Can you give me a reference for this rather remarkable statement? I recall that in controlled listening tests, high THD amps are indeed distinguishable from low THD amps. Now, below detection threshold, there is indeed no correlation, but that's how one defines detection threshold, n'est-ce pas? If you honestly think that there's no correlation, you should run your amps at clipping and beyond.
Good THD performance is necessary and is concommitant to linearity. But no-one argues that it's sufficient. Straw man.
As a general rule, I try and never think that I am clever!fdegrove said:
I would posit that the distortion of the speaker drivers is the #1 distortion affecting sound quality in any system, so I would say we are measuring the wrong thing when trying to make a piano disappear. (of course someone will prove me wrong, but I am used to that 🙂 ) Until you get a perfect driver, you will always be able to "hear" the speakers, regardless of the distortion in amplification stage (let alone cables, but that is another discussion entirely!).
Perceptions of distortion go back to my earlier argument--regardless of your opinion of how distortion sounds (whatever the harmonic), it is not accurate relative to the original recording. No assumptions there. I can measure it. If you do not like the original recording and would like additional effects to make it sound better, I see nothing wrong with that at all. If you do not like perfect amplification, that is perfectly fine! (and I almost missed that pun. . . 😀)
If you think that "perfect" amps sound bad and you like tube sound better, great! Use the tubes, but acknowledge that what you are enjoying with the tubes is an additional layer of sound not in the original. I happen to like more clean amplification, but like I said, that is my opinion.
One thing that I think is pretty interesting though, is that I have seen a bunch of people arguing for tube amps, and that the distortion that they add is musical, but I have never heard anyone argue for using sonically rich hard woods for speaker cabinets. Just an observation to add fuel to the fire. . . .
Au contraire. Spencer Hughes is one top speaker designer who comes to mind. And Dudley Harwood. There are no doubt others.
SY said:
I stand corrected!
Though I was under the impression that MDF/Plywood were fairly ubiquitous nowadays. I thought all of the hardwood designed had been abandoned. I guess that goes to show what I know. 🙄
Hi,
I don't think "perfect" amps would sound bad at all.
It's just that I don't have a clue what this "perfect amp" should measure like.....
What do you think tubes add? An additional layer of what exactly?
If there is such thing as an addition to the original signal then that would be distortion whether that distortion is generated by a tube or a semi-conductor is really not relevant.
Either way it should be measurable.
Last time I checked tubes are not generic distortion generators. Quite to the contrary, they are infinetely more linear devices than semi-conductors requiring much less correction for the same reduction in distortion.
Whatever distortion tubes generate however is intrinsically less abrasive to the human ear.
Add to that a softer way of going into clipping and it's not too hard to see why they're popular devices for amplifying music.
IOW to my ears tube amplifiers sound more musically correct whereas semi-conductor amps sound more errr..........musically corrupt.
As I said, to my ears. So pick your poison.
Cheers, 😉
I don't think "perfect" amps would sound bad at all.
It's just that I don't have a clue what this "perfect amp" should measure like.....
What do you think tubes add? An additional layer of what exactly?
If there is such thing as an addition to the original signal then that would be distortion whether that distortion is generated by a tube or a semi-conductor is really not relevant.
Either way it should be measurable.
Last time I checked tubes are not generic distortion generators. Quite to the contrary, they are infinetely more linear devices than semi-conductors requiring much less correction for the same reduction in distortion.
Whatever distortion tubes generate however is intrinsically less abrasive to the human ear.
Add to that a softer way of going into clipping and it's not too hard to see why they're popular devices for amplifying music.
IOW to my ears tube amplifiers sound more musically correct whereas semi-conductor amps sound more errr..........musically corrupt.
As I said, to my ears. So pick your poison.
Cheers, 😉
No correlation with how an amp sounds, is what I said. Meaningless would be have been a better word. Useless also comes to mind.
I won't deny that the are distinguishable..but only under very specific circumstances. (Such as when the distribution of THD is similar.)
However even when THD is above the audible threshold it has no correlation to how an amps sounds...or virtually no correlation. For instance an amp with 5% THD can sound less distorted than an amp with 1% THD. Simply because THD does not tell us how the THD is distributed.
It is the de-facto method by which tube based amplifiers are dismissed. Not a straw man at all. Instead the people who fence with THD are the ones who are creating a straw man.
