[bhobba]

I have heard a digital amplifier with virtually identical high frequency resolution and air compared to an analogue amplifier with a full power bandwidth of 250kHz.

I have been suspicious for some time the supposed good sonics of digital amps may be due to residual unfiltered HF - which fits in with Hugh's suggestions quite well.  I am not sure however that like the type of distortion produced by valves it is enjoyable to listen for long periods.

Thanks
Bill

[ginger]

Well - I haven't had time to sit a digest all of the above BUT always happy to throw some cats amongst the pidgeons.

1/ Harmonic distortion - The largest source of harmonic distortion in your sound system is your speakers - always. If we want to limit our discussions to just amplifiers then I'm in a reasonable position to make some SUBJECTIVE assessments.
My least favourite amplifier I've ever owned was a top of the line ROTEL at 210W per channel and 0.001% ditortion figures - it was jusy plain cold, sterile and boring.
One of my favourite amplifiers is a ECL86 (6GW8) Ultralinear Push pull with some trick shunt feedback and a whole 6dB of global feedback, developing a massive 8 Watts per channel.
I like my VERY expensive 845 Single Ended Triode Amp (20W per channel) so long as I don't turn it up to much, passed a certain volume level it turns into a 2nd harmonic distortion genertaor with bloated (artificial) bass.
The amp I'm working on now is a parallel push pull triode mode EL34, no global feedback. It is starting to sound quite stunning although all of the theory says it should'nt. It has a damping factor of about 1.3 at 1kHz BUT the Output Impedance changes with frequency and just happens to track the speaker Impedance variation with frequency quite well so the DF is 1.3 across 20Hz to 100kHz. To indulge in the latest technobable that means I'm not using Voltage Drive of the Speakers, nor am I using Current drive, what I have is just about pure POWER drive.

2/ Frequency response - 20Hz to 20kHz right? WRONG I think. That will certainly reproduce the frequency information BUT it won't cut it for the spatial information in a stereo source. If you have -3dB points at 20Hz and 20kHz that means that you have 45 degrees phase shift at those frequencies. For best stereo IMAGE its my view you need less tghan 5 degrees phase shift in the audio band and that means a full decade frequency response extension in each direction, that is, 2Hz to 200kHz. Of course if you are using a CD as your source (digital sampling frequency is 44kHz) you are wasting your time. Nyquist tells you that at a sampling rate of 44kHz you can't reproduce any more than half that i.e. 22kHz - Thats why folk are looking a 96kHz SACD etc. I recently modified my phono preamps RIAA response. In the standard RIAA the high frequencies are rolled off at 6dB/octave for ever. This is to oppose the boost which was applied to the cutting head amplifier. Obviously they did not boost it forever. Adding a 3.18us time constant (50kHz) high pass to limit that roll off made an incredible difference detail and imaging (BTW this can be done to your GK-1 Phono stage by the addition of 1 resistor).

3/ Detail or "Downward Dynamic Range" to use the modern BUZZ Words - this is simply the old Signal to Noise Ratio. Anyone who has done a Nivarna Plus upgrade to their AKSA will uderstand what this is about. Its not so much that we added detail, the detail was always there, what we did was remove a veil of noise to expose that detail. For Power Amps and line level preamps there is nothing to say that tubes are inherently better than transistors or vice versa. Tubes however have a certain irreducable noise floor and are NOT suitable for Moving Coil Phono Preamps. Much loss of detail occurs due to residual noise on power supply rails and the rails varying with the audio signal as the current drawn varies. This is why some audio nutters now suggest the use of a "Super Regulator". This is jus one scheme (there are other ways to do the same thing) to keep rail noise to absolute minimum and not let it vary with the audio signal. Have a look at the second half of the output tube in your GK-1 to see another way of doing this.

4/ Clipping and all that, Transistor Amps hard Clip and Tube Amps Soft Limit RIGHT, Well that depends entirely upon the circuit. Put 20dB of global feedback around a tube amp and it will clip as hard as ant transistor amp. The harder the clip the most higher order harmonics are produced. When ever the transfer function (Vout vs Vin) is not a straight line we call it non-linear and it will produce harmonic distortion. The more abrupt the curve away from a straight line the higher the order of harmonics to be produced. When you have a straigh line up to clipping point and then it goes flat you are at the worst case condition.
The above presupposes symmetrical clipping. Any guitar amp designer can tell you that if you want gobs of 2nd harmonic distortion just clip one side. Clip it hard and you'll get some 3rd harmonic (which gives the sound some "edge") and then severely limit the frequency response so that the more objectionable higher order harmonics don't get through the amp. This is NOT an acceptable method for HiFi however.

