Transparent amplification

[amirm]

That's interesting, thanks. I wonder how much difference the lack of a back-emf generator would be in practice. I really don't know. I did wonder if taking a 'normal' difficult 'speaker crossover and substituting fixed resistors for the drivers as the circuit you linked to would be good enough an emulation.

S

I don't know either and unfortunately because of that one component, it may be difficult to replicate their their results.

Thinking out loud, I was wondering about using a real speaker for the test, but placing it in a box or far away as to it being silent for the purposes of testing.

 

[Blumlein 88]

http://audiojudgement.com/speaker-equivalent-circuit/

This short article on loudspeaker modeling has useful info about how to include back EMF.

 

[sergeauckland]

I don't know either and unfortunately because of that one component, it may be difficult to replicate their their results.

Thinking out loud, I was wondering about using a real speaker for the test, but placing it in a box or far away as to it being silent for the purposes of testing.

One straight-wire bypass test using loudspeakers I'm aware of, did just that, put the 'dummy' load in another room, well away from the listening room. Possible I suppose in an academic or research environment, rather more difficult at home.

I'm happy to go with measurements that indicate that the amplifier should be transparent, but clearly it's not proof, just expectation from experience.

S

 

[fas42]

That Swedish experience with the Bryston amplfier demonstrates that the test methodology is effective, even if not perfect. I've mentioned several times hearing the amp in question at the last audio show, and it was easily superior to anything else I've heard in unmodified form. In particular, it had zero problems driving a conventional speaker at intense SPLs, showing none of the normal compression and loss of quality that nearly all amplifiers exhibit when asked to do this.

But the real point is that the system has to be transparent, rather than just the amplifier. If a second amplifier is inserted in a normal chain, how does one know that the extra electronics are not unduly affecting at least one of the other components, say through the extra noise on the mains feed, from the additional power supply. Just saying that this shouldn't happen is not the same as being certain that such is the case.

Which is why the best method is going backward, when assessing SQ. What matters is every little thing that the rig is getting wrong - know what the sound should be like, and take note of all the failings of the system in front of one. How does one know what the sound should be like? It's a combination of all the best qualities that you have ever heard a particular recording show, extracted from every time you've heard it being played, on your systems or others - that's the actual, intrinsic nature of the recording. Which is why my ears pricked up on hearing the Bryston - it absolutely nailed the track being played, that I had heard so many times before.

 

[fas42]

Having said all that, I'm feeling a bit uneasy arguing that amplifiers may not be transparent, as I don't think any tiny lack of transparency in modern well-designed amplifiers will be noted when listening to music.

For actual listening, I suspect that the S/N-ratio might be the most important factor, plus providing sufficient power for dynamic peaks. When listening to very dynamic recordings with loudspeakers that are able to go loud in a quiet room - which I will do when I move to a new apartment in February - the noise floor of the electronics becomes one of the limiting factors. I'd like a S/N ratio of at least 100 db for the first watt. So my requirement is essentially that the amp should provide lots of power, and also be very quiet. A difficult combination for many amps. Modern class d might be the best bet for achieving that, even though those Swedes don't think it's a technology which is completely transparent at this point.

Sorry to disappoint you, but every "tiny lack of transparency in modern well-designed amplifiers will be noted" - they're much better than a couple of decades ago, but still quite obviously show problems when carefully assessed. Yes, you can call it a S/N problem - the system overall creates a masking, blurring layer of 'noise', which hides critical detail in the recording - when you evolve a rig to be devoid of this usual crud, then true "transparency" is finally heard.

Lots of power? A very useful first grab, but no guarantee. Very quiet? Mechanically of course, and no audible hiss, on maximum gain - pretty easy these days; the killer is signal modulated 'noise', crud that is generated typically at greater and greater levels as the amplifier is asked to work harder - the absence of the latter is a true revelation, when heard.

 

[Wombat]

Ideally the amplifiers should be tested using a variety of sources and loudspeakers to allow for input and output interface impedence matching affects. Interconnects and loudspeaker cables could also be added to the variables.

I think we are able get a damn good idea of an amps performance in-situ by simply attenuating the output level to match the input level and then doing an (A-B), (B-A) and (A-B)-(B-A) signal summation.

 

[Bjorn]

I think it's the same thing, really. Sean Olive showed in his listening tests that the Harman employees always preferred their own high end speakers over the speakers of the competitors - but this preference was much stronger when the comparison was done sighted. Unsighted, the mean preference for the Harman speakers was subsantially diminished. Similarly, I would assume that the preference for Ncore amplifiers among objectivists to be substantially stronger than it would have been absent the knowledge about how it measures. It may be that people would have preferred Ncore anyway. But I'm pretty sure that the knowledge that it measures better than most other amps conditions people to like it even more.

