound you hear is made of sine waves. See Fourier et al. That is why engineers are so focused on ensuring their circuits minimally distort a single (or multiple) sine wave(s).
I'm just having a laugh at the "high-end" engineering within audio comments in another thread--pretty sure the amount of engineering in an opamp eclipses them by a few orders of magnitude.
It wasn't for pleasure. It was originally for making recording and mixing decisions, which is better classified as work. I think I could hear such differences before on some Etymotic ear buds, which can reproduce HF quite well, and do so over very short distance. Ear buds don't work for mixing though, can't tell if EQ, balance, and other mixing decisions will translate to other systems.
My point is a difference exists. Some people can hear it in some situations. It is not impossible, so people who insist is it completely impossible should tone back their claims a little and we could probably find some areas of agreement. I agree it probably doesn't matter at all for most music listening half way across a room from some speakers. But if you want to mix for for a wide, unknown audience possibly including very picky people with good hearing who might be wearing earbuds with excellent HF and/or transient response, you need to allow for those situations as well. Or younger people who listen in the near field to good playback systems, but a little farther away than I have to now.
The other thing is that when trying to make the best possible recordings, everything matters. Things that degrade the sound only a little in one instance tend to add up as more tracks and processing steps are added. You can end up with a horrible sounding mess if not careful. Doing the best you can at each stage seems to help achieve the best results at the end.
C'mon Jay, I don't believe some of these statements.
For #1, please share the details . . . I am all ears.
As for #2, that's absolute nonsense. Its perfectly feasible to stabilize any commercially available opamp. I've worked with LF355, 356, SA5532/34, AD797, AD744, LM308, LM301 LM741, LM748 and others I cannot even remember anymore. If you cannot stabilize an opamp, then there's something wrong with the design. Feedback theory is not black magic. Its a well documented, well tried and tested branch of engineering. You should remember that next time you or anyone else that has doubts about feedback and stability catches an airplane!
😉
For #1, please share the details . . . I am all ears.
As for #2, that's absolute nonsense. Its perfectly feasible to stabilize any commercially available opamp. I've worked with LF355, 356, SA5532/34, AD797, AD744, LM308, LM301 LM741, LM748 and others I cannot even remember anymore. If you cannot stabilize an opamp, then there's something wrong with the design. Feedback theory is not black magic. Its a well documented, well tried and tested branch of engineering. You should remember that next time you or anyone else that has doubts about feedback and stability catches an airplane!
😉
Why do you call it "mis-information"? When people say that opamps sound different, we have to accept that the condition in which the opamp is implemented should be assumed as part of "the opamps"...
Of course, there is a possibility that when the circuit or implementation is "proper", then there will be no audible difference. But that's irrelevant with the discussion, especially when no one is willing to provide/show the "proper" implementation...
For perfectly linear systems, if integrated over all time. An approximation otherwise, albeit often a highly useful one. Not always useful for accurately testing every possibly distortion mode of an amp, though.
What engineers do is simplify very complex physical reality into approximations manageable by human brains, and by practical computational constraints. By using simplified models of reality, it becomes possible to build what we think of as some pretty sophisticated devices. Not nearly as sophisticated and complex as a human brain though.
long time using opamp for driving headphone amplifier ,and as i have few different kind of opamps on hand it cost me nothing to test them.
tested(into 56R headphones,separated by cushin) 3 different and one that was distorting most(and lowest slew rate) sounded best to me.
my point, we should ditch any kind of listening tests as our hearing is not instrument, and "sound" is no more than our brain playing games.
edit: was using creative sound blaster xfi at 24/96 and rightmark aa for testing
tested(into 56R headphones,separated by cushin) 3 different and one that was distorting most(and lowest slew rate) sounded best to me.
my point, we should ditch any kind of listening tests as our hearing is not instrument, and "sound" is no more than our brain playing games.
edit: was using creative sound blaster xfi at 24/96 and rightmark aa for testing
Attachments
Last edited:
What is the point? Too many people give their opinions based on nothing... I will share it only when people are ready to listen, which could be the case if I pass the test.
So, what do you think is the cause or reason for audible differences between opamps? That's why I ask YOU to provide the files (and of course your OWN circuit)...
Please follow the link in my signature for circuits.
My comments are more directed at claims that there are big differences between devices. Not if they are applied correctly. Maybe on a very high resolution system, but then would someone pass an ABX test? Maybe, but in most cases I think not.
Someone did an experiment here a while back where they put a whole lot of opamps (the same ones) in series. IIRC they could only hear a difference after 5 or some number like that.
My comments are more directed at claims that there are big differences between devices. Not if they are applied correctly. Maybe on a very high resolution system, but then would someone pass an ABX test? Maybe, but in most cases I think not.
Someone did an experiment here a while back where they put a whole lot of opamps (the same ones) in series. IIRC they could only hear a difference after 5 or some number like that.
If that's the concern, fine, I think I can agree with that. So, you want the circuits to be measured and show very low distortion (first) then and only then people can judge whether they can hear differences?
There was a thread where PMA compared his opamp-based pre with direct wire. I mentioned that for me the difference was like "heaven and earth" or there was "HUGE" difference. What do you think is the cause of the audible distortion I perceived?
I have seen your opamp circuit. My current circuit is close to the TIDU034...
So we're back to ground zero. Most people perform so poor in blind test. I perform well above average. But what is the point?
That's why my hypothesis is important: "People failed in ABX is because they have poor skill, it doesn't mean that when they can not hear difference between amp A and amp B, both amps should be equally good for them in the long run. They just don't have the skill for doing it well in an ABX."
That's why the ABX test (amongst others) is so good. Claims of differences are soon laid to rest.
I can only take your word for it that you have the ability to hear night and day differences and are above average.
I can only take your word for it that you have the ability to hear night and day differences and are above average.
The argument for comparison of opamps by simply rolling one opamp for another in the exact same circuit and board layout is the flaw in the argument given for the difference in sound between devices. As Scott and others have pointed out so many times until you actually design each circuit for a particular opamp and have the board design correct for that circuit and proper layout there is no point in comparisons. Once you have each opamp designed for its optimum parameters and get the layout correct then I would consider comparison of opamps for sound quality, otherwise the test protocol is seriously flawed.
In an automotive sense this would be like trying to compare camshafts by simply substituting one for another without doing any optimization for each. Put a cam in for a dragster and compare that to another for a truck application and of course you will see gross differences, but this has nothing to do with one camshaft being superior to another, just misapplication of devices. No different than someone rolling a 5532opamp for one produced for RF amplification, just a simple case of misapplication for the given purpose.
In an automotive sense this would be like trying to compare camshafts by simply substituting one for another without doing any optimization for each. Put a cam in for a dragster and compare that to another for a truck application and of course you will see gross differences, but this has nothing to do with one camshaft being superior to another, just misapplication of devices. No different than someone rolling a 5532opamp for one produced for RF amplification, just a simple case of misapplication for the given purpose.
N Heineken makes me throw up just as heavily as N-1, also sounds identical (merely the quantity is monotonous asymptomatic)
Sorry, but this is wrong. Whether you are talking about an aperiodic impulse or an LVDS clock at 300 MHz, they are all built from or decomposed into sine waves. Linear (or not) doesn't change this.
N>4 Oh yes, it is !
(N=2 is like a sheit opamp. From my childhood memory bank, that would be the Boerebond store brand : buy two crates, barf the 3d for free. N>20 would be the equivalent of a 797)
Last edited:
- Status
- Not open for further replies.