Amplifier distortion characteristics.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have been on a mission for the past year or so for ultra low distortion and ended up with DSPs LME49830s and active monitors with the flagship scan speak drivers which are all very good, I wouldn't have it any other way. Some others would. The reason I went with low distortion is because i don't think a system should have any characteristics, if it does (in my opinion) it isn't doing its job properly.

Another reason is because those who like distortion in their products, so they sound different, don't have any way of getting to their goal whatsoever, they work 100% blind and so most are messing around with combinations like a blind person trying to find a needle I a haystack. Someone will make an amp, someone will build it, some will like it and others won't. But inevitably the amps start life with the electronic engineer, who in my experience don't seem to have a grasp on what people want as an amp, the designer will try and make an amp with as little distortion and best performance as possible (because he is an engineer!) but because of costs and old stereotypical ways of doing things in audio they will end up with a mish mash of an amp.

So they will end up with an amp that was designed to be good but ended up not. This amp will then be connected to some speakers they make be totally inappropriate for it perhaps due to damping, voltage, current capabilities or more. This is what gives us the total lottery of hi-fi, which ends up being not really hifi.

On a side note I think very few people know what hi-fi is. Hi-fi is high fidelity or being true to the original signal, this does not include the listeners opinion! If it's true to the source then it's true to the source! You get what your given! For this reason tubes, passive crossovers or anything with naturally high distortion are not hi-fi. They are enjoying music with a different perspective.

This is what leads me to think integration is absolutely key to a achieving true hi-fi, at least with current driver technology.

What if we could design an amplifier to driver one woofer, no other, we could tailor if perfectly so we need no more or less power, voltage current etc. This way the designer knows exactly what he's doing and so doesn't have to make sacrifices in performance to try and make the amplifier compatible with the majority of speakers, this also reduces cost for unnecessaries.

So what i am saying is this forum does things wrong! We shouldn't have a amplifier section nor a speaker section, or line level and so on, we show have two sections. Audio systems and everything else!

I hope this opens eyes for some people or at least gives a spin on things!
 
So what i am saying is this forum does things wrong! We shouldn't have a amplifier section nor a speaker section, or line level and so on, we show have two sections. Audio systems and everything else!

I hope this opens eyes for some people or at least gives a spin on things!
I'm on your side, Boscoe, I would agree that the audio industry in general is approaching the goal of optimum sound in the wrong fashion, that being, to be fiercely obsessed with refining the performance of one element therein, and ignoring the rest of the ensemble. Imagine if a Formula 1 car was developed by distinct groups of people, each of which was totally dedicated to extracting the most from either the engine, the suspension, the control systems, etc, who never talked to each other, consulted on anything. Until just before the race, the bits were quickly tied together as best they could -- just the way to create a real winner, I reckon!!

A typical example: bloke creates the "ultimate" amplfier, hooks it up, and it sounds rather non-special. Aahh, he says, everyone knows speakers are the weakest link, what hope have I got when you have to deal with such a device slugging the sound quality ...

Many, many years ago I realised that the "only" way to get optimum sound, the only way, is to treat everything in it as a gesalt, as a single entity, as a unity. Doing otherwise is always going to create an under performing compromise, the sort of setup that does some recordings brilliantly, and makes a mess of a whole lot of others ...

Frank
 
Another thing that the audiophile circles seem far too interested in is spending lots of time/money obsessing over things that are largely unimportant, or are actually completely pointless. Of course one needs to be quite capable to be able to build/mod things in any semi decent (read competent from an engineering point of view) way*, but even so, people end up spending silly money on areas that don't really count, where really they'd be far better just spending that money on music and enjoying their system!

* of course one area where most people could easily build excellent quality parts is with interconnects/speaker cables.
 
But inevitably the amps start life with the electronic engineer, who in my experience don't seem to have a grasp on what people want as an amp...