What do I use for a Reference Amp while designing, building, modifying all these tube amps - why my AKSA55N+ of-course.

Well lunch time is over and I've painted enough targets on the front of my shirt for one day - I best get back to the design of the digital imaging system for the survey aircraft.

Cheers,
Ginger

[gonefishin]

Ginger...I'm curious.  What speakers do you use as your reference?


  thanks,

  dan

[ginger]

VAF Research DCX - the old ones not the new model.
96dB/W/m , nominally 6 Ohms.
Cheers,
Ginger

[_scotty_]

jules, the term "lapse" was first used by Hugh, tongue firmly in cheek,
in a reply to me in his second post on page three of this thread.
As far as emotion and electronics is concerned I don't expect the electronics
to be entirely responsible for my emotional involvement in the music.
I have been moved to tears by songs played on my car radio, an admittedly
flawed  playback system. I think emotion is a variable that the listener must be mostly responsible for. What expectations you bring to the listening experience are at least as important as the choice of music you listen to and the equipement it is played back through. I don't expect electronics to transport me emotionally. I think that is best left to the artist I am listening to.

bhobba, as happens I listen to a DIY Digital amp everyday frequently off and on for more than 8hrs at a time and no digital nasties are present.
It should be noted here that my speakers are an 8ohm resistive load above 3.5kHz and are an ideal termination for the filter network used to remove the carrier frequency the amp employs. I also probably use less
than 3watts on a continous basis which minimises the amplifiers tendency to ring on transients.  
 
ginger, you might try extending that phono response out to 100kHz.
This would be 5 times 20kHz and would reduce the inband phaseshift down in the midrange to almost nothing. If you are using a moving coil cartridge only the phase shift due to the inductance of the coils would be left.   As far frequency response in power amps is concerned I would agree with in regards to SS analogue amplfiers, when digital amplifiers are considered frequency response to 200kHz does not seem to be necessary
to have excellent imaging.  More than simple frequency extension parameters  seem to apply to this phenomenon.  Very low propagation delay in the feedback network may also be important for good imaging
to be realized.
Scotty

[ginger]

Scotty,
Just for info - I ditched my Moving Coil Phono and have gone back to Moving Magnet for exactly the reason you highlighted. Inductance in the coils is VERY difficult to manage - although I did see one bog simple, extremely clever idea just recently - the fellow used a second identical moving coil cartridge wired in the feedback loop around an op amp. Gloriously simple and gloriously elegant. What his source cart. introduced his feedback cart. took out again. Expensive if using one of those $500 cats thou'.
Cheers,
Ginger

[Geoff-AU]

First if recording cuts off higher harmonics then the obvious answer IMHO is to use higher resolution formats.

I thought the problem was to do with distance from the mic and being off-axis from the mic, not the actual recording format itself.

Audiophiles who believe even power cords can make a difference to sound

They're full of shit, no 2 ways about it.  How can 2 metres of magic cable make a difference when there are hundreds of kilometres of old wiring back to the power station?

I have to totally agree with this one although there's also the possibilty that musos are either a. too poor to own hi fi gear or b. consider making music to be a few steps above listening to it.

Ah yes that was the other point.  I agree with all answers to that question, really

[Geoff-AU]

For best stereo IMAGE its my view you need less tghan 5 degrees phase shift in the audio band and that means a full decade frequency response extension in each direction, that is, 2Hz to 200kHz

I don't think phase matters so much for bass, as our ability to perceive the direction of bass is not so good.  But certainly for the top end of the frequency scale I can see where you're coming from!

[jules]

gotcha Scotty and sorry I missed the back-reference  

If there's one more confession here of being brought to tears by a car radio we might all have to throw out the hi fi and get out on the road.

