Natural Sound

Concerning SET amps, it is interesting to note that Arthur Salvatore, a big fan of (some) SET amps, has recently changed his speakers to Sadurni Acoustics Staccato horn speakers. A few things to note:
- those speakers have active subwoofers
- he's been fairly critical of many horn speakers, some of which are mentioned in this thread

This is what he has to say about SET amps:

"SET amplifiers have important and fundamental sonic advantages in the midrange and highs over any other amplifier design in my experience, especially with acoustical music. They have the lowest sound-floor and also are the best "organized" (and music is simply "organized sound")."

I find that to be an interesting way to present things.

His point of view is not based on "theory" and technological supremacy, it just reflects his experience.

I quoted his ideas about sound reproduction (what he calls his "audio philosophy") at length in one of my modest reviews and he wrote back:

"It now appears that the interest in the ideas, concepts and theories I discuss in my reviews will outlast the performance details of the actual components. This is not an accident. I’ve always believed the components are reference points only, which are still required to provide a common perspective, but not as important, in the long run, as the ideas."

That's a positive attitude!

hopkins said:

Concerning SET amps, it is interesting to note that Arthur Salvatore, a big fan of (some) SET amps, has recently changed his speakers to Sadurni Acoustics Staccato horn speakers. A few things to note:
- those speakers have active subwoofers
- he's been fairly critical of many horn speakers, some of which are mentioned in this thread

This is what he has to say about SET amps:

"SET amplifiers have important and fundamental sonic advantages in the midrange and highs over any other amplifier design in my experience, especially with acoustical music. They have the lowest sound-floor and also are the best "organized" (and music is simply "organized sound")."

I find that to be an interesting way to present things.

His point of view is not based on "theory" and technological supremacy, it just reflects his experience.

I quoted his ideas about sound reproduction (what he calls his "audio philosophy") at length in one of my modest reviews and he wrote back:

"It now appears that the interest in the ideas, concepts and theories I discuss in my reviews will outlast the performance details of the actual components. This is not an accident. I’ve always believed the components are reference points only, which are still required to provide a common perspective, but not as important, in the long run, as the ideas."

That's a positive attitude!

Click to expand...

Largely agree with Arthur but I like to delve deeper into the why SET has the advantages Arthur (and myself) observes.

Atmasphere said:

'Natural' and "Accurate' are exactly the same thing as I described earlier. Synonymous in every way. If its not accurate its not natural and if not natural certainly not accurate. What we cannot do is be truthful to the musical event. That is because whatever it was, the recording is all we have. But we can be truthful to that, so however the recording is, the reproduction will be as true to the recording as possible if its 'accurate' or 'natural'.

Click to expand...

I think we are talking past each other, perhaps because you are not recognizing that my objective of high-end audio is different than your objective of high-end audio. I believe your objective is: Objective 2) reproduce exactly what is on the tape, vinyl or digital source being played.

My objective, for the purpose of this discussion, is: Objective 4) create a sound that seems live.

I cannot put a vinyl LP up to my ear and hear music. I take the recording as a starting point. I am not trying to maximize the fidelity of what comes out of my stereo to what is encoded on the LP. Is this what you mean by accuracy? That is your objective, correct?

I am taking that LP and doing whatever I have to do to extract from it the particular sonic cues which subjectively for me cause my stereo to output a "sound that seems live" to me. Whatever I do with the signal extracted from the grooves and amplified by my system to achieve a sound that is most natural to me -- a sound that seems live to me -- is what I want my stereo to do. If I achieve for my ears a sound that seems the most live -- a sound that seems the most natural to me -- by running the signal through a distortion generator and by having an oompa-loompa whistle at the same time through a piece of woo-woo wood then I am happy. To me this has little to do with a notion of "accuracy" from the medium.

morricab said:

Largely agree with Arthur but I like to delve deeper into the why SET has the advantages Arthur (and myself) observes.

Click to expand...

He's got more to say about the topic in his section about dynamics, which he sees as the main limitation of audio systems ("their inability to produce lifelike dynamics"):

"Most audio systems are simply pathetic when it comes to imitating what occurs in real life. Any live concert, even for a solo flute, or just a passing high school band, makes this unavoidably clear."

