An interesting thread regarding amplifiers.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I used to think that most amps sound different, but not to the extent that it could make or break a listening experience. ie. my official stance was that "Yes there's a definite difference, but that's only with A/B switching. In the long term my ears would probably get used to it, so there's no big deal"

It some ways that sat well with me because i wasn't really keen to spend big bucks on amplifier- with an electronic engineering background and having a friend who works at electronics wholesaler, I know that these things cost peanuts to make, there's no new design or R&D to speak of (class A, A/B amps) and I didn't want to spend thousands of dollars just because it was "brushed aircraft aluminium" or "retro-modern" exterior.

But with more experience in equipment and listening, now I really think that with some speakers, the amp can be a make-or-break factor. Perhaps not as important as the acoustics of the room, but there can be extremes- some amp/speaker combination which is unlistenable after about an hour (or having to turn the volume down, when it's not very loud to begin with), and switching amps, and all of a sudden having an involving and emotional musical experience.

I still don't know why or how that happens, and what amps will suit which speakers- it's a real black science/art?

I think double-blinded testing is valuable tool, but we should keep in mind that it's not the be-all-and-end-all of testing. With any kind of test, we have to be careful when judging the results for external validity. That is, just because one can/cannot reliably detect an audible difference in situation A (eg. under some particular controlled test conditions, doesn't necessarily mean the same will hold true for situation B (under relaxed conditions in one's own home.
The same is true for the opposite situation- just because I can hear a difference in in my own home at my own enjoyable volumes doesn't mean I should be be able to detect a perceived difference/better/worse in situation A.

The next leap of faith is that "Well I couldn't hear a difference in a double blind test, so well then must be no difference"

This idea was thrust on me when I studied a bit of statistics at the undergraduate level. The stastical jargon is called "internal validity" - how valid, or correct, a particular test results are; versus "external validity"- given a test with internal validity, the degree to which the results can now be generalised for broader situations or populations. That is, validity is two different things- and given a test which has internal validity, we cannot merely assume that it has external validity.

The other thing with listening is that it's a pretty poorly developed sense. At least in humans. The difference in sound is not imediately apparent, even if there a TRUE difference.
So asking a participant whether he/she can detect for, or prefer some true difference is quite a tough test.

From first hand experience I can say that detecting subtle differences in sounds are much easier when subjects have been are TRAINED- ie. they've heard or know what the differences ought to be.

eg. As a student, it's very hard to discern between sounds heard through a stethoscope. The minute differences in volume and subtle differences in character are next to impossible to discern. Only after being told what they sound like, then it's a bit easier (but even then, not always possible) to discern-

Teacher: "Can you hear that?"
Student: "No"
Teacher: "It's very faint."
Student: "Hmmm"
Teacher: "It's a kind of rustling noise"
Student: (thinking of and listening for leaves in the wind) "Err..."
Teacher: "The kind of sound you get when you touch your hair"
Student: {moves stethoscope to another position to try again}
Teacher: "Hold your breath, you should be able to hear it"
Student: {now feeling very embarrassed} "Oh ok!"

A person who has not been trained to hear the differences will have difficulty detecting subtle differences.

I'm rambling but the short version of my story is that
Being able to hear a difference in a test is not a reliable predictor whether one can hear a difference in an in-home (or other) listening sitation. And whether one can hear a difference in the comfort of their home doesn't mean one will be able to detect it in some kind of test. Drawing conclusions like "There's vey little difference in amps" based on "I couldn't hear it in a double blinded test test sitation" is a false conclusion.The only thing you can conclude is that you couldn't hear it in that particular test"

To detect subtle differences you need training.

YMMV.

regards,
Thanh.
 
I have been designing amplifiers for some 17 years now, and I believe I can make a very good case for the fact that it is far easier to design a specific amplifier for a specific task (or system type) than it is to design the perfect all-round amplifier. For example, I design both valve and transistor amplifiers, and although in a general way the valve amplifier is better (lower distortion, better transients etc.) it will only work better when coupled to a system with the correct driving impedance, which means both D/A converter output impedance and volume control. For this reason it was necessary to design matching D/A + volume solutions for both the valve amp and the transistor amp - although all the equipment will function either way, coupling the correct equipment (in the correct way, using good quality interconnect) will always work better than the wrong way round (especially trying to drive the low-impedance transistor amplifier with the high-impedance valve D/A, for example).

