Amplifier Sonic Signatures

[Dimitrov]

I had a debate with a friend of mine yesterday about the role of tone controls in amplifiers and his position was that tone controls mess up the signal and he prefers to listen without them. Fair enough.

So it got me thinking. If tone controls mess up the signal to reach a desired preference then to me the idea of an amplifier sonic signature is similar in that (I assume) the waveform is distorted/messed up in various ways to achieve a "flavour" of sound. Is this line of thinking correct?

I'm no amp designer so I would appreciate comments from those more experienced. If an amp is claimed to have "warmth" or "recessed mids" or whatever, and assuming these characteristics are limited to the amp itself, is that not deviating from the core operation of an amp?

If this is true, then purists are in a bit of a pickle ie tone controls mess up the signal, amps that have a "sound" mess up the signal.

 

[DonH56]

An age-old debate that for me falls into the "religious" category so not worth the effort to participate. Here are my off-the-cuff responses:

1. EQ is used extensively throughout the recording and playback process. Tone controls are just one more step of EQ. Since the consumer probably does not listen in the same room using the same components as the final mastering/mixing engineer then it is unlikely the consumer is hearing the exact same final result anyway. Tweaking tone controls to taste and to help correct room response and so far doesn't seem any worse to me than all the bazillions of things done up to that point.
2. An amplifier that deviates from ideal is, well, non-ideal. The fact is no amplifier is perfect, and amplifiers all have some load interaction, so there will always be some sort of "signature" for a given amplifier/speaker combination. (Or preamp/amp combination, but interaction between preamp and poweramp is usually insignificant unless something is badly mismatched.) I hold the opinion that most SS amps sound similar if not the same driving most speakers so long as the amp is not clipping or distorting heavily. It is pretty easy to tell a tube from a SS amp, and speakers with wider impedance excursions tend to be more sensitive to amplifier output impedance. Differences among amplifiers exist but IME/IMO are overblown most of the time. That is not saying they aren't there, but they are often more subtle than people think and influenced by folks' own bias. Mine too.
3. Purists who feel a "straight wire with gain" is the ideal system are looking through rose-colored glasses IME/IMO. They'd probably choke at the amount of processing a typical recording goes through on the way to the final product, much of it done (gasp!) digitally, and at the end they are likely listening mostly to the impact of the room and their speaker choice rather than the amplifier.

I decided to stop at three since others will chime in with better and more detailed responses and I have to get back to work.

FWIWFM - Don

 

[RayDunzl]

An amp should measure flat. There's no good reason I know of for them not to be.

Adjustments should be applied to the signal fed to the amplifier (which is what a "tone control" does)

---

I guess I'm no purist if that means don't mess with anything.

Maybe I am a purist if I go by measurable results.

---

I send my digits (signal) to DSP before they go to the amps.

What would be a flat frequency response looks something like this leaving the preamp after DSP is applied:

Red - no DSP - flat signal
Blue - DSP "tone" adjustment for JBL LSR 308 (left)
Green - DSP "tone" adjustment for MartinLogan reQuest with some cheap subs (left)

Note that after the speakers and room apply their contribution, the measured audible result approaches "flat" - that suits the "purist" in me.

Here is a recent in-room measurement of the MartinLogans, left and right using signal correction similar to if not the same as that above:

 

[Dimitrov]

An age-old debate that for me falls into the "religious" category so not worth the effort to participate. Here are my off-the-cuff responses:An amplifier that deviates from ideal is, well, non-ideal. The fact is no amplifier is perfect, and amplifiers all have some load interaction, so there will always be some sort of "signature" for a given amplifier/speaker combination. (Or preamp/amp combination, but interaction between preamp and poweramp is usually insignificant unless something is badly mismatched.) I hold the opinion that most SS amps sound similar if not the same driving most speakers so long as the amp is not clipping or distorting heavily. It is pretty easy to tell a tube from a SS amp, and speakers with wider impedance excursions tend to be more sensitive to amplifier output impedance. Differences among amplifiers exist but IME/IMO are overblown most of the time. That is not saying they aren't there, but they are often more subtle than people think and influenced by folks' own bias. FWIWFM - Don

Thanks for your comments. I read the other two points but felt I had to quote your 3rd point. Aren't ideal amplifiers common-place nowadays? When you say no amp is perfect, I take that to mean that no amp is sonically transparent because a perfect amp is one that behaves as, in my view, a voltage multiplier. Correct me if I'm wrong here. Is it not true that amps that are ideal/perfect in intended operation have been built?

You say there will always be a sort of signature, but blind test studies support the notion that if two amps are operating linearly into loads, at low distortion there is a high level of confidence they will be indistinguishable to ears.