Can I ask a question here .... Right back near the start of this thread Hugh made the point that some very high quality german mikes use pencil tubes as part of their workings. Does this mean that even at the start of the recording process certain harmonics have already been lost and are being re-introduced in a generic form to create the desired timbre? One way or another if tubes are inserted this early in the chain it does have significant ramifications for understanding the rest of the process.

jules

[Rhythm Willie]

Assuming that one would want to introduce harmonics via valves or digital methods somewhere in the reproduction chain due to shortfalls in recorded material through to source, I would have thought the logical place to insert a change would be in the pre-amp stage - not the power amp. Proviso is that your good  power amp can accurately reproduce the desired high frequency harmonics.

Wish list for such an expensive pre-amp might include switched options such as a straight passive attenuator(for great recordings from good CD/SACD players) to class A transistor(for dynamics/drive) to a couple of valve options(for harsher bad recordings and those who just prefer that valve sound) and maybe a digital processor.

$ and market forces dictate such a product is most likely just a pipe dream.

[Geoff-AU]

yeah you're essentially talking about 3 pre-amp circuits in one with that, the only common aspect being the input switching device.  You'd have to add another switching device to select which input you were using (and also switch the correct pre-amp to the output jacks).  Also, you'd likely need 2 power supplies in the unit as a typical SS pre-amp uses a +/- 15V supply and valve pre-amps could use something else entirely for all I know

The use of tubes in recording mics is either for their ability to add low-order harmonics (thereby making the voice sound better) or the soft-clipping ability, which would produce sonic benefits when trying to cleanly record vocals and also allow interesting effects if the mic was driven into overload.  A SS mike treated the same way would sound shit, but we're talking about creativity and expression here not fidelity.

[bhobba]

How can 2 meters of magic cable make a difference when there are hundreds of kilometres of old wiring back to the power station?

Well said.  But to be fair some claim to be able to know if bybee filters (which I think are of rather dubious value - although I am willing to be proven wrong) are present in blind tests
http://www.enjoythemusic.com/magazine/equipment/0702/bybee.htm

So my position is I will accept it when and if blind listening tests show it.

Thanks
Bill

[audiojerry]

Since I'm reading posts from a number of apparently very knowledgeable folk, I have a question about power cords. (Sorry Hugh, but I'm afraid your thread is drifting to several different topics). I thought that power cords can make a difference because a "well-designed" pc helps to filter out nasties such as RF that can be carried on the 60hz power stream. They are least effective on power amps because amps have highly regulated power supplies, but on other equipment - especially digital, a pc can make an audible difference on sound quality. In my experience, I have noticed lower noise floors and improved low-level resolution when using better pc's. Is there any validity to this?  

Personally I am not one for judging equipment on how it sounds alone. I also want to see vanishing small distortion and other objective criteria.

Given this mandate, if you were auditioning two amps, would you favor the amp that measured better, even though subjectively you ears told you the other sounded more pleasing?  If I had to rely on the measuring criteria currently available today, I choose to trust my ears over measurements.

[andyr]

Since I'm reading posts from a number of apparently very knowledgeable folk, I have a question about power cords. (Sorry Hugh, but I'm afraid your thread is drifting to several different topics). I thought that power cords can make a difference because a "well-designed" pc helps to filter out nasties such as RF that can be carried on the 60hz power stream. They are least effective on power amps because amps have highly regulated power supplies, but on other equipment - especially digital, a pc can make an a ...

Hi audiojerry,

PCs are a "religious" argument, mate ... best just to experiment and not post your results!  

However, I believe your statement that "They are least effective on power amps because amps have highly regulated power supplies" is actually not correct - bcoz power amps typically have not-very-regulated power supplies (bcoz my understanding is that regulation typically increases the impedance which the amp "sees" from the wall supply ... hence this degrades power amp transient response??).

However, knowledgable people please put me right, if the above is a misconception!  

Regards,

Andy

[Geoff-AU]

linear power supplies as found in power amps are not regulated.  I'm pretty sure andy's on the money with the look-back impedance being quite high, and a degradation in transient current response.

as for power cords, if a "well designed" power cord suppresses RF and EMI then it obviously has filtering elements built in.  Such filtering elements could just as easily be placed inside the amplifier (eg a filtering IEC socket), and if there is a noticeable sonic difference then it might mean the amp is not well enough designed to begin with.