The point I am trying to make is that far too often end-users make all kinds of assumptions about equipment but may be coupling in inappropriate pairings, resulting in substandard setup. Without at least a moderate understanding of the design parameters of given equipment, it is unlikely that you will arrive at appropriate matching by luck. Given this fact, many amplifier or D/A appraisals are not worth much.

Finally, the greater majority of "audiophiles" tend not to understand the importance of certain forms of source material (not only recording quality, but also instrumental content). Therefore, many comparisons are made using source material that does not prove much.
 
tktran said:
I still don't know why or how that happens, and what amps will suit which speakers- it's a real black science/art?

Thanh, I recall you saying that your L18 system sounded better with one amp than the other. That system has a rather large impedance swing right in the midrange. It didn't really affect much with the 3 amps I tested with, but I can't help but wonder if some amps are audibly sensitive to that. (and I mean some solid state amps - I don't design for tubes) So, I've been considering posting an optional compensation network that flattens the impedence curve to almost a straight line. What I'm about to ask may interest you also.

[email]clive@tsrg.net[/email] said:
I have been designing amplifiers for some 17 years now...

That makes you just the person to help us out in this discussion. :) See above and let us know what you think. Note that I'm a member of the "Amps Don't Make Much of a Difference Club" but my ears are open to ideas.
 
The thread at Madisound is the same Great Debate I've been reading, over and over, since 1975. Some people hear valuable differences in the sound of SS amps, some don't.

What bothers me about ABX testing is that it advocates the substitution of an artificial set of circumstances (i.e. a double blind test), for reality. "Reality" is me, listening long-term, to my system, in my room, with my music. If I can hear a difference under those circumstances, then it matters, because it's reality. If I can't hear the same difference in a DBT, the the test is flawed. It's as simple as that.
 
Ps. I took a quick glance at the Madisound thread. I'd just like to mention that Ken Kantor is one of the industry professionals that I respect the most. In fact, the last commercial speaker I ever bought many years ago was an NHT from back when Ken was in charge there.

So it's no surprise that my views almost perfectly match Ken's.
 
audiobomber said:
..."Reality" is me, listening long-term, to my system, in my room, with my music. If I can hear a difference under those circumstances, then it matters, because it's reality. If I can't hear the same difference in a DBT, the the test is flawed. It's as simple as that.
You can do double-blind tests in your own home, or under any circumstances you so desire. As long as neither you nor the experimenter know what you are listening to then it is double-blind.

Merely removing the knowledge of what is being listened to will have an impact on the perceived sound. I think it's a shame that there is not serious study done to correlate the appearance of equipment with its perceived sound - it would be most enlightening (I'm sure it would prove that my desire to fill everything with LEDs has a scientific basis!).
 
Just a small addendum. It is unfortunately true - possibly due to economic reality, possibly due to lack of knowledge, maybe other factors too - that the average amplifier out there is quite compromised, and that very often a comparison serves only to compare one unfortunate compromise with another, different one. It is true that there are amplifiers that are much better than certain others, but again this does depend on the system you are hooking them in to.

All this said, the comment above about the 'real test' and personal judgement is very, very valid - if the equipment does not do it for you, in your room, what is the meaning of 'better'?
 
There is and will always be a personal validation required for the purchase of any esoteric device and audio has been at the forefront of this. It is one of the reasons many folks are intimidated & do not enter this great hobby/vocation. And that is the real shame of it.