 

[DonH56]

Thanks for your comments. I read the other two points but felt I had to quote your 3rd point. Aren't ideal amplifiers common-place nowadays? When you say no amp is perfect, I take that to mean that no amp is sonically transparent because a perfect amp is one that behaves as, in my view, a voltage multiplier. Correct me if I'm wrong here. Is it not true that amps that are ideal/perfect in intended operation have been built?

I'm an engineer so "ideal" and "perfect" are not words I usually/willingly use for any real device. The rest of the paragraph is important; many amps are audibly indistinguishable IME/IMO and based on various reported test results, but with the right (wrong) load and such or when overdriven then they can usually be distinguished. There are many things that can make amplifiers sound different. Most amps sound the same driving most speakers but the more tightly you constrain the test cases the less relevant is the design in the real world.

Most amplifier designs are to emulate an ideal voltage source, so they do multiply voltage, but also current, providing power gain (not just voltage gain). Nit-picking, but more than just a voltage multiplier. A transformer can provide voltage multiplication but would be a poor power amplifier.

You say there will always be a sort of signature, but blind test studies support the notion that if two amps are operating linearly into loads, at low distortion there is a high level of confidence they will be indistinguishable to ears.

Wonderfully weasel-worded. I think we are saying the same thing. Measurements will probably show a difference among amps even at low output levels but those differences are typically inaudible. What tends to distinguish amplifiers is how they operate at their extremes. Studies (which I have no hope of finding at the moment; Google them) show even tube and SS amps are indistinguishable in blind tests when operated well within spec and into average ("reasonable") speakers. They are plenty of unreasonable speakers out there...

One of my articles plots the simulated frequency response into various speaker load models. Many others have done the same. Much easier to see in a picture.

 

[Dimitrov]

I'm an engineer so "ideal" and "perfect" are not words I usually/willingly use for any real device. The rest of the paragraph is important; many amps are audibly indistinguishable IME/IMO and based on various reported test results, but with the right (wrong) load and such or when overdriven then they can usually be distinguished. There are many things that can make amplifiers sound different. Most amps sound the same driving most speakers but the more tightly you constrain the test cases the less relevant is the design in the real world.

Most amplifier designs are to emulate an ideal voltage source, so they do multiply voltage, but also current, providing power gain (not just voltage gain). Nit-picking, but more than just a voltage multiplier. A transformer can provide voltage multiplication but would be a poor power amplifier.



Wonderfully weasel-worded. I think we are saying the same thing. Measurements will probably show a difference among amps even at low output levels but those differences are typically inaudible. What tends to distinguish amplifiers is how they operate at their extremes. Studies (which I have no hope of finding at the moment; Google them) show even tube and SS amps are indistinguishable in blind tests when operated well within spec and into average ("reasonable") speakers. They are plenty of unreasonable speakers out there...

One of my articles plots the simulated frequency response into various speaker load models. Many others have done the same. Much easier to see in a picture.

Thanks for the informative reply. So in your view if an amp has a characteristic "sound" as a design feature (limiting to solid state) then is that usually a factor of a non-flat frequency response? Or a higher than ideal output impedance? Whatever it is would you agree that it violates the definition of an amplifier?

 

[A.wayne]

Yes and no , Amplifier tone is usually load dependent even on well made amplifiers because all amplifiers ( except boulder ) increases distortion as you lower drive Resistance , distortion increases in itself is not the issue, as all sounds we hear in nature has THD.

As to tone controls they bring a whole other bag of nails to the party , additional signal complexity aside , their turn over frequency has sweeping effects and does offer a change of "tone" across a broad range..

You should really playback flat to truly hear what was intended by the Recording engineer and producer, in a proper system , no tone circuit is really necessary ...


Regards ..

 

[A.wayne]

This is interesting Ray .... thanks for posting ....How are you guys measuring below 300 hz in room ..? how much smoothing are you using ..?


Regards .

 

[Mivera]

Trust your ears. Engineers that have no noteworthy achievements in the field of high end audio are often heavily biased from textbook material that was pounded into their heads back in general EE school. Your ears are a far more valuable tool for finding the answer to this question.

 

[RayDunzl]

This is interesting Ray .... thanks for posting ....How are you guys measuring below 300 hz in room ..? how much smoothing are you using ..?

It's a umik-1 perched on the back of the couch, at the listening position.

Smoothing is 1/6 octave above (in the labels at the bottom).

 

[DonH56]

Thanks for the informative reply. So in your view if an amp has a characteristic "sound" as a design feature (limiting to solid state) then is that usually a factor of a non-flat frequency response? Or a higher than ideal output impedance? Whatever it is would you agree that it violates the definition of an amplifier?

Not sure where you are going with this... Most amplifiers measure flat on the bench into a resistive load. Output impedance higher than zero and having any reactive terms will change the frequency response into a reactive load that is less than infinite in impedance. "Ideal" is a standard unobtainable in the real world but many amplifiers get very close, and in the audio world close enough that the non-ideality is insignificant (not audible). No real-world amplifier is "ideal" in the simplest technical sense of the definition. Don't read too much into my responses. Do you have a technical background? Not trying to be an arse, trying to understand my audience.