In most companies an amp starts life in Marketing where some guy in a suit says, "I need an amp that does this and costs this much". The engineer is just a hired gun. If he can't meet those goals in the allotted time, there is somebody just out of college that will be hired to do the job. When an engineer is allowed to design with a clean sheet of paper, most often the result is something that takes too long to finish the design, costs too much and nobody will buy it. If you want such a design, you are in the right place. :)
 
Member
Joined 2010
Paid Member
Much of the attraction of DIY is for the very reasons belittled in the OP. Many guys here are indeed seeking that particular quality sound in the electronics that tickles their fancy rather than the purist notion of "no characteristic sound". I think the popular term is "high end" sound. In fact, most commercial offerings will often do straight, clean amplification better and cheaper overall than typical DIY efforts, so why would we bother reinventing that wheel?

I suggest there is no basis in the field of personal entertainment to say that an audio device should or should not have a sonic character, colour or whatever effect the user wants. If we look at the entertainment field, we find huge variation in the form and function of products intended to fill basic needs so why should amplifiers be any different? Also, in the High End market, we find that the popular big bucks are in the bent, not straight commercial amplifiers. I should think it only natural for DIYs to be attracted to such High End, high profile products which can also be cheaply emulated.

Now, 'still wondering why when someone posts a new design, after the salutary peer remarks, the first interested DIY post always asks "How does it sound?" :D
 
This seems a simplistic statement and one that ignores reality to some extent. While I agree in principle that the audio system should be designed holistically, you would be better off preaching to manufacturers. They are the reason audio has hardly progressed since the 80s, and yet they are presumably stuck with the need to design and sell equipment that people want to buy. The main problem seems to be that the general public with MP3 etc is just not interested in high quality sound, leaving only the audiophile community that buy high performance audio gear, and audiophiles are generally deeply suspicious of integrated systems, especially when DSP is involved. They also like to be able to tweak things, and an integrated system leaves no room for this.

Many people building amps and speakers on this forum probably don't have the breadth of knowledge, experience or test equipment required to construct a full system holistically. What's more many will already have some parts of the system and only be interested in fiddling with the rest - for example one might have an expensive set of passive speakers and therefore will only be interested in constructing the best possible power amp to drive them. Furthermore DSP filters, the digital - analogue interface and so many other factors make the whole thing extremely complex and therefore beyond the scope of many hobbyists. Concentrating on one part of the system, eg the power amp, at least keeps the scope manageable and the project is more likely to be successful.

I also second Gootee's post that THD is a simplistic measure of audio performance and concentrating on that at the expense of everything else could be seriously missing the point. I have built and measured many amps and preamps with very low THD and yet have found that this parameter alone does not correlate well to the "fidelity" of the resulting sound. It seems akin to designing louspeakers with perfectly flat frequency response - to achieve this you have to compromise on other factors and the end result is generally not as pleasant or accurate as a design that considers all the parameters that might affect what we hear and perhaps compromises on measured response accuracy. And by the way: ALL loudspeakers (active or passive) following the design process have to be fine-tuned by ear as well as measurement before they sound correct. Why? Because there are so many variables that affect the final sound that we hear and good engineering is about managing the inevitable compromises.
 
Let's think outside the box

I have been on a mission for the past year or so for ultra low distortion... The reason I went with low distortion is because i don't think a system should have any characteristics, if it does (in my opinion) it isn't doing its job properly.

those who like distortion in their products, so they sound different, don't have any way of getting to their goal whatsoever, they work 100% blind

On a side note I think very few people know what hi-fi is. Hi-fi is high fidelity or being true to the original signal,

For this reason tubes, passive crossovers or anything with naturally high distortion are not hi-fi. They are enjoying music with a different perspective.

Let's think outside the box (and be open minded).

So you think hi-fi is being true to the original signal? How about hi-fi is being true to the original sound? Is it not a better "definition"?

If we can accept the above proposal, isn't a high distortion tube amp that can produce sound closer to the original sound more hi-fi than a typical low distortion chip amp? (also don't forget that real sound, even if distorted, tends to be non-fatiguing in nature).