I have been involved with audio for just over 3 decades, and amps have always been my favorite piece of the puzzle. That is odd, as they have the least influence over the final sound delivered to the listener. The source & acoustic output devices, due to their transducer requirement, are the most critical, other than the acoustics of the environment. The conversion of mechanical to electrical & then back is very challenging. At least the newer digital recording techniques are doing away with one of those problems. The latest A to D conversion devices are very good. And no matter what anyone says, I am grateful for not having to set up turntable, arm & cartridge and treat/dust my records. They were too much work. They can sound great, but a CD is good enough for me. I have high hopes for the next audio media, it will surely outperform any analog available.

I was an EE in college, as long as I lasted, and now I work in pro audio/video system integration. I deal with high end balanced equipment everyday. In my youth, I was trained as a classical and jazz musician. I have tasked talented musicians to find the golden ear details that are promulgated about with abandon and they have a hard time hearing this stuff. Many ask why. That is a valid point as the music is the thing. I love a great listening experience. I was fortunate enough to have recently heard the Vienna Symphony Orchestra @ Stern Hall @ Carnegie (I do work there) and I was in awe of the sound.

That is another point I like to make with hardware obsessed people. What is your benchmark? Only witnessing live, un-reinforced music gives you that. When you do start to listen at events such as these, it becomes very apparent that acoustics is the key. That leads me to another point.

The acoustic environment is the most critical task to overcome. That is why important music venues spend so much on acousticians and implementing their recommendations. They are magicians and have a black art all of their own. That is one of my current technical passions.

After you deal with the source & output transducers, there is the small detail of the wire with gain we call amplifiers. Valve amps have much higher distortion numbers. We all know that. They do sound great due to how they distort and the psychoacoustics of their performance. Solid state amps are better in every way. But they are clean until they start to distort and then they sound very poorly very quickly, for the opposite reasons of why bottle amps sound better in this situation. A well designed amp, of any type, has certain intrinsic inductive and capacitive reactance issues with the speakers and to a lesser extent the cables that carry this signal. We all know this and the subtle differences between these 3 devices are what this thread is about. There is little new in this field. The newer digital amps are about it. The designs from AT&T labs from 60 years ago still hold true for sand amps & the classic designs from the dawn of electronics still hold true for bottle amps. The facts are well known about how all of these devices work and this website delineates this information very well. There is a cornucopia of talent here, and I always learn something while tooling around here.

But let’s call a spade a spade and amps are what they are. If you want to play, go for it. If you took a peek @ the professional world of audio recording and live performance, most golden ears would be shocked at the fact that the NE5532 passes most of the audio you have heard in the past couple decades. And none the worse for it, IMHO.

Thanks for the great read and info. See you folks @ AES in October.
Warren
 
Warren,

In principle I agree with much of what you say. The issues of live performances (I have attended literally thousands), acoustic environment (mine cost 3 times what I have spent on my electronics and speakers, which is considerable) and pointless comparison (listening to music is, after all, what we are attempting to do, not?) are all extremely valid.

There are a few things I would like to mention, though.

First, the fact is that valve amplifiers can be designed to be extremely distortion-free. Although my semiconductor amplifiers are better than most (as in significantly lower in distortion of every kind) my valve amplifiers are even better. Most valve amplifiers distort because their designers either desire this or do not know how to design. Designing distortion-free valve amplifiers (mine deliver harmonic distortion 62db below the signal level, totally without feedback) is very difficult indeed, and generally not attempted.

Second, op-amps are not as good as they are thought to be. An average op-amp (like a 5532), in an average circuit, will deliver quite poor results as compared with carefully designed discrete componentry in an appropriate topology, and the difference will, indeed, be audible under correct circumstances. To provide a concrete example of this : one of my regulated power supplies requires a comparator as one of the functional elements (no surprise there). What is surprising is that the very best op-amp money can buy delivers very average performance in terms of load regulation in this topology, whereas the discrete circuit I now use (constructed solely of transistors out of desperation when all op-amps tested would not shape up) delivers amazing regulation - approximately 20 times better, at a fraction of the cost. This is not an "I think this sounds better" issue, but concrete, clinical measurement of regulation parameters. Interesting, no?
 
[email]clive@tsrg.net[/email] said:
...and I believe I can make a very good case for the fact that it is far easier to design a specific amplifier for a specific task (or system type) than it is to design the perfect all-round amplifier.

