THE OPAMPs WAR differences between different types of opamps

Frequent asked about building blocks: gain stages, buffers, clipping configurations, ...
User avatar
DarkRain
Solder Soldier
Information
Posts: 233
Joined: 10 Feb 2009, 08:32
Location: Bucharest
Has thanked: 13 times
Been thanked: 181 times
Contact:

Post by DarkRain »

@Emu2

- 300mVpk wich is quite a good average of a guitar signal (not talking about active pickups .. etc).
- I do not push anyone NOT to experiment :) as I said ... you don't know me but i'm all for experimenting and alternative solutions :) so again (maybe i wasn't clean from the first time): i'm not trying to convince nobody of anything :)

And again i'm telling you (without being rude) - if you like to have an experiment that will talk by numbers and not by ears and personal perception i'm more then glad to talk about it; but numbers and graphs :) and as i said many times (in this topic too) i invite anyone who would like to make some test to make them and share the results, and if someone thinks that 30Hz to 40KHz is not enough in terms of analog bandwidth we can go over that but i don't think it would be relevant (same goes for FFT resolution :) )

Regards,
DeX

PS: just to be clear - i invite anyone and everyone toward experimentation and to listen with their own ears not with someone else's golden ears.

PPS: I hope you didn't register this account only to post on my topic :D as i can see the only posts around here were made on this specific topic :))

User avatar
Emu2
Information
Posts: 8
Joined: 05 Mar 2016, 17:39
Been thanked: 4 times

Post by Emu2 »

I would like to briefly comment on measurement verses "golden ears." It is possible to make mistakes using either approach. It is possible to convince ourselves that our present test technology can measure everything that exists. As test technology has evolved, experience has shown that beliefs change along with it. It used to be thought that humans could not hear less than about 0.1% harmonic distortion, which has since been shown incorrect. The reliance on available measurements and errors resulting therefrom has not been confined to audio or to engineering. You also see that in various fields such as medicine. For one example in audio, it used to be thought that existing tests for harmonic distortion and intermodulation distortion covered all bases for distortion. Then people started to realize that transient intermodulation distortion was a problem that was not being caught by the existing tests. Anyway, my point is that engineering and design does rely heavily on mathematical modeling and measurement, but not to the exclusion of all else. The models always have limitations and so do the measurements. Use them, but understand their limits.

Regarding claims of persons to the effect that they can hear things you can't measure. That can be true in some cases, when measurement technology and practice is imperfect . Also true that a sort of audio placebo effect exists and people can fool themselves into believing they hear something that isn't really there.

However, in such cases we do have double-blind testing methodologies. If someone can reliably sort out by ear random A/B substitutions to adequate statistical significance, then they are most likely hearing something real. Nobody has passed an honest test of that type to differentiate oxygen free verses standard speaker cable. However, just as some people have perfect pitch and some are tone deaf, some people are very good at hearing tiny amounts of distortion. With perfect pitch and distortion perception, the best evidence we have is to the effect that people are born with some natural propensity to be be good and one or the other, or maybe even both, but then learning plays a subsequent role in developing those abilities to fruition.

Finally, I would just note human brains are subject to many well researched subconscious biases. Confirmation bias is one of the major ones. It is the tendency to form a belief first, then search for information in support of it being true. Once sufficient information is found, the search stops and the conclusion is taken as confirmed. In this process, people rarely, almost never in fact, seriously search for information having the effect proving that the initial belief is wrong. If anything, such information is ignored. For one hypothetical example, not saying it occurred here, if someone believed or suspected a priori that opamps don't make any difference in TS-808 distortion circuits then went out and did some tests confirming that, the tests would have to be viewed skeptically, at least initially. Because humans tend think as humans do despite best efforts to be objective, that's why we have to have peer review and other methodologies to help counter some of the limitations of how human brains function. But now I am delving too much into neuroscience and cognitive psychology for a stompbox forum, so I will quit here.

User avatar
MoonWatcher
Diode Debunker
Information
Posts: 715
Joined: 28 Jul 2008, 12:27
Has thanked: 40 times
Been thanked: 52 times

Post by MoonWatcher »

Emu2 wrote:Finally, I would just note human brains are subject to many well researched subconscious biases. Confirmation bias is one of the major ones. It is the tendency to form a belief first, then search for information in support of it being true. Once sufficient information is found, the search stops and the conclusion is taken as confirmed.

