What is data-data?
I thought it was all data?
Data doesn't care if it works or doesn't work.
It's still data, yes?
Sometimes we have data and we don't know what it means until later.
It is what it is.
HOWEVER,
If we make it with unobtainium...then it has to be the greatest, right?
Cheers,
Sync
Okay, that's pretty contentious with actually presenting a leading argument style and projecting your opinions on the discussion. Which one will it be? If we look at what you've written, then it seems you're less interested in seeing where people stand than projecting yours forward. There's nothing wrong with that, but transparency is appreciated.
Why do you take such exception to the word "defensible?" Perhaps disprovable or falsifiable would work better for you? (I really should have used falsifiable from the word go) A position that can either be proven or disproven, with its constant force to refinement. Likewise, it's a little quick to cite scientism if you read the back and forth and the outstanding nature of some claims. Would you consider it a bit exceptional if I claimed I can see and identify orbiting satellites from my back porch? I would hope that you would be highly skeptical.
I summarily disagree about your assertion as to severely curtailing creativity and making progress in this hobby. Especially within context of audio electronics, where a modest understanding of the electronics can get you pretty far, especially with a little effort to learn and understand some of the free tools at our disposal. Why should it limit creativity?
Fastl's book is a good start.
What are you trying to say? Are only people with biases towards hearing something supposed to be party to the conversation?
If it's that I'd be disqualified from a listening test due to bias, wouldn't that disqualify everyone for some reason or another? That's the point of needing thorough test protocols and limited experimental scope. (Positive predictive value)
Do you want to determine "what's wrong with op-amps" or just further this back-and-forth? The rubber has to hit the road somewhere, whereupon something must actually be settled. Words clearly aren't doing the trick.
Or just to insult members by saying they're there to impress one another? Or the previous insinuation that it's simply prove they're right. That comes across pretty dang duplicitous as to intentions.
Nor this current line.
I should be quicker to follow the admonishment to avoid eristics.
Last edited:
Absolutely. 😀
A consistent anecdote can provide a good starting hypothesis, but it should be treated with skepticism, due to its quality. More than enough has been spilled about quality of data. (It depends)
Gerrymandering at its best. Selected amps. All male audience past a certain age and with a bias. C'mon guys can't anyone propose "random"?
Last edited:
All these challenges to do abx testing. How would this happen? Realistically! Who flies where? Who sets this up? The real devil in the.... Who certifies the validity of the mechanics/electronics of the test?
Not that I care since I listen to every unit I build before it goes out. That is pushing something on the order more than 1k units by a very healthy amount.
So these dares? To me it's like an ISIS guy in Syria goading someone here in the western world to sue them.
Not that I care since I listen to every unit I build before it goes out. That is pushing something on the order more than 1k units by a very healthy amount.
So these dares? To me it's like an ISIS guy in Syria goading someone here in the western world to sue them.
Last edited:
In my experience, what is wrong with op-amps is: The datasheet!
Not only are the sample schemas works of omission, but also there's not many clues provided on what makes a particular model of op-amp special or how to optimize it.
Maximizing the amount of possible applications in datasheets/advertisements serves also to maximize available customers; however, the cost is omission of optimized examples/applications. Thereby, the provided sample schema is likely to be really bad junk or at least rather uninformative.
Making your own schema from scratch is doable but it takes a long time; however, there is a shortcut to be had with a reconnoiter of service manuals.
This reminds me of the JRC4580, and a google search reveals many who seek to replace it. However, it is also the output of the #1 voted best quality sound cards on DIYaudio.com. The usage is considerably different. Most of the time the JRC4580 does hi-fi, it is used as a parallel amplifier; and, most of the time its replacement was sought, it was used as a buffer. Other replacement-sought applications include overcurrent, which is misuse. Provided documentation won't tell you which is a better use for it.
Also datasheets would be much too bulky if they started out with: "Not for use as an outdoor barbecue grille." And, 1600 more things that you shouldn't use it for. Such guidance wouldn't increase popularity at all, and it might reduce the number of customers. And that is what you won't get from a datasheet. They won't tell you the one best use for fear that's the only (few) customers they get. So, there is some omission in the documentation.
Another thing wrong with op-amps is commonplace usage of an amplifier design set on zilch for gain to make a poor buffer. That's just crap, even though the practice is really popular. A good buffer never involves a built-for-gain design. Screwdriver and hammer are not well interchangeable. So, don't use an amp as a buffer with much expectations of it going well.
Wrong application or even sub-optimal application, of an op-amp, results in an effects box set on bad; and, that's really popular. But, that's the only thing wrong with an op-amp.
I actually can't find fault with op-amps, except that the documentation is just awful.
Not only are the sample schemas works of omission, but also there's not many clues provided on what makes a particular model of op-amp special or how to optimize it.
