Thoughts on the future of HiFi

[E-Stat]

That was a post and a half Maybe more!

He had lots of good points although you'll find a couple of overstated ones. Tube amps, especially pentode/tetrode designs, don't vary their bench tested flat frequency response with all speakers. Many Audio Research, Atma-Sphere, Joule Electra and VTL amps are quite neutral sounding with certain speakers. You must choose your speaker wisely, however, when using tubes or you will get the results he mentions. My amps, for example, sound decidedly colored driving the double Advents while they do not when driving the electrostats. It has to do with the impedance curve of the speaker. Wilson speakers, for example, work very well with tubes due to their flat profile.

I applaud Dr. Toole for his wanting to correlate measured response and listener preference. His innovative lab and its moving "speaker shuffler" is unique, but suited to monopole speakers only. Its indifferent consideration of the back wave is not suited for dipoles. Sean Olive acknowledged they do not believe in system optimization factors like choice of amplifier, cabling or speaker position - which works fine if you don't care about such things.

Shuffler video

rw

[poppachubby]

So while it might be feasible that certain, specific tube designs are closer to the mark than SS, clearly all (or even most) tube designs are not...

What do you mean by this? I am not understanding.

[E-Stat]

I use NOS dacs. What do you mean by this?

With the CD Redbook format and its limited bandwidth, one must choose between one of two evils: perfect phase response or perfect frequency response. If you choose perfect FR, it requires brickwall filtering which causes ringing. My GamuT CD-1 takes the other approach and trades some HF roll off for signal integrity. Consequently, it's response is 1 db down at 20 kHz. As to which choice is better, that depends upon preference. As for me, the last time I heard a pure 20 kHz tone was quite a few years ago.

Which relates to the Nyquist theorem debate as well. In the theoretical and perfect world, a 44.1 kHz sample rate is sufficient to provide response to half that frequency and achieve 20 kHz.. In the real world, however, such really isn't possible. While no one truly needs response out to 192 kHz, such a sample rate ensures enough room to filter appropriately without compromising the signal integrity in the time domain. A similar situation exists with the word size. 16 bits provides the incredible dynamic range - if and only if - the output level is high. In the real world of music, however, quiet passages cause quantization errors where far fewer bits are firing and digital goes deaf. Which is why you really need the equivalent of 24 bits if you wish to duplicate analog tape.

rw

[Sir Terrence the Terrible]

With the CD Redbook format and its limited bandwidth, one must choose between one of two evils: perfect phase response or perfect frequency response. If you choose perfect FR, it requires brickwall filtering which causes ringing. My GamuT CD-1 takes the other approach and trades some HF roll off for signal integrity. Consequently, it's response is 1 db down at 20 kHz. As to which choice is better, that depends upon preference. As for me, the last time I heard a pure 20 kHz tone was quite a few years ago.

Which relates to the Nyquist theorem debate as well. In the theoretical and perfect world, a 44.1 kHz sample rate is sufficient to provide response to half that frequency and achieve 20 kHz.. In the real world, however, such really isn't possible. While no one truly needs response out to 192 kHz, such a sample rate ensures enough room to filter appropriately without compromising the signal integrity in the time domain. A similar situation exists with the word size. 16 bits provides the incredible dynamic range - if and only if - the output level is high. In the real world of music, however, quiet passages cause quantization errors where far fewer bits are firing and digital goes deaf. Which is why you really need the equivalent of 24 bits if you wish to duplicate analog tape.

rw

Could not have said it better......thanks!

[E-Stat]

Could not have said it better......thanks!

We really do share some common ground.

rw

[Ajani]

What do you mean by this? I am not understanding.

The debate was based upon the assertion that all SS (or at least High neg feedback/class A/B/whatever SS) sounds similar enough as to be indistinguishable from each other in DBT, while tube amps vary greatly: some are ultra-sweet, some sound remarkably like SS, etc...

So using gun analogies with both SS and Tubes shooting at a target (reproducing the recorded signal exactly), I theorized that Tubes are clearly shooting wildly (since they vary so much) while SS is clearly aiming at the target (since they have such high level of consistency)...

