What's wrong with Class-D?

[qusp]

did you really just miss the irony/sarcasm after he just confirmed he was being ironic just before?

I gather he has been gathering measurements to explain and explain well what he means.

this is completely tongue in cheek and that is perfectly obvious, inperfect English or not.

Originally Posted by AP2
Today just read an audio forum, to know that:
Class D is immune from the component on the vcc and voltage variations.
The class AB (for use audiophile) works just fine even with a smps for lamps.
amplifier with the same chip for television, is perfect for audiophiles.
Maybe there is something wrong?

Class D is immune from the component on the vcc and voltage variations.

a Myth stated here in the thread, you really think Roberto would say that? he makes excellent high grade PSUs for Class D and in the same post mentions how very important the PSU is. it seems you just want to argue at all costs...

The class AB (for use audiophile) works just fine even with a smps for lamps.

a passlabs project using lightbulbs for current sources and wallwart psu for the amp camp

amplifier with the same chip for television, is perfect for audiophiles.

again this and several thread hyping the Ti Eval boards as the next amazing thing for modding

Maybe there is something wrong?

pure sarcasm

Sure Roberto is prone to some hand-waving every now and then, you yourself are not free of guilt, but generally if you pay attention he really does want to understand why and isnt afraid of measurements to get there. I dont see you contributing towards the greater understanding...

 

[mr_push_pull]

did you really just miss the irony/sarcasm after he just confirmed he was being ironic just before?

I wasn't talking about that. it's partly a coincidence that I replied after you did.
what I wrote stands, regardless the irony/sarcasm thing.

 

[qusp]

huh, yes you were... you took the paragraph and replied to every line

 

[AP2]

I find that (generally and particularly in this thread) AP2's posts are hard to decode.
moreover, he uses a lot of hand-waving. the "this is difficult to explain" above. I like to think this is a technical forum with a relatively large average user "quality", not a gathering of easy to manipulate zombies.
stuff like "this is hard to explain" should not be accepted and responded with in a critical way. I've witnessed cases here where people seemed to be able to follow pretty complex explanations. I trust that there are such people reading this thread,


Today just read an audio forum (like this thread on this audio forum), to know that:
Class D is not immune from the component on the vcc and voltage variations.
The class AB (for use audiophile) doesn't work just fine even with a smps for lamps.
amplifier with the same chip for television, is not perfect for audiophiles.

and now I say... Maybe there is something wrong?

hope you get my point.


and...
I see this thing repeated from time to time, that the ear can't viewed as a frequency-domain analyzer and that there are some other phenomena at play. I once used to believe that, because some very vocal users here who also project a self-confident image kept repeating it. meanwhile I did a bit of homework. some people seem to believe that there are some mysterious time-domain phenomena that hide from the typical measurements. some even mention studies that show such phenomena. when asked for such studies, they answer that they can't do one's homework, that if one isn't aware of them then he's not up to the discussion level etc. some of them even sell devices that cure such mysterious phenomena.
I repeat, this is a technical forum. science doesn't work like that. I urge people to ask for hard facts and if such hard facts are not brought to light, they should try and draw some conclusion from it.
this is not politicians gathering votes, but nevertheless resembles that situation.

I have no problem with you.
I know my English is not perfect, but how is it that you just do not decode my word or concept. (Usually the concept can be decoded from our brains, not by the written word.)
my word "hard to explain" refers to my difficulties, not because I think the people on this forum do not have the correct knowledge).
I'm sorry that you think in a negative way about me.
I think the problem is reversed. explain right now clear to you.
I do not think in this thread, my posts are not readable. the concept is clear.
It's not me I want zombies, I want the fans understand well.
All serious and well qualified, they know the problem of power supply.
I solved after a long research and the difference is huge, perhaps as you've never heard him play an amplifier, for the simple reason that not exixt a psu with 2.7V of voltage drop at 500W burst) there is only the DPS-500 / S in the world...please buy one and test.
I never said (I can not develop new technologies, because I have not found on google! as other Guru). then push a power supply for lamps. and convinced all,that it is not necessary to have a fast response. but how many laughs make the engineers, when they read the story "convenient"?
----------------------------------------------
Why, an class D amplifier, must have huge nfb? necessary for good audio fidelity..or?
Why, an amplifier must have exaggerated DF? necessary for good audio fidelity..or?
please, you answer these questions and explains with good technical articles, so that I can understand well.
my F3 have only DF200 at 100Hz ... it's not good?