Maximizing the amount of possible applications in datasheets/advertisements serves also to maximize available customers; however, the cost is omission of optimized examples/applications. Thereby, the provided sample schema is likely to be really bad junk or at least rather uninformative.
Making your own schema from scratch is doable but it takes a long time; however, there is a shortcut to be had with a reconnoiter of service manuals.
This reminds me of the JRC4580, and a google search reveals many who seek to replace it. However, it is also the output of the #1 voted best quality sound cards on DIYaudio.com. The usage is considerably different. Most of the time the JRC4580 does hi-fi, it is used as a parallel amplifier; and, most of the time its replacement was sought, it was used as a buffer. Other replacement-sought applications include overcurrent, which is misuse. Provided documentation won't tell you which is a better use for it.
Also datasheets would be much too bulky if they started out with: "Not for use as an outdoor barbecue grille." And, 1600 more things that you shouldn't use it for. Such guidance wouldn't increase popularity at all, and it might reduce the number of customers. And that is what you won't get from a datasheet. They won't tell you the one best use for fear that's the only (few) customers they get. So, there is some omission in the documentation.
Another thing wrong with op-amps is commonplace usage of an amplifier design set on zilch for gain to make a poor buffer. That's just crap, even though the practice is really popular. A good buffer never involves a built-for-gain design. Screwdriver and hammer are not well interchangeable. So, don't use an amp as a buffer with much expectations of it going well.
Wrong application or even sub-optimal application, of an op-amp, results in an effects box set on bad; and, that's really popular. But, that's the only thing wrong with an op-amp.
I actually can't find fault with op-amps, except that the documentation is just awful.
Marce spent 6 years of his life working on the next generation vehicle comms. for the US military, he also spent numerous years working with a company that does noise cancelling headphones for such systems (so spent many a happy day with acoustic engineers as well as the usual EEs, programmers etc. that make up a design team). I took part in many controlled listening tests where my hearing capabilities were tested so that my results from the tests could be further refined. I have also done the odd test with friends where we have tried to follow the recommendations for DBT.
Blind testing or unsighted testing is interesting... Try it with an open mind and you'll find some of the rubbish promoted by true believers regarding night and day differences disappear. It changed my views big time as I realised such things as the results of some mods I had done were perceptual; and actually expectation bias rather than day and night changes.
Just doing even the most basic unsighted test (eras only) can be a big step in proper control...
Of course as a few on this thread are promoting, it goes against the true believers dogma, if you believe something will change the sound, it will, so many rely on easily fooled perceptions and sighted testing; oh and the odd encouraging word from some guru to tell them they are correct and the rest are just stubborn deaf people with inadequate systems....
ABX is only one possible format. There are many others.
You can set up actual listening tests on your own just to screen out what turns out to be incorrect from the get-go. If after that, you feel there's a there there, then write up what you did and put it out there for review and replication- honest researchers are completely transparent in that regard.
The important thing that too many are missing is step one- define specifically in advance what question your experiment is to answer, what are the criteria for a positive and negative result, and stick to it. Don't expect one experiment to answer a whole multitude of questions or to answer a question which is vague and undefined or to answer a question that was not the one it was designed for. A good experiment gives you actionable data regardless of the outcome- as soon as you hear people say "pass" or "fail," you know they miss the point.
If you tightly define a question to be answered by a listening test and you want help in designing a good one, I'll be happy to do so.
Usually the data sheets are ok, maybe with minor omissions like the
noise hill at 300 KHz or that unity gain stable has to be taken with a grain of salt
if the -inp net has half a pF too much.That is what parts qualification is for.
Applications belong into app notes, that is where the name comes from.
The real offenders with bad data sheets are JFETs. That begins with the Siliconix
barn doors and Interfet gives you half a page. Emails for more info
are silently ignored.
I just got bitten by space qualified JFETs where everything worked nicely with
the engineering samples, and the flight units featured 2V more Vgs. I could
have used BF862 for engineering just as well. What do I need engineering
samples for, when they are not even remotely similar to the flight parts that just
have gone through the ESA quality assurance system at a few hundred times
the price?
Probably we should have used BF862 for flight, but they won't let me do that.
Gerhard
So do I, space and near space is easy, found an even more pedantic group to work for, its unreality design based on 50s ideas of layout with modern components!
With reference to the comments on op-amp data sheets as someone else pointed out there are hundreds of application notes out there as well as other sources of related information. My view is if they can work for the likes of LIGO and similar levels of design they can work for audio....
With reference to the comments on op-amp data sheets as someone else pointed out there are hundreds of application notes out there as well as other sources of related information. My view is if they can work for the likes of LIGO and similar levels of design they can work for audio....