He identifies the following limitations:

"The causes are many and various, but the two main culprits are amplifiers and speakers. For an amplifier to reproduce IDR, it must have high voltage and power swing, and this gives the advantage to tube designs, which should be no surprise to experienced audiophiles. However, amplifiers with truly high voltage swings are very rare because of their extra difficulty and expense to build."

Some other factors may be in favor of tube amps, such as the absence of "thermal memory distortion" or the speaker steering mechanism (from what I understand).

Rensselaer said:

No, I understand that bit. I think what you are saying is that the system doesn’t make music sound natural, but that a good recording will sound most natural from a system that is most accurate (least distortion), right?

But, if I get that part correctly, having an accurate system will not make a crap recording sound natural (as in the sound of real acoustic instruments in situ), but will accurately play what was recorded. That is where I have trouble reconciling your comment that natural and accurate are the same thing.

I accept that a natural sounding system must necessarily be an accurate system if the naturalness is in the recording. That is if it is only heard when accurately pulled from a stellar recording and not something in the combination of analogue front end, DH SET’s and certain horn speakers putting out some sort of auditory clues similar to live recordings so that even lesser quality recordings somehow sound more real, some psychoacoustic effect of the system on the listener.

Click to expand...

Yes natural and accurate are not always the same. Do SET/Horns make unnatural recordings sound more natural, probably.

tima said:

What an entre to Beethoven's 6th Symphony, "The Pastorale". I can't resist - it's one of my very favorites. Here is true program music, illustrative music. B wrote a title at the beginning of each movement.

Mvmnt 1: Awakening of cheerful feelings on arrival in the countryside
Mvmnt 2: Scene by the brook
Mvmnt 3: Merry gathering of country folk
Mvmnt 4: Thuder, storm
Mvmnt 5: Shepherd's song. Cheerful and thankful feelings after the storm

Beethoven writes: "One leaves it to the listeners to discover the situations."

The 2nd's 'Scene by the brook' and the 4th's 'Thunder, Storm' are the most memetic of the piece.

You can hear the murmur of the brook in the repeated triplets of strings. Trills and melodic figures in higher registers speak of birds flying across the scene. B even names the birds -- solo flute is the nightengale, the oboe is the quail, and two clarinets play the cuckoo.

The storm that chases the merry gatherers to shelter develops realistically. Tremolo from lower register strings sound distant thunder. The music builds as the storm comes closer, you can almost feel the rain drops pelting you. A timpani roll and raspy trombones signal the thunder and heavy rain which passes. Sunshine breaks out and the peaceful pastoral feeling is restored.

Böhm on DG is excellent. DG 2530 142

My fav is Cluytens. HMV ASD 433

View attachment 133786

Click to expand...

Fantastic stuff thanks Tim… and yes it was extra kind of LVB to provide us with his Natural Sound Symphony. Love it when we get the chance to talk about the music, much appreciated… Cluytens Berlin is a benchmark in the 6 for sure and so for me is Klemperer and the Philharmonia. For those looking for a strangely still natural but contemporary LVB 6 Fischer and the Budapest do a fabulous take…

Ivan Fischer in an interview in Gramophone added his thoughts on the LVB 6 and the freedom of the natural…

Most of Beethoven’s works have the dual theme of tragedy and jubilation; and he was, of course, very preoccupied with ideas of freedom and liberation. The end of the Fifth Symphony is akin to the end of Fidelio, but this is a visionary, green symphony. It represents a different type of liberation, from beginning to end: a liberation by and through nature. The final happiness is a bit pantheistic, influenced perhaps by the philosophy of Spinoza.

Beethoven stepped out of the Classical tradition between his Second and Third Symphonies, but there is a relationship here with the ‘pastoral’ musical forms of the 18th century, a century that was just coming to its close. The Ninth Symphony is about the continuation of the French Revolution; and Beethoven found a way of putting into music what the crowd felt, the sense of freedom that came with storming the Bastille and throwing away the aristocracy and the feudal system. The aspiration of being freed involved overturning the social order. That was freedom in the city; this symphony is about the very different freedom we can experience by leaving the city altogether. The Ninth Symphony literally did change the world by being so grand, so jubilant: it steps out of the boundaries of music. In the Sixth Symphony Beethoven explores a simpler kind of freedom, one involving total harmonic happiness.

