Toward a Theory To Increase Mutual Understanding and Predictability

[Bobvin]

Whoa baby, this thread really has gotten a lot of attention in a few days... try reading from the top it'll take a while. And in the process cover ground that has come up in so many other threads.

I appreciate those folks who are so passionate about this hobby, I learn a lot from them. And many of you have far more years experience than I, or theoretcal backgrounds which guide you. I fear, for myself, of getting too deeply into thinking about this hobby that I'll become like Mark Twain as the riverboat captain who, once having learned every eddy and sand bar in the river, discovered the magic and allure of the river, experienced when young and naive, is lost to him.

I think what Ceasar stated, way back in post 24(?), resembles my goals, which are mostly to create a system that allows me to get lost in the music. I don't listen to music daily, and listen to very little background music (except in the car). But after a week or 10 days I need a fix, and then I open a nice bottle of wine and retire to my music room. I want a baratone sax to grumble, a bass guitar to thump, and drums to give me a bit of a punch. I want to 'get' what Miles was playing when he was one with his trumpet. Interestingly, I can sometimes get the meaning of the music listening to my factory car radio. I want to be moved by and astonished by the sound. I have no illusions of being able to remember small details of the sound of live symphony weeks after attending. My cousin is a conductor and played and listened to music his entire life, He knows what live music sounds like.

As for gear choices, I like some things other might call colorations. There are systems too dry and analytical for me, others too warm and syrupy. And that can change for me as I am exposed to new gear, and is also limited by budget. I know damn well upgrading my cables, both power and signal, will bring the music a little closer (getting rid of noise). I know my DAC is a weak link, and I want to hear what my system will sound like with more power or different amplification.

But most of all I want to be wowed by music, and enjoy watching the little smile that crosses guests faces who've forgotten the joy of listening to records. (And a confession, I want my audiophile buddies to honestly tell me my system sounds amazing.)

I suppose therefor, I am a 3 wrt the original question.

 

[bonzo75]

I guess it's ok to chase either of the 3 or more points as long as people are consistent with what they are doing and what they think they are doing

The problem in audiophilia is that people delude themselves.

 

[microstrip]

Sorry, early 2000s is not old for scientific research...it is highly relevant to today's designs. Not much has changed in SS or tube design in the last 20+ years, despite weekly claims for "new" and "amazing" breakthroughs by the audio press. There are plenty of examples...just go look for yourself...I am not going to spoon feed you.(...)

Just one point - IMHO a lot changed during the last 20 years. We have many new devices and, mainly, a new way of designing circuits. The computer aided design techniques and the powerful electronic simulators have opened doors to a better knowledge and understanding of the circuits. Very small effects, that would be impossible to study empirically, can be simulated and analyzed. Some time ago, Dan Agostino explained to a small group how he designed his new amplifiers - I was there and it was really interesting. I am sure he will not publish a paper on it, but it was enlightening. The younger generation learned from their masters and have access to new tools. They can control critical aspects of performance in new ways. Many other designers have their own ideas and we just see the tips of the iceberg on them.

I repeat myself, but considering we can keep on using just the same old over beaten information easily available on the net because nothing has changed in the last 20 years will not lead to a better understanding of the psychoacoustics of electronics. Specially when these articles are just essays, that should be debated as so, not accepted as general knowledge. YMMV.

Release the schematic of a power amplifier on the net and ask a few people to build it. If properly built they will all measure the same in terms of classical measurements, including distortion spectra. But they will probably sound very dissimilar. I know about it, long ago I assembled a perfect copy of the famous Quad 405 - even the critical coil was carefully copied. In case of doubt a better specification component was chosen. But it did not sound like the original at all.

Interviews with designers are interesting but in some sense always frustrating - they give us enough to make us understand there is something really interesting behind there products, but never release the critical part.

 

[DaveC]

What I am saying about electronics is anything but gibberish. It is clear that different electronic circuit topologies create different distortion patterns. It is also clear that some of these are better tolerated by the human ear/brain than others. Read Geddes distortion papers and the Cheever thesis for why this is likely to be true. Jean Hiraga noticed this as well in the 1970s that a monotonic distortion pattern was a preferrable pattern sonically. That means a regular order of even and odd harmonics in an exponential decay. This is what Cheever showed to be the same pattern as what the human ear/brain self-generates. He argues that if the electronics mimic that pattern, at ALL SPL, then it will sound essentially pure becuase it will be perfectly masked.