regards
Roberto

 

[mr_push_pull]

I dont see you contributing towards the greater understanding...

you make the typical mistake of implying that I pretended the above.
I just pointed out some issues, no logic requires me to be a Nobel prize winner to be able to do it.
and I never said he is afraid of measurements.
but I still have a big problem taking seriously stuff like "this is hard to explain", especially when the person writing it sells stuff.

 

[fastvideo]

To answer the thread title: nothin' wrong with class D. I think it was unfortunate that they named class D "D" because it happened to follow "C", and people think the D stands for 'Digital'. Would have been better if it was class K for 'Kool".

I like this very much.
Cars with 20 air bags today, all reviews say they are very safe.
From what I see, cars that were built 30 years ago are much safer .

It is all about cost and marketing more than what is products are for, no?

 

[mr_push_pull]

Why, an class D amplifier, must have huge nfb? necessary for good audio fidelity..or?
Why, an amplifier must have exaggerated DF? necessary for good audio fidelity..or?
please, you answer these questions and explains with good technical articles, so that I can understand well.
my F3 have only DF200 at 100Hz ... it's not good?

regards
Roberto

the burden of proof lies with you, not with me.
I'm not a believer of extremely high DF either, I think that beyond a certain figure it's only a selling point.
an amplifier may use huge NFB in order to be linear, which is one of the measures of how good it is at following its input.
I could ask... why does and amplifier need to have 200kHz bandwidth, when most tweeters start to drop sharply below 25 kHz? I didn't even quote the ultrasonic hearing studies.
why does and amp need to have huge slew rate, when it's a known fact that real music rarely contains more than 1V/us and that with RedBook there's a low-pass filter (actually lots of them) in the chain that can't pass sharp steps unaffected?
but I won't ask those things because we aren't even talking about them

all I asked was: less hand waving, more facts. if you don't intend to explain things fully, just refrain from even mentioning them.

another user recently wrote:

[...]this is a technical forum. You get technical questions. You answer them using data, plausibility arguments from first principles, simulation, or handwaving. It's usually pretty clear which you're doing.

 

[AP2]

the burden of proof lies with you, not with me.
I'm not a believer of extremely high DF either, I think that beyond a certain figure it's only a selling point.
an amplifier may use huge NFB in order to be linear, which is one of the measures of how good it is at following its input.
I could ask... why does and amplifier need to have 200kHz bandwidth, when most tweeters start to drop sharply below 25 kHz? I didn't even quote the ultrasonic hearing studies.
why does and amp need to have huge slew rate, when it's a known fact that real music rarely contains more than 1V/us and that with RedBook there's a low-pass filter (actually lots of them) in the chain that can't pass sharp steps unaffected?
but I won't ask those things because we aren't even talking about them

all I asked was: less hand waving, more facts. if you don't intend to explain things fully, just refrain from even mentioning them.

another user recently wrote:

Why is speed so necessary?
Because live music and speech are both essentially made of transients. And transients are only sounding accurate when they are rebuilt into the ear with exactly the time coherence they had initially. Only extra-large amplifier bandwidth can rebuild time coherence accurately.
The audible effect ? A real frequency-balanced time response brings Intelligibility which is extremely important for the ear to recognize musical instruments, and understand speech.
------------------------------------------
A perfect audio component will maintain the coherency of a signal over the full range of frequencies. Phase distortion can be extremely difficult to reduce or eliminate. The human ear is largely insensitive to phase distortion, though it is exquisitely sensitive to relative phase relationships within heard sounds. The complex nature of our sensitivity to phase errors, coupled with the lack of a convenient test that delivers an easily understood quality rating, is the reason that it is not a part of conventional audio specifications.
--
A system may have low distortion for a steady-state signal, but not on sudden transients. In amplifiers, this problem can be traced to power supplies in some instances, to insufficient high frequency performance or to excessive negative feedback. Related measurements are slew rate and rise time.
Distortion in transient response can be hard to measure. Many otherwise good power amplifier designs have been found to have inadequate slew rates (ref to a module of sound envelope from real instruments).
-------------------------------------------------------------

