So show us an input stage that obeys the RIAA record curve, if that’s what you’re talking about, with flat response to 50Hz, then rising to 500Hz, then flat to 2122Hz, then rising from there on. Please exhibit just one. Very hard to make sense of any of this. Impossible, actually.
THD is only one of the many factors to judge an amp i think. It's too much stressed upon by objectivist, and to much neglected by some audiophiles. Not all distortion has the same effect on the sound, and some sound actually nice for many. I have both tube and class D NCore amps and both i love for the right purpose.
But it's only one of the many factors i think, group delay, step response and frequency response are also important, just like how it can deal with impendance changes and low impendance.
But it's only one of the many factors i think, group delay, step response and frequency response are also important, just like how it can deal with impendance changes and low impendance.
True, its not good to go overboard with distortion. It sounds muddy with very complex orchestral music.
When i use wolverine, with almost unmeasurable distortion, at least by my equipment, distortion definitely lower than your composite amplifier, i must use something-something preamp in order to achieve pleasing musical presentation. Without something-something its just boring. And i do not like it boring.
When i use first watt amplifiers, or an39 or other classA amps, i am not missing something-something.
What i want to convey is that those who believe that low distortion amplifier automatically sounds great, and calling me names, like i am putting shades on mona lisa, you know who you are...are wrong. Low distortion amplifier can sound great and can sound like nails on glass. Truth is, we have not figured out what parameters predict how amplifier will sound.
Just like you can not predict how vine will taste based on gas chromatography profile.
What should a recording sound accurately is very difficult because mostly we do not have the comparison.
So we used our taste. It is not 100% subjective but there always the part of subjectivity and objectivity.
The most difficult to understand is to know our taste and other people taste and related to the measurement.
In my opinion, very low distortion amplifier is the best because it can be cost effective. We can add "effect" to pre-amplifier to change the sound.
Pre-amplifier cost usually much lower than amplifier cost.
So we used our taste. It is not 100% subjective but there always the part of subjectivity and objectivity.
The most difficult to understand is to know our taste and other people taste and related to the measurement.
In my opinion, very low distortion amplifier is the best because it can be cost effective. We can add "effect" to pre-amplifier to change the sound.
Pre-amplifier cost usually much lower than amplifier cost.
This myth is like a multi-headed hydra where the heads keep growing back!
It is plainly false that a speaker's distortion is baked-in. Plug a speaker into 2 different "0.001% THD" amps, whose own distortion can be safely ignored, and the speaker could produce 1% distortion in one, and 0.1% in the other. A 20dB difference. Seriously, start asking the right questions, people! How or why is this possible? What causes these variations?
If I can add a little to this, please. Having (what I assume to be) this precise amplifier that you're mentioning and its twin has been a delight. It's an excellent example of two listeners / two experiences. We never got the chance to listen in the same room, with the same speakers, at the same time, to the same music, but I've never heard anything that I'd describe as 'not pleasing' (even with complex music). Sure, for some reason or another (it can't be simply DF b/c I've tried other amplifiers with even lower DF) perhaps the bass isn't as "tight" or may not have as much "impact" in some cases with some speakers, particularly when run balanced / bridged (which may even be expected). So, I've given them a little 'help' when paired with some speakers in the form of an SS <gasp> buffer. To me, that combination with those speakers is about as good as I've ever had in my room. The main disadvantages (to me) are size/space, complexity, and heat (along with that sinking feeling of having perhaps at some point to purchase two matched sets of 300Bs). But pulling them off the shelf and hooking them up is almost like a ritual... so, I'm bound to enjoy them, because I make an entire week-long experience out of listening to them.
The emphasis (by some, not you) on this type of amplifier or that ... or this distortion profile or that one... is interesting to me. Even though there are amplifiers that are (by some metrics) 'better' than your / my DG300Bs, I just can't see a situation where they'll ever leave my home. Are they always in the rack, no. Do they always seem to find their way back into the rack, yes. It's the je ne sais mesure....
I still need to get off my rump and build one.
I can say this without hesitation, your 'very low distortion' headamp makes me grin from ear to ear. That doesn't mean I still don't adore my little Bottlehead Crack and the plethora of other headamps I have sitting around.
We can all like all sorts of different things, as you said.
This is just an opportunity to say thank you again... and to share with the crowd yet another experience that even with the same amp... we all can have different impressions of even the most subtle differences.
I just don't understand all the passion and the direct statements (not from you) bound by what seems to be no evidence whatsoever re: what people should or should not enjoy or what is or is not "correct" or "important" about amplifier design.
But... I'm still relatively new to the hobby, and from what I've read, this is a topic as old as amplifier design (and marketing).
Cheers!
Last edited:
The unwanted harmonics have an equivalent voltage equal to the square root of the sum of their squares. The THD is simply the ratio of this equivalent voltage to the voltage of the wanted signal. The THD figure is a good indicator of the cleanliness of an amplifier.
We forget an important thing - people hear music with different goals!