If "my" idea of ignoring/masking makes little sense to you it is becuase you have been ignoring about 60+ years worth of research into the area of human hearing. Suffice to say it was not my idea but has come out of research and it is a real phenomena. It is well known that the harmonics closest to the fundamental are masked and preferrentially in the direction of the higher(overtone) not lower (sub) harmonics and that the degree of masking is SPL dependent. So, what I am describing has absolutely nothing to do with the recording and engineering choices...I am talking about something much more fundamental and you clearly didn't realize it.

Yes, most of the tubes are the same but I think the power supplies are far more stable and this reduces the distortion. Tubes are more linear than any transistor, which is why no feedback as a concept can work with tubes and is dubious at best with SS. However, a single ended MOSFET, running Class A, should only produce simple and low order harmonics. This makes it potentially a good output device for a single ended Class A amplifier without feedback. I have had such an amp and while it did not sound exactly like a normal SET it was still quite good and had better grip on the bass than most (but not all) SETs.

I never said noise control did a magical anything. You were the one banging on about noise. I simply agreed with you that it is important to control. I run my sources off of a Monarchy Audio Power regenerator for just that reason...it reduces noise and increases transparency and resolution of my sources by a significant margin.

Yes, power supplies in SE tube amps are ultra-important, maybe the most important factor...

The largest advance modern tech has given tube amps are volumetrically efficient and ultra-low ESR film capacitors. In the last preamp I built the final CR filter used 150 uF film caps and they are smaller than a 12 oz beer can! In the past people have built power supplies using older tech like oil-filled motor run caps, but see what it takes to get to 150 uF / 600V using them! You'd need a whole separate chassis to house them, they'll weigh a ton and cost a lot too.

 

[amirm]

You are wrong about strerile electronics and high order harmonic distortion. Go look at measurements at Soundstage and Stereophile. You will see clearly that most amps still generate significant amounts of high order harmonics.

I'd love to see a reference to some of these. Do you mind providing an example of this?

 

[DaveC]

Can't agree, Folsom - morricab is expressing an opinion that I agree with.

It all boils down to psychoacoustics & the mechanisms by which auditory processing works.
Room reflections are distortions in so far as they are not part of the original recorded sound but we are accommodated to them from our experience of how acoustic spaces behave & our auditory processing is able to analyse & separate these reflection distortions into a separate acoustic stream. We are therefore able to focus on other aspects of the soundstream & virtually ignore room reflections. Exactly the same as we do when listening to & following a single conversation in a room full of babbling voices - we are analysing the nerve pulse signals arising from ALL the vibrations hitting the eardrum & grouping a certain subset of vibrations into the stream that represents the conversation we are trying to follow. By focussing on this conversation we are consciously oblivious to the babble in the room unless it intrudes our consciousness & it does sometimes. But even though we are not conscious of this babble it is affecting our listening - we are straining more than normal & given a chance we will try to avoid the additional energy necessary to do this.

But what happens if we eliminate these room reflections altogether by listening in an anechoic chamber? The most often expressed opinion is that it sounds weird & unnatural i.e our auditory processing recognises this as a sound not encountered in nature. Again people do not want to stay in this environment for too long.

The lesson seems to be that our auditory processing system is comfortable with a level of distortion that we have been exposed to in our contact with the world of sound & we know we are comfortable with.

Again all instruments have amplitude & frequency modulations in their sound envelopes as a result of the non-linear aspects of these instruments - it's what gives them their timbre. I believe it's the accurate reproduction of these fine FM & AM signals (low level signal linearity) that is at the heart of realism & immersion. These low level signals have to be preserved in their journey through the audio electronics on the way to the speaker. it's what can often differentiate one instrument among a number of the same between one type of Unnaturally high levels or distortion that

This is exactly what morricab is saying & which I agree with - it's our auditory processing mechanisms that have to be satisfied if we want believability & envelopment (immersion) from our audio playback.

Accuracy is a wasteful goal, IMO - firstly we have failed many times in the past with measures of accuracy in audio & I have no doubt we are currently still off the mark. Secondly, input to output is not 100% accurate - the judgement of 'accuracy' is premised on the errors measured being below audibility so we are already talking about a qualified accuracy. Now there are a lot of holes in this qualification - are the measured errors fully characterising the playback system in all its performance, including it's behaviour with a dynamic, non-repeating, chaotic signal such as music? Can this be truthfully guaranteed?