Regards
Roberto

 

[radiosmuck]

Does anyone here actually listen to music and are you all equipped with the intruments to test for all the stuff you write about?
I was a lad back in the 60's and listened to pop music with either a little transistor radio tuned to an am pirate radio station or a one note bass radiogramme. The music was still great.
Listening to these old recordings on 128kbs internet radio with a cheap d amp and good speakers is a real revelation.
Ignorance is bliss.

 

[Ken Newton]

It's fun to see capitalism failing just because one of its main features was not sustainable (it topped and popped): arbitrarily inflating the price of things, the dissociation between the price and the actual value (in terms of the substances and the energy expended to produce something).

In terms of the basic macro-economics involved, money used to purchase (consume) luxury goods actually isn't the problem. This is because all money used to purchase goods and services passes through as income to those suppling the purchased goods and services. As an extreme example, imagine someone were to spend $1M on a scoop of the world's finest ice cream, as low as the utility value of such a purchase may be! The $1M passes to the ice cream maker and their employees, and then to the milk, sugar, and fruit suppliers and their employees, and so on. As long as everyone in the chain uses the income they receive to then consume other products and services the rotation of the whole economic wheel would be relatively self-sustaining.

The problems begin when that wealthy person instead purchases a $5 scoop of ice cream, and simply accumulates (invests) the remaining $999,995 - the accumulation of wealth is the accumulation of investment capital. What's wrong with this, some may ask? Investment for purpose the creating supply will follow demand, or expected demand, and not the other way around, as 'supply side' proponents would have us believe. Without sufficient demand (consumption) enabled by sufficient income, workers get furloughed, which exacerbates the lack of aggregate consumption until economic recession or depression results. Sound recently familiar to anyone?

In short, a major functional flaw in capitalism stems from the greed motivation pushing the accumulation of capital to far outgrow the wealth returned to the economy in the form of the consumption of goods and services. This is why sufficiently taxing the wealthy becomes so important, it redirects some of the wealth destined for capital accumulation, and redirects it to the purpose of consumption. Such taxation isn't merely a question of fairness, morality, or political ideology, it's a simple necessity based on the mechanics of capitalism.

This leads us to a proper form of 'trickle-down' economic theory, where government policies promote a much greater 'trickling down' of INCOME due to the wealthy spending a much larger portion of their investment based incomes on purchasing goods and services, no matter how poor a value some given purchase may appear. This would be far different than is the current trickle-down theory, where government polices instead promote further capital accumulation and concentration. So, please wealthy folks, by all means, buy more Levinson, Burmester, Audio Note, etc.

 

[Kindhornman]

I find that Roberto's English, though a second language to him is fairly easy to follow. Any kind of farcical language that is done without facial cues is always hard to realize without words that tell the reader that this is happening, it is not always self evident that is being done, hence the smiley faces that some of us use.

I very much agree with your simple explanation of the mechanism you just stated Roberto of why we need wide bandwidth and very fast slew rates and rise time. I actually am of the same mind when it comes to speakers, that this is often the missing link to reproducing the magic, the realism that we are all looking for. Impedance rise turning to reactive loads, slow ponderous cone materials such as polypropylene with its signature slow decay times, crossover design with non-linear phase response at crossover and beyond all contribute to the smear that we hear in reproduced musical instruments leaving us knowing we are hearing a reproduction of those instruments and not the disbelief that we aren't hearing the real thing.