I'm not quite following you logic. You're referring to a preference test, but what I'm talking about is transparency based on objective methods (measurements). That's very different.

Also: Assume that the LTS engineers are right that there is something in class d amps that don't show up in measurements (because of the filter that gets applied to measurement gear in order to be able to measure class d, as I've understood it). Assume that this becomes known in systematic tests. I am very sure that people then will start to "hear" these things, which they didn't hear before.

I don't see the point in assuming and speculating. Bring the goods on the table.

This is all speculation, of course. But I think we should be honest enough with ourselves to acknowledge that listening bias goes in all directions. It's not only about the flowery language of the audiophools, listening bias can also arise from knowledge about measurements and objective performance.

Of course we're all proned to bias. However, what you're basically indicating here is that LTS knows something the rest of the industry doesn't know and based on you assumption that they have testet a Ncore amp that are supposedly considered transparent by certain objectivists. A lot of speculation.
Actually I think it would be fun if was to be true and we could learn something new. But considering the measuring methods we have today and the understanding of what's audible, I think the chances are more than slim.

However, it should be noted that it's not a problem to get a Ncore amp to add audible distortion. So even testing one wouldn't necessarily give us the answer. The input buffer, the power supply and even the cooling are examples of what effect how it performs. If a Ncore amp is pushed with a very heavy load and cooling and power supply isn't up to match it, the distortion will certainly raise.

 

[Bjorn]

I don't know either and unfortunately because of that one component, it may be difficult to replicate their their results.

Thinking out loud, I was wondering about using a real speaker for the test, but placing it in a box or far away as to it being silent for the purposes of testing.

Interesting idea.

I believe the major challenges of such test is the following:

1. The acoustics of the room
In many ways it's quite ridiculous of us (including me) to talk about the transparency of electronics before we have removed at least a minimum of room anomalies. The "distortion" of the room is simply much higher and will mask subtle differences in electronics unless it's dealt with. Both in low frequencies and the area above the Schröder.

It's worth mentioning the example here that when the LEDE acoustics principle was developed they found that phase issues from speakers became very audible in such environment.
So the question of how good does the acoustics needs to be, becomes very relevant.

2. Trained listeners or not
Without some training before doing a test, it's very difficult to distinguish smaller differences. When I've just jumped into to a blind test with MP3 of 256 or 320 kbps vs lossless I almost always fail the test. But if I train my hears some before the test and know what to listen after, I've been able to pass the test.

3. Music material
If one is going to hear the differences, the right music material needs to be chosen. I personally have chosen music after much listening and picked out the material where I felt I heard differences. Seldom this is in the beginning of a song. And I also found it crucial to listen to only a very short passage (4-10 seconds). Any longer than that is difficult both because the audio memory is short but also due to the fact that the material may not reveal differences of much longer than a few seconds.

While some of the challenges are also the case with measurements set up against listening tests with headphones, I personally think the latter is a safer way. Listening tests with headphones over years with trained listeners should give us a very good indication of the audibility. Some few listening tests in a room with maybe not so optimal acoustics is more likely to have weaknesses.

But hey, I think it's still a great initiative! Nothing needs to be perfect.

 

[oivavoi]

I'm not quite following you logic. You're referring to a preference test, but what I'm talking about is transparency based on objective methods (measurements). That's very different.

Of course we're all proned to bias. However, what you're basically indicating here is that LTS knows something the rest of the industry doesn't know and based on you assumption that they have testet a Ncore amp that are supposedly considered transparent by certain objectivists. A lot of speculation.
Actually I think it would be fun if was to be true and we could learn something new. But considering the measuring methods we have today and the understanding of what's audible, I think the chances are more than slim.

Bjørn, I don't think you're doing a very good job of trying to understand my point now...

There are two different issues here. I'll try to say it again.

1) The first is whether the Ncore is transparent or not. I have no stakes in this. I was simply referring to what engineers associated with the LTS have stated on Swedish forums. They have said - several times - that they don't think Class D as a amplifier class has matured to the point where it is transparent, with no exceptions. They have said that they have looked at many well-regarded Class D amps which were said to be transparent, and deemed them to be non-transparent. They also have a policy of mostly writing about products they think are good enough, and don't see their job as putting down products they don't like. I think that's regrettable, though, and several people on Swedish forums have encouraged them to publish the results from non-satisfactory results as well. But so far, they stick to their policy of trying out products, and only continuing to write about the products they think are good enough.

