Atma-Sphere Class D Mono blocks

[DasguteOhr]

Only when you inject enough to be audible. Studies have shown that up to 2% 2nd order is inaudible. The only odd order harmonic that is not unpleasant is 3rd.

The problem only arises with speakers under 100hz, which usually produce 3-5%k2 thd. when the two are added together it becomes clearly audible. For comparison, a completely overdriven amplifier makes 10% thd. Therefore, choose a speaker that produces little thd in the bass. Unfortunately, manufacturers do not specify this. Exceptions Neumann, Geithain but unfortunately mainly active speakers..pity

 

[DonH50]

Professor Don. I prefer to the Pass sound as polite.
i tried to resist.
Dr. Olive has claimed to have proven that listeners prefer lower distortion. You proffer that people like distortion. You can understand why I find those positions at odds.

My position is that the lower the distortion the better. I prefer to not change the source material. Nelson Pass claimed low levels of even-order distortion was preferred by listeners. There was an JAES paper, by a Japanese group I think, with similar claims using a small (~10) group of listeners. There were a lot of challenges to the study, among them dependence upon the material and listening environment, and small sample size. This is not my field so I am speaking (writing) from memory.

I probably should have used the word "claimed" instead of "shown" listener preference; I do not know how well-proven those claims are. And I do mean that literally -- I do not know. I have read other sources (Olive, Toole, Everest, etc.) that say lower distortion is preferred.

I did wade through the 100+ papers on the computer I am on now but did not find the distortion articles I had in mind. My notebook (not accessible at the moment) has many more so they are likely there, but I am disinclined to go hunting for them. This is Ralph's thread and has already diverged mightily, plus he is well aware of the research (marketing, whatever it is). Sorry, I'll go back to lurking; this is not the place for me, as I do not know much about psychoacoustics and such.

 

[Gregadd]

Nothing to apologize for. I can only account for what I kike. I am not even sure why. Did not mean to scare you away.
Merry Christmas.

 

[D.Duttilleux]

Dr. Olive has claimed to have proven that listeners prefer lower distortion. You proffer that people like distortion. You can understand why I find those positions at odds.

The problem with modern "textbook science" is that it's largely based on short-term ABX listening tests. 90% of the grey sounding (MOR) high-end gear is based on such science. Real science includes intangible aspects that are more difficult to measure. Unfortunately, we live in a world dominated by 'scientism'.

The (older) research/findings referenced in this thread are (imo) more valuable.

 

[D.Duttilleux]

The problem only arises with speakers under 100hz, which usually produce 3-5%k2 thd. when the two are added together it becomes clearly audible. For comparison, a completely overdriven amplifier makes 10% thd. Therefore, choose a speaker that produces little thd in the bass. Unfortunately, manufacturers do not specify this. Exceptions Neumann, Geithain but unfortunately mainly active speakers..pity

If you want low THD in the bass, choose bass horns.
Btw, our hearing (ear-brain) isn't particularly sensitive to distortion in the bass. However, wrt IMD it's another story.

And here too, many misconceptions are dominant.
I refer, among other things, to the Grimm white paper on bass reflex versus closed...

 

[Gregadd]

The problem with modern "textbook science" is that it's largely based on short-term ABX listening tests. 90% of the grey sounding (MOR) high-end gear is based on such science. Real science includes intangible aspects that are more difficult to measure. Unfortunately, we live in a world dominated by 'scientism'.

The (older) research/findings referenced in this thread are (imo) more valuable.

There is nothing wrong with science per se. The issue is what can I sell? How can convince the consumer to buy my product. My science says my product is best. Why will you not buy it?
Anytime you follow a formula the results will not be perfect. You are going to have to apply some art and trial and error to get it "right."

 

[D.Duttilleux]

There is nothing wrong with science per se. The issue is what can I sale? How can convince the consumer to buy my product. My science says my product is best. Why will you not buy it?
Anytime you follow a formula the results will not be perfect. You are going to have to apply some art and trial and error to get it "right."

The dominant trend in hi-fi since the 1960s has been miniaturization. In terms of speaker technology, this is the root for 'the modern approach'.
A similar trend is also visible in pro audio (smaller, more portable systems), albeit to a lesser extent.
These developments are at odds with, or at least at the expense of, a number of fundamental issues directly related to - what can be referred to as - 'natural sound'.