Edit: Regarding your question about "characteristic sound", there are many ways to achieve a desired sonic profile, and frequency response is only one parameter that a designer can vary though it is perhaps the easiest and most obvious one. But there are many ways to vary the frequency response at the circuit level, some of which will impact how it interacts with the load more than others.

 

[RayDunzl]

I suppose my little Krells have a "characteristic Krell sound"...

I don't see it in the specs, other than I would think they should at least be competent.

I haven't measured them myself, but they don't sound broken.

A similar amp was tested by JA...

https://www.stereophile.com/content/krell-full-power-balanced-350mc-monoblock-amplifier-measurements

Maybe not terribly impressive when compared to more moderne designs 17 years later.

 

[Dimitrov]

Not sure where you are going with this... Most amplifiers measure flat on the bench into a resistive load. Output impedance higher than zero and having any reactive terms will change the frequency response into a reactive load that is less than infinite in impedance. "Ideal" is a standard unobtainable in the real world but many amplifiers get very close, and in the audio world close enough that the non-ideality is insignificant (not audible). No real-world amplifier is "ideal" in the simplest technical sense of the definition. Don't read too much into my responses. Do you have a technical background? Not trying to be an arse, trying to understand my audience.

Edit: Regarding your question about "characteristic sound", there are many ways to achieve a desired sonic profile, and frequency response is only one parameter that a designer can vary though it is perhaps the easiest and most obvious one. But there are many ways to vary the frequency response at the circuit level, some of which will impact how it interacts with the load more than others.

Hi Don. I don't have a technical background, but I am very interested in the technical aspects of the conversation. Where I am going with this ... if tone controls mess up the signal then an amp that has a deliberately designed "sonic signature" is in some way also messing up the signal in various ways. That's what I was trying to get at.

As you say, amps can be designed that are sonically transparent - they just amplify voltage/current with no discernible "sound" of their own. This is the definition of an amplifier - it amplifies, so if you deviate from this "ideal" then you have an effects box. The amp then behaves like a non-configurable, non-defeatable tone control EQ.

I just wanted to know if I was missing some larger picture here. I know perfection is illusive in the real world.

 

[watchnerd]

Hi Don. I don't have a technical background, but I am very interested in the technical aspects of the conversation. Where I am going with this ... if tone controls mess up the signal then an amp that has a deliberately designed "sonic signature" is in some way also messing up the signal in various ways. That's what I was trying to get at.

Yes. Also true of DACs.

 

[Cosmik]

Trust your ears. Engineers that have no noteworthy achievements in the field of high end audio are often heavily biased from textbook material that was pounded into their heads back in general EE school. Your ears are a far more valuable tool for finding the answer to this question.

Some people trusting their ears:
http://www.bbc.co.uk/programmes/p01tds0d

(from 40.40-ish if the link takes you to the full programme)

 

[watchnerd]

Trust your ears. Engineers that have no noteworthy achievements in the field of high end audio are often heavily biased from textbook material that was pounded into their heads back in general EE school. Your ears are a far more valuable tool for finding the answer to this question.

This being the science forum, it should be pointed out that your statement pretty much contradicts the existing science research on auditory perception and how easily our ears can be fooled.

 

[Mivera]

This being the science forum, it should be pointed out that your statement pretty much contradicts the existing science research on auditory perception and how easily our ears can be fooled.

There's 2 options:

1: Have the credentials to know what matters based on years of experience developing noteworthy gear in this specialized field

2: Trust your ears

Anything in between will only end with disappointing results.

 

[Cosmik]

There's 2 options:

1: Have the credentials to know what matters based on years of experience developing noteworthy gear in this specialized field

2: Trust your ears

Anything in between will only end with disappointing results.

You make it sound as though this stuff doesn't exist in text books and also that previous great work disappears into the ether when the designer pops his clogs. Surely knowledge is built upon the foundation of what went before and can be written down.

The people who believe that cables sound different have a difficulty, though: none of it exists in text books because there is nothing to write down, or the ideas are too ridiculous to actually put down in print. Instead they must pretend that they have "proprietary" secrets presumably kept in a vault in the basement.

 

[Mivera]

You never truly know anything until you have first hand experience. This must be obtained by personal trial and error. Reading textbooks and believing what you read is no different than believing in Santa Class or the Tooth Fairy as a child. It's only a belief at this stage.

Me personally don't believe anything until I have personally conducted my own testing. For many layman audiophiles, they simply don't have the knowhow to conduct the necessary tests to truly know what matters. So the next best thing is to just trust your ears.

 

[watchnerd]

There's 2 options:

2: Trust your ears

This one isn't actually an option.

Ears can't be trusted.