Rich suggested that all SS might just be off the mark. So presumably all the shots are to some degree off target.

My response (which you queried) simply means that with tube shooting wildly, then it is only possible that some of them (if any) are closer to hitting the mark than SS....

Summary (without gun analogies):

Since Tubes can sound totally different from one another, then only some of them can actually be sounding like the original recording (the rest must clearly be altering the sound in some way).

[Feanor]

Sorry Bill, little to none. Nonsense? Not really. Extreme linearity especially with triode designs is the whole appeal of a class A tube amp. Any tube amp that a knowledgable enthusiast would consider, will most certainly have properly implemented NFB. I doubt if even the most golden eared audiophile could hear any distortion in the well made designs.

NFB in Class A isn't nonsense. Nonsense is thinking that they are all implemented properly or the same. This is what sets good apart from great.

"Linear" is not the same as "zero distortion", however.

[E-Stat]

Since Tubes can sound totally different from one another, then only some of them can actually be sounding like the original recording (the rest must clearly be altering the sound in some way).

As with many audio issues, it boils down to individual preferences. I have yet to hear a solid state amp that provided a "you-are-there" sort of see through to voice, woodwinds, strings, etc. in the midrange. They are measurably flat and work with a wider range of speakers. Harmonic integrity, however, is a different aspect of accuracy that is different from flat frequency response. Forget the static benchmarks, since they are next to useless in providing any sort of relevant correlation to what we hear.

I have the greatest respect for this designer who has striven to address the underlying differences in the amplification paradigm - voltage vs. current - in innovative ways. While Nelson Pass is better known for his high powered class A / AB designs, his First Watt "kitchen table" projects are quite innovative. The F3 uses power JFETs in a manner that more closely replicates the tube performance paradigm. With its 1 ohm output impedance, however, this sophisticated, yet simple little SS amp will behave more like a tube amp with some speakers. The important factor here is that he recognizes that simpler circuits usually provide more faithful reproduction of harmonic tones than complex designs having lots of stages and multiple levels of NFB, despite their "superior" metrics.

rw

[YBArcam]

With the CD Redbook format and its limited bandwidth, one must choose between one of two evils: perfect phase response or perfect frequency response. If you choose perfect FR, it requires brickwall filtering which causes ringing. My GamuT CD-1 takes the other approach and trades some HF roll off for signal integrity. Consequently, it's response is 1 db down at 20 kHz. As to which choice is better, that depends upon preference. As for me, the last time I heard a pure 20 kHz tone was quite a few years ago.

Mr. Terrible said that NOS DACs have to use brickwall filters and that they ring. You stated that the decision comes down to choosing perfect frequency response or perfect phase response, the former of which requires the use of brickwall filters, which causes ringing. The implication with both of these posts in mind is that NOS DACs use brickwall filters, which ring, and this achieves perfect freq. response. The alternative, implying this is the route chosen by OS DACs, is perfect phase response, and a rolling off of high frequencies.

Maybe I'm just misreading this, and of course I shouldn't take what two entirely different people say and frame it as if it's coming from one source. But it reads like NOS DACs must use brickwall filters - however, some clearly do not. I'm definitely interested in the MHDT Havana DAC, which from my readings does not oversample nor use brickwall filters. I don't know how it handles phase response (to be honest I'm not even sure what phase response is). The reviews of MHDT DACs however, of which there are many, are positively glowing. How bad can this ringing be?

[Ajani]

As with many audio issues, it boils down to individual preferences. I have yet to hear a solid state amp that provided a "you-are-there" sort of see through to voice, woodwinds, strings, etc. in the midrange. They are measurably flat and work with a wider range of speakers. Harmonic integrity, however, is a different aspect of accuracy that is different from flat frequency response.

So do all or even most of the tube amps you've encountered preserve harmonic integrity? If they do, then that would seem to imply that the differences in the sound of one tube design versus another is really attributable to differences in frequency response. If that's the case then I'd still expect all tube amps to sound similar just with notable differences in frequency response (like changing the settings on an equalizer)...