Translated into technical terms, this concerns topics such as:
- Efficiency/sensitivity in relation to (lower) impedance (> amplifier power became increasingly cheaper, which fueled miniturarization).
- Excursion > most subwoofer drivers resemble air pumps.
- Materials: High-tech and exotic materials may lead to 'better' measurements, but say little about the 'listening experience'.

Etc.

The promise of Class D is to (also) achieve enormous efficiency gains in 'amplification'. Imo, 2x 40 watts should be more than sufficient (for home use).

 

[Atmasphere]

If you have only odd order distortion they are unmasked, Period. Show me literature to the contrary or stop making the claim. 3rd doesn’t “stretch” to cover 5th etc. This is why it’s so audible and Keith Howard commented as he did.

What makes makes you think I haven’t heard differential circuits? I have owned all kinds of gear…including true balanced stuff like BAT and Einstein and circlotron amps. I find SET to be more authentic and less synthetic sounding.

Occam's Razor is the oldest principle in science, being about 850 years in use. I'm guessing that since you made the argument above, after my suggestion to use it, that perhaps you don't know what it is (no judgement).

Occam's Razor tells us that, given two explanations for a thing, the simpler one is likely the correct one. Our amps and preamps are known for being smooth and not irritating. Since they are fully differential and thus make mostly odd ordered harmonics and as you suggest, the 3rd can't mask them, they must somehow be smooth on some other account that no-one has really figured out. OR, the simpler explanation is that 3rd really can mask higher orders and how ever you got the idea that it can't is simply wrong.

FWIW, its well-known in the speaker world that a 3rd harmonic will mask higher orders. Why is that so for loudspeakers and not amplifiers?? Occam's Razor again...

My proof has been in the pudding for a long time (Atma-Sphere is one of the older manufacturers in high end audio). You have to demonstrate that somehow in the face of all this that what you are saying is really true. You might ask Keith about this instead of me- he might be able to convince you better if he's properly informed. I've been offered teaching positions in the past but I prefer the work I'm doing instead. It reaches more people.

I've spent a lot of time working with differential amplifiers. I know you've not kept up with everything I write so I'll put it in a nutshell: most differential amplifiers leave performance on the table because the CCS used is poor performance. This is very common in solid state amplifiers where you have an issue if the differential input has too great performance- quite often they have to tone it down a bit to avoid other (unrelated) problems. Link- its all good but the relevant text is on page 12. At any rate, we're not constrained by that issue and so have had better CCS circuits in our tube stuff for a really long time. Once they are set up right you unlock what the differential amplifier can really do: wider bandwidth, lower noise, lower distortion and greater gain all at the same time (which is rare; often one of those parameters is sacrificed for another...). I suspect the amps you've heard really don't demonstrate what this is about. Heck, our amps didn't either until I sorted this out in the late 1990s.

 

[Atmasphere]

The promise of Class D is to (also) achieve enormous efficiency gains in 'amplification'. Imo, 2x 40 watts should be more than sufficient (for home use).

People have asked why we didn't make a more powerful amp but this comment is really what its all about. We figured that if 100 Watts wasn't enough, you would really need 1000 Watts at that point, since you need 10X more power to only sound twice as loud. Speakers that need so much more power really aren't practical.

 

[DasguteOhr]

If you want low THD in the bass, choose bass horns.
Btw, our hearing (ear-brain) isn't particularly sensitive to distortion in the bass. However, wrt IMD it's another story.

And here too, many misconceptions are dominant.
I refer, among other things, to the Grimm white paper on bass reflex versus closed...

Unfortunately it's far too big to produce really deep bass, I once had 2 corner horns. my next project ripol sub 2× 15" -3db at 22hz.
Living space friendly, very quick due to dipole radiation, installation uncritical in the room.
Wide, deep and height dimensions under 20 inches in all active up 80hz.

 

[DonH50]

A pair of 85 dB/W/m speakers out in the room need about 1 W to achieve 80 dB SPL for a listener 8' away. Depending on the source you need 17~20+ dB headroom for peaks, and a 20 dB increase means you need 100 W from the amplifier. Obviously horns need only a fraction of that power, but I suspect for many of us 100 W is a good number for an amplifier. No matter the topology, type of active devices, etc.

 

[morricab]

Occam's Razor is the oldest principle in science, being about 850 years in use. I'm guessing that since you made the argument above, after my suggestion to use it, that perhaps you don't know what it is (no judgement).

