Class D high frequency problems.

[Old_School_Brad]

I’m not recommending anything specifically. I’m questioning some of the practices used on here that don’t make sense to my simple laymans brain. In the pursuit of truth my objective is to learn about a subject about which I am passionate.

I would recommend against Tubes, pass, vinyl and other verifiable inferior coloration factories. The music is beautiful enough by itself and if you are relying on those to get your sonic desires fulfilled you haven’t explored camp rational enough.

I think it's unnecessarily harsh to dismiss people's personal preferences like that, especially when responding to someone who genuinely expressed interest and curiosity in a friendly manner.

Additionally, I feel your comments in this thread are looking for an issue that isn't really there -a point that several others have already mentioned.

 

[JRR]

I don’t mean to come across as harsh. I’ve been down that same road myself spending tens of thousands of dollars on various tube gear, megabuck record players, Silver cables etc. If I can help steer someone away from what I see as the wrong path. I will. He asked for recommendations, the best one I can give is to stay away from colored gear because, in my experience, it results in a never ending and expensive replacement hamster wheel.

My point, if it isn’t clear, is that I believe there may be more to explore with modern class D measurements. They are all on the top of the all important score board for SINAD. This only represents one frequency at which they clearly excel. Multitone tests are done at 5w at or below at which they all excel. But what about a multitone tests with all 32 tones being between 5 and 20khz and at 15w. In my optic there are indications that class D amplification starts to experience issues in this range.

 

[Doodski]

I don’t mean to come across as harsh. I’ve been down that same road myself spending tens of thousands of dollars on various tube gear, megabuck record players, Silver cables etc. If I can help steer someone away from what I see as the wrong path. I will. He asked for recommendations, the best one I can give is to stay away from colored gear because, in my experience, it results in a never ending and expensive replacement hamster wheel.

Good stuff. You are mostly all the way there. Being cautious and aware is a good thing but reading graphs and determining what is audible and what is not audible are two different things.

 

[JRR]

Good stuff. You are mostly all the way there. Being cautious and aware is a good thing but reading graphs and determining what is audible and what is not audible are two different things.

Do you think an IMD product at 7khz 70db down could be audible?

 

[Doodski]

Do you think an IMD product at 7khz 70db down could be audible?

70 dB down, no. Why are we rehashing the same stuff experts have already advised you on?

 

[NTK]

Here is Buckeye Amp's 2-tone 19/20 kHz IMD test at 50 W into 8 ohms.

Buckeye Amps: 2nd Generation Purifi EIGENTAKT 1ET9040BA Amplifier

From my point of view, this is not enough. I would expect at least the 5-Why Method here. The RC could also have affected other deliveries/products. I get it. I've been a participant and driver of dozens of '5-Whys.' Worthy efforts all, but they are expensive, intrusive, and often result in no...

www.audiosciencereview.com

 

[oleg87]

Audible, possibly, if you knew what to listen for with the right kind of material, but far from obviously.
Significant - well, I recommend simulating it for yourself. I reckon I could maybe just about pick out some nastiness -70dB down from a pure test tone. Masked by music, it'd be hopeless.

 

[JRR]

How Much Distortion Can We Hear With Music?

How much distortion can we hear with music? An experimental study at Axiom.

www.axiomaudio.com

During my own klippel tests I found test tones to be less revealing of distortion than music. Anomaly, maybe but in my case the numbers didn’t lie. I’m not fully convinced that an experienced listener couldn’t hear some odd order distortion 70db down in the high frequency bands above 10khz. Anyway, I find the resistance to measuring the top performers outside their obvious comfort zone quite interesting.

 

[oleg87]

Anyway, I find the resistance to measuring the top performers outside their obvious comfort zone quite interesting.

The information is in Amir's reviews, it is being measured. I think it's fine to care about it and make purchasing decisions accordingly. What Amir chooses to highlight in his evaluations of measured performance and where he chooses to measure the multitone is up to Amir (I would personally prefer to see a sweep, but I suppose that's hard to visualize without some fancy web app) - I personally don't invest much value in the SINAD charts or equipment clearing targets I know don't make much difference to my ears and listening setup.

 

[DavidMcRoy]

M

May I ask what you would recommend or that you use? Does it actually bother your listening experience?