Yes, those components (like compressors) are used to modify the sound. Nothing to do with sound reproduction, which is what I assume this thread is supposed to be about.morinix said:
DA is linear. The most it can do is make a small difference to LF rolloff. Anyway, thanks for confirming that you are not interested in hi-fi.
Could this perhaps be related to the fact that almost all recorded sound will already have passed through lots of very unfashionable opamps in the studio? If you are looking through 10 windows it is daft to claim that the last window (the only one under your control) must be of a particular type in order to see clearly. If any decent window is good enough then the last one only has to be any decent window too. If ordinary decent windows are not good enough then the problems introduced by the first 9 cannot be undone however good is the last window - and putting the last window in an expensive frame changes nothing.mmerrill99 said:
No, what is happening here is that people get excited about the only changes they can make. In some threads it is cables. Once people learn how to solder they do cap rolling. Some do opamp rolling instead, and usually remain ignorant of stability issues etc. When the only thing you can change is a cable, all audio problems look like cable problems. However, once you have progressed to designing circuits some form the notion that the harder it is to design a circuit the better the circuit must be - discrete must be better than chip, and expensive chip must be better than cheap chip. 1GHz GBW must be better for audio than 200MHz, surely? Let's forget that the signal has already passed through quite a few 20MHz GBW chips in the studio.
When an extraordinary scientific claim ('this opamp modifies an audio signal in a readily audible fashion, even though accepted theory says that it cannot') is made only a scientific test can back it up. Burden of proof is the issue. We don't have to prove anything, because accepted knowledge is on our side.
Burden of proof is the issue. If I claimed to see fairies at the bottom of my garden it would be up to me to prove it. If I said that all the people demanding proof have never had an eye test which involved seeing fairies so they should not demand proof from me, then people would rightly laugh at me. Truth is the issue, not my or your experience.
If I claim that a resistor more or less follows Ohm's Law I would not expect people to jump up and demand proof, because my claim is fully in line with accepted knowledge.
To suppose that audio electronics might be outside the domain of science is pure silliness at it's most overt. Thanks for declaring yourself.
Is this a manifestation of feedback denial? I find it difficult to think of a good buffer which starts off with little or no gain. An emitter follower, for example, works as a buffer precisely because a BJT has massive voltage gain.danielwritesbac said:
The simplicity of analogies always lets it down in the end but let's stick with this analogy. You are assuming that every window is performing the exact same role in the exact same circumstance - is this the case for the many opamps in the recording/playback chain? To stretch the analogy somewhat, if one window is in a steamy environment will it fog up or will it remain clear because of a surface treatment? This treatment wouldn't be noticed ordinarily - only the steamy environment differentiates one window from another. What about other window treatments that are intended to retain transparency in other environments?
Perhaps discrete allows optimisation of aspects of performance that better suit the role in which they are intended to be used - you know, like the window in a steamy room.
Hmm, 'accepted theory', 'burden of proof', 'extraordinary claim' - are the well-worn, over-used scientism phrases from most audio discussions. Sorry but not interested in that discussion.
But of course you have no example of this 'perhaps' other than armwaving? op-amps are exceedingly good at clear windows in the domestic environment.
And you are not interested in discussing the scientific method as its 'scientism'. Can't see much open mindedness here.
And you are not interested in discussing the scientific method as its 'scientism'. Can't see much open mindedness here.
Does that mean that you would agree that in some specific circumstances or circuits or use an opamp can be transparent? That would be progress!
Jan
Why do you look for examples of opamp weaknesses from me when you ignore previous examples given in this thread? Perhaps you just want to hone your debating skills?
As nobody is able to confirm the perfectness of any reproduction, a valid hypothesis is that a lot of components "modify the sound" and within the framework of a reproduction that per definitionem is based on the imagination ability of the listener there is room for different but equal directions.
Did you ever had the chance to compare some "raw" recorded material to the end product i.e. the cd or vinyl album?
There is so much difference that you should not loose anymore quality at the reproduction at home. So, if the last window lowers the quality it should be avoided because that is possible.
Btw, there are a lot of people in the recording business who are very interested in quality of their gear and of course modifications are quite common.......
Quite often it´s just a hypothesis that a certain effect can´t be audible but nevertheless it is presented as fact. Therefore any test that confirm that believe is accepted despite any methodological flaws (even sighted listening is obviously totally acceptable as long as the result is "did not hear a difference") while any test with diverging results is critizised....
It´s human but don´t call it scientific.
And if you really care about science you should be the first to find potential flaws in your theory.....
Because you said 'Perhaps discrete allows optimisation of aspects of performance that better suit the role in which they are intended to be used'. So I assumed you had an example. I don't recall any particularly obvious examples shown that had any substance. Not trying to hone debating skills at all. I won awards for that when younger then gave up when I saw politicians at it.
- Status
- Not open for further replies.