But now I am delving too much into neuroscience and cognitive psychology for a stompbox forum, so I will quit here.
The biases are what always intrigue me the most, at least with stuff like this (not just audible sounds from pedals, but really from anything that is pleasing or 'engaging').

I've just always struggled with trying a sort of "X Files bias," in that I approach this stuff as "I want to believe," rather than taking the Myth Busters approach. It's just a bit more exciting than approaching it by "the labcoat method."

But with my testing (or lack thereof) with different capacitor composition types, there was never an instance where I found one type to be audibly better than another. I actually included some pedals that had a potpourri of different types, and again - none were either really better or worse. If anything, it seemed like some pedals just sounded better because of what were probably a bunch of small differences in tolerances. Oddly, those were almost always examples that had what would probably be considered to be inferior composition caps.

You hit on a lot of points about how or when an op amp probably might assert its characteristics without being masked by clipping diodes (or lowpass filters, or whatever). And that's what I hope to have more A/B testing to get more results, even if they might be biased one way or the other.

User avatar
mojah63
Breadboard Brother
Information
Posts: 144
Joined: 11 Apr 2008, 20:08
Location: East Coast US
Has thanked: 40 times
Been thanked: 24 times

Post by mojah63 »

Yawn... These type arguments have been going around forever. It wasn't that long ago people were arguing over capacitor dielectrics.. Are all the opamps from the same manufacture too??

User avatar
DarkRain
Solder Soldier
Information
Posts: 233
Joined: 10 Feb 2009, 08:32
Location: Bucharest
Has thanked: 13 times
Been thanked: 181 times
Contact:

Post by DarkRain »

Nope ... totally random and i do think that one of them is a CHN knockoff :)
mojah63 wrote:Yawn... These type arguments have been going around forever. It wasn't that long ago people were arguing over capacitor dielectrics.. Are all the opamps from the same manufacture too??

User avatar
DarkRain
Solder Soldier
Information
Posts: 233
Joined: 10 Feb 2009, 08:32
Location: Bucharest
Has thanked: 13 times
Been thanked: 181 times
Contact:

Post by DarkRain »

And a very nice and constructive discussion started here. I would like to thank Emu for taking the time to post a lot of valuable info even if it is in the neuro domain (not a problem with that because ... i do think that all variables must be taken into consideration).
At last but not at least ... the truth is always in between the shades ... so ... this experiment was not to prove a point but to show a finding and also to experiment and not to dismount a theory.

Regards,
DeX

User avatar
electrip
Information
Posts: 24
Joined: 29 Aug 2011, 11:57
Been thanked: 4 times

Post by electrip »

And I think this belongs to TGP. :roll:

electrip

User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

Still - I'm happy that people are still seeking out the 1970s JRC version of the 4558 - I still have several tubes of NOS ones and a boxful of "pulls" and still release then to Ebay in small quantities to provoke bidding wars amongst the deluded! My beer money supply is assured for years to come!
"Why is it humming?" "Because it doesn't know the words!"

User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

Emu2 wrote:It used to be thought that humans could not hear less than about 0.1% harmonic distortion, which has since been shown incorrect.
Total nonsense! The audiophools are frequently convinced that valved amplifiers have "lower distortion" than semiconductor-based types - they just (subconsciously) like the kind of concordant distortion they produce. There is a fundamental maxim in electronics - if you can't measure it, it doesn't exist. The differences you claim to hear are entirely your personal, subjective bias. In a proper, blind A - B test, you would not be able to tell the difference, much less identify the type of device used. I'd be happy to stake a house on that - I could use another house!

In the same way - A - B blind testing of a slightly modified (base - collector capacitors) silicon Fuzz Face is indistinguishable from a 1968 vintage germanium one.

Listeners usually "hear" what they want or expect to hear. Remember - there's no fool like an audiophool!
Last edited by mictester on 13 Jul 2016, 07:01, edited 1 time in total.
"Why is it humming?" "Because it doesn't know the words!"

User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

mojah63 wrote:Yawn... These type arguments have been going around forever. It wasn't that long ago people were arguing over capacitor dielectrics.. Are all the opamps from the same manufacture too??
Nope - you've missed the point. As with capacitors, the op-amp changers "hear" what they want to hear. After all, when they've been stupid enough to pay $100+ for a $0.02 dual op-amp, they've got to justify the expense - even to themselves!