I ask because I prefer valve amps. I'm one of those who crave distortion (as some people put it) and because of that I can listen for hours on end vs
30-45 minutes at most with amps like SST or the zero feedback models; Ampzilla being one of those.

At minimum a valve preamp. It seems the older the better. Add tone control and I can listen for days on end and do from time to time.

EX: C20, C11, C22 or even a Cary Audio SLP-05 or 98 with class Ds. I use NC500 oem with Nord Buffers (which I hear very little difference if at all)

The early ICE products EX: Wyred4Sound SX1000 (no R) are like taking a shower with a raincoat on, they are so sterile and the background is so dark.

Pass designs and the newer Atmasphere class Ds are simply music to my ears and one of the few SSs that cause ZERO fatigue as long as I can tweak the
HF a bit on the types of speaker drivers I use in a moderately treated room for HF boil.

What would you say is the ultimate amp is in your opinion? OR is this just an observation with no practical purpose other than noticing the chip in the paint
on a fender-skirt (so to speak) while looking at a Rolls Royce Silver Shadow?

I'd be interested in hearing your preference, vs looking at a chart.

With great regard! The OHM

I completely understand where you're coming from. A sprinkling of just the right added distortion components is more "entertaining" to my ear than a squeaky-clean signal. My Pass Aleph 3 circuit-based monoblocks with their two (2) amplification stages consisting of a single VFET front end followed by a pair of power MOSFETS, driving ESS Heil AMTs above 1200Hz, are engaing to listen to. My Fosi V3 Monos, being cleaner, sound more lifelike, though. Real sound doesn't have mist and sparkle enhancements. Still, as Nelson Pass notes, home audio is part of the entertainment industry. It's nice to be able to easily swap amps out at will. I've dabbled with the more convenient route of a using a selector and a line-level tube stage, but "the right one" remains elusive. Maybe I need a Pass Labs VFET preamp to switch in and out.

 

[Old_School_Brad]

Do you think an IMD product at 7khz 70db down could be audible?

In practical listening conditions, distortion at -70 dB is highly unlikely to be audible.

You referenced the Axiom Audio study, which found that distortion needed to exceed 1% (-40 dB) to be perceptible. You also mentioned the Klippel distortion tests -what were your personal results from those?

 

[OldHvyMec]

PASS labs. A famous producer of “flavored gear”.

I know that's why Pass is so unpopular with many individuals. LIKE ME. (pun intended) It makes most gear "LIKE" the older Klipsch horns, corner horns, Jubilee and just about anything produced in the 50s-80s listenable for me. It takes a little more than that for the most part, like room treatment, knowing the room, seated position placement of the speakers L/Pads and definite tone controls on the preamp. The older Mcintosh happens to check all the boxes (especially with the WAY they have manufactured their tone control)

If I had my rathers, a well built MC240/C11 or C22 with all factory recommended valves. Telefunken, Blackburn Mullard, or GOOD RCA BP (if you can find a pair or two)
make for a wonderful experience.

So as I suggested, it's not an actual issue for you, it's a question you're posing because you noticed IT (being the operative word) in a graph or two. As I said I have used
the same reasoning my whole life, "if it SOUNDS good it is good, if it measures good but sounds bad you're measuring the wrong thing OR worse there is something
missing in the actual design that tickles one's fancy. Some call it distortion and others call it flavor, I'm pretty sure it's akin to timbre and the individual's EAR preference.
My preferred gear has a definite personality even the same models are different in some cases.

For me the test is simple, do I get up and leave the room after 20-30 minutes or not? THAT is the measurement. If I can sit without pulling my own hair out the gear has
passed the most important test. Good subjective listening session (S) that I can repeat over and over to make darn sure if there is a problem like with Ampzilla ribbons/planar combinations. It just happens that Pass and Atmasphere's newer class Ds keep me from leaving the room. It's strange also that the combination
of a valve preamp like the aforementioned Macs and Carys added to the mix can soothe the nasty beast to a degree of keeping me in the room for a considerably
longer time. BTW the Cary is only tenable because of a bass management system that I use from 300hz <. It also requires a good source recording to begin with.
Some music reproduction mixes are just all F&^%ed up, to say the least. 90% of all Janis Joplin just happens to be one of those. I've never found a great recording yet
except on a RtR or two. I love her voice and passion but the recording just plain sucks!