The Pastoral Symphony happens inside us. There is an internal, mental exercise at work, as the simplicity of nature creates strong feelings – and Beethoven was interested in those feelings that nature awakens in us.

For those looking more for whole cycles of the Beethoven Symphonies (which is an extraordinary way to view the lifetime growth and interrelationships within the composers output within a discipline) I’d suggest there are some marvellous 6’s below in these Beethoven symphony cycles as well… for analogue and digital options on highly recommendable LVB cycles … amongst others…

Gunter Wand NDR

Szell Cleveland

Eugen Jochum Concertgebouw/BPO/Rundfunks

Barenboim Staatskapelle

Kletzki and the Czech Phil

Freedom from reason… the half human half beast gods, the satyrs like Pan (as a god of music) are representations that we are more than just obvious Apollonian reason (Apollo as another god of music) and in the importance of letting go to explore the experiential Dyonesian and instinctual natures within the simpler natural landscape of music.

In Peter’s world of experience of his sound and within his own system he is the only truly real and valid expert. This isn’t ultimately about engineering… it is about experientiality… Peter’s experience of his own system. Peter’s not lying to us or making this up… this is his experience of his system and that’s the only validity that there is ultimately in this. As much as some have spent three years trying to pound Peter into letting go of his sound world view within his own system thread to accept and be bound to their particular experiences of what sound can be like. We do seem to share huge common grounds in perceptions but these perceptual abstracts are very fine distinctions within the rough and ready reason and within the greater schema of things. Only Peter can say what those fine distinctions are ultimately for him.

If you look at preamplifiers, dacs, phonostages you will see the output stage is almost class A.
The most linear topology is Class A.
Low/zero feedback SET tube amplifiers are Class A (very simple and linear).
Compression drivers for horns (like vitavox s2) are fast.
Horns like impedance matching devices need less power than direct drive speakers.

When you use SET with Horns you will have maximum speed plus maximum linearity.
I think good horns have wider dynamic range and are also fast (live sound).

hopkins said:

He's got more to say about the topic in his section about dynamics, which he sees as the main limitation of audio systems ("their inability to produce lifelike dynamics"):

"Most audio systems are simply pathetic when it comes to imitating what occurs in real life. Any live concert, even for a solo flute, or just a passing high school band, makes this unavoidably clear."

He identifies the following limitations:

"The causes are many and various, but the two main culprits are amplifiers and speakers. For an amplifier to reproduce IDR, it must have high voltage and power swing, and this gives the advantage to tube designs, which should be no surprise to experienced audiophiles. However, amplifiers with truly high voltage swings are very rare because of their extra difficulty and expense to build."

Some other factors may be in favor of tube amps, such as the absence of "thermal memory distortion" or the speaker steering mechanism (from what I understand).

Click to expand...

Again, don't disagree with him...I think he is speaking from a lot of listening experience AND a critical mindset...this second part is what is mostly missing from the vast majority of audiophiles (and humanity in general). It is not easy to be critical...although to some it comes more naturally than others. It is not always a good trait to have in "polite" company.

Just this past weekend I was inundated with non stop live band music. My daughter plays trumpet and is in a horn orchestra. Her orchestra was playing in the Welt Jungendmusik Festival band competition in Zurich. From Friday through Sunday I was listening to several hours a day of live, brass, woodwind and percussion based orchestras. Both in concert halls and outdoor...all without amplification (none needed with such loud instruments).

Arthur is absolutely right about the dynamics being the weakest limitation. Even a solo violin is too much for most systems to do realistically. I know this as well from when I was recording my ex as she practiced for her Paganini Concerts (she performed the 24 Caprices in 4 concerts around Switzerland many years ago). I was recording to R2R tape and it was very difficult in my smallish space to keep the dynamics within the limitations of the tape...for ONE instrument! The way the Strad she was playing loaded the room literally made my ears pulsate.

