Class A/AB vs D -- Is it Audible in 2021?

[MakeMineVinyl]

Class A/B amplifiers are still made because a very significant segment of purchasers demand them. We make both class A/B and D.

 

[Hipster Doofus]

Don't know much about geography
Don't know much trigonometry
Don't know much about algebra
Don't know what a slide rule is for

Don't know much about history
Don't know much biology
Don't know much about a science book
Don't know much about the French I took

but just added my 12ax7 Chinese clone of a Marantz 7 pre amp ….to my class D amp and it sounds a bit nicer and less sterile.
Dare I say more tube like. Ha ha

With linkwitz speakers and a dsp crossover.

hows that for melting pot or Gumbo of audio theories.

 

[Head_Unit]

Why are most AVR companies still cramming A/B amps in their products?

I believe it has nothing to do with any sonic superiority. I believe it is because the radiated noise is not easy to design around, especially when @Amir is measuring them to keep them honest. And I dunno, class AB is still cheaper I'm guessing? And like tradition, or in other words, that's all they know how to do and it's a big hump to incorporate a whole new technology. Plus, most consumers wouldn't know AB versus D if it stuck shapr Jaws teeth into their privates, so there is negligible marketing advantage because many (most?) marketing departments have a lack of technical understanding and imagination to actually promote such.

Back to the original thread concept, let us for the sake of argument say that amplifiers CAN sound different below the clipping point (and if their frequency response is the same).* It's still a flawed premise to say "AB" versus "D" because there can be great and crappy implementations of either.

*On the one hand I don't think all amplifiers sound the same. On the other hand, there's this
https://forums.stevehoffman.tv/threads/denon-vs-parts-express-round-1.984507/
On the other other hand, we swapped the Denon AVR-X3600H for an Integra 6.5 + ATI 525nc. There's definitely an image difference, noticeable, but not huge by any means. And possibly explained by different Audyssey, since the -3600 lets you use the Audyssey App to for instance turn off their midrange dip which I had done.

 

[Descartes]

Audioholic had a great video on these issue’s

 

[MakeMineVinyl]

It has nothing to do with companies who make class A/B amplifiers 'that's all they know how to do and it's a big hump to incorporate a whole new technology". Both classes have their uses and strengths. They are different in meaningful ways, and the user can choose which approach matches their own priorities.

Myself, I don't go either way. I use vacuum tubes so shut up and get off my lawn.

 

[Alice of Old Vincennes]

I believe it has nothing to do with any sonic superiority. I believe it is because the radiated noise is not easy to design around, especially when @Amir is measuring them to keep them honest. And I dunno, class AB is still cheaper I'm guessing? And like tradition, or in other words, that's all they know how to do and it's a big hump to incorporate a whole new technology. Plus, most consumers wouldn't know AB versus D if it stuck shapr Jaws teeth into their privates, so there is negligible marketing advantage because many (most?) marketing departments have a lack of technical understanding and imagination to actually promote such.

Back to the original thread concept, let us for the sake of argument say that amplifiers CAN sound different below the clipping point (and if their frequency response is the same).* It's still a flawed premise to say "AB" versus "D" because there can be great and crappy implementations of either.

*On the one hand I don't think all amplifiers sound the same. On the other hand, there's this
https://forums.stevehoffman.tv/threads/denon-vs-parts-express-round-1.984507/
On the other other hand, we swapped the Denon AVR-X3600H for an Integra 6.5 + ATI 525nc. There's definitely an image difference, noticeable, but not huge by any means. And possibly explained by different Audyssey, since the -3600 lets you use the Audyssey App to for instance turn off their midrange dip which I had done.

The average consumer drop off point is a Bose soundbar. When you pay past a grand or two for AVR the seller knows your not a hick from French Lick.