At this juncture, in audio playback development, we have peeled off enough layers of the onion that we are now at a crossroads - we know that the illusion of realism & immersion/envelopment are possible - a lot of us (I won't say all) have heard it or heard hints of it - do we endeavour to achieve this in a stable way by improving 'accuracy' even though it is considered by many to be 'perceptually accurate', anyway? There's a disconnect here, I believe.

As morricab eloquently says - much better to ignore those distortions that are in the blind spot of our perception & expend our energy on those aspects that have more significance (even though they may well be considered

Wow, I replied to this and it just disappeared into the ether, that's frustrating because I hate doing things twice!!!

On the bolded portion... only to a certain degree and not nearly as much as it seems you believe.

The time of arrival of the reflection is very important, longer reflections can make things sound more natural but shorter reflections smear the fine detail that creates an immersive soundstage.

I've experimented with this by placing the same driver on a flat baffle vs a waveguide. The only difference was the driver on the flat baffle had a dustcap designed to control the highs, the waveguide linearizes the driver w/o cap.). The driver in the waveguide produces better detail, better clarity and is much more immersive vs the driver on the flat baffle.

 

[morricab]

Yes, power supplies in SE tube amps are ultra-important, maybe the most important factor...

The largest advance modern tech has given tube amps are volumetrically efficient and ultra-low ESR film capacitors. In the last preamp I built the final CR filter used 150 uF film caps and they are smaller than a 12 oz beer can! In the past people have built power supplies using older tech like oil-filled motor run caps, but see what it takes to get to 150 uF / 600V using them! You'd need a whole separate chassis to house them, they'll weigh a ton and cost a lot too.

Aries Cerat is using these so-called "super caps" which have ultra low ESR and weigh a whopping 3kg each!! (Diana integrated uses three of these).

 

[jkeny]

Wow, I replied to this and it just disappeared into the ether, that's frustrating because I hate doing things twice!!!

On the bolded portion... only to a certain degree and not nearly as much as it seems you believe.

The time of arrival of the reflection is very important, longer reflections can make things sound more natural but shorter reflections smear the fine detail that creates an immersive soundstage.

I've experimented with this by placing the same driver on a flat baffle vs a waveguide. The only difference was the driver on the flat baffle had a dustcap designed to control the highs, the waveguide linearizes the driver w/o cap.). The driver in the waveguide produces better detail, better clarity and is much more immersive vs the driver on the flat baffle.

Yes, I agree - I kept it simple to make the point but it is all premised on the fact that the room is a sonic normal room in which speech is intelligible
I've read David Greisinger's work on room(hall) reflections, intelligibility, engagement & envelopment & to summarise, his LOC measurement of room intelligibilty plots the power output of direct sound Vs the total power output of the reflected sound. The measure of intelligibility/engagement is shown on this plot.

Apart from the power, I believe the arrival time of reflections determines whether it is perceived as part of the original sound or as an echo - the Haas effect - somewhere above 100mS delay results in perception of an echo, below this perception of ambience

 

[DaveC]

Yes, I agree - I kept it simple to make the point but it is all premised on the fact that the room is a sonic normal room in which speech is intelligible
I've read David Greisinger's work on room(hall) reflections, intelligibility, engagement & envelopment & to summarise, his LOC measurement of room intelligibilty plots the power output of direct sound Vs the total power output of the reflected sound. The measure of intelligibility/engagement is shown on this plot.

Apart from the power, I believe the arrival time of reflections determines whether it is perceived as part of the original sound or as an echo - the Haas effect - somewhere above 100mS delay results in perception of an echo, below this perception of ambience

I'd also say I completely disagree with Harmans findings and call into question the testing and the quality of their test systems. Their findings that reflections with shorter delay times, such as 1st reflections of sidewalls and floor are actually beneficial is simply a reflection (oh no...) of their predisposition and acclimation to that kind of sound.

As I've mentioned, a controlled directivity speaker can sound odd after getting used to a direct radiator and vice versa... this is a really big deal and can confound these kind of tests if you're not careful. No mention of that so I disregard Harman's results as deeply flawed. IME 1st reflections are one of the largest acoustic issues with direct radiating speakers and what requires them to be placed in a carefully controlled acoustic environment for best results. Combined with the typical xo points and modern hard diaphragm materials it makes it very difficult to get decent results. This is the main reason so many high $ systems sound very ordinary imo. Modern speakers start out choosing the wrong set of compromises and make the most important stuff much more difficult to achieve.