This is the main reason I migrated from large planar speakers (I have owned Apogees, Acoustats, big Infinities, DiY hybrids etc.) to high sensitivity speakers (all my systems are >97dB sensitive.). This allows me to use what I consider to be superior amplification types that also have relatively unrestricted dynamics (within their power limits of course). Do I get lifelike dynamics? Maybe close for solo or small ensembles...no for larger works...however, nearly all large orchestral recordings are dynamically limited due to the limits of the recording media to capture the full dynamics of such large ensembles. The recording itself is a big bottleneck and I don't believe anyone who claims otherwise. I think you can capture realistic dynamics on something like a small Jazz ensemble or string quartet but that is about it.

Amir said:

If you look at preamplifiers, dacs, phonostages you will see the output stage is almost class A.
The most linear topology is Class A.
Low/zero feedback SET tube amplifiers are Class A (very simple and linear).
Compression drivers for horns (like vitavox s2) are fast.
Horns like impedance matching devices need less power than direct drive speakers.

When you use SET with Horns you will have maximum speed plus maximum linearity.
I think good horns have wider dynamic range and are also fast (live sound).

Click to expand...

Exactly and after hearing a full weekend of horn orchestras it only reinforces that dynamics are the final frontier of audio reproduction (and recording!). A lot of systems get the soundstage, the tone etc. correct...nearly all fail miserably on dynamics...even with a great recording.

The less a driver actually has to move the lower it's distortion. The best drivers move hardly at all but the distance they do move is under extreme acceleration from a huge magnet field strength and low moving mass.

PeterA said:

We could not know what instruments captured on the recording sound like unless we had the reference of the actual sound of those real instruments heard live in a space. Without that reference, how could we possibly know if the recording is being presented naturally, or realistically, or convincingly? Even Ralph when making a recording is referring to the actual sound of the instruments that he is recording to judge if the recording is any good or not.

Click to expand...

I agree. An acoustical event can be recorded in a myriad of ways, some better than others. But once that is done, the recording is all we have. By then, it is up to our systems to render, and in those terms, accurate is natural, IMO. From what I can hear from the very limited videos/recordings, your system sounds very accurate/natural within its passband. The level of energy, transient capacity and tonal correctness are apparent, and they are reference level as far as I can see. You can refer to what I thought was interesting in Munich this year here, and get a feeling as to where my tastes drift. If it didn't mention it, it is because I didn't think it accurate, hence not natural, and there where lots of systems there with the fame of being extremely accurate.

The only incongruence I see is, because recordings are all we have for most music, that we feel that matching the rendering of our system to what we hear live is a guarantee of success, for all recordings. As they are all potentially different, and the event is lost to time, this is problematic. It maybe be the source of the proposed decoupling between accurate and natural, and the more I think about the more it looks like the typical circle of confusion paradox with a nicer jacket. I'm ok with that, I know what I value in a reproduction chain. Keeping the equivalence accurate=natural and a wide curated list of reference recordings helps cut though the issues, the audio lingo, the unrealistic standards, and provides me with what I want.

Atmasphere said:

IME most solid state amps ever made, while lower distortion, are not accurate because the distortion they make is easily heard as harshness and brightness, especially when you turn it up. Its a coloration plain and simple. It does not matter that the distortion is very low on paper because the ear doesn't care about that. Its tuned to be very sensitive to higher ordered harmonics- its more sensitive to them than almost anything else! So when this quality of the ear is ignored by designers and spec sheets, we get sound that isn't natural and certainly not accurate.

Click to expand...

PeterA said:

The reference for natural sound is one’s memory of a live acoustic music event.

The reference for accurate sound is a recording of a music event or assembly.

Click to expand...

Al M. said:

My point was that per "Audio's Circle of Confusion" it is objectively unknowable what is on the recording (including actual tonal balance; look at the in-room frequency graph for studio monitors used for monitoring recordings in the link above), certainly for those not present at the recording process, and potentially beyond.

Therefore, if what exactly is on the recording is unknowable, then what is objectively accurate reproduction of the recording is unknowable as well. If things sound subjectively Accurate to you that is a different thing. I strive for that kind of "accuracy" too, but that's not really Accurate.

Click to expand...

