I doubt that the average DIYer can easily build discrete circuits that outperform modern high-quality op-amps. Those "skilled in the art" may of course be able to do it.
I do use op-amps but I try to use the least number possible to achieve the desired function. I also try to use them in inverting topology as often as possible since many of these devices show less THD that way.
I once thought that I was brave and wanted to try a subtractive-delay crossover at 300 Hz approx. For this I would have to build an analog delay line using a high-order Bessel allpass. I dropped the idea without trying it at all since I assumed that these 16 to 20 op-amps would have a higher sonic influence than the less steep slopes (that are causing more lobing and higer driver distortion) of the subtractive crossover that I am unsing now. But I wouldn't hesitate to use such a crossover above 2 kHz since a) less op-amps would be needed and b) the critical midrange would not be handled by them.
It is like anywhere else - it often maters less IF you use them or not but HOW you use them. And it makes a big difference which op-amp you use for what purpose.
Regards
Charles
I do use op-amps but I try to use the least number possible to achieve the desired function. I also try to use them in inverting topology as often as possible since many of these devices show less THD that way.
I once thought that I was brave and wanted to try a subtractive-delay crossover at 300 Hz approx. For this I would have to build an analog delay line using a high-order Bessel allpass. I dropped the idea without trying it at all since I assumed that these 16 to 20 op-amps would have a higher sonic influence than the less steep slopes (that are causing more lobing and higer driver distortion) of the subtractive crossover that I am unsing now. But I wouldn't hesitate to use such a crossover above 2 kHz since a) less op-amps would be needed and b) the critical midrange would not be handled by them.
It is like anywhere else - it often maters less IF you use them or not but HOW you use them. And it makes a big difference which op-amp you use for what purpose.
Regards
Charles
The AD8066 and THS4631 surely has good static measurement data. As do most OPamps even thought they do not have such a high OL gain a these two. I have them both and have tried them in both IV and buffer applications.
Many here say that discrete sounds better than OPamps. I agree to that but only and only if the design is very good and the components used are also very good and as long as the PS is very good etc etc.
But, I am curious to why OPamps sound different?
I mean, many of the audio ones show extremely good measurement results in the datasheets.
Or do all audio OPamps sound the same?
If OPamps do sound different, can it then not be so that a discrete design can sound different from OPamps and that this sound is then better than any OPamp?
This is kinda a logical assumption. The fundament is of course that OPamps do sound different, and that the measurements we see in the datasheets are relavant to sonics.
I must admit that I have found different audio OPamps to sound different by doing swapping of OPamps in a given design. I cannot proove by science why they sound different.
Sigurd
Many here say that discrete sounds better than OPamps. I agree to that but only and only if the design is very good and the components used are also very good and as long as the PS is very good etc etc.
But, I am curious to why OPamps sound different?
I mean, many of the audio ones show extremely good measurement results in the datasheets.
Or do all audio OPamps sound the same?
If OPamps do sound different, can it then not be so that a discrete design can sound different from OPamps and that this sound is then better than any OPamp?
This is kinda a logical assumption. The fundament is of course that OPamps do sound different, and that the measurements we see in the datasheets are relavant to sonics.
I must admit that I have found different audio OPamps to sound different by doing swapping of OPamps in a given design. I cannot proove by science why they sound different.
Sigurd
It seems odd to me how some of the tenets of good audio design get brushed aside when it comes to talking about sound quality of op-amps.
Why does anyone think an op-amp ought to sound better than a discrete circuit?
What happened to the importance of capacitor and resistor materials? What about psu isolation in different stages? What happened to the importance of minimizing the number of semiconductors in the signal path? What about reducing cross-talk? What about optimising stability?
As I see it, the ability of an op-amp to sound nearly as good as an optimised discrete circuit is quite an accomplishment. Not the other way around.
Why does anyone think an op-amp ought to sound better than a discrete circuit?
What happened to the importance of capacitor and resistor materials? What about psu isolation in different stages? What happened to the importance of minimizing the number of semiconductors in the signal path? What about reducing cross-talk? What about optimising stability?
As I see it, the ability of an op-amp to sound nearly as good as an optimised discrete circuit is quite an accomplishment. Not the other way around.
Well put, Traderbam. If I could use IC's as building blocks everywhere, as I have attempted to do for the last 40 years, and make the same audio quality, I would do so immediately. I just have not found it to be as good sonically.
When designing our own discrete, class A op amps, even with sophisticated topologies not easily possible in IC design, we found that transconductance amps, or even open loop designs sounded better, even if the distortion starts to rise toward perceptible levels and great care has to be put forth to keep the distortion under control, if only to have acceptable measurements in the marketplace.
This is where Charles Hansen (Ayre) and I (CTC and Vendetta) are at, today. We are both in ruthless competition with each other and the rest of the world to make the best sounding designs, just like F1 racing teams in auto racing. Yes, we work at that level of audio design, and have done so for decades. Once again, if we could just use IC's to get the job done, we would do our designs with IC's and pay more attention to other factors, like making a living or sophisticated xover networks, etc, etc.
When designing our own discrete, class A op amps, even with sophisticated topologies not easily possible in IC design, we found that transconductance amps, or even open loop designs sounded better, even if the distortion starts to rise toward perceptible levels and great care has to be put forth to keep the distortion under control, if only to have acceptable measurements in the marketplace.
This is where Charles Hansen (Ayre) and I (CTC and Vendetta) are at, today. We are both in ruthless competition with each other and the rest of the world to make the best sounding designs, just like F1 racing teams in auto racing. Yes, we work at that level of audio design, and have done so for decades. Once again, if we could just use IC's to get the job done, we would do our designs with IC's and pay more attention to other factors, like making a living or sophisticated xover networks, etc, etc.
scott wurcer said:
It may not be an issue of enough resolution.