 

[rwortman]

What I see as the biggest difference now when I test the AIYIMA A07 is the modulation of frequency response by speaker impedance. This is due to interaction of the output LC filter with speaker impedance. I have measured it with JBL Control1 Pro. The capacitive impedance above 3kHz adds to C of output LC and changes the filter response. This should not happen with PFFB amps and should not happen with NC400 and Purifi, at least I hope so. This peaking and decay never happens with a "normal" class AB amp. The issue is audible and creates harsh upper midrange.

View attachment 104598

View attachment 104599

More here:

https://www.audiosciencereview.com/...-measurements-lm4562-option.19155/post-629539

This the reason why I do not like simple generalizations about class XY sound. It always is related to the design.

5khz is not upper midrange, not sure a peak that high would cause a harsh upper midrange, and it‘s only half a db. Audible but just barely. There are peaks much higher than that in any in room speaker response. Operatic soprano can maybe reach 1khz.

 

[Waxx]

I have class A, class AB and class D amps, and the main difference in sound with good models is het harmonic distortion. Class D does not have harmonic distortion (or so low it's not relevant). Class AB has a little, and class A or tube amps (especially SE tube amps) have a lot. Some do like that (i'm on of those) and think class D is liveless and clinical perfect. Some like that clinical perfect sound and think the harmonic distortion is a defect. An other difference is efficiency in electricity use (class D use much less per watt output) and heat dispensation (class A or tubes have a sidejob as heating systems).

But the days that class D was sounding bad by definition are gone (luckely) and high quality class D (like NCore, Purify and so) are the most neutral amps arround without discusison. That most neutral does not means best sounding for everyone, that is the only point of discussion i think. So in my main setups i use class A or class AB amps (tubes and transitors), but for speaker measurement i prefer class D, and also use it in less important places (because it's cheap, even for good models) like my kitchen or workshop. And even the cheap "bread and butter" class D amps for pa use are on a level today that they can be used in hifi systems without sounding bad. And many do that today. You don't have to spend a lot to get a good amp these days.

 

[MakeMineVinyl]

I have class A, class AB and class D amps, and the main difference in sound with good models is het harmonic distortion. Class D does not have harmonic distortion (or so low it's not relevant). Class AB has a little, and class A or tube amps (especially SE tube amps) have a lot. Some do like that (i'm on of those) and think class D is liveless and clinical perfect. Some like that clinical perfect sound and think the harmonic distortion is a defect. An other difference is efficiency in electricity use (class D use much less per watt output) and heat dispensation (class A or tubes have a sidejob as heating systems).

But the days that class D was sounding bad by definition are gone (luckely) and high quality class D (like NCore, Purify and so) are the most neutral amps arround without discusison. That most neutral does not means best sounding for everyone, that is the only point of discussion i think. So in my main setups i use class A or class AB amps (tubes and transitors), but for speaker measurement i prefer class D, and also use it in less important places (because it's cheap, even for good models) like my kitchen or workshop. And even the cheap "bread and butter" class D amps for pa use are on a level today that they can be used in hifi systems without sounding bad. And many do that today. You don't have to spend a lot to get a good amp these days.

I agree with all you've said, but I've read on the internet that class A/B amplifiers can have distortion just as low or lower than a class D. Its all in the implementation / design. At least that's what my internet machine said.

 

[Waxx]

I agree with all you've said, but I've read on the internet that class A/B amplifiers can have distortion just as low or lower than a class D. Its all in the implementation / design. At least that's what my internet machine said.

That is probally an opinion based on older class D amps, as they were in general noisy untill at least 10 years ago, but tech advanced and with the Philips UcD and Ncore (now called hypex) the time started that they became quiet decent to very good. In the meantime the rest of the better ones also moved forward so they are in general with less distortion. Class AB has long been the best measured sound, but not anymore.

 

[Sam Ash]

That is probally an opinion based on older class D amps, as they were in general noisy untill at least 10 years ago, but tech advanced and with the Philips UcD and Ncore (now called hypex) the time started that they became quiet decent to very good. In the meantime the rest of the better ones also moved forward so they are in general with less distortion. Class AB has long been the best measured sound, but not anymore.