I know many people don't have a mechanics-trained ear, they have to rely on certain instruments just as I do AFTER I leave a room, but if I can't get by sitting through
a concert, whether with records, RtR, cassettes or CDs what's the difference? Live concerts for the most part are even worse, It only took me 300+ concerts to finally
quit trying and at least a toilet paper roll of TP to figure it out, IF quality was what I was after. Prince, Carlos, and very few others could actually pull a good concert off
in my day of going to concerts.

If that distortion in many of the class D is a real problem, (in the real world, which it's not) a lot of people have been fooled and it's simply NOT the case. Most people
whether they like it or not couldn't tell 30% distortion from 20-250hz until it was actually removed and they were taught what to listen for. It's why they call it "a TRAINED EAR." I always use the tools AFTER I find or hear a problem NEVER before.

As I said before, With the greatest regard. The OHM

 

[RandomEar]

Just to clarify this again: THD at 10 kHz test frequency means harmonics at 20, 30, etc. kHz. That's inaudible anyway.

At 5 kHz, you might be more concerned but have to remember the Fletcher-Munson curve. The audibility threshold at 10 kHz (first harmonic) is roughly 15 dB higher than in the most sensitive range of our ears from 2-4 kHz. So you could effectively "discount" distortion at that frequency by another 15 dB from the already low levels they are measured at.

From a purely technical point of view, I agree that auch rise in distortion in the graphs doesn't look pretty, though.

 

[antcollinet]

They are given at 5W which coincidentally is the threshold at which THD+N starts to rise dramatically for several class D designs

You sure about that. Even the lowly V3 mono just starts to level off at this point (where it is no longer noise dominated). It still goes down from 5W then levels, then starts to rise again. It does't climb up to the 5W level until it is around 70W. Good luck hearing -100dB of distortion when your amp is pumping out 70W.

More up market designs, such as the Nilai 500 are, as you'd expect, better still. It doesn't climb back up to the 5W level of distortion (this time around -112dB) until it gets to 300W. Sort of puts to bed your claim that distortion rises from 5W. Unless (again) you are talking about high frequency distortion - in which case I'll state (again) that you can't hear the distortion products of anything above 10kHz.

And let's also point out that 5W is chosen since it represents a really very high average volume level for typical systems. Few people are going to be listening significantly higher than that. (peaks/transients can obviously go higher).

For example - with my system, 5W would represent about 95dB SPL. Some 15dB higher than my typical "high" level.

It is my understanding that IMD indeed can produce harmonics much closer to the fundamental than simple harmonic distortion

If it is IMD, they are not harmonics they are difference tones from the modulating frequencies.

In any case all the audible band IMD components, together with the HD components are shown in the multitone plot. And at levels much higher than with real music, since all the test tones are at the same level, rather than falling off at higher frequencies as music does. So the test actually exaggerates the impact of high frequency distortion.

in my optics the test sampling seems rather selective.

I'm going to suggest that is because you're not really understanding what the tests currently used are showing. The multitone test is a particularly stringent one that really shows all (and more) of the real world impact of distortion in the audible band.

 

[voodooless]

Let's borrow some data from Erin from his WiiM amp test (but you can also check others):

Like magic, we see THD+N go down past about 6.5 kHz. Is that real? No, it's obviously not. However, it does reveal that the distortion components live in the ultrasonic. The measurement bandwidth that Erin uses is quite a bit smaller (about 15 kHz).

 

[JRR]

From the snake oil peddlers at Benchmark audio (sarcasm):

“Unlike class-D amplifiers, the AHB2 is noise free over its entire 0.1 Hz to 200 kHz frequency range. Don't settle for the limited performance of class-D amplification”

 

[voodooless]

From the snake oil peddlers at Benchmark audio (sarcasm):

“Unlike class-D amplifiers, the AHB2 is noise free over its entire 0.1 Hz to 200 kHz frequency range. Don't settle for the limited performance of class-D amplification”

I’m sure the bats will be ready to cheer them on.