 

[jkeny]

I'd also say I completely disagree with Harmans findings and call into question the testing and the quality of their test systems. Their findings that reflections with shorter delay times, such as 1st reflections of sidewalls and floor are actually beneficial is simply a reflection (oh no...) of their predisposition and acclimation to that kind of sound.

As I've mentioned, a controlled directivity speaker can sound odd after getting used to a direct radiator and vice versa... this is a really big deal and can confound these kind of tests if you're not careful. No mention of that so I disregard Harman's results as deeply flawed. IME 1st reflections are one of the largest acoustic issues with direct radiating speakers and what requires them to be placed in a carefully controlled acoustic environment for best results. Combined with the typical xo points and modern hard diaphragm materials it makes it very difficult to get decent results. This is the main reason so many high $ systems sound very ordinary imo. Modern speakers start out choosing the wrong set of compromises and make the most important stuff much more difficult to achieve.

Have you read & how do you find Greisinger's work?

 

[YashN]

Yes, power supplies in SE tube amps are ultra-important, maybe the most important factor...

Indeed. Line level PSUs are super important too since whatever occurs there is then amplified in the chain.

Actually, all the PSUs are important, both within the chain, and also out of the audio chain if you haven't worked on proper isolation or a dedicated line.

 

[microstrip]

The largest advance modern tech has given tube amps are volumetrically efficient and ultra-low ESR film capacitors. In the last preamp I built the final CR filter used 150 uF film caps and they are smaller than a 12 oz beer can! (...)

Aries Cerat is using these so-called "super caps" which have ultra low ESR and weigh a whopping 3kg each!! (Diana integrated uses three of these).

Can you give us the typical values of the ESR per capacitor? Or even better than a few figures the graph of ESR versus frequency.

 

[DaveC]

Can you give us the typical values of the ESR per capacitor? Or even better than a few figures the graph of ESR versus frequency.

http://www.claritycap.co.uk/products/tc.php

Subjectively, the sound is smoother, less noise, much cleaner and more dynamic. It's a pretty massive difference vs electrolytics.

 

[DaveC]

Have you read & how do you find Greisinger's work?

Only some of it, I'll admit I'm not as familiar as I probably should be. I think there are some fundamental differences between music halls and stereo systems but there's certainly some overlap as well.

 

[microstrip]

http://www.claritycap.co.uk/products/tc.php

Subjectively, the sound is smoother, less noise, much cleaner and more dynamic. It's a pretty massive difference vs electrolytics.

Thanks. Conrad johnson had since long been using this type of capacitors - I remember that the main difference from the MV100 to the MV125 was just the change from electrolytic capacitors to polypropylene ones. I had both side be side for some time and the difference was nigh and day! The amplifier with the electrolytic sounded slow and less detailed that the MV125.

Currently Audio Research uses a different technique to avoid the "slow" effect due to electrolytic capacitors - a mix of bypass film capacitors and a serial inductor with each of them bypassed by capacitors with some resistors.

 

[bonzo75]

The UK restorer uses Clarity caps in Apogees, which Justin (User 211) owns. But Henk uses Mundorf, Mundorf Supreme, Obbligato, etc, and whether its the caps or the crossover points that sound better, I do not know.

 

[microstrip]

Have you read & how do you find Greisinger's work?

I went through it some time ago. As far as I remember the acoustics part was mainly focused in great concert halls and classrooms, not audiophile listening rooms. One of the few people to address vertical localization of humans clearly.

 

[DaveC]

Thanks. Conrad johnson had since long been using this type of capacitors - I remember that the main difference from the MV100 to the MV125 was just the change from electrolytic capacitors to polypropylene ones. I had both side be side for some time and the difference was nigh and day! The amplifier with the electrolytic sounded slow and less detailed that the MV125.

Currently Audio Research uses a different technique to avoid the "slow" effect due to electrolytic capacitors - a mix of bypass film capacitors and a serial inductor with each of them bypassed by capacitors with some resistors.

Yeah, it's possible to come really close by bypassing electrolytics but with the Clarity TC you just need the one cap and while I have not played with it, apparently it's possible to use much less capacitance vs electrolytics as well. That would be a key point for manufacturers, but for me I just bought the same value as you'd use in electrolytics.

I used to work for Vestas Wind Turbines, that industry is actually the largest motivator for development of volumetrically efficient film caps and there are lots of massive low-ESR film caps available now as a result that would work great for tube amp power supplies. A new turbine coming out when I left the company filled a very large volume of the Nacelle with capacitors... the entire floor of the Nacelle was thicker than normal and completely filled with caps.