[_scotty_]

Information deleted and PMed to ginger. Sorry, Scotty

[jules]

quote:-

"The use of tubes in recording mics is either for their ability to add low-order harmonics (thereby making the voice sound better) or the soft-clipping ability, which would produce sonic benefits when trying to cleanly record vocals and also allow interesting effects if the mic was driven into overload."

thanks Geoff ... this would seem to indicate that the use of a tube type mic. would be almost mandatory for something like a solo trumpet. Fidelity with a trumpet can be strident [would you flinch if your child took it on as an instrument   ] and massaging to remove  higher order harmonics [or maybe just not pick them up in the first place] is probably the norm.

jules

[Geoff-AU]

I played the trumpet... very briefly.  I didn't want to subject the neighbourhood to the quantity of tortured sound it would take to make me any good  

[bhobba]

I thought that power cords can make a difference because a "well-designed" pc helps to filter out nasties such as RF that can be carried on the 60hz power stream. They are least effective on power amps because amps have highly regulated power supplies, but on other equipment - especially digital, a pc can make an audible difference on sound quality. In my experience, I have noticed lower noise floors and improved low-level resolution when using better pc's. Is there any validity to this?

There may be.  IMHO power chords do not make any difference but the hallmark of the scientific method is to keep an open mind and base your views on observation.  In this case the answer is simple - perform some blind tests not just with yourself but with friends, work colleagues, fellow audiophiles etc.  Be very weary of those that try to weasel out of blind tests by making claims like the switch that switches the component in and out makes a difference - they remind me too much of those that use similar reasoning to bolster discredited claims about the paranormal.  It is unfortunate people of dubious integrity are active in audiophile circles and have even attracted the attention of debunkers like James Randini -
http://www.randi.org/jr/120304youve.html
The number one weapon you have is blind listening.  I immediately discount claims that cast doubt on the methodology itself.

if you were auditioning two amps, would you favor the amp that measured better, even though subjectively you ears told you the other sounded more pleasing?

Sigfried Linkwitz claims all modern amps with distortion below 0.1% sound about the same, all other things being equal.  Blind listening tests have tended to confirm it.  Note the caveats - 'all other things being equal' and 'tended'. I know from experience with my LS 88's and other speakers low distortion amplifiers sometimes sound the same - but not always.  Some of the time it is corrected by ensuring they are spl matched (what some dealers will do to sell product is sickening really) but again not always.  Those cases are where the caveats come in.  First amplifiers have more objective measurements than just distortion and they can sometimes be audible eg current capability into difficult loads.  I know that low distortion amps feeding my Axis Ls88's do sound different even with matched SPL - an old Pioneer with low distortion sounded thin compared to a Rotel amp and an ME sounded better again.  Best was when they were bi amped.  The cause was LS88's go down to 2 ohms at cross over - a difficult load for any amp.  So if I did find a difference I would investigate other causes like what kind of load the speaker is.  If that still did not explain it then blind listening tests have also shown another phenomena - practiced listeners can sometimes detect subtle differences most can not - so you may be one of those who actually can.  For example I have no doubt that Hugh's comments about the AKSA are valid - but Hugh is a very experienced listener - others (including me) may not be that good.  My opinion, for what it is worth, is if the amplifier is below 0.1% distortion (and the AKSA is) and other factors are precluded and you can detect a difference then one can safely be guided by ones hearing - but always remember the comparison must be done blind.  If the distortion is not below 0.1% then I would be inclined to say the audible difference is distortion and even though I may prefer it I would avoid that amp because I am only too aware how easily we can be fooled by other things - but that is just me.  If you are a very experienced listener and believe you can be objective then you may take a different view - which is OK.  You will be surprised how often differences that you thought were apparent without it being blind disappear once it is and SPL levels are matched (I have proven such many times at Hi Fi stores by taking along an SPL meter).  This is not just an audio phenomena.  I was heavily into wine tasting at one time - blind tastings are always an eye opener.  It is generally recognized the only valid wine tastings are blind - IMHO it is the same for audio.

If I had to rely on the measuring criteria currently available today, I choose to trust my ears over measurements.

Fine - just make sure the comparison is done blind and with matched SPL.

Thanks
Bill

[AKSA]

Folks,

I've just got back from a 600 mile round trip to South Australia, so I'm all in and will add my 2c tomorrow.  I've read through the posts;  some really interesting comments......  thank you!

Cheers,

Hugh