I'm an sound engineer, I search for truth. Hard membranes, close listening distances, controlled room and - an amplifier with no influence. When something sounds bad I want to hear it and not bury in 1% H2. Many forget that NATURAL instruments often don't sound nice at all! Violine ... these are not "creamy" when you stand close. Drumset - MOST people don't know how a real drumset sounds, it's far from what we hear live or on recordings.
I want to hear what the mixing and mastering engineer has done. And when a record sounds bad I'm interested in how they f?d up. And I want to hear all the parts of a complex piece of music to be able to follow the piece. Understand the meaning.
But in my living room home theatre ... I use a soft membrane ring radiator hf driver. Bigger membrane midrange. Higher bass bump.
I don't want to hear streaming artefacts there and every single bad artificial reverb on voice. It's listening for fun. And I would say MOST people are listening for fun and enjoyment. They don't care about the details and often find them distracting (for a good reason). "Polishing" with some nice sounding 1% H2 is probably a good idea (until music get's more complex).
And then there is the mix engineer. When I tell you how much distortion I use for a production ... we would not need to continue this discussion 🤓
In old times we used huge mixers with 100s of (cheap) active parts in the signal flow. They added something to the over all signal.
With in box mixing (mixing completely digital in the computer) we don't have this influences, we get a perfect clean signal when we want. But do we? These mixes often miss the "glue" between the instruments, it sounds like they where recorded by their own and just put somewhere with a pan pot. (Cause that's what we do, more or less 😏). The distortion of a big mixer helps to "cover up" and fill the lowest details and make everything sound "round and together" again.
I use the Softube Console 1 system for exactly that - make the mix more "analog". You can simulate imperfect gear and the summing of a big desk - and even push it harder to get more distortion. Which is a GREAT tool to influence sound, put instruments at the right position in the mix and even influence tonality without EQing the signal (adding higher order THD to the whole signal or lower order to only lower frequencies - that's what a pushed transformer does).
It depends on the music if this effect is wanted/needed or not. You don't need it with a classical recording with a main mic and spot mics - cause there is already low "channel separation". You want it with normal multi track recording which should sound loud.
And you have to decide for every track and piece individual what's the best to transport the vision of the musician and producer.
And no - THESE are not the "bad people". They work their a?? off with long and low paid working hours to get everything together and try to get the best sound they can achieve/the musicians are able to deliver/the composition is able to deliver/the market needs/the budget allows/....
As long as there are different GOALS for sound quality ... we can discuss endless ... and we will 😎
I'm an sound engineer, I search for truth. Hard membranes, close listening distances, controlled room and - an amplifier with no influence. When something sounds bad I want to hear it and not bury in 1% H2. Many forget that NATURAL instruments often don't sound nice at all! Violine ... these are not "creamy" when you stand close. Drumset - MOST people don't know how a real drumset sounds, it's far from what we hear live or on recordings.
I want to hear what the mixing and mastering engineer has done. And when a record sounds bad I'm interested in how they f?d up. And I want to hear all the parts of a complex piece of music to be able to follow the piece. Understand the meaning.
But in my living room home theatre ... I use a soft membrane ring radiator hf driver. Bigger membrane midrange. Higher bass bump.
I don't want to hear streaming artefacts there and every single bad artificial reverb on voice. It's listening for fun. And I would say MOST people are listening for fun and enjoyment. They don't care about the details and often find them distracting (for a good reason). "Polishing" with some nice sounding 1% H2 is probably a good idea (until music get's more complex).
And then there is the mix engineer. When I tell you how much distortion I use for a production ... we would not need to continue this discussion 🤓
In old times we used huge mixers with 100s of (cheap) active parts in the signal flow. They added something to the over all signal.
With in box mixing (mixing completely digital in the computer) we don't have this influences, we get a perfect clean signal when we want. But do we? These mixes often miss the "glue" between the instruments, it sounds like they where recorded by their own and just put somewhere with a pan pot. (Cause that's what we do, more or less 😏). The distortion of a big mixer helps to "cover up" and fill the lowest details and make everything sound "round and together" again.
I use the Softube Console 1 system for exactly that - make the mix more "analog". You can simulate imperfect gear and the summing of a big desk - and even push it harder to get more distortion. Which is a GREAT tool to influence sound, put instruments at the right position in the mix and even influence tonality without EQing the signal (adding higher order THD to the whole signal or lower order to only lower frequencies - that's what a pushed transformer does).
It depends on the music if this effect is wanted/needed or not. You don't need it with a classical recording with a main mic and spot mics - cause there is already low "channel separation". You want it with normal multi track recording which should sound loud.
And you have to decide for every track and piece individual what's the best to transport the vision of the musician and producer.
And no - THESE are not the "bad people". They work their a?? off with long and low paid working hours to get everything together and try to get the best sound they can achieve/the musicians are able to deliver/the composition is able to deliver/the market needs/the budget allows/....