Occam's Razor tells us that, given two explanations for a thing, the simpler one is likely the correct one. Our amps and preamps are known for being smooth and not irritating. Since they are fully differential and thus make mostly odd ordered harmonics and as you suggest, the 3rd can't mask them, they must somehow be smooth on some other account that no-one has really figured out. OR, the simpler explanation is that 3rd really can mask higher orders and how ever you got the idea that it can't is simply wrong.

FWIW, its well-known in the speaker world that a 3rd harmonic will mask higher orders. Why is that so for loudspeakers and not amplifiers?? Occam's Razor again...

My proof has been in the pudding for a long time (Atma-Sphere is one of the older manufacturers in high end audio). You have to demonstrate that somehow in the face of all this that what you are saying is really true. You might ask Keith about this instead of me- he might be able to convince you better if he's properly informed. I've been offered teaching positions in the past but I prefer the work I'm doing instead. It reaches more people.

I've spent a lot of time working with differential amplifiers. I know you've not kept up with everything I write so I'll put it in a nutshell: most differential amplifiers leave performance on the table because the CCS used is poor performance. This is very common in solid state amplifiers where you have an issue if the differential input has too great performance- quite often they have to tone it down a bit to avoid other (unrelated) problems. Link- its all good but the relevant text is on page 12. At any rate, we're not constrained by that issue and so have had better CCS circuits in our tube stuff for a really long time. Once they are set up right you unlock what the differential amplifier can really do: wider bandwidth, lower noise, lower distortion and greater gain all at the same time (which is rare; often one of those parameters is sacrificed for another...). I suspect the amps you've heard really don't demonstrate what this is about. Heck, our amps didn't either until I sorted this out in the late 1990s.

Nice attempt at pretending to not insult. I guess you will get a pass on your snide remarks because you are a manufacturer. I am quite sure my knowledge on philosophy and history of science is quite up to snuff (Ph.D in Analytical Chemistry and minor in philosophy from undergrad). I am sure Wikipedia was your


The actual formulation is not to make things more complicated than necessary. Also, this is philosophy and not a scientific principle. And while true you often start with as simple a hypothesis as possible, that doesn’t make it more or less right.
Just FYI, your amps (OTLs) actually have a reputation of being very transparent but somewhat lean sounding.

Speakers are different because they have only primarily low order distortion. Otherwise we wouldn’t hear any of the small (absolute) distortion levels as they would be masked by the much larger speaker distortion. 3rd might mask the 4th but not the 5th…especially if the third is low in level to begin with. You want to stretch the blanket too far to justify your design decisions.

Keith doesn’t have a vested interest in the outcome…you clearly do.

 

[Gregadd]

KISS

 

[morricab]

A pair of 85 dB/W/m speakers out in the room need about 1 W to achieve 80 dB SPL for a listener 8' away. Depending on the source you need 17~20+ dB headroom for peaks, and a 20 dB increase means you need 100 W from the amplifier. Obviously horns need only a fraction of that power, but I suspect for many of us 100 W is a good number for an amplifier. No matter the topology, type of active devices, etc.

Are you counting the fact that a pair of 85dB speakers is actually louder per watt as it’s a pair? What rate of drop off with distance are you using? You realize in a real room the drop is often significantly less than theoretical because of reflections and room resonance? I would expect something closer to 83-85dB at 8 feet for one watt. Most popular recorded music has quite limited dynamic range…only some jazz and classical have significantly larger range. I would expect 30 watts to be plenty.

 

[Dogberry]

I never expected Ralph to get as rude a reception here as he did at ASR. Please remember your manners.

 

[DonH50]

Are you counting the fact that a pair of 85dB speakers is actually louder per watt as it’s a pair?

Yes, that is why the sentence starts with "A pair". That provides a 3 dB increase in SPL at the listening position and is included in the numbers.

What rate of drop off with distance are you using?

I said "out in the room" so that is without room reinforcement, a worst-case scenario, 6 dB loss for each doubling in distance. Typical rooms exhibit somewhere between 3 and 6 dB depending upon treatment and stuff in the room. My room is heavily treated and is very close to 6 dB. Upstairs, the family room is fairly live, and closer to 3 dB for each doubling in distance from the speakers.

You realize in a real room the drop is often significantly less than theoretical because of reflections and room resonance?

Yup, and it also depends upon the radiation (dispersion) pattern of the speakers. My old Maggies send most energy forward and back once you get above the bass region (where they devolve to a point source), little to the sides, so pick up a little in sensitivity (requiring less power) compared to my current Revel Salon2 speakers. Room modes and SBIR (speaker boundary interference response) will peak (and reduce) the response but that is usually undesirable, since it changes the frequency response, so such peaks are often reduced via EQ, room treatment, positioning, and so forth.