There should be as much thought to applying the same analysis that looks for .0001% distortion factors in the electronic amplification side as there should be on the loudspeaker side. We should not have to accept 10% distortion factors as the limit of maximum excursion to use a loudspeaker, why are we not aiming for that less than .1% factor?

 

[mr_push_pull]

ok. let's step back and try to extract something meaningful from this conversation. I'll start.

are you talking about this?

ok. as far as I recall you said that you don't want to say what amp was it because it's made by competition or something. and you didn't post "regular" measurements (e.g. the ones that Stereophile does) but I don't recall any motivation. but, nevertheless.
at the most basic level what I see there is a device being unable to track fast changes. a full-scale 20kHz sine is fast enough, actually faster than what is visible in those pics. any disability of said device to output an undistorted 20kHz sine without any significant delay would show not directly but indirectly in a CCIF IMD test and a bode plot, wouldn't it?

 

[GoatGuy]

Why we Argue against D

Does anyone here actually listen to music and are you all equipped with the intruments to test for all the stuff you write about?
I was a lad back in the 60's and listened to pop music with either a little transistor radio tuned to an am pirate radio station or a one note bass radiogramme. The music was still great.
Listening to these old recordings on 128kbs internet radio with a cheap d amp and good speakers is a real revelation.
Ignorance is bliss.

I too used to listen to mono-channel radio using a self-built 4 transistor "radio" (it was pathetic, but I was proud of it as a teenager, since it was NOT built from a kit, nor was it built from a cookbook recipe). I also rode a fat-tire bike, rusty as hell, and used nearly busted Zippy Womper sleds in the snow.

Now, I enjoy my very-budget priced (all in all) 18 speed bike, I ski on lovely parabolic skis, I drive a BMW ... and I listen to audio - great music - on a pretty darn nice sound system. Moreover, since I'm not a machinist, I can't really make parts for a bike, or my car. And I have a mental allergy to grease and grime ... so my "hobby" is building novel electronic stuff.

Because video has become unreachable, and computers have become totally unreachable to intellectual-inquiry, abstraction, synthesis and build-out, the only thing left for us engineer-savant types ... is audio. So here we endlessly debate things, not unlike the forums of Athens and Rome.

Class D can sound just fine. I've listened to a lot of D that is pretty convincing, and it sure is nice to have 100, 200 or 500 watts of per-channel power in a box that in total weighs less than a single power-transformer on my SET amplifier, monoblock, left-channel. Its pretty amazing.

So what then do I have to say to all this real- and pseudo-intellectualism? Us "old timers" who disdain Class D mostly disdain the fact that soon, it too will be like video, logic and electronics in general: the mechanics of doing it will be completely ensconced in little chips from AMD, Linear, Nation, TI, and the rest, and our eletronics "skillz" will be reduced to mere printed-circuit-board layout, drilling, and soldering. Class D, almost by definition, removes the "need to calculate, desire to perfect" out of things, and reduces audio to magic.

I posit then, that that conversion to blind-magic ... is what we're alarmed at, and in some sense, "fighting over". Nothing more. Even my own quite carefully written Class-D objection a few pages back ... shows that this Conversion to Magic is the problem.

So ... ya, I get what you're saying. LISTEN to the music. But for makers of violins, tuners of pianos, carvers of wood, blowers of glass, and designers of audio electronics, the interest lies not in setting up machines to mindlessly mill out exact duplicates of everything, but producing a few at a time, each somewhat different, not just to "see" (hear) what is different, but to confirm that our minds are capable of great things, in the artistic domain.

GoatGuy

 

[mr_push_pull]

I find that Roberto's English, though a second language to him is fairly easy to follow.

it's actually Google Translate's English but I'm ok with that, to a point.

we need wide bandwidth and very fast slew rates and rise time.

Charles Hansen of Ayre believes in high bandwidth.
Yves-Bernard André of YBA doesn't.
Nelson Pass of Pass Labs doesn't believe in very high slew rate.
Yves-Bernard André doesn't believe in balanced electronics.
Accuphase does.

if these are converging opinions, it fails me how.

