@ vacuphile,
the simple answer would be, " we are measuring the same things as ever, with better measurment gear thus higher resolution, but these measurements obviously do not fully reflect the way our perception works"
Of course a good amplitude linearity is mandatory and low THD+N too, but beside that the relationship between perception and usually measured numbers becomes more and more "unstable" .
An additional factor is the somewhat artifical arrangement for measurement which does not really reflect the normal usage. As mentioned already often in the cable discussions, measuring a single amplifier with an audio analyzer on a bench with transformer coupled inputs (and maybe isolated mains connections) is a quite different condition.
Some combinations with other preamplifiers, loudspeakers and sources might be less ideal than others.
But, despite the overall good measured performance, the slope of the THD+N curves in the upper frequency region could be a hint why the results in real world setups might be different.
the simple answer would be, " we are measuring the same things as ever, with better measurment gear thus higher resolution, but these measurements obviously do not fully reflect the way our perception works"
Of course a good amplitude linearity is mandatory and low THD+N too, but beside that the relationship between perception and usually measured numbers becomes more and more "unstable" .
An additional factor is the somewhat artifical arrangement for measurement which does not really reflect the normal usage. As mentioned already often in the cable discussions, measuring a single amplifier with an audio analyzer on a bench with transformer coupled inputs (and maybe isolated mains connections) is a quite different condition.
Some combinations with other preamplifiers, loudspeakers and sources might be less ideal than others.
But, despite the overall good measured performance, the slope of the THD+N curves in the upper frequency region could be a hint why the results in real world setups might be different.
Last edited:
None of them. As usual, you're attributing statements to me that I never made.
Dear Sy,
I attributed nothing to you. I asked a question to clarify your statement.
Sy: ".... the measurements .... quite nicely show the capabilities of this amp."
T: "Can you actually tell me which of these measurements is the one that shows that the Amplifier will sound good?"
Sy: "None of them."
To me at least now everything is very clear and there is no possible misunderstanding.
Ciao T
POST SCRIPVM: the .... above indicate cutting out a sub-section of the statement that is superflous to the point, however to avoid being accused of editorialising, here the full quote:
I attributed nothing to you. I asked a question to clarify your statement.
Sy: ".... the measurements .... quite nicely show the capabilities of this amp."
T: "Can you actually tell me which of these measurements is the one that shows that the Amplifier will sound good?"
Sy: "None of them."
To me at least now everything is very clear and there is no possible misunderstanding.
Ciao T
POST SCRIPVM: the .... above indicate cutting out a sub-section of the statement that is superflous to the point, however to avoid being accused of editorialising, here the full quote:
This is IMHO very well said. THD and IMD measurements cover very long time period of settled non-changing signal. The ear is most sensitive to transients, to the rising part of transients. Thus the methodology of THD and IMD is wrong regarding human perception. And it does not explain differences in sound of amplifiers at all, as the speaker distortion is about 20x higher at any level than that of well engineered solid state amplifier. Only some tube amps and poorly designed solid state have higher non-linear distortion than the speaker.
The only reason why THD/IMD is so popular is that it is easy to measure. Engineers love simple measurements that can be simply evaluated and compared. But, the ear is the judge and engineers tend to overlook principles of human hearing, or they rather do not know it at all and simply do not care. They usually use the equation lowest THD = best amp. Not many changes since 1970-ties.
What was unclear? The output replicates the input to a high degree of accuracy even at the high powers needed to show any artifacts. The power is high enough that the amp is unlikely to clip. Frequency response does not vary with load.
Dear Sy,
To some the term "capabilities of an Amplifier" may be taken relating to the resulting sound quality. I wanted to clarify if there was anything of the like in your statement.
And I thank you that you made clear that for you "capabilities of an Amplifier" does not relate to sound quality.
Ciao T
To some the term "capabilities of an Amplifier" may be taken relating to the resulting sound quality. I wanted to clarify if there was anything of the like in your statement.
And I thank you that you made clear that for you "capabilities of an Amplifier" does not relate to sound quality.
Ciao T
FWIW, and if anyone cares - I've measured amps that have a changing harmonic structure depending on frequency and impedance. E.G. the forth harmonic might vary depending with load while the others don't change much. H3 might change more depending on frequency than other harmonics.
Not things that show up in the typical FR & THD specs. Maybe IM.
Not things that show up in the typical FR & THD specs. Maybe IM.
I don't remember if I ever listened to a Cary amp. Krells are pretty lifeless, but I do like the solid state amps designed by the guys who haunt this forum, namely Parasound, Ayre and Pass. Siegfried Linkwitz used Pass amps this past RMAF and they improved his system considerably.
Most tube amps have distortion so low that you never hear it, when paired with speakers of sufficient sensitivity. It's really hard to say what amps sound good and which ones don't because a lot of systems on the whole are not set up properly. Local hi-fi salons are usually a poor place to assess performance of an amp.