So have they tested the Ncore or not? I don't know. If they have, it is also difficult to know whether their initial testing was systematic and blind. I assume that even the LTS listeners have bias, and if they are biased against class D (which they seem to be) it may very well be that the initial sighted listening of Ncore (or other class d amps) yielded "non-transparent" results that would disappear in blind listening. That's why I wrote about this here: I hope other people would do similar tests. The reason is that I have great respect for the LTS and their testing, and I find it very intriguing that they hold opinions in this regard which as you say run counter to most of the industry.

2) The second point is about bias. This point is actually very straight-forward. When people judge subjectively whether an amplifier is "transparent" or not, the same bias mechanisms are at play as with preference. Nobody has a magical connection to the original material they listen to. Any claim that they perceive amplifier A as more transparent than amplifier B - in the absence of systematic before/after tests - is a subjective judgment which rests on shaky grounds. The fundamental bias mechanisms that always operate in humans will - obviously I would say - influnce such judgments as well. If they know in advance that an amplifier has certain measurable characteristics, this will bias their perception into perceiving it this way.

Is this really a controversial point to make?

 

[Bjorn]

oivavoi:
You're basing everything on your assumptions and pure speculations. There's no proof in the pudding here.
Most of the reviews from LTS are negative by the way. Very few of their published tests come out as being transparent. I know since I've read the magazine for several years.

Combined with knowing well how a Ncore based amp built properly can measure, I find your speculation quite controversial.

FIY: I'm planning to send our coming Ncore amp to LTS for a before/after test. As long as they want to test it of course.

 

[oivavoi]

oivavoi:
You're basing everything on your assumptions and pure speculations. There's no proof in the pudding here.
Most of the reviews from LTS are negative by the way. Very few of their published tests come out as being transparent. I know since I've read the magazine for several years.

Combined with knowing well how a Ncore based amp built properly can measure, I find your speculation quite controversial.

FIY: I'm planning to send our coming Ncore amp to LTS for a before/after test. As long as they want to test it of course.

Sigh. You're really not a good reader in this thread Bjorn. "Controversial" referred to the second point about bias. Making this point is NOT controversial. It is just about pointing out a fundamental feature of human perception which applies to every human being on the planet, objectivist or not.

With regards to LTS, I am just referring to what I have seen people associated with them state on Swedish forums. That is all.

EDIT: I really respect that you intend to send your amp to testers such as Amir and LTS, btw. Hats off.

 

[Bjorn]

Sigh. You're really not a good reader in this thread Bjorn. "Controversial" referred to the second point about bias. Making this point is NOT controversial. It is just about pointing out a fundamental feature of human perception which applies to every human being on the planet, objectivist or not.

With regards to LTS, I am just referring to what I have seen people associated with them state on Swedish forums. That is all.

I'm reading what you're saying but simply don't agree with you reasoning.

First you're indicating that Ncore may not be transparent and bases in on your assumption that LTS has tested it, but the fact is that you really don't know if they have. So pure speculation. And based on the assumption that Ncore may not be transparent, you start to talking about bias and that objectivists that believe Ncore is transparent also can fall into this with examples of listening tests. Well, of course they can. That's nothing to discuss. We are all proned to bias as mentioned, that's exactly why we blind test. However, the objectivists believe Ncore is transparent due to how it measures and what we know about the audibility, which is again based on blind tests.

If you're trying to make a point that all are being biased, there's nothing to discuss. We already know that. But I wouldn't say that believing in measurements together with multiple of blind tests is really biased. You can't really get more objective than that!

Speculating that these measurements with our knowledge of audibility is bias based on a belief that LTS has done a test you don't know they have done, thus having knowledge that rest of the industry doesn't have become moot IMO. I think we should leave the speculation on side till we know something for certain.

 

[oivavoi]

Bjorn, one last try: You misunderstand my points. So I'll try to say it again. But please put some more effort into attempting to undestand what I'm saying, ok?

When I started to talk about bias, it was because you introduced the point that people are subjectively perceiving even Hypex UCD amps to be more transparent than Bryston. My point is that these are not good data points, because perceptions of what is transparent or not are prone to bias. I am NOT saying that the perception that Ncore is transparent IS biased or wrong. I think this is what gets you going here. But this is not what I'm saying. I'm just saying that subjective perceptions of transparency don't hold much weight at all, so they should be left out of the discussion.

