I would add 4th to that, which is the room acoustics. Listeners will hear difference even from the same source when they move their ear position slightly. I often read posts describing how the listener switched the components (amps, cables, disc players ...etc.), he got up from the chair, walked up to the component, switched, came back and sat on the chair. When I tell them that their ear position has changed from previous spot, they get all defensive about their own (humanly) physical ability. 🙄 Not to mention their faded aural memory...
did you miss a negative in there?
because as I read it is completely wrong - no loudspeaker speaker system reproduces audibly indistinguishably the sound field coming from live groups in style appropriate performance venues at the listening position
we have extensive threads on loudspeaker room interactions - arguments over what is desired, if impossible in current few channel commercial music reproduction
but even a 18" woofer in a 400 sq ft x 10' ceiling domestic room isn't the Timpani at Symphony Hall
the test of "realism" that may make some sort of sense would be to take a live mic feed from the next room, then ask if differences between amplifiers are audible driving the speakers in the room with you
but mics, amplifiers and loudspeakers simply aren't literaly "realistic" in today's technology - anyone can tell there are audible differences between being in the room with the live performers vs any current commercial musical recording in the general case
Last edited:
Exactly. Yet some would claim that this perfect system is just over the horizon waiting to be discovered.
Shoog
Shoog
Why should they not claim that? If both are inaudible, or at least benign, what difference does it make?
DF96 used the word "preferred." If it's inaudible, it can't be preferred over no distortion by definition. If it's audible, no matter how "benign," it's not hifi by definition.
That's one explanation, and may be DF96's explanation. But I read a lot of posts on this forum and others that claim tubes are inferior because of the high measured distortion. That high measured distortion may not be audible.
We just want to be careful that we don't assume people prefer a certain measured distortion, when they can't even hear it!
Yes it can. All you have to do is look at distortion specs and you can easily prefer some device over another. You don't have to hear it. Gold plating and pretty lights can have the same effect. 😉
We just want to be careful that we don't assume people prefer a certain measured distortion, when they can't even hear it!
Psychologically, it has been stated that a little second harmonic "sweetens" the sound (whereas third tends to sharpen the sound).
If one had a harsh sounding recording, then arguably, although not scientifically a correct reproduction, the small amount of second harmonic would make a flawed recording more listenable...
Is an amplifier that is pretty transparent to good recordings, but slightly flattering to poor recordings an inherantly flawed system? And is it non hifi?
Wrong if you mean the audibility. Try some experiment yourself if you've never been fooled by recorded sound. Big part of it has to do with sound waves and room interaction.
yes there have been a few carefully crafted "tests" with speakers in the same room, near same placement as the performers/instruments
but for most of us the most relevant, my "general case", is home small domestic living room reproduction of a recording of a live performance in vastly differing size, acoustic properties, venue with music containing varying frequency content both exciting room's modal frequencies and givng strong specular reflections
then just turning your head samples quite audibly different sound field due to speaker directivity and small room boundaries interactions compared to a listening position in the live performance venue
"being fooled" is infrequenct in my experience and I expect not robust to repeat listening
but for most of us the most relevant, my "general case", is home small domestic living room reproduction of a recording of a live performance in vastly differing size, acoustic properties, venue with music containing varying frequency content both exciting room's modal frequencies and givng strong specular reflections
then just turning your head samples quite audibly different sound field due to speaker directivity and small room boundaries interactions compared to a listening position in the live performance venue
"being fooled" is infrequenct in my experience and I expect not robust to repeat listening
Last edited:
Nothing wrong with saying that amp "x" is measureably inferior to amp "y" if it is so.
That's called marketing.
Whether they can hear it or not would have to be tested. When I say tested, I don't mean some casual subjective listening.
As far as I can tell ... everything 'has a sound'. It is all but impossible for this not to be the case.
The microphones are composed of gossamer films that vibrate (one hopes) faithfully in accordance to the thousands of infinitesimal pressure changes that constitute sound. Yet, they have mass, and thus they are as much displaced as accelerated. This odd 2nd order behavior gives all microphones a remarkably different "sound", all other factors being equal.