 

[JRR]

In practical listening conditions, distortion at -70 dB is highly unlikely to be audible.

You referenced the Axiom Audio study, which found that distortion needed to exceed 1% (-40 dB) to be perceptible. You also mentioned the Klippel distortion tests -what were your personal results from those?

The distortion I could hear was in the bass where we are less sensitive. I know what distortion sounds like in the midrange and up. Its a much more unpleasant sound.

 

[antcollinet]

The distortion I could hear was in the bass where we are less sensitive. I know what distortion sounds like in the midrange and up. Its a much more unpleasant sound.

View attachment 428025

Pretty good result - is that with headphone, or speakers?

 

[JRR]

I know that's why Pass is so unpopular with many individuals. LIKE ME. (pun intended) It makes most gear "LIKE" the older Klipsch horns, corner horns, Jubilee and just about anything produced in the 50s-80s listenable for me. It takes a little more than that for the most part, like room treatment, knowing the room, seated position placement of the speakers L/Pads and definite tone controls on the preamp. The older Mcintosh happens to check all the boxes (especially with the WAY they have manufactured their tone control)

If I had my rathers, a well built MC240/C11 or C22 with all factory recommended valves. Telefunken, Blackburn Mullard, or GOOD RCA BP (if you can find a pair or two)
make for a wonderful experience.

So as I suggested, it's not an actual issue for you, it's a question you're posing because you noticed IT (being the operative word) in a graph or two. As I said I have used
the same reasoning my whole life, "if it SOUNDS good it is good, if it measures good but sounds bad you're measuring the wrong thing OR worse there is something
missing in the actual design that tickles one's fancy. Some call it distortion and others call it flavor, I'm pretty sure it's akin to timbre and the individual's EAR preference.
My preferred gear has a definite personality even the same models are different in some cases.

For me the test is simple, do I get up and leave the room after 20-30 minutes or not? THAT is the measurement. If I can sit without pulling my own hair out the gear has
passed the most important test. Good subjective listening session (S) that I can repeat over and over to make darn sure if there is a problem like with Ampzilla ribbons/planar combinations. It just happens that Pass and Atmasphere's newer class Ds keep me from leaving the room. It's strange also that the combination
of a valve preamp like the aforementioned Macs and Carys added to the mix can soothe the nasty beast to a degree of keeping me in the room for a considerably
longer time. BTW the Cary is only tenable because of a bass management system that I use from 300hz <. It also requires a good source recording to begin with.
Some music reproduction mixes are just all F&^%ed up, to say the least. 90% of all Janis Joplin just happens to be one of those. I've never found a great recording yet
except on a RtR or two. I love her voice and passion but the recording just plain sucks!

I know many people don't have a mechanics-trained ear, they have to rely on certain instruments just as I do AFTER I leave a room, but if I can't get by sitting through
a concert, whether with records, RtR, cassettes or CDs what's the difference? Live concerts for the most part are even worse, It only took me 300+ concerts to finally
quit trying and at least a toilet paper roll of TP to figure it out, IF quality was what I was after. Prince, Carlos, and very few others could actually pull a good concert off
in my day of going to concerts.

If that distortion in many of the class D is a real problem, (in the real world, which it's not) a lot of people have been fooled and it's simply NOT the case. Most people
whether they like it or not couldn't tell 30% distortion from 20-250hz until it was actually removed and they were taught what to listen for. It's why they call it "a TRAINED EAR." I always use the tools AFTER I find or hear a problem NEVER before.

As I said before, With the greatest regard. The OHM

I get that the flavoring can add dimensionality. Two examples: tubes distort transients but softly enough to merely emphasize dynamics. Second order harmonics adds a sense of space due to subtle octave doubling. However this always carries secondary issues. Some recordings sound great where that same flavor doesn’t go with others. Consequently people subconsciously end up with a library that favors their system.

A clean system in an unforgiving room can sound worse. But with a clean signal, even dispersion, proper acoustic treatment, and cross talk cancellation nothing in the traditional subjectivist ether comes close. IMO.

At low levels my current chain is the best I’ve heard. My only gripe is that the Topping B200 power amplifiers don’t seem to have the power handling capability that the purify based amps do. Subjectively I do prefer the Topping A/B amplifiers.