Gotta call "shenanigans" on that one. I have in front of me a classic solid state amp with vanishingly low THD (0.002%), dominated by third. I also have an SET with massive amounts of THD(6.5%), dominated by second. Are you saying that because the harmonic distribution is different that these amps can't be distinguished in a blind test?
Or did you mean the opposite of what you wrote, that an amp with 6% third won't sound different than an amp with 0.01% third?
Or did you mean something different, that an amp with 5% THD, dominated by second, will be indistinguishable from an amp with the same THD but dominated by fifth?
You've got me buffaloed.
Not at all. I'm saying exactly what I wrote. And that is that those figures do not tell us which one sounds better. You might know which sounds better because you've listened to them. But measuring THD beforehand cannot predict by any stretch of the imagination which amp will sound better.
And verbatim : "For instance an amp with 5% THD can sound less distorted than an amp with 1% THD"
Well..I've gotta skedaddle out of here, because I've had enough of your tomfoolery and well..its late.
I have stayed away from these type of debates because they usually just cause controversy. Tubes VS SS, ears VS test equipment, SE VS P-P........which is best. Who defines best? My best or yours? Of the things that I just mentioned, I believe in all of them. Tubes, yes. SS, yes. Ears, yes, yes again, if your ears are not satisfied the rest doesn't matter! Test equipment, yes. Good test equipment is not needed to replicate a good amp design. It does come in handy to design them, or quantify the performance of an amp. Great measurements do not guarantee a great sound, although bad measurements do usually identify a bad sounding amp. High THD does not always correlate with bad sound. I have heard some amazing sound come out of an amp that measured 2% THD at 1 watt and climbed gently to 8% at 10 watts. The FFT revealed that the distortion was almost all 2nd harmonic up to 5 watts. Sometimes we get all caught up in the measurements and forget about the sound. I don't believe that the abillity to produce a full power square wave at 10 KHz means much to the sound. I have a large rack full of test equipment, but sound comes first.
There are obviously many different opinions here, some of them strong. This can lead to these debates, but it this is really a good thing. Why? Because it leads to some different thinking, which could lead to some new ideas. New ideas lead to inovation.
We must understand that some people do have strong opinions that may never change. This is OK, they are entitled to their opinion, and the right to share it. It does limit their options though. My father always believed that Volvo made the best cars. He was so strong in this belief that he never considered any other brand. This severely limited his choices when it was new car time. In his later years a smaller car may have been a better choice for him, but he wouldn't consider them. Many die hard tube heads will never stand for any "sand" in their amp. This just limits their choices.
In my case, I can't make up my mind what my best amp is, so I have a setup with 4 different amps, each can be switched in when needed. There is provisions for a 5th one, whichever amp I happen to be testing at the time. I have a P-P 300B amp, a small SE amp that uses 45's, 2A3's or 300b's. The 3rd amp is an 845SE and the 4th? An old SS amp made in Italy in 1969, brand name Voxson. Different amps for different moods, or music. Yes they DO sound different even at low volume. The differences become more obvious as the knob gets turned up.
As my tastes change, and my ears age, I may change amps. I am just starting to play with speaker building. I don't have room for 5 sets of speakers though.
Yeah, Pearl Jam, Nirvana, Soundgarden. The 300Beast (P-P) turned up to just below clipping! No? If you don't like MY choice in music, why would YOU let ME tell YOU what kind of amp to buy? Trust YOUR ears with your music through your speakers, then compare the measurements. To quote a popular TV show "Trust no one, the truth is out there" There are several popular truths. Which one do you want?
There are obviously many different opinions here, some of them strong. This can lead to these debates, but it this is really a good thing. Why? Because it leads to some different thinking, which could lead to some new ideas. New ideas lead to inovation.
We must understand that some people do have strong opinions that may never change. This is OK, they are entitled to their opinion, and the right to share it. It does limit their options though. My father always believed that Volvo made the best cars. He was so strong in this belief that he never considered any other brand. This severely limited his choices when it was new car time. In his later years a smaller car may have been a better choice for him, but he wouldn't consider them. Many die hard tube heads will never stand for any "sand" in their amp. This just limits their choices.
In my case, I can't make up my mind what my best amp is, so I have a setup with 4 different amps, each can be switched in when needed. There is provisions for a 5th one, whichever amp I happen to be testing at the time. I have a P-P 300B amp, a small SE amp that uses 45's, 2A3's or 300b's. The 3rd amp is an 845SE and the 4th? An old SS amp made in Italy in 1969, brand name Voxson. Different amps for different moods, or music. Yes they DO sound different even at low volume. The differences become more obvious as the knob gets turned up.