As long as there are different GOALS for sound quality ... we can discuss endless ... and we will 😎
What about the harmonics that some people (even me) seem to like sometimes? If we can only factor the unwanted ones... fantastic!
Perhaps we can come up with the mathematics that allows people to come up with their own version of "adjusted" HD for their own preferences, which by definition would remove all their "wanted" harmonics. All the magazines will beat a path to our door! Dang it... I may have given the snake-oil sales folks an idea.
Everyone could then agree to their own 'normalized' level of HD, and all arguments would cease.
Person 1 - I like 0% AHD.
Person 2 - HEY, ME TOO!
This is of course sarcasm. Your overall point is well-taken. I take your meaning to be exactly as intended. It is a mathematical representation of the harmonics not present in the original signal, and thus "unwanted".
I just could not resist the temptation to continue the discussion re: whether any distortion is "good" or "bad" / "wanted" or "unwanted", but perhaps adding a little levity.
THD simply is... what it is. People seem to use it for various reasons and attach some level of merit based on their individual needs / biases / goals.
Cheers!
Last edited:
Could you give me a technical reason for the same speaker produces 0.1% in one amp and 1% in another applying the same output voltage/frequency?
As you said, we are assuming the 2 amps have negligeable THD (<0.001%).
We can go to the minimum details, since I design class AB discrete amplifiers and I would like to understand your statement.
Sure, I'm considering real reactive load.
If both amps produce a clean sine wave with 10Vrms and THD = 0.05%, why a speaker would produce more THD in one amp and less in the other?
I'm assuming we are considering class AB amplifiers with minimum good quality (high damping factor, stable, no phase shifts etc).
That would be the trick.
Perhaps abstract did not assume this. However, I don't want to assume what they assumed. 😉
@ron68 -
To me, that's one of the challenges of discussions like this. THD is almost always measured with a pure resistive load. Heck, people are 'mocked' when they don't use non-inductive resistors for their dummy loads on occasion.
Then... when the talk of THD comes into play, (some) people may assume that the THD at 1kHz is "the same" as the THD across the audible frequency range for all amplifiers. For some beloved amplifiers out there... that's not the situation. I like a few amplifiers that do not have 'ruler flat' THD vs. frequency plots. Some people may further assume that the THD characteristics measured at 1W / 8R (purely resistive) hold true into any load the amplifier sees. That's not true for what I'd consider a large number of amplifiers out there.
Engineers that seek the "perfect" amplifier may strive to create amplifiers that negate those 'limitations'. I think it's phenomenal, and I've even built / bought a few. However, those that simply look at an FFT of the behavior of one amp at one frequency at one power level into one (purely resistive) load and say "that's the kind of amp I like"... may (or may not) be ignoring how that one specific amplifier is behaving at different frequencies, different power levels, and into reactive loads.
To me, that's one of the challenges of discussions like this. THD is almost always measured with a pure resistive load. Heck, people are 'mocked' when they don't use non-inductive resistors for their dummy loads on occasion.
Then... when the talk of THD comes into play, (some) people may assume that the THD at 1kHz is "the same" as the THD across the audible frequency range for all amplifiers. For some beloved amplifiers out there... that's not the situation. I like a few amplifiers that do not have 'ruler flat' THD vs. frequency plots. Some people may further assume that the THD characteristics measured at 1W / 8R (purely resistive) hold true into any load the amplifier sees. That's not true for what I'd consider a large number of amplifiers out there.
Engineers that seek the "perfect" amplifier may strive to create amplifiers that negate those 'limitations'. I think it's phenomenal, and I've even built / bought a few. However, those that simply look at an FFT of the behavior of one amp at one frequency at one power level into one (purely resistive) load and say "that's the kind of amp I like"... may (or may not) be ignoring how that one specific amplifier is behaving at different frequencies, different power levels, and into reactive loads.
@ItsAllInMyHead
Sure! Single number at a specific situation doesn't mean much.
We need to consider the curves along the frequency range and with real load. When simulating, use a load composed by an arragement of capacitors, inductors and resistances to model the impedance of the speaker. In the workbench real testing, better use a speaker while wearing earplugs.
So considering real load, THD along a curve for both the amp and the speaker, the question: is it worth to design a amp with THD 0.001% (along the curve) if the THD speaker curve is between 0.1% to 1%?
We are merging 2 real curves along the frequency range which includes all factors: amp THD curve + speaker THD curve.
Would the merged curves (the final curve) different much from the speaker curve?
Sure! Single number at a specific situation doesn't mean much.
We need to consider the curves along the frequency range and with real load. When simulating, use a load composed by an arragement of capacitors, inductors and resistances to model the impedance of the speaker. In the workbench real testing, better use a speaker while wearing earplugs.
So considering real load, THD along a curve for both the amp and the speaker, the question: is it worth to design a amp with THD 0.001% (along the curve) if the THD speaker curve is between 0.1% to 1%?
We are merging 2 real curves along the frequency range which includes all factors: amp THD curve + speaker THD curve.
Would the merged curves (the final curve) different much from the speaker curve?