Hi @Waxx - have you had the opportunity to carry out comparisons between AB and D?.

Have you measured a class D amp that you think has surpassed AB, I'm told ATI make nice Ncore based class D amps.

Would be nice if @amirm can give us his views and input.

 

[MakeMineVinyl]

That is probally an opinion based on older class D amps, as they were in general noisy untill at least 10 years ago, but tech advanced and with the Philips UcD and Ncore (now called hypex) the time started that they became quiet decent to very good. In the meantime the rest of the better ones also moved forward so they are in general with less distortion. Class AB has long been the best measured sound, but not anymore.

You probably didn't notice in my signature that I work for a manufacturer of power amplifiers and we make both class D and class A/B.

 

[JSmith]

Class AB has a little

Unsure if you're aware the amp with the top spot here on the rankings (Benchmark AHB2) is an A/B hybrid H class design.


JSmith

 

[dlaloum]

If amp avoids clipping I doubt one can discern difference. Crown XLS will not clip.

Crown XLS will clip - it just typically has substantially more power than most amps, so you are less likely to reach the limit - but the limit is there nevertheless.... (yeah so you have to go to 440W @ 8ohm to clip an XLS2500.... but a lot less for the XLS1000 - 215W @ 8ohm)

 

[Chrispy]

LOL like it wasn't in 2020?

 

[Alice of Old Vincennes]

Crown XLS will clip - it just typically has substantially more power than most amps, so you are less likely to reach the limit - but the limit is there nevertheless.... (yeah so you have to go to 440W @ 8ohm to clip an XLS2500.... but a lot less for the XLS1000 - 215W @ 8ohm)

If you are clipping Crowns, call the fire department.

 

[dlaloum]

So if I don't care about heat, cost, size, or efficiency, or SINAD differences that are below the audible threshold....

Is there any real audible difference between Class AB or D?

I've read some claim that Class D has "better bass control".

Is there any truth to this?

I think I've heard this, but perhaps it's just placebo / cognitive bias after being heard to listen to it.

I've also heard some claim that A/AB (well, A, mostly) has "wider sound stage" or "warmer" or "more analog".

Which, *maybe* I've heard, but, again, can't rule out cognitive bias, once again.

Is there any objective truth to any of these stereotypes?

Or, really, do they both sound the same when operating way below redline?



Full disclosure:

My "big rig" passive spaker amp is a Devialet Expert 400, which 'ADH", some kind of A/D hybrid, while all my active speakers use Class D. I can't say I've noticed some kind of measure difference...maybe the Devialet sounds a tad 'wetter'. Maybe.

Different speakers and rooms have different issues.... a high efficiency (say 96db/wm) speaker with an impedance of 8ohm and no dips in impedance that drop below 6ohm.... will probably sound the same with almost any amp....

The amps are unlikely to clip or run out of current, unlikely to get strained by driving a low impedance load.... - all amps will run optimally - with negligible audible distortions - functionally they will all act as a wire with gain.

Then you take a speaker with a nominal 6 ohm impedance (shouldn't be too difficult for those amps) - but it has a funky tweeter with a capacitive response (that does things to phase) and impedance that drops below 2 ohm.... and a woofer crossover that also drops impedance down to around 2 ohm, and a 4ohm woofer....and make it average sensitivity... 86db/wm... now it is a whole different ball of wax!

The speakers will sound reasonable with most amps, they are sufficiently sensitive that huge power is not a requirement... but the low impedance of the tweeter and the woofer, is going to require a lot more current - and these aspects will put strain on the amp.
Under strain different amps respond differently - and their distortions vary in the frequency domain too - so very different sounds, drivent by differing distortion profiles.

If you provide this speaker with a mid power amp - say 100W @ 8ohm, but which is capable of driving low impedances with something around say 300W @ 2ohm - it will handle such speaker easily.