The same applies to "audio grade" capacitors. The "audio" electrolytic caps from one manufacturer had a notably inferior specification to much cheaper "ordinary" types from another manufacturer. The "audio" ones had a much lower temperature rating and would frequently fail in a hot amplifier, but the audiophools insisted on telling us that these components were "better".

My supply of beer money is safe as long as there are enough deluded clowns happy to bid up the price of NOS junk devices!
"Why is it humming?" "Because it doesn't know the words!"

User avatar
DarkRain
Solder Soldier
Information
Posts: 233
Joined: 10 Feb 2009, 08:32
Location: Bucharest
Has thanked: 13 times
Been thanked: 181 times
Contact:

Post by DarkRain »

Not to mention NOS devices that are advertised as the best parts you can buy... how the heck ... after spending 30 years in a storage ... not that is an affirmation full of sh%$
And no ... other then tubes (because i do have quite a stash) i'm not using NOS components.

Regards,
DeX

User avatar
Emu2
Information
Posts: 8
Joined: 05 Mar 2016, 17:39
Been thanked: 4 times

Post by Emu2 »

mictester wrote:
Emu2 wrote:It used to be thought that humans could not hear less than about 0.1% harmonic distortion, which has since been shown incorrect.
Total nonsense! The audiophools are frequently convinced that valved amplifiers have "lower distortion" than semiconductor-based types - they just (subconsciously) like the kind of concordant distortion they produce. There is a fundamental maxim in electronics - if you can't measure it, it doesn't exist.
Uh, actually, you can measure well below 0.1% distortion and hear it too. There are limits of course, but lower than 0.1%. The reason some audiophools used to think humans couldn't hear less than 0.1% was because the audiophools didn't know how to measure anything less than that, so they assumed it couldn't exist. There are limits both to hearing and to measuring. As it turns out, people can hear some things that are pretty hard to measure, but that's just a limitation of existing test technology.

However, you are correct that people also can easily convince themselves that they hear things that don't really exist. Special speaker cables being a prime example of something that does not sound different from normal speaker cable.

Regarding hearing a difference between distortion pedals, as you turn up the gain or drive and increasing the amount of clipping, the more they all sound the same. If you keep the drive minimal so the wave peaks are only slightly clipped then you hear the most difference. Some audiophools turn up everything to 10 and then wonder why they can't hear any nuance.

User avatar
MoonWatcher
Diode Debunker
Information
Posts: 715
Joined: 28 Jul 2008, 12:27
Has thanked: 40 times
Been thanked: 52 times

Post by MoonWatcher »

Emu2 wrote:Uh, actually, you can measure well below 0.1% distortion and hear it too. There are limits of course, but lower than 0.1%. There are limits both to hearing and to measuring. As it turns out, people can hear some things that are pretty hard to measure, but that's just a limitation of existing test technology.
I think we are allowing for huge generalizations by merely saying "able to hear below .1% distortion."

I'd also argue that there are probably fewer instances of limitations of existing test technology vs. establishing the best testing methods with the given technology. Testing for our specific purposes is a fitting example - using 'other audio testing standards' is often insufficient. Most of these guitar-specific circuits are implemented in ways that are at odds with electronic audio design - lower supply voltages, unintentional or intentional misbiasing of amplifiers, inducing clipping, purposefully filtering/bandpassing the signal in ways that are meaningless with 'full-spectrum' audio, and oftentimes choosing components with no real regard to the specs listed on datasheets.

Getting back to the point of people being able to hear things that are hard to measure, I think that is often due to a combination of phenomena occurring. In other words, someone merely hearing .008% 'harmonic distortion' probably doesn't occur. But if you configure a circuit so that some op amps will exhibit ultrasonic oscillations with 'some common electric guitar signals,' it's no longer just the distortion on its own. If the grounding/negative feedback is poorly designed so that it also exacerbates oscillations/other phenomena related to distortion, it can make it detectable to the human ear.

