There must be some form of correlation between measured and audible performance, but much of our more traditional (simplistic) published measurements haven't provided an especially useful correlation.
My interpretation of what Bruno has written (I can't speak for him) is that he believes he can demonstrate an improved correlation between measured and perceived performance than has existed previously. What I find interesting about his approach and developments is that they they appear to be driven from the basis of hearing things first then trying to determine what to measure and how to design to achieve those audible results.
From a philosophical point of view I've come to the opinion that the role of a power amplifier should probably be the distortionless, frequency-independent control of any load it's connected to. if there are euphonic distortions that can increase our listening pleasure (though I'm not convinced there are) then these should be generated before the power amp. That is the roll of eq and effects within the recording chain.
Different amps do sound different and do affect my (our) pleasure with this illusion engine. But it always strikes me as strange that the recording chain uses so many devices and processes that we believe we shouldn't dare use in the replay chain
My interpretation of what Bruno has written (I can't speak for him) is that he believes he can demonstrate an improved correlation between measured and perceived performance than has existed previously. What I find interesting about his approach and developments is that they they appear to be driven from the basis of hearing things first then trying to determine what to measure and how to design to achieve those audible results.
From a philosophical point of view I've come to the opinion that the role of a power amplifier should probably be the distortionless, frequency-independent control of any load it's connected to. if there are euphonic distortions that can increase our listening pleasure (though I'm not convinced there are) then these should be generated before the power amp. That is the roll of eq and effects within the recording chain.
Different amps do sound different and do affect my (our) pleasure with this illusion engine. But it always strikes me as strange that the recording chain uses so many devices and processes that we believe we shouldn't dare use in the replay chain
there is a lot of subtext in your messages, but you pretend otherwise. one can clearly read "I think UCD is crap and most likely nCore will be too". what I do not like is that you are trying to present things like you didn't think that.
phrases that slip "by accident", like that one "let's look at the past performance of class D amps" have an obvious purpose and that is to instill the idea that UCD or nCore are nothing but successors in a line of known poor performing products.
for one, I'm sure you are incapable to prove or explain the supersonic performance degradation.
for me it's ok that you like AP2's products, although I would think that you've listened to his products before holding on to an opinion, based on the fact that I haven't and never saw any measurements, relevant or not. repeating statements like "I like his approach more" adds no information. his products may indeed be better but at this point I have no basis whatsoever to believe that. OTOH I have more reasons not to take the guy seriously and believe Bruno.
I would not be surprised if you'll think that nCore sucks while I will disagree. just to emphasize the futility of this discussion.
phrases that slip "by accident", like that one "let's look at the past performance of class D amps" have an obvious purpose and that is to instill the idea that UCD or nCore are nothing but successors in a line of known poor performing products.
for one, I'm sure you are incapable to prove or explain the supersonic performance degradation.
for me it's ok that you like AP2's products, although I would think that you've listened to his products before holding on to an opinion, based on the fact that I haven't and never saw any measurements, relevant or not. repeating statements like "I like his approach more" adds no information. his products may indeed be better but at this point I have no basis whatsoever to believe that. OTOH I have more reasons not to take the guy seriously and believe Bruno.
I would not be surprised if you'll think that nCore sucks while I will disagree. just to emphasize the futility of this discussion.
I am curious to learn more about that.
What I have missed until now is a FFT distortion plot (Bruno was asked for a distortion spectrum back in June) because this tells us at least something about the distortion spectrum (2nd, 3rd a.s.o. harmonics). These spectra can provide useful information; FFT plots are pretty common nowadays, and seem to have more correlation with sound quality.
The THD curves which Bruno provides in his white paper look very good, but it is also known that low measured THD and perceived sound quality not seldom don't correlate at all.
Yes completely agree.
Look at sound reproduction as having microphone(s) at the beginning, and loudspeakers at the end.
These are generally the weakest elements in the chain (being converters from acoustic to electric and v.v.), so their quality matters most IMO.
Everything in between is for processing, amplifying a.s.o. and this is in the electrical domain where distortions are a magnitude lower.
So also the power amplifier should merely do his (electrical) job by amplifying the signal and feed it to the loudspeaker with minimal loss of quality.
I have been lucky a couple of times to have been present at recording sessions of the Dutch Radio Philharmonic Orchestra, where the "minimal" microphone technique approach was used (two microphones in stead of a large number for each group of the orchestra, so very much depends on microphone placement and communication between recording engineer and conductor). I can tell you this is art, and very much of the quality we hear at home depends on the quality of the recording.