If we can accept the above observation, then there could be something wrong with the assumption that "hi-fi is low distortion amplifier". What could be the case? May be THD is not 100% parameter that determines system fidelity? Then what, or how many percent is the THD contribution, 80%? 50%? 10%?

Tube amp is so simple that it doesn't alter many aspects of the sound. It just distorts the idealized sine-wave negatively, but it doesn't "touch" the other aspect of the sound.

It is different with complex circuitry that "touch" almost everything. Isn't it unfortunate, that designers are not smart enough to think of better measures for fidelity other than THD?

SECOND OBSERVATION: A VIEW ON INTEGRATED SYSTEM.

In line with your opinion (and I'm agree with it) that a hi-fi system should be viewed as an integrated system from source to speaker, isn't THD is only a small part of the chain?

What does that mean?

I think we can be agree on the concept that a hi-fi system must have low THD amplifiers. But the integrated system fidelity can only be slightly affected by that. It is imo possible that a certain high distortion amplifier will have better synergy with everything else in the signal chain, than a low distortion amplifier.

This is in accordance with some people experience including me, that it is equivalently okay to pick any good low distortion amplifier, as long as the speaker and the other chains (such as power supply) are made perfect.

But wait! Isn't this like comparing class-A amps with class-B amps? Nelson Pass seems to leave class-B amps behind for sound quality reason. I have such experience that I still find class-A amps less fatiguing.

So? How can we at least put this fatiguing variable into the grand fidelity equation??
 
So you think hi-fi is being true to the original signal? How about hi-fi is being true to the original sound? Is it not a better "definition"?

If we can accept the above proposal, isn't a high distortion tube amp that can produce sound closer to the original sound more hi-fi than a typical low distortion chip amp? (also don't forget that real sound, even if distorted, tends to be non-fatiguing in nature).

If we can accept the above observation, then there could be something wrong with the assumption that "hi-fi is low distortion amplifier". What could be the case? May be THD is not 100% parameter that determines system fidelity? Then what, or how many percent is the THD contribution, 80%? 50%? 10%?
This is where I think the industry has gone astray -- everything in my experiences over the last 20 years or so has said very clearly that lowest possible distortion is the goal, but being sloppy about achieving real, low, distortion will always push the achievement of satisfying, non-fatiguing, natural sound further away.

And what do I mean by "sloppy"? Well, sorry to say, the pulling a component out of an audio system, sticking it on test bench, attaching an impressive array of test gear to it, running a standard regime of procedures on it, and then pronouncing it competent is never going to do it for me. This is like someone complaining that a car doesn't perform right, so the engineers unbolt and remove the engine, stick it in a test jig, have a robotic procedure run it over a few rev ranges, and then claim that there are no issues. Would you believe them, and be confident that the "problem" was all in your head?

Frank
 
Last edited:
Thanks for the replies. I didn't mean to belittle it! If that's what your into so be it! Even in that case I don't think pele have a goal whe designing an amp, they just hit and hope.

My goal is and always has been to have the most-accurate reproduction, whether I "like the sound" or not.

(It turns out that I do like the sound when it's accurately reproduced. But I know people who don't like it unless it has hugely-overblown bass, and boomy box-speaker bass artifacts, for example. I happen to call those people "idiots".)

It makes ZERO sense to want it any other way than as accurately-reproduced as possible. Any desired "coloration" should be provided by the original artists. Otherwise, you won't be hearing the music as they intended it to sound, which I thought I could presume was the goal of a "sound-reproduction system". Otherwise it should be called a "sound-changing system". For example, fun-house mirrors are interesting. But we usually call them "fun-house mirrors", not just "mirrors". And I wouldn't want one as the mirror that I normally use every day.
 
AX tech editor
Joined 2002
Paid Member
So you think hi-fi is being true to the original signal? How about hi-fi is being true to the original sound? Is it not a better "definition"?