You mentioned that it's hard to distinguish these effects at high gain/clipping settings, and that it's much more detectable at low(er) gain settings. While I overall agree with that, I've done a fair amount of testing with what I refer to as "a threshold point," and it's typically "higher gain than low," but it absolutely is dependent on the circuit's construction. Some circuits can be implemented so that they continue to sound pleasing even at (sometimes much) higher gain settings, and some op amps will allow for greater or lesser acceptable upper limits to the gain.

...And to be clear, I'm not just talking about doing crude design stuff like simply adding more lowpass filtering to nix unpleasing harmonics. A big part of it is simply experimenting with the proximity to clip, usually with diodes, and typically finding a decent balance between clipping from the op amp and clipping from the diodes. After that is done, multiple and precisely chosen lowpass filters will only roll off the minimum amount of higher frequency harmonic content. It really does make it possible to (potentially) have a more distorted effect that will sound considerably different from a pedal with a 'boilerplate distortion design.'

Coming full circle back to op amps, I think that what I described in the above paragraph should be included in testing and determining which op amps are detected to both sound (to the human ear) and measure differences. And if there is the insistence that some chips of the same designation, but by different manufactures, do indeed sound different, I think it would be a great way to expose it. This need not be limited just to different 4558s, as there are different 072-manufactured chips, and there are obviously numerous other examples.

I still think that the Tubescreamer represents an overall terrible choice for such testing, despite whatever gain/distortion it is set for. As I thought about it, I realized that the Bluesbreaker (1st generation, 1st year of production) is very possibly a better test candidate. IC1A contains no clipping diode circuits, and the lowpass filter is almost entirely negligible. IC1B's highpass filter is considerably lower than a Tubescreamer, and the effect of the clipping diodes are minimised by the 6k8 series resistance. And for good or bad, the Bluesbreaker has no input or output buffer. Whether or not the buffers might have no audible effect is not what's important - eliminating them from potentially being variables is considerably important.

User avatar
Emu2
Information
Posts: 8
Joined: 05 Mar 2016, 17:39
Been thanked: 4 times

Post by Emu2 »

It's refreshing to hear from someone knowledgeable and serious. That being said, I have a TS-9, have swapped opamps, and hear a difference. And the the difference isn't subtle to me. However, neither is the difference in sound between things like DACs such as Lynx 2, Cransong HEDD, and Benchmark DAC-1. Or different preamps. Also have had the experience of training neophyte sound engineers to hear the difference, and once you point out to them what is different, they can reliably hear it too. People do tend to be oblivious to subtleties of distortion at first though. Just as some people naturally have perfect pitch and some are tone deaf, most people can get better at discerning relative pitch with some training. Same for distortion in my experience.

User avatar
Emu2
Information
Posts: 8
Joined: 05 Mar 2016, 17:39
Been thanked: 4 times

Post by Emu2 »

I guess I would add some commentary about a "threshold point" concept. You are probably right that there is some intermediate sweet spot gain or drive level producing distortion where the differences between components are the most apparent. To me however, I look for settings that are generally euphonic and that are most responsive to playing differences. It other words, I want to play the effect. Also, for me a tonal requirement is how the resulting sound will fit into a mix, not how it sounds in isolation, unless it is to be used exclusively for soloing. A pedal that sounds great in a more static way can be useful if switched in and out frequently, but not so much for extended use. In brief, I am not usually as interested in differences of some set of components at the point of maximum difference, so much as in terms of breadth of capacity for expressiveness with different playing techniques, constrained by at least adequate tone and suitability for fitting a mix.

User avatar
okgb
Diode Debunker
Information
Posts: 768
Joined: 16 Mar 2010, 03:58
Has thanked: 226 times
Been thanked: 38 times

Post by okgb »

I DO believe some caps are better than others,
i think comparisons need to be recorded so real time feelings are not a factor, and levels matched as close as possible
[ flipping the phase of one and matching level in mono for maximum cancelation when 180 degrees polarity ]

I recapped a high end hybrid tube compressor with the Elna electrolytics, one channel at a time and the difference was very obvious
not brighter as I imagined , but more clear as if there was a slight narrow cut between 200-300 hz

now , I'm not going to pay for Voodoo a 2% difference for 200% cost......there will be other things to negate this!

User avatar
theblackman
Resistor Ronker
Information
Posts: 250
Joined: 27 Jun 2007, 00:43
Location: Western Australia
Has thanked: 10 times
Been thanked: 3 times

Post by theblackman »


Post Reply