Forget the static benchmarks, since they are next to useless in providing any sort of relevant correlation to what we hear.

I actually believe that forgetting the benchmarks is the major problem with hifi. I think we should be encouraging designers to find ways to measure and understand harmonic integrity and ever other aspect of why we hear what we hear... This is science and not witchcraft, so even though we don't know how to measure everything yet, with enough dedication we will eventually be able to do so... But just ignoring the measurements seems a backwards step to me... I want to see more attempts to do like Harman and find the correlation between measurements and sound (I'd like to see it done by a major tube or planar manufacturer so we could get a different perspective on the situation)...

I have the greatest respect for this designer who has striven to address the underlying differences in the amplification paradigm - voltage vs. current - in innovative ways. While Nelson Pass is better known for his high powered class A / AB designs, his First Watt "kitchen table" projects are quite innovative. The F3 uses power JFETs in a manner that more closely replicates the tube performance paradigm. With its 1 ohm output impedance, however, this sophisticated, yet simple little SS amp will behave more like a tube amp with some speakers. The important factor here is that he recognizes that simpler circuits usually provide more faithful reproduction of harmonic tones than complex designs having lots of stages and multiple levels of NFB, despite their "superior" metrics.

rw

Nelson Pass has a legendary reputation: anyone with enough confidence to give away his old schematics to the DIY community must know what he is doing....

[poppachubby]

"Linear" is not the same as "zero distortion", however.

Of course not, but in the case of SET amps they are tied together and inter relate.

[poppachubby]

How bad can this ringing be?

That's what I am asking too. Thanks for the run down guys, but since I know there's no audible "ringing", does this mean the ringing is actually diminishing what I am able to hear?

BTW, my DAC in the H/T is a 4 x TDA 1543 using a DIR 9001. I just love it to bits and have done tons of comparisons with it. Gives the music such a natural sound, it's mind boggling.

[E-Stat]

But it reads like NOS DACs must use brickwall filters - however, some clearly do not.

New old stock or not - whatever relevance that has - one has to make a choice. There is an Ayre player that provides both profiles and you may switch between the two. You really must oversample in order to preserve proper response in the time domain. The Crystal 4390 in the GamuT provides 128x oversampling.

The reviews of MHDT DACs however, of which there are many, are positively glowing. How bad can this ringing be?

I cannot comment directly on that which I haven't heard, but ringing is not a good thing.

rw

[E-Stat]

So do all or even most of the tube amps you've encountered preserve harmonic integrity? If they do, then that would seem to imply that the differences in the sound of one tube design versus another is really attributable to differences in frequency response.

Overall, yes. The biggest challenge is not the FR of the amplifiers themselves. Because of their relatively high source impedance, it is their interaction with speakers having roller coaster impedance curves where the end product gets screwed up a couple of db in several octaves. The quality of the output caps and power supply stiffness also matter.

I think we should be encouraging designers to find ways to measure and understand harmonic integrity and ever other aspect of why we hear what we hear... This is science and not witchcraft, so even though we don't know how to measure everything yet, with enough dedication we will eventually be able to do so...

Spectral analysis helps, but is typically run with a single fundamental frequency like 50 hz.

(I'd like to see it done by a major tube or planar manufacturer so we could get a different perspective on the situation)...

Ralph Karsten of Atma-Sphere has a white paper on his website that addresses this question.

Nelson Pass has a legendary reputation: anyone with enough confidence to give away his old schematics to the DIY community must know what he is doing....

His products survive the test of time, too. I have a 29 year old Threshold Stasis amplifier which still offers exceptional performance (after having replaced the Mallory computer grade electrolytics in the power supply).

rw

[poppachubby]

New old stock or not - whatever relevance that has -

In this context, it's Non-Over-Sampling.

[E-Stat]

In this context, it's Non-Over-Sampling.