I would also say the difference is night and day . The driver section is key too, and usually a bottleneck in a typical SET amp, and can also be responsible for slow, syrupy sound from a SET... My SET, with a good driver using it's own power supply and output tubes with their own PS, is hardly recognizable as a tube amp anymore, but retains the many advantages through using very few parts of very high quality. I've heard SS amps that sound more "tubey" than my SET amp...

 

[Folsom]

Sorry, early 2000s is not old for scientific research...it is highly relevant to today's designs. Not much has changed in SS or tube design in the last 20+ years, despite weekly claims for "new" and "amazing" breakthroughs by the audio press. There are plenty of examples...just go look for yourself...I am not going to spoon feed you.

"At what point have I said amplifiers don't sound different, despite their similar distortion measurements? I propose reasons for that all the time, and say that distortion figures are almost meaningless for how good something will be since it's subjective and shows pretty much no correlation."

What do you propose then other than an amps deviation from linearity for its sound character?

Yes, THD and IMD as single numbers are meaningless...Geddes demonstrates as much, showing no correlation between THD and IMD and listening tests. I am not talking about nor have I been talking about those metrics. The content behind those metrics matters though. The metric of Geddes or Cheever show much more promise...hell even the metric from the 1950s by BBC engineer D.E.L Shorter works better.

"I just said exactly that designers aim for 3rd harmonic or lower"
What does this mean? You are not writing clearly here. 3rd harmonic or lower what?? Go look at amp measurements, you will see many amps with harmonics out to at least 20th as common. These negatively impact sound quality. Antiquated?? I think not. BTW, Pass talks the talk, but measurements of his amps shows he is not walking the walk.

"It's not exactly hard, or new, to model the types of harmonic distortions topologies will produce. "

Apparently you are wrong. And it seems to be very hard to design amps to fit to psychoacoustic hearing models...only one mfg. even claims this.

"Atkins says despite the higher order harmonics than you'd expect, they don't appear to play any role"

He has not based his statement on anything. Studies show that they will matter and will influence the outcome of the sound. Look at the comparison MF did between the Lamm ML3 and the big Dart. You will see that the descriptions of the Dart have everything to do with "speed" "transparency" "slam" but not about correct tonality and image/soundstaging , for which he praises the ML3. Clearly those harmonics matter. High order harmonics are known to provide an "edge" to the sound that can be perceived as greater detail or "speed". They also have the effect of making things sound closer because our perception of loudness is tied to higher order harmonics. The perception of louder also makes things sound closer...this impacts image 3d and soundstaging.

I don't care so much about the topic per se, only about 10% of the posts in this thread are strictly on topic. It does tie in; however, because how things are presented affects the perception of realism (point 1) and ties in to enjoyment as well (point 3). Whether or not it meets (2) depends on if you define accuracy from a strictly mechanical/measurement POV or if you take matching output with a psychoacoustic model as being a definition of accuracy. I take this as a definition of accuracy rather than a strict measurements basis. Measurements matter but not in the way they are generally being used. If a poorer measuring amp does things more correctly psychoacoustically than a better measuring one then that psychoacoustically correct amp will likely deliver better upon point 1 even though it is worse on point 2 and will likely deliver better on point 3 as well.

Spinning wheels here. I still haven't seen all these horrible harmonic amps.

And if you don't understand what 3rd harmonic or lower designing means, I don't think you really get the subject matter well. It means people purposely design for dominate harmonics to be 3rd or below. Not sure how else to phrase that. And do you distortion past 3rd, higher, sure, but at what dB? Voltage only divides so there's always limitations.

Why don't manufacturers advertise their harmonic distortion choices? Because it's about as interesting as a plumber designing a house's pipes to not leak.

I eagerly await your LTspice models along with topology that has perfect distortion patterns.

 

[Robh3606]

As I've mentioned, a controlled directivity speaker can sound odd after getting used to a direct radiator and vice versa... this is a really big deal and can confound these kind of tests if you're not careful. No mention of that so I disregard Harman's results as deeply flawed. IME 1st reflections are one of the largest acoustic issues with direct radiating speakers and what requires them to be placed in a carefully controlled acoustic environment for best results.

Hello Dave

JBL Synthesis and Revel are two examples of their designs directly driven by their research. I find it odd you don't realize the implications of your statement compared to what they market and manufacture.

Rob