I would expect something closer to 83-85dB at 8 feet for one watt. Most popular recorded music has quite limited dynamic range…only some jazz and classical have significantly larger range. I would expect 30 watts to be plenty.

Well, I have a lot of jazz and classical music, but even some well-recorded rock music has dynamic range pushing 20 dB.

I provided an example, so it certainly does not apply to all, but over the years ~100 W is what I found worked for me. I have usually had rather insensitive speakers and listened at 10~15 feet away. The past decade or so I have tried to meet the THX spec of 105 dB/speaker at the listening position and that has required more power than that in my system. If you only need 30 W that is great. I know others who need much more since they sit much farther away with similarly (in)sensitive speakers and listen louder than I, and on the flip side also have friends with highly-sensitive horn systems who need much less power. I provided a very specific example but one that has served me well for decades as a rule of thumb. Everyone has different thumbs, no worries.

 

[D.Duttilleux]

One issue (of several) related to miniaturization (from a Purifi AES paper):

"A major breakthrough was the work by Cunningham in 1949, who analysed the inherently non-linear response of the motor due to magnetic effects with surprising depth of insight.

The drive towards speakers with long strokes, full audio range and high linearity in very small form factors makes Cunningham’s analysis even more important today than in 1949 when low power amplifiers dictated the use of speakers with large diaphragms and low excursion. Force factor modulation is a much greater problem today."

Not every technological development automatically translates into progress.

 

[Atmasphere]

Nice attempt at pretending to not insult. I guess you will get a pass on your snide remarks because you are a manufacturer. I am quite sure my knowledge on philosophy and history of science is quite up to snuff (Ph.D in Analytical Chemistry and minor in philosophy from undergrad). I am sure Wikipedia was your


The actual formulation is not to make things more complicated than necessary. Also, this is philosophy and not a scientific principle. And while true you often start with as simple a hypothesis as possible, that doesn’t make it more or less right.
Just FYI, your amps (OTLs) actually have a reputation of being very transparent but somewhat lean sounding.

Speakers are different because they have only primarily low order distortion. Otherwise we wouldn’t hear any of the small (absolute) distortion levels as they would be masked by the much larger speaker distortion. 3rd might mask the 4th but not the 5th…especially if the third is low in level to begin with. You want to stretch the blanket too far to justify your design decisions.

Keith doesn’t have a vested interest in the outcome…you clearly do.

I mentioned Occam's Razor since your posts suggested no familiarity. When there is a convoluted explanation that doesn't make sense, I tend to suspect a made up story; misapplied studies, theories and so on.

Your comment about 'lean sounding' is contradicted customer feedback. But I've always been careful to vet the customer's speakers so as to make a sale in which everyone benefits. Since our OTLs have no feedback and a very high output impedance, choice of the loudspeaker is really important for success. I've no doubt they can sound lean on certain speaker loads, just as an SET (pot, meet kettle...).

You put your finger on why our OTLs work the way they do with your comment about speakers Apparently, no need to talk to Keith then- you sorted that there are exceptions.

 

[morricab]

I mentioned Occam's Razor since your posts suggested no familiarity. When there is a convoluted explanation that doesn't make sense, I tend to suspect a made up story; misapplied studies, theories and so on.

Your comment about 'lean sounding' is contradicted customer feedback. But I've always been careful to vet the customer's speakers so as to make a sale in which everyone benefits. Since our OTLs have no feedback and a very high output impedance, choice of the loudspeaker is really important for success. I've no doubt they can sound lean on certain speaker loads, just as an SET (pot, meet kettle...).

You put your finger on why our OTLs work the way they do with your comment about speakers Apparently, no need to talk to Keith then- you sorted that there are exceptions.

What do you mean Keith? You mean the ZOTL? That’s not a real OTL. It uses some kind of high frequency transformer and carrier signal…something funky but not OTL.
Keep grasping at straws Ralph.

 

[Atmasphere]

What do you mean Keith? You mean the ZOTL? That’s not a real OTL. It uses some kind of high frequency transformer and carrier signal…something funky but not OTL.
Keep grasping at straws Ralph.

No, I was referring to the same Keith as you.

Despite the IMO misleading acronym, Berning's amplifier topology is brilliant.