It may be an issue of learning.
This is something R. Heyser taught me.
Like phase in speakers.
Phase between speaker drivers was once thought inaudibe. As was time arrival. He showed us otherwise.
Warning:
Once you're taught to hear a difference, it's really hard to go back.
R. Heyser taught me as well about the problems of negative feedback, as well as Charles Hansen and many others. He found that negative feedback was problematic in the 1950's!
The Variations in the JC 2 thread here should show you how to make or buy a transconductance amp that should be OK, IF the output current is set high enough.
The Variations in the JC 2 thread here should show you how to make or buy a transconductance amp that should be OK, IF the output current is set high enough.
I finally finished my first-ever chip amp, 3875 gainclone last night and the sound was nice (I wouldn't say the best among amps I have built though).
I'm looking forward to being 40's next year.
I heard from several sources that men starts to produce some feminine hormones (vice versa) as we get older.
And I also saw many articles that environmental hormones affected manhood, so bunch of guys (even in the States) are driving pink Toyota Prius in fear of wasting too much $$$ when they drive 427ci big block muscles. (I also sold my Corvette C6 Z06 last year and thinking of getting that sissy MB SLK's)
So, am I getting too old or did I eat too much environmental hormones already?
I built a chipamp.
Doug
I'm looking forward to being 40's next year.
I heard from several sources that men starts to produce some feminine hormones (vice versa) as we get older.
And I also saw many articles that environmental hormones affected manhood, so bunch of guys (even in the States) are driving pink Toyota Prius in fear of wasting too much $$$ when they drive 427ci big block muscles. (I also sold my Corvette C6 Z06 last year and thinking of getting that sissy MB SLK's)
So, am I getting too old or did I eat too much environmental hormones already?
I built a chipamp.
Doug
I do not like Integrated circuits, nor op amps, nor power audio amps
They are power operational amplifiers.
I think sounds strange (with exceptions)... small distance inside forces us to slow down the unit using capacitors...if not already "controled" internally.
I feel the "toy" sound when i listen.... something i cannot explain...but blind tests made me perceive i can detect that "toy" sound very well.
Even those top ones sounds this way.
Well.... i use to be here, at Solid State.... i feel this is the place to the ones loves discrete.
Carlos
They are power operational amplifiers.
I think sounds strange (with exceptions)... small distance inside forces us to slow down the unit using capacitors...if not already "controled" internally.
I feel the "toy" sound when i listen.... something i cannot explain...but blind tests made me perceive i can detect that "toy" sound very well.
Even those top ones sounds this way.
Well.... i use to be here, at Solid State.... i feel this is the place to the ones loves discrete.
Carlos
Re:
Hehe, Carlos
You may not know this???
... but DX-Amp is one typical Power Operational Amp
- with transistors in discrete .. the normal Class AB output.
In fact DX-Amp is the very basic topology for op-amp.
Without much new thinking.
The old and good op-amp design. Not much more.
What blind tests ???
made you spot Op-Amps/Power Op-Amps have TOY sound.
'Very well'
And what the this ABX give for validation value?
Close to 1.0 you were not guessing, I suppose
you maybe can recall this.
as you may not have saved the papers with test results.
Another question,
If for example LM3875 & LM3886 have 'toy sounds',
then maybe this is what we should be looking for.
As so many reviews tell they Sound Good.
Even compared with the most wellknown good designs using fully discretes.
Like Leach, AKSA, PASS etc.
destroyer X said:
Hehe, Carlos
You may not know this???
... but DX-Amp is one typical Power Operational Amp
- with transistors in discrete .. the normal Class AB output.
In fact DX-Amp is the very basic topology for op-amp.
Without much new thinking.
The old and good op-amp design. Not much more.
What blind tests ???
made you spot Op-Amps/Power Op-Amps have TOY sound.
'Very well'
And what the this ABX give for validation value?
Close to 1.0 you were not guessing, I suppose
you maybe can recall this.
as you may not have saved the papers with test results.
Another question,
If for example LM3875 & LM3886 have 'toy sounds',
then maybe this is what we should be looking for.
As so many reviews tell they Sound Good.
Even compared with the most wellknown good designs using fully discretes.
Like Leach, AKSA, PASS etc.
I have asked a friend his opinion about ABX versus normal listening evaluation. He very simply said: ''Why you wanna evaluate something in a way you don't normally listen to music?''. I said: ''For being fair''. He said: ''Is it fair to exclude the person? Its key part of the experience''. I thought... Fair is, I can only really recommend to those who know me. ABX is for NOT recommending to those who don't know me. It's just being politically correct in the end. Thus flat.
salas said:
I would absolutely agree to this view, provided that you start all your statements with "In my opinion...", "According to my taste...", "It is in my experience...", or any other introduction that excludes generalizations. As flawed as you may think, the ABX test results are at least statistically valid.
no two people hear the same thing the same way. if they did, then we could design a "perfect amplifier" and EVERYBODY would buy it. but there is no "magic bullet" amplifier, and that is why ABX tests are a good idea. you take a significant sample of people, eliminate the emotional variables, and have the people judge for themselves which amplifier sounds better. there is enough of a statistical cross section of the population to get a good basic idea how an amp sounds compared against another. elimination of the emotional variables is done by a true ABX test done double-blind, so no emotional information at all is conveyed to the listeners. while the ears and brain can do some wonders of signal processing, they adapt to and tend to ignore errors, especially if the error is continuous or gradual.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.