 

[mr_push_pull]

So ... ya, I get what you're saying. LISTEN to the music. But for makers of violins, tuners of pianos, carvers of wood, blowers of glass, and designers of audio electronics, the interest lies not in setting up machines to mindlessly mill out exact duplicates of everything, but producing a few at a time, each somewhat different, not just to "see" (hear) what is different, but to confirm that our minds are capable of great things, in the artistic domain.

GoatGuy

at the risk of side tracking the discussion, if people only listened to music there wouldn't be a concept of being an audiophile to speak of. it's like saying to a super car owner "do you really need 300 km/h?". obviously not all the time and it's not a matter of needing it but why not, especially if one affords it?

but I believe we should focus on this post: http://www.diyaudio.com/forums/class-d/228047-whats-wrong-class-d-4.html#post3404416 there lies the essential problem IMO.

 

[GoatGuy]

but I believe we should focus on this post: What's wrong with Class-D? there lies the essential problem IMO.

I think that the answer is crystal clear: nothing if done really, really well. But how much is "really, really well done", when so many designs are tuned to be linear, whereas music (especially) is wickedly logarithmic? Ah, there becomes the focus of the piece de resistance.

GoatGuy

 

[radiosmuck]

at the risk of side tracking the discussion, if people only listened to music there wouldn't be a concept of being an audiophile to speak of. it's like saying to a super car owner "do you really need 300 km/h?". obviously not all the time and it's not a matter of needing it but why not, especially if one affords it?

but I believe we should focus on this post: http://www.diyaudio.com/forums/class-d/228047-whats-wrong-class-d-4.html#post3404416 there lies the essential problem IMO.

What if I cannot find anything wrong with class D, does that make be a bad person?
We are lucky, life is full of choices, imagine only eating bamboo shoots.

 

[AP2]

ok. let's step back and try to extract something meaningful from this conversation. I'll start.

are you talking about this?

ok. as far as I recall you said that you don't want to say what amp was it because it's made by competition or something. and you didn't post "regular" measurements (e.g. the ones that Stereophile does) but I don't recall any motivation. but, nevertheless.
at the most basic level what I see there is a device being unable to track fast changes. a full-scale 20kHz sine is fast enough, actually faster than what is visible in those pics. any disability of said device to output an undistorted 20kHz sine without any significant delay would show not directly but indirectly in a CCIF IMD test and a bode plot, wouldn't it?

First of all, (sorry at all for this new argument by MrPush), first pic refer to musical transient in any time segment,as x-time show. and not a sin input. as him said, the difference is visible in relation when it is fast (transient) , we can call "phase error" but is not this the problem.
see angle of peak, structure of envelope changed, then musical instrument are modified. obvius that we listen mp3 also and? study is not important becouse we listen even a chip 1€?
This measures refer to just at start of study,i mean is not recent.
A lot of job to realizze today a better integrated amplifier with special smps capable of deliver performances as audiophile expected.

 

[xrk971]

Maybe a car analogy would be a class D amp is like a Nissan GT-R, very easy to drive fast as it is computerized and takes care of balancing everything. But feels sterile. On a track, it can keep up with many supercars but lacks soul and looks kind of industrial and ugly. A nice handbuilt class A is like a nice older Ferrari that takes a lot more skill to drive fast on a track, may have strange behaviors, costs a lot more, but has soul and people love them and wish they could have one. Just the sound from the Ferrari's exhaust note is unmistakable. I use and like my class D's but I confess that I would also love to have a nice hand built boutique class A room-heater and would not turn it away if someone gave it to me.

 

[qusp]

I believe thats a pretty outmoded cliche. nothing wrong with class A, but I believe the gap is closing, or perhaps closed when it comes to the high quality (still a bit expensive and not in a single chip) Class D offerings. besides, why havent you built yourself a big class A amp? they dont have to be OTT expensive, not cheap to run though, depends on how much power you actually need though