Ah. I confess that all of the best CDPs / DACs I've heard use oversampling in one flavor or another. The EMM Labs SACD players upsample Redbook to twice the DSD rate. My previous reference, the magnificently built and gorgeous sounding Burmester 969/970, also upsamples. It was those that I directly compared to the GamuT CD-1 before purchasing. Which uses an upsampling Crystal chip as does the Manley DAC I use in the vintage system.

rw

[Feanor]

Ah. I confess that all of the best CDPs / DACs I've heard use oversampling in one flavor or another. The EMM Labs SACD players upsample Redbook to twice the DSD rate. My previous reference, the magnificently built and gorgeous sounding Burmester 969/970, also upsamples. It was those that I directly compared to the GamuT CD-1 before purchasing. Which uses an upsampling Crystal chip as does the Manley DAC I use in the vintage system.

rw

The EMM, Burmester, and GamuT are all many times to price of a basic NOS of the type Poppachubby is referring to. So many be it's expensive to implement upsampling really well, (though there a plenty of quite good upsamplers around, e.g. Cambridge, Musical Fidelity, and since a price reduction, the PS Audio).

It's perhaps harder to justify the Audio Note NOS models that range from quite expensive to very expensive. The least expensive AN, in kit form, runs $1500; you can buy DAC based on the same digital chips for about 1/10th that price on eBay. Of course the eBay examples lack the pricey analog components, capacitors, etc., found in the AN.

[pixelthis]

Its kind of ironic, a thread on the future of "HI-FI", and a lenghty discussion that
is a case in point as to why its dying.
When someone in the know asks me what is so special about a SET, a design that
was basically obsolete by 1950, I don't know what to tell them.
When the ask about the gobbledegook concerning brick wall filters and various
DAC'S, NONE OF WHICH MATTER, because most GERMAN SHEPARDS
couldn't tell the difference, I tell them its philosopical, like how many angels can dance on the head of a pin.
USED TO BE you could compare turntables, various combinations of equipment, etc,
but today it doesnt matter.
The only thing that will make a real difference is type of speakers, and decent power output(which a tube set seldom has, btw).
The perfection of digital media has left HI-FI types to argue about the most extraneous
of things.
Is there a difference between tube and solid state?
Yes, tubes are obsolete, solid state is not.
Why is it that the more you pay for "high end" gear the more obsolete its bound to be?
You tell me.
Which answers the question that is the subject of this thread,
the "future" of HI-FI.
irrelevance, in one word.
EXPLAINING why music in its purest form in a home listening enviroment is such a joy
to the great unwashed is hard enough, without explaining why gear that was obsolete
in 1934 is the only way to do it.

[audio amateur]

I have the greatest respect for this designer

He definitely needs to get out of his cave me thinks

[E-Stat]

... because most GERMAN SHEPARDS couldn't tell the difference...

Why would you consult a German Shepherd on audio matters? Somehow, I don't think they appreciate music very much.

rw

[E-Stat]

He definitely needs to get out of his cave me thinks

He's never seemed to find much use for scissors whether its the 70s or today, but he does have a nice cave. Follow the article through for more pics. I especially like the TAS poster of Leopold Stokowski hanging on the wall (you'll find a pic of the same one hanging in my listening room)

rw

[YBArcam]

Why would you consult a German Shepherd on audio matters? Somehow, I don't think they appreciate music very much.

rw

My old German Shepherd used to howl at opera music. Even that part before the guitar solo on Bohemian Rhapsody! He was clearly into it.

[poppachubby]

When someone in the know asks me what is so special about a SET, a design that
was basically obsolete by 1950, I don't know what to tell them.

Hey I have an idea, why don't you tell them you don't know what you are talking about?!? Try a history book, oh right...you've opted to write your own.

[Ajani]

Why would you consult a German Shepherd on audio matters? Somehow, I don't think they appreciate music very much.

rw

It makes perfect sense to me...

Every time I turn on my SET/Horn system, my dog howls as if I'm trying to neuter him with a rusty spoon and no anesthetic... But whenever I put on my megawatt SS/Cone speakers, he dances around like snoopy (and occasionally humps the coffee table)... So this is clear evidence that SS is superior to tubes...

[E-Stat]

... he dances around like snoopy (and occasionally humps the coffee table)... So this is clear evidence that SS is superior to tubes.

LOL! I got a little bit of Diet Coke up my nose with that one.

rw