The point about LTS: Again, I am NOT "indicating" that Ncore may not be transparent. This is your own (mis)interpretation. I wrote that LTS engineers have claimed that class D as such - with no excpetions, so it includes Ncore - is not ready for prime-time yet when it comes to transparency. As I wrote early on, I don't know whether this is correct or not. It is above my paygrade, as I'm far from being an audio engineer and I don't regard myself as an expert listener. My hunch would if anything be that the LTS guys are mistaken in this regard, especially when it comes to the Ncore. My simple point was that I would love to see other people employing the same tests as LTS for listening to Ncore (and other class D amps), since I have become intrigued by the fact that LTS seems to hold opinions in this regard which are fairly contrarian in the world of hifi.

 

[Bjorn]

Bjorn, one last try: You misunderstand my points. So I'll try to say it again. But please put some more effort into attempting to undestand what I'm saying, ok?

When I started to talk about bias, it was because you introduced the point that people are subjectively perceiving even Hypex UCD amps to be more transparent than Bryston. My point is that these are not good data points, because perceptions of what is transparent or not are prone to bias. I am NOT saying that the perception that Ncore is transparent IS biased or wrong. I think this is what gets you going here. But this is not what I'm saying. I'm just saying that subjective perceptions of transparency don't hold much weight at all, so they should be left out of the discussion.

I've read your posts again, but you didn't refer anywhere to this being a point about the UcD vs Bryston comment.

Obviously the UcD subjective comment is purely subjetivism, so no disagreement there. But you can't expect me to understand something that isn't communicated properly

 

[oivavoi]

I've read your posts again, but you didn't refer anywhere to this being a point about the UcD vs Bryston comment.

Obviously the UcD subjective comment is purely subjetivism, so no disagreement there. But you can't expect me to understand something that isn't communicated properly

Ok, I see. Well, then I didn't express myself clearly enough! Peace out

 

[oivavoi]

I'm just saying that subjective perceptions of transparency don't hold much weight at all, so they should be left out of the discussion.

Addendum to my own post: this came off too strong. I do think that subjective listening impressions hold value, they may tell us something. Persponally, I trust what my ears tell me concerning sound reproduction in my own home. If I like something, I will include it in my system. If I don't like it, then I won't. This is also because I think that a bias/preference, once formed, for any reason whatsoever (the nice color on the speakers for example), is likely to become sticky. So if I have formed a preference for something, I can just as well go with that, even if it's based on non-objecetive reasons.

If a lot of people have perceptions which point in the same direction it may indicate something about objective reality.

The problem is just that subjective impressions can be all over the place. In the case of Ncore, for example, there are many people who like them very much and/or perceive them as transparent. But there are also people who report that they prefer other amps and/or perceive them as not 100% "right". Who is right then?

Our best bet, I think, is to strive for linearity as a design goal as far as possible. And include simple tone controls in the final system for adjusting the result to one's taste. In the case where there's scientific/rational disagreement over whether a certain technology is indeed linear or not, then systematic listening tests which investigate transparency may be needed.

 

[amirm]

While some of the challenges are also the case with measurements set up against listening tests with headphones, I personally think the latter is a safer way. Listening tests with headphones over years with trained listeners should give us a very good indication of the audibility. Some few listening tests in a room with maybe not so optimal acoustics is more likely to have weaknesses.

That was my idea, to use headphones. Using closed back ones would help in not hearing the "dummy" speaker playing elsewhere.

 

[fas42]

Interesting idea.

I believe the major challenges of such test is the following:

1. The acoustics of the room
In many ways it's quite ridiculous of us (including me) to talk about the transparency of electronics before we have removed at least a minimum of room anomalies. The "distortion" of the room is simply much higher and will mask subtle differences in electronics unless it's dealt with. Both in low frequencies and the area above the Schröder.

It's worth mentioning the example here that when the LEDE acoustics principle was developed they found that phase issues from speakers became very audible in such environment.
So the question of how good does the acoustics needs to be, becomes very relevant.

This shows that a) people listen for different things, and b) most people have never experienced a system that projects an illusion so strong that it completely dominates the listening area - the "distortion" of the room becomes a tiny, tiny pimple when this happens - and the difference between one status of the electronics, and an alternative one, screams at one.

 

[Bjorn]

The old hifi way seemed to be finding you favorite coloration from electronics to suit your speakers, acoustics, mood of the day and music choice.
I believe this idea is outdated to large degree and EQ and harmonic distortion added combined with transparent electronics is a better option. But audiophiles seem to enjoy the hunt for a new box. The thrilling excitement that often doesn't last long.

Also, I believe the need to add coloration will greatly be minimized when people get speakers that actually measures well in the important areas.

 

[watchnerd]

Can't they just do a null test like the old Bob Carver Stereophile challenge?