But many high quality high power amps aren't up to such a task... you will see specs that look like 150W@8ohm, then 220W @4ohm, but not rated and recommended for 2 ohm.... and although nominally higher power than the previous amp I described... this kind of amp will probably sound worse on the difficult low impedance dips speaker.... again how it sounds worse, will vary dramatically by the design of the amp itself.

Your Devialet amp is a very interesting design... it is a relative of the vintage "Current Dumper" design by Peter Walker - still used today in Quad amplifiers - this technique is also called "Feed Forward" - a small high quality amplifier, ensures high quality sound, while a large economical amplifier backs it up with plentiful current.

The Benchmark AHB2 (current flavour of the month/year here!) - is a similar Feed Forward design too

What Devialet has done differently, is that the small high quality amp is standard class A (AB?) - but the high power current support amp, is Class D.... an interesting design with a lot of great potential.

If you look at the amplifiers over the last 50 years that have used this type of topology, there are a lot of very highly regarded amplifiers - Quad 405/606/707/909/QSD/Artera, Nakamichi Statis, Threshold, Benchmark AHB2.

The Devialet brings the principles up to date by using Class D for the heavy lifting, makes perfect sense to me! - It would be interesting to get it measured and tested, like the AHB2 has been ....

 

[Alice of Old Vincennes]

You probably didn't notice in my signature that I work for a manufacturer of power amplifiers and we make both class D and class A/B.

Which do you prefer?

 

[Alice of Old Vincennes]

Different speakers and rooms have different issues.... a high efficiency (say 96db/wm) speaker with an impedance of 8ohm and no dips in impedance that drop below 6ohm.... will probably sound the same with almost any amp....

The amps are unlikely to clip or run out of current, unlikely to get strained by driving a low impedance load.... - all amps will run optimally - with negligible audible distortions - functionally they will all act as a wire with gain.

Then you take a speaker with a nominal 6 ohm impedance (shouldn't be too difficult for those amps) - but it has a funky tweeter with a capacitive response (that does things to phase) and impedance that drops below 2 ohm.... and a woofer crossover that also drops impedance down to around 2 ohm, and a 4ohm woofer....and make it average sensitivity... 86db/wm... now it is a whole different ball of wax!

The speakers will sound reasonable with most amps, they are sufficiently sensitive that huge power is not a requirement... but the low impedance of the tweeter and the woofer, is going to require a lot more current - and these aspects will put strain on the amp.
Under strain different amps respond differently - and their distortions vary in the frequency domain too - so very different sounds, drivent by differing distortion profiles.

If you provide this speaker with a mid power amp - say 100W @ 8ohm, but which is capable of driving low impedances with something around say 300W @ 2ohm - it will handle such speaker easily.

But many high quality high power amps aren't up to such a task... you will see specs that look like 150W@8ohm, then 220W @4ohm, but not rated and recommended for 2 ohm.... and although nominally higher power than the previous amp I described... this kind of amp will probably sound worse on the difficult low impedance dips speaker.... again how it sounds worse, will vary dramatically by the design of the amp itself.

Your Devialet amp is a very interesting design... it is a relative of the vintage "Current Dumper" design by Peter Walker - still used today in Quad amplifiers - this technique is also called "Feed Forward" - a small high quality amplifier, ensures high quality sound, while a large economical amplifier backs it up with plentiful current.

The Benchmark AHB2 (current flavour of the month/year here!) - is a similar Feed Forward design too

What Devialet has done differently, is that the small high quality amp is standard class A (AB?) - but the high power current support amp, is Class D.... an interesting design with a lot of great potential.

If you look at the amplifiers over the last 50 years that have used this type of topology, there are a lot of very highly regarded amplifiers - Quad 405/606/707/909/QSD/Artera, Nakamichi Statis, Threshold, Benchmark AHB2.

The Devialet brings the principles up to date by using Class D for the heavy lifting, makes perfect sense to me! - It would be interesting to get it measured and tested, like the AHB2 has been

 

[Alice of Old Vincennes]

3 or 4 way high quality speakers will strain. Not necessarily audible absent comparison.