Yes, indeed.

[email]clive@tsrg.net[/email] said:
Finally, the greater majority of "audiophiles" tend not to understand the importance of certain forms of source material (not only recording quality, but also instrumental content). Therefore, many comparisons are made using source material that does not prove much.

Amen. :angel:
 
That is another point I like to make with hardware obsessed people. What is your benchmark? Only witnessing live, un-reinforced music gives you that. When you do start to listen at events such as these, it becomes very apparent that acoustics is the key. That leads me to another point.

Hmm, I wouldn't exactly say acoustics is the key. It's definitely better to hear unamplified music in a well designed space, if you can, but what about other sounds?.

When I walk outside my door I still know and love the sounds of cars, birds, leaves rustling etc. I am as fond of every day sounds as I am of music, in the ever changing acoustic space of the world at large.

As I've said before, Stereo recording is and will probably remain for some time, the biggest problem in accurate audio event reproduction. Other measures may get us closer to the music, but none will bring such a radical step up as the total disbanding of stereo recording altogether. Unfortunately, I can't see this happening any time soon...
 
derf said:


Other measures may get us closer to the music, but none will bring such a radical step up as the total disbanding of stereo recording altogether.

Just curious, aside from attending live performances (which is rather a hit and miss affair; when will I have the next opportunity to see a live performance of Bach's Goldberg variations? Especially by Gould, who is now dead?) what would you suggest as a means of recording / reproducing music?

I was sort of under the impression that good stereo is pretty fantastic - remarkably similar to the real thing, I thought?

I am at the moment listening to du Pré (also now dead) playing the Elgar cello concerto with the London Symphony (situated on another continent) conducted by Sir John Barbirolli (dead) recorded nearly 40 years ago, and the performance is stunning and really enjoyable.

Not trying to be sarcastic, just curious what the alternatives would be?
 
[email]clive@tsrg.net[/email] said:


Just curious, aside from attending live performances (which is rather a hit and miss affair; when will I have the next opportunity to see a live performance of Bach's Goldberg variations? Especially by Gould, who is now dead?) what would you suggest as a means of recording / reproducing music?

I was sort of under the impression that good stereo is pretty fantastic - remarkably similar to the real thing, I thought?

I am at the moment listening to du Pr?(also now dead) playing the Elgar cello concerto with the London Symphony (situated on another continent) conducted by Sir John Barbirolli (dead) recorded nearly 40 years ago, and the performance is stunning and really enjoyable.

Not trying to be sarcastic, just curious what the alternatives would be?

As different cars will feel different based on the drivers experience and skill, so will differences in audio judgement. From a listerner point of view, if the enjoyment of music gives you the feeling that you are in that live performance, you will know if you really enjoy live performances very much.
 
what would you suggest as a means of recording / reproducing music?

I'd suggest a Soundfield microphone(my avatar), UHJ(horizontal surround) or even better B Format(surround sound with height) recorded onto a suitable digital medium, then reproduced with an Ambisonic decoder(I'll explain this further a bit later)

I was sort of under the impression that good stereo is pretty fantastic - remarkably similar to the real thing, I thought?

It can provide *some* of the cues of live music, but there will always be the phase problems from mic spacing, also the incorrect mapping of sounds from all directions to two front speakers. Alan Blumlein aswell as Bell Labs knew this as early as the 30's, yet settled on Stereo as being "close enough" and convenient(only needs two speakers, formats of the time were mono, so when stereo vinyl came along in the 50's?, only then could it be implemented etc)

I am at the moment listening to du Pré (also now dead) playing the Elgar cello concerto with the London Symphony (situated on another continent) conducted by Sir John Barbirolli (dead) recorded nearly 40 years ago, and the performance is stunning and really enjoyable.

Being a classical fan, you've really lucked out. The way most classical concerts are mic'ed is by hanging a stereo pair in front/above the orchestra, this is a very good way to accurately capture the sound of the orchestra/concert hall(in stereo). Add to that the fact that the music you listen to will have gone through no/very little processing(eq, compression etc), I'm sure your collection is quite possibly among the most accurate reproductions stereo recordings have to offer.