As my tastes change, and my ears age, I may change amps. I am just starting to play with speaker building. I don't have room for 5 sets of speakers though.
Yeah, Pearl Jam, Nirvana, Soundgarden. The 300Beast (P-P) turned up to just below clipping! No? If you don't like MY choice in music, why would YOU let ME tell YOU what kind of amp to buy? Trust YOUR ears with your music through your speakers, then compare the measurements. To quote a popular TV show "Trust no one, the truth is out there" There are several popular truths. Which one do you want?
richwalters said:
The most accurate.
My point was, while I acknowledge humans have an excellent auditory system, I seriously doubt it's good enough to detect events to the same accuracy electronics can. I didn't so much mean, a perfect amplifier; that's going into subjective, qualitative realms, of coarse.
The harmonic patterns have already been discussed and I don't think we need to carry on with the discussion with regards to the fact that a certain level of THD with a different HD pattern is not necessary going to sound better, or worse, than another. Pretty much everyone agrees, THD is not all of it.
This is where my point is rooted. I'm reasonably confident that the waveforms and FFT analysis you're seeing on your scope does actually have almost all of the information in it you need to tell a good distortion from a bad distortion. The problem is, how to interpret it? I'd go further than saying certain patterns of HD sound better than others and suggest that the relationship probably isn't linear. As THD, signal frequency, signal form, loudness level, and so on, change or increase, I wouldn't be suprised to see the HD pattern preference change at least slightly. This would start to make things very complicated to understand when all you're doing with your scope is blasting the amp with a one waveform, one frequency and watching for one snap shot of the distortion.
I'm thinking about this in relation to, how could you improve THD ratings? I thought, perhaps you could quote some of the harmonics in there as well, or provide a red, yellow, green system of THD, with red being huge amounts of high harmonics and green closer to the fundamental, as a very rough idea.
But you see the split between SS and valve HD patterns, which suggests that measurements this precise, whilst possible, are falling into the tolerance commonly known as, personal preference.
I was kind of suprised when I started reading these forums how little people seemed to care about where their distortion was coming from, and just quoted THD numbers a lot of the time, or a few harmonic peaks. I started looking at amps as a guitarist, where distortion is something you usually have loads of, compared to a Hi-Fi anyway. One thing that immediately became apparent to me, despite not knowing anywhere near as much as some of the guys on this board at the time, was that where electrons are coming from and what they're traveling through makes a big difference to the sound.
Rectifiers are an example. No one seems to really think about how they effect the sound beyond their noise input, concentrating more on the total THD of the amplification stages etc. Those who like valve rectifiers have a disturbing habbit of not admitting to the fact that a valve rectifier is just added supply resistance; signal compression. In guitar amps, rectifier type makes a big difference, perhaps because the amps are usually run at high distortions, close to their maximum power capacity. That takes the sound from the capacitors (They're more and more depleted with higher volumes) and towards the rectifier (The amp now more heavily relies on the rectifier's resistance to 'recharge' it's self, you're listening to this instead of the capacitor's resistance). Whereas Hi-Fi designers tend to use lower power stages and build supplies with higher current to capacitance ratios, so the rectifier tends to play less of a part in how those stages sound. That, however, is most definitly not to say that it doesn't play a part. I think it's that kind of assumption, that it just sounds better without considering why, is where this idea of 'immeasurable quality' comes from; too much emphasis on the valves and OPT and not enough on the entire system.
With regards to why some people prefer the valve effect, the added colouring, of coarse ignoring the smell, appearance, heritage of valves, status and just familiarity and expectations of what one will hear when turning on an amp they've owned for a while, I think a percentage of this need for colouration may be coming from something to do with counterbalancing any compression or homogenisation that's occured during recording.
An example, low quality MP3s are missing quite a lot of data, easily enough that the average guy on the street can hear it sounds 'bad' by comparison with a high data rate compression. Studios have a tendancy now to rape the music with compressors and harmonisers to get that glossy sound. By reproducing this with perfect fidelity, you just reproduce the lack of original data (The new streams put out by a harmonisers are just slightly copies of the original). I think that by adding some HD colouration from a valve, you can selectively mask over and 'trick' the brain into thinking it's listening to something closer to the original, kind of like aliasing but a bit more complex. The distortion counters the lack of life that's been created by the original signal compression. If you're going to add this distortion / colour, you'll usually want it closer to the fundamental such that it makes more musical sense. You're creating 'new data' in a kind of harmoniser mask, no linear behavior of that distortion pattern helps prevent the mask looking fake; listen to an Eminem track, all modern artists have this but he seems to have it the most, and you'll hear the layers and layers of harmonised junk they've added to make his voice sound 'phat'. Randomising those harmonising voices (By singing each of them independantly for example), rather than just pitch shifting them up and down a few cents with a processor, would help prevent that, but it takes time and megastars don't have any spare time.