The input to your hifi is not a 'sound', it's an electrical signal. The only thing you can hope to achieve is that the signal is amplified and reproduced as faithfully as the input. The alternative is to 'massage' the signal in the hope to get closer to the original. The problem here is that 1 - you have no idea how the original sounded, and 2 - you're second guessing the recording engineer and you become part of the musical performance. Could be an interesting and rewarding activity but it is not high fidelity reproduction.

Isn't it unfortunate, that designers are not smart enough to think of better measures for fidelity other than THD?

I think they are smart enough. Over the decades many smart measurement techniques have been developed beyond a simple THD number. IMD, TIM, sine chirp, slew rate mesurements, asymmetrical pulse, DIM-30, DIM-100, the 31-frequencies music-like ISO tones, you name it.
Typically, what you find is that they all are in some way related, meaning that a very good performance on one test also shows good performance in another. Also, designers found that the very good, neutral, hi-fidelity amps don't necessarily sell well. As you mentioned, many people enjoy the coloration from some types of amps so for the designer it is counter productive to build 'better' amps.
The situation is that if you want a neutral amp that faithfully reproduces whatever is thrown at it, with no coloration or audible distortion, you can go out and buy one. If you want an amp that gives a specific character to the sound, you can go out and buy one. So I don't see what's missing here.

jan didden
 
Last edited:
Hi,

My goal is and always has been to have the most-accurate reproduction, whether I "like the sound" or not.
What a pity! :rolleyes: I let my ears decide about that. :D
What could make You listen to something You dislike if You´re not forced to?
Sound like some weird kind of masochism to me. :p

HiFi is Lifestyle and should be enjoyment and fun and give You good times.
If my ears (or brain) tell me that I dislike the sound, if they tell me: "no, that doesn´t sound authentic and real to me", then any discussion about simple numbers and figures is obsolete and the device is either put away or goes back to the workbench.
The classic distortion measurement doesn´t correlate to sonic impression. Why then should I regard it as the most serious parameter? Unfortunately more promising types of measurement are more complex and it take ages before they become standards if ever.
Of course do I also look at THD figures and else. But as soon as certain limits for example say -60dB THD are reached, the importance of other parameters surpasses. Of course would I prefer the lower THD-circuit over the higher-THD circuit if the sonic impression were the same.
I regard THD measurements as a technicians/engineers tool for proof of working and quality control, but certainly not a tool that decides about sound quality.

It makes ZERO sense to want it any other way than as accurately-reproduced as possible.
I disagree. I could indeed fully agree with this statement if our hearing sense and brain would be anywhere close to a measurement system. But this is by far not the case. Accurate to the ear is certainly something totally different than accurate in the technical sense.

jauu
Calvin
 
I disagree. I could indeed fully agree with this statement if our hearing sense and brain would be anywhere close to a measurement system. But this is by far not the case. Accurate to the ear is certainly something totally different than accurate in the technical sense.

jauu
Calvin

That is not the task of the amplifier to do. The amp just needs to amplify exactly the input signal.

The 'humar-ear' correction you talk about took place long before you get to feed the signal to your amp: The microphones that recorded the live singer, instruments. They converted soundwaves as they are and as we hear them, into electrical waves. Those are not the perfect "original" waves, they are as a listener would hear it at a certain location. Mics are aurally placed, much like ears. So what you get, is an electrical signal that has 'encoded' into it the way we should hear the music, when that signal is exactly reproduced into a converter that's the exact opposite of the microphone: The speakers, positioned such way, the soundwaves it reproduces would be the exact sound waves in the air as were present at the time of recording.

A LOT somewhere in between there is altered, and one end happens in the studio, the other end in your living room. The amp? It has just the 'dumb' task of making stronger that what's thrown at it.

Edit: Basically I could say that I would not want an amp to tamper with the existing 'encoding' in the music signal by allowing to distort it, whether it's the nice one or the fatiguing one :)

Another: The ideal reproduction would be if you could position your speakers as the mics were and equalize the signal so that the transfer curve for both the recording mics and the speakers would be identical.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.