Unfortunately, a lot of other music(pop,rock,some jazz) is pan-potted close-mic'ed mono in disguise as stereo, which has none of the cues of live/unamplified music. Add to that possible eq, compression etc and it becomes a sad state of affairs, as far as "recreating the event is concerned".


Not trying to be sarcastic, just curious what the alternatives would be?

The alternative/better would be Ambisonics.

Developed mostly by Academics this side of the pond, it's a way of capturing sound accurately within 3 dimensions(front/back,left/right and up/down) and then mapping it correctly to your speaker setup(regardless to some extent of speaker position, you can tell the decoder where the speakers are and it will compensate)

Even though it's been around for 3 decades, very few people have even heard of it. Mainly thanks to the people entrusted with getting Ambisonics out there, they were really only equipped to find a sole company to market Ambisonics, when what it needed was many, many companies getting involved.

As proud as I am to know it was mainly a British thing, if it had been in the hands of the Americans or the Japanese, the situation could of been a lot different and many of us might be enjoying the superiority of Ambisonic recording/reproduction.

A few links for enquiring minds:

Ambisonics main site

A short introduction to Ambisonics

A very good indepth, but not too heavy article, reccomended reading for all

Soundfield microphones website

Funny, I didn't latch on to any sarcasm in your post ;)
 
As different cars will feel different based on the drivers experience and skill, so will differences in audio judgement. From a listerner point of view, if the enjoyment of music gives you the feeling that you are in that live performance, you will know if you really enjoy live performances very much.

Hmm, I'm not sure I fully understand your post, but I think I know what you're trying to say....

Stereo recording/reproduction can be enjoyable, some may even prefer it to Ambisonic recording/reproduction, it's ultimately down to them and their preferences.

What isn't down to preference though is reality/fact. The reality of the situation is that an Ambisonic setup will provide more cues, better localisation and ultimately a better representation(reproduction is perhaps too big a word for our current state)of an audible event. Just as taking a paracetamol will cure a headache faster than eating a boiled newt, I doubt anyone here would baulk at such a statement, as it's an established fact.

Audio recording/reproduction is no less of a science than producing paracetamol, all the mysticism and magick some people like to pump into audio, cannot detract from fact.

Fact cannot be changed by preference. Everyone can continue to have their preferences, just please don't refer to things as "matter of fact", if they aren't "matter of fact".
 
derf said:
I'd suggest a Soundfield microphone(my avatar), UHJ(horizontal surround) or even better B Format(surround sound with height) recorded onto a suitable digital medium, then reproduced with an Ambisonic decoder(I'll explain this further a bit later)

You seem to be spreading your 'ideas' as if you have the divine truth.
If you enjoy all that parafernalia... doesn't mean hat others HAVE TO.
Stereo can be FANTASTIC, on a good system.
It has height, depth, wide soundstage, a clear, focused center stage...
Yes, it can be very similar to a live performance, which can be unamplified or amplified, on a closed space or on free air.
I suppose you are after the sound reflections of the trees too? :cool:

Also, the more channels you add to a system the more compromises you have, and this goes down to every considerations, including internal layout of source components, preamps, power amps.
Two channels is enough to hear with very good realism, and it captures the ambience of the room or even the studio.

Listen to Jeff Buckley's 'Grace' CD.
On track 4 ('Lilac wine'), it starts with voice and a simple accoustical guitar.
When Jeff sings louder, you can hear an echo of his voice on the studio.
But you can only hear that reverberation on a very transparent system.
Go and test it.
No need to inject reverberations and fireworks all around, it's all there.

:cool:
 
Talking about amplifiers in a Loudspeaker section is weird enough, but getting into recording techniques in an amplifier thread in a loudspeaker section is more difficult to understand. I'm interested in recording techniques, but this does not seem to be the place for it.

derf, maybe you care to open a recording techniques thread?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.