I suspect that distortion, especially distortion that isn't perfectly linear, tricks the mind into subconsciously concentrating on the music more, making it feel more alive. The human brain has precisely zero tolerance for anything that's repetative or homogenous; we're preset to pay more attention to things that change or that are different, period! I do believe we're capable of hearing quite minute changes in the distortion spectrum changes, if only at a subconscious level. Such distortion changes then creates a subtexture that keeps the mind concentrating on it more closely. This is my explaination for why so many people say digital and SS have a tendancy to be tiring, not inspiring, lifeless or that they 'stop listening' mentally to the music. The added texture of a valve's distortion keeps you're mind switched on to it.
Obviously, there's lots of room for movement in those comments. None of it will be ideal. You'll want some of each; e.g. linear behaviour to reproduce what's on the medium, distortion to mask over what isn't, it's just that sometimes the distortion also masks too much of what is there.
Then you just have personal taste. Sometimes you want a brighter, clearer sound because you're feeling happy and jumpy. Other times you want it smooth and soft because you're tired and want to relax.
In conclusion, I think that the tolerances with which we can measure these parameters is high enough. Our problem is that we don't consider a wide enough bad of interactions and, possibly more importantly, that our own person tolerances on what we want from something are larger than those of the test equipment. People can have a general preference for a particular sound, but I know from personal experience that some days I get up in a different mood and want to hear something else.
Generalisations...
My ultimate conclusion? If you want to know why sometimes we like certain spectra and sometimes we don't, you'll first need to put down your scope and take up neurological, psychological and philosophical science. Just staring into the CRT glow for a pattern will never tell you why.
tubelab.com said:
Simple, but 100% correct. And something far too many seem to forget.
Do you have a Delorean?
Another aspect to this discussion is that of the speakers we use. Although I haven't had the experience of high end commercial systems that some of you have, I have listened to a lot of studio and PA speakers, and one thing I notice again and again is how a system can sound completely rubbish, but if you turn it up, all of a sudden the sound seems to gel and all the things that were wrong suddenly become so right. This is especially noticable on the B&W Nautilus series. At low to normal domestic levels, they just sound rough and disjointed, and frankly, not worth the money, yet turn them up to studio monitoring levels and they sound superb. Is this true to increasing distortion of the drivers? I certainly don't know, yet I'm sure many of you are familiar with the phenomenon.
SY said:
Earl Geddes (gedlee.com?) has been doing these experiments. His experiements (pretty rigorous knowing him) have shown no correlation of THD to sonic merit. He has/is developing a metirc wich does.
Expect an AES paper. I wouldn't be surprised, given Earl's influence in the AES, that if his metric holds up it could well become an AES standard measure.
dave
dfdye said:
MDF is anyways... it is used first off because it is cheap and is easy to build with. secondly because everyone seems to use it, so it is expected (ie it is the same reason that MS Word -- a 3rd rate word processor -- or AutoCad have become the defacto standards -- not because they are good, but because that is "what everyone else uses")
dave
dfdye said:
And why is that... as Frank implied, the most linear amplifying device invented my man to date is the triode, so it has a natural advantage when it comes to accurate. Furthermore... how do you decide whether an amplifier is accurate?
dave
Is it just that? I am convinced..although I have not tried it..that it will still sound different than ss with a same value resistor.
Exactly. How does one interpret it. Later on in your post you say that tubes sound coloured because of distortion. To me tubes sound far more uncoloured than most SS equipment. I suspect therefore that you interpret distortion "wrongly" as well. Because to the human ear and brain the type of distortion I suppose you are referring is less audible and therefore less significant. Without me wanting to put words in your mouth, that is what you appear to be saying.
pinkmouse said:
Good point... an evaluation of an amplifier independent of the speaker it is driving leaves a whole lot out of the picture.
dave
Don't hold your breath. THD is the only proof that tubes are inferior. There is no way they are going to get rid of the "evidence".
- Status
- Not open for further replies.