I confess to some confusion about this discussion of "natural" and "accurate" when it leads to a conclusion that seems to say 'by definition natural and accurate as adjectives attached to the word 'sound'' characterize the same thing .

Accurate sound, Natural Sound and natural sound. I used the capitalization I did to help me clarify my thoughts.

A. I understand 'natural sound' as the sound of live acoustic music.

B. I understand Natural Sound as: a) the proper name of this thread and b) as an audio phrase for the experience of reproduced music as evocative of or reminiscent of natural sound. It is an audio phrase about reproduction (though perhaps not an audiophile phrase.)

C. I take 'accurate' to describe identity in measurement of audio system input and output. (When the in-out measurement is not accurate I understand that as 'distortion'.) I'm not sure what 'accurate sound' means unless it is defined as what is heard in reproduction when there is no distortion, ie. the input and output measure the same.

C cont. I think (but am not sure) that Ralph adds the condition "and when the reproducing equipment is designed in conjunction with the rules of human hearing." I don't how those rules are codified or described in measurement. If the test of conformity to the rules of human hearing is Natural Sound then okay, but it seems the argument (natural=accurate) becomes circular or tautological.

Toole aside, I agree with Al M. that we cannot know by listening what is on the recording. The recording is the ding a sich -- the content only known through the sensory apparatus, the experience of the source player/reader. Maybe if we measure the original master tape or original source and compare those measurements with measurements of the record (or whatever) that is used for reproduction we might say the measurements are the same and thus the record is an accurate capture of the original source. But that is not listening.

RCanelas said:

I don't see any incongruence if one admits that the recording is all we have, and the only actual reference. Under those terms, natural and accurate are the same thing, no space for debate, it is a direct derivation of the logical proposition.

Click to expand...

Will you please share your syllogism?

PeterA said:

We could not know what instruments captured on the recording sound like unless we had the reference of the actual sound of those real instruments heard live in a space. Without that reference, how could we possibly know if the recording is being presented naturally, or realistically, or convincingly? Even Ralph when making a recording is referring to the actual sound of the instruments that he is recording to judge if the recording is any good or not.

Click to expand...

Let me give you a detailed example, with reference to @morricab 's reference to the violin Paganini Caprices, which are challenging dynamically.

One of the better violinist's resident in London is Alina Ibragimova. Her mother teaches at the Menuhin School and she played at his funeral when a student. My wife calls her a force of nature. She is now world renowned, first really hit the public's awareness when she performed Bach's Partitas and Sonatas at the Proms. That's two long concerts from memory solo in front on 9,000 people and live broadcast around the world. They are online. One is here:


The earliest performances I recall attending were of the Beethoven sonatas. They were recorded live at the Wigmore Hall and are available online:

Around that time I heard her play Mozart's 4th violin concerto under Haitink with the LSO at the Barbican. This is a large hall with a rather dry sound.

I also heard her play all the Mozart sonatas. She performed them with Cedric Tiberghein at the Wigmore Hall, then drove up the M4 and recorded them at Wyastone Hall, a superb hall but much larger (owned by the Nimbus Foundation, used by Chandos and many other labels).


More recently she recorded the Paganini, in July 2020 during extreme lockdown, at the Henry Wood Hall in London, for many years a favourite venue for violin and piano records by many great artists. It has a wonderful acoustic. It is not a performance venue, sadly.


I've heard her at other venues, quite a lot in 2020 and 2021 at the Fidelio Cafe, to an audience of about 40. Here she is warming up with Samson Tsoy. This is an up-close-and-personal venue, with a lot of passing traffic.


Samson arranges a festival that used to be at this venue, The Ragged School. It does not look promising, the sound is hard, but it's exhilarating.


I've also heard Alina play here, at Peckham Levels (yes, it's a car park). This is Johan Dalene, another very fine violinist. On this occasion wind was an issue and his score ended up in my wife's lap, but he continued from memory (it was Arvo Part, so not much to remember).


All these performances were in a "live space". Most were recorded and some can be listened to online. But which one is the reference for Alina? The Wigmore Hall has a distinctive warmth (I'm going there tonight). Henry Wood Hall could be considered a recording reference for solo violin, but give me The Ragged School any day.