Maybe good to realize when arguing about amplifiers....
Have we lost the topic of this thread? Gentlemen, please.
of course it has crossover distortion, like any class D amp in existence. it's only that measurements have been faked. didn't you know?
guys, I'm unsubscribing from the thread too. I used to think of DIY audio as the only forum I know where even a heated discussion can be pleasant. I've come to change my mind. I'm absolutely convinced that this topic is poisoned with people that have other intentions than to reveal truth.
I think the last answer was uncalled-for.
The poster asking the question about crossover distortion didn't make any posts in this thread before (so neither ranting ones). As far as I could see he was not even posting in class-d at all so one can assume that it was 1.) a honest question and 2.) he isn't aware of the special crossover distortion mechanisms that come into play with class-d amplifiers.
Class-d in general do have a very different crossover behaviour than linear amps. There is always an alternating idling current flowing through the output inductor at no or low output power. This current is greatly supporting the switching process. At the very point when the output current in the load approaches the value of the idling current a class-d amp's crossover takes place. This is usually at a few watts. There are amps where this is intentionally pushed to higher output power levels by the use of smaller inductors (and therefore increased idling current). IIRC it was mentioned within the thread where this takes place for the ncore and if my memory serves me right it is between 20 and 30 Watts. Or in other words: At quite high output levels from a domestic point-of-view.
Regards
Charles
The poster asking the question about crossover distortion didn't make any posts in this thread before (so neither ranting ones). As far as I could see he was not even posting in class-d at all so one can assume that it was 1.) a honest question and 2.) he isn't aware of the special crossover distortion mechanisms that come into play with class-d amplifiers.
Class-d in general do have a very different crossover behaviour than linear amps. There is always an alternating idling current flowing through the output inductor at no or low output power. This current is greatly supporting the switching process. At the very point when the output current in the load approaches the value of the idling current a class-d amp's crossover takes place. This is usually at a few watts. There are amps where this is intentionally pushed to higher output power levels by the use of smaller inductors (and therefore increased idling current). IIRC it was mentioned within the thread where this takes place for the ncore and if my memory serves me right it is between 20 and 30 Watts. Or in other words: At quite high output levels from a domestic point-of-view.
Regards
Charles
Agreed
Which possibly/probably explains the 'bump' in the THD plots for the ncore at about 30W into 4ohms
http://www.hypex.nl/docs/nc1200%20folder%20web.pdf
And if so probably also the bump at 60-70W on the 8R plot.
(same current flow but double the voltage...)
The nice thing is the very low distortion plateau beneath these bumps which are very useful for most domestic listening
- on the assumption that low distortion relates to good sound of course
cheers,
I will
probably build up a nCore amp with the hopes that it will sound good enough for me to make the switch from A and A/B amps to class D for good. I sure would like to save the energy.
I have heard the uCD 400s here (with the Hypex deluxe linear supplies and HrX regs) and have a pir of ICE ASP 500 based monoblocks around as well. The uCD 400 was the best sounding class D amp I have heard, but still lost out to a traditional class A/B amp (Pass Labs X 150.5) for listening.
I look forward to the nCore DIY modules availability with high hopes.
probably build up a nCore amp with the hopes that it will sound good enough for me to make the switch from A and A/B amps to class D for good. I sure would like to save the energy.
I have heard the uCD 400s here (with the Hypex deluxe linear supplies and HrX regs) and have a pir of ICE ASP 500 based monoblocks around as well. The uCD 400 was the best sounding class D amp I have heard, but still lost out to a traditional class A/B amp (Pass Labs X 150.5) for listening.
I look forward to the nCore DIY modules availability with high hopes.
Charles,
Very interesting; I had thought that crossover distortion wasn't an issue for class D because they don't construct the output waveform from two halves like AB design.
I guess this would be true for the IRS2092-based modules?
The only evidence I see in the curves starting on p.8 here
http://www.inductor.com/irf/iraudamp7s.pdf
is a small bump at 20W in some of the THD+N vs. power graphs, though curiously only in one channel.
What is the freq of this alternating idling current?
If the same as the PWM freq wouldn't it be inaudible?
Last edited:
What kind of installation did you use for the listening test? Any infos on the release date ?
Ask Bruno for specific details, for it was organized by Hypex together with AES section NL.
Active loudspeakers: SEAS tweeter and mid; bass: Peerless
(think it was...)
- Status
- Not open for further replies.