John
Also @ Thorsten and PMA.
Of course these measurements do not reflect the way our perception works. The are meant to measure the accuracy of an electronic amplification device, nothing else.
From the measurements of this Boulder amp, you can see it is straight, has insignificant distortion, and produces a very, very nice 10Khz square wave. In other words, what goes in, comes out, with little added nuisance. Perhaps a bit too much noise for an amp of this class, but that's about it.
It might go terribly wrong on a highly reactive load, but that can easily be measured, no voodoo here.
On PMA's remark that distortion and audible quality are loosely coupled, agree. I had a directly heated SET for a while, which appeared to sound great, but measured like crap. After a while I developed an allergy against all that distortion and sold the thing. Below a certain figure and depending on the composition of the distortion products, it does not seem to play a very large role. The Boulder is way below this certain figure at all power and frequency measurements shown.
Then on Thorsten's -quo- To some the term "capabilities of an Amplifier" may be taken relating to the resulting sound quality. I wanted to clarify if there was anything of the like in your statement.
And I thank you that you made clear that for you "capabilities of an Amplifier" does not relate to sound quality. -unquo-.
I refuse to accept that there is some sort of magic ingredient that cannot be measured, but which is responsible for 'sound quality'. We are talking engineering here, not religion.
I am open to the possibility that our measurement suite is incomplete and that we miss out on an important potential yardstick for audio quality. However, with the measurements we do know how to make, we can already make fair statements about how accurate an amplifier is capable of reproducing an input. The Boulder is very good at this.
Thorsten, if you think this is not a good indicator of the capabilities of an amplifier, what then is it we might be overlooking?
vac
Indeed true in a philosophical sense. What it will take is actual evidence that a box of gain which measures well can be distinguished by ear from other boxes of gain that measure well, assuming the usual qualifications (e.g., neither box clipping, levels match...). What we get is hand-waving, speculation, and assertion, but no evidence to date. Given the 30+ year history of many loudly asserted claims but no evidence, I think it's highly unlikely, but any open-minded person must allow for the possibility that someone will produce a good experiment showing the contrary.
I'm not holding my breath.
Who knows, but maybe some EE's are also sound engineer's that has quite a good knowledge about human hearing?
I don't say that THD/IMD is telling everything about an amp, by no way, but it tells alot if you look at other measurements at the same time, like frequency responce, phase responce, and I also like to look at the IHD distribution, just to mention some.
BTW: the measurements of the mentioned Boulder amp, doesn't look that good.
Cheers
S
Was this the proverbial "everyone in the room"? Some of Nelson's amps have easily measurable distortion, even he speculates that the different harmonic structures give each amp its fan base.
BTW don't bother answering the first question, the psychology of peer pressure and crowd mentality let alone a comment by one or two revered "experts" present could easily get everyone's head bobbing.
jlsem said:
Yeah, I you got that right. You can't take much from listening impressions at a hi-fi salon. If it sounds awesome then you know the whole system including the room is working well together. If it doesn't then something is amiss.
I've heard system with the Boulder that sound great, and others that sound bad.
If I turn down the volume just 0.1dB on our subwoofer crossover, which is connected to just the woofers and subwoofers, as our speakers are bi-amped, my dad complains that the sound is bright and irritating and hurting his ears. But, where it is now the sound is glorious.
The voicing, sensitivity and impedance of some speakers will complement some amps better than others.
I think you take high quality components that work well together and set them together properly in a good room you're going to get good sounds.
The Boulder amps don't suck, just like a lot of other amps don't suck if set up properly in a complementary system.
John
Last edited:
Depends on the application. For loudspeakers intended to be driven by voltage sources (99% of what's out there), the Stereophile measurement suite is a pretty good one. I can think of two or three things that could be added, but the basics (frequency response into test loads and speaker loads, distortion spectra, noise, stability) are good guarantees that amps will not impart coloration. If I were in charge of their testing, the first thing I'd add is something like Audio Critic's "Power Cube" measurement, but I'm not, so it won't be.
It would also be nice to run a controlled listening test to ensure there's nothing weird going on- it will be null most of the time, but it would be interesting to see when it isn't.My personal hobby horse is overload recovery, which is a moot point for kilowatt monsters but more important for the smaller amps that we tube guys build.
Actually I had some Crown MA5000 amplifiers driving tweeters. They sounded absolutely horrible when they clipped! The tweeters had a circuit inserted that caused them to resonate and produce more output with the amplifier the designer used to test them.
I have mentioned before that an enlightening test is to use one channel of an amplifier to drive a real loudspeaker and have a second channel drive a resistor. Looking at the difference on music with a digital scope set to trigger on transients shows a whole lot of difference on amplifiers with similar THD or IM plots.
Home audio stuff?
Well, yes, they'd also need to have the same output impedance to parts-per-million. It's not hard to measure differences that are inaudible.
- Status
- Not open for further replies.