Talk of "reference" sound quality seems as elusive as the pot of gold at the end of the rainbow. I can listen to Alina on YouTube at the Proms and I know it's her. When she recorded the Paganini Caprices there was no one else in the room (only an engineer and producer in separate boxes, and they were never in the same room as it was severe lockdown), so only she will ever know what it really sounded like.

So I just try and enjoy my hifi sound reproduction without these nebulous intellectual exercises.

RCanelas said:

You can refer to what I thought was interesting in Munich this year here, and get a feeling as to where my tastes drift.

Click to expand...

Interesting system described here, for those curious about it:

PeterA said:

We could not know what instruments captured on the recording sound like unless we had the reference of the actual sound of those real instruments heard live in a space. Without that reference, how could we possibly know if the recording is being presented *naturally*, or realistically, or convincingly? Even Ralph when making a recording is referring to the actual sound of the instruments that he is recording to judge if the recording is any good or not.

Click to expand...

Now … Were you to simply leave out the adjective that I have highlighted ! then I for my part would agree entirely with the remaining consideration.
Your Lamm amplification , or any other for that matter , is not a device that has the functionality to receive , amplify and forward on , a recorded of a complex audio signal *naturally* … However the same Lamm amplification does have the functionality designed into its construction to receive , amplify and forward on the same complex audio signal in a *realistic* or *convincing* way based on how little damage the circuit does to the integrity of said audio signal , which can of course be verified by measuring the level of integrity.

morricab said:

The less a driver actually has to move the lower it's distortion. The best drivers move hardly at all but the distance they do move is under extreme acceleration from a huge magnet field strength and low moving mass.

Click to expand...

Sensitivity is important but the amp/speaker combination seems to play an important role as well, as one would expect it should. Here's an interesting comment from my amp manufacturer (as posted on their website) about some of the issues at play:

"As the voice coil resonates within a strong magnetic field it will generate an AC voltage similar to a dynamo or generator. Most conventional amplifiers get disrupted (feedback loop out of control) by this unwanted AC voltage injection into the amplifier output stage. One side effect is excessive cone movement during bass reproduction. Needless to say that this also creates distortion across the complete audio spectrum, the louder the speaker plays, the more AC it generates the more the sound colours or distorts."

I can't have a critical opinion about this, not having the technical background, but my ears would tend to agree.

From what I understand there are a variety of amplifiers that address this - see the Thrax link I mentioned above - and some tube amps as well. More technically savvy forum members could offer explanations.

Argonaut said:

Now … Lamm amplification does have the functionality designed into its construction to receive , amplify and forward on the same complex audio signal in a *realistic* or *convincing* way “based on how little damage the circuit does to the integrity of said audio signal , which can of course be verified by measuring the level of integrity”.

Click to expand...

This states (as a matter of fact) that Lamm designs are based upon how (siq: accurate) it’s playback is compared to the input (said audio signal), verified by measurements (not listening). Would you please site your reference from Lamm that states this (otherwise it looks like something you made up)?

hopkins said:

Sensitivity is important but the amp/speaker combination seems to play an important role as well, as one would expect it should. Here's an interesting comment from my amp manufacturer (as posted on their website) about some of the issues at play:

"As the voice coil resonates within a strong magnetic field it will generate an AC voltage similar to a dynamo or generator. Most conventional amplifiers get disrupted (feedback loop out of control) by this unwanted AC voltage injection into the amplifier output stage. One side effect is excessive cone movement during bass reproduction. Needless to say that this also creates distortion across the complete audio spectrum, the louder the speaker plays, the more AC it generates the more the sound colours or distorts."

I can't have a critical opinion about this, not having the technical background, but my ears would tend to agree.

From what I understand there are a variety of amplifiers that address this - see the Thrax link I mentioned above - and some tube amps as well. More technically savvy forum members could offer explanations.

Click to expand...

Most SETs have no feedback loop to allow back EMF from the speaker to reenter the amp and get reamplified. This type of distortion was first discussed by Matti Otala in the early 80s. A SET is insensitive to back EMF if it doesn’t have a feedback loop.