Microphone signals - especially for quality ensemble recordings - are very low indeed. They need amplification of hundreds to thousands of times in amplitude. The nature of ALL amplifiers is that they're only modestly linear near that zero-amplitude point. Moreover, the 2nd order behavior of microphone diaphragm mass becomes greater (proportion of the signal) as the sound gets more quiet overall. Hence, what is being amplified changes dynamically per the coloration due to the microphone, and the coloration due to the zero-crossing characteristics of the mic preamps used. Its unavoidable;
Then you have all the work done by the recording engineers and artists to shape what is recorded in the raw, to what they want to represent on the final recording. For modern genre music, the modifications are so pervasive and massive, that whatever is on the track is literally only a harmonic approximation of whatever the performers played and were first recorded. Artificial echoes are added, pitch correction done, timing correction done (by repeating microslices of the recording), whole drum kits are substituted for the original drummer's kit. Things are compressed, companded, expanded, dynamics are squashed, harmonics are enhanced. Wow. But anyway, you get a recording. How on earth is one supposed to know it was INTENDED to sound?
For classical and "live" recorded music, its nearly as badly misrepresented. Its not like those overhead and surface-effect microphones are mixed together, and recorded straight to CD tracks, you know. No ... at best, they're very carefully preamplified, then transferred to a "multitrack" medium, whether in analog to tape, or digitally to a hard disk. From there - again - the recording engineers ply their trade, setting mix levels, adjusting compandors and compressors. They'll play up far-field mics with time-advances (making up for the distance, digitally), but with attack-gates and other things to make the enhanced sound natural. By the time its all mixed down ... does that recorded waveform really faithfully represent the sound of the original performance(s)? Nah... not by a long shot.
Then though, never mind the fact that it's all quite far-removed from the original "live" performance. It is made nice, and recorded, and I have a copy. Now, I play it ... and sure enough, the various nonlinearities of my system kick in, and further color the picture in their own (often irreducible) way. Devices such as digital-to-analog decoders might have "a sound", but really they tend not to. Or, when they do, it is because - again - they're introducing some kind(s) of signal transform functional "corrections" to fix something. Something perceived, something real, something mythical, something popular.
Plenty of words go into describing the exquisite engineering and exacting manufacture of wickedly cool little pieces of the chain - needles, D/A digitization techniques, aligned cryogenically treated copper crystals in cables, massive power supply isolation filtering, and more. The hand-made discrete op-amps, or the purely "Class A" (sounds good) amplification. Whatever. Truth is, like DF96 often says, "if there IS a noticeable sound difference, that means the device is acting as an effects box, too". Yep, it is true. Maybe some of the stuff is worthwhile, maybe not. Maybe some of the "Tube Sound" is undoing the compressive effect of real-world microphones and preamplifier compression. Maybe not.
Yet, I'm not done! Because following analog waveform reconstruction, the signal now passes through perhaps more preamplification (exerting its inevitable slightly non-linear first and second order transfer functions onto the signal), then the amplifier; Except for the physics of the microphone, and the correspondingly first, second and third order behavior of speaker cones in resonant boxes, the amplifier has its own effect, and the effect is very, very real.
A lot of "effect" of that amplifier can't be measured as "faithfulness to the incoming signal", because, well, that's what both linearity and feedback conspire to make so. Instead, it is the case that the very last output stage can have its own remarkable interaction with both the forward and reactance-caused reverse EMF of the speakers and speaker cables themselves.
On the one extreme, an amplifier with no "damping factor" just presents voltages to the output pins, and makes no adjustment for counter-force EMF coming back from the speaker. On the other, the amplifier is tightly trying to CONTROL that voltage to very tight coherence to the amplified input signal. Between, amplifiers do a bit of both. And "damping factor" is just one factor. Another is "stored energy", which is the effect that tube-amplifiers have that solid state simply do not (at least they don't without an output transformer!) ... The inductor stores substantial energy, which the output tubes are constantly modulating as a function of the voltage of the signal.
However, the actual amperage or current flow is quite removed (waveform wise) from the original signal! There are the speaker's dynamics, the stored energy of the inductor, the tightness of its magnetic couplings to the output, and the step-up or reverse-emf of the linear-motor aspect of practical loudspeakers. Oh... then there are those pesky crossovers, in the speaker cabinets.
What I'm trying to say here - is that it is REALLY straight forward to spec the steady-state behavior of each element, and say that it is 0.013% varying from linear (as an example). But by the time the real-world, first-, second-, and third-order differential response of the whole system is taken into play, the signal can be quite different - at the speaker cone - than envisioned by the data encoded in that primary recording.
And then, to add insult to injury ... the damned speaker cones also have masses, and thus have all sorts of resonances, eigenvalues and nodes which they compress and rarefact the air, every moment during the recording. In essence, what we actually hear is a "reasonable approximation" of the spectral dynamics of the recording, which is itself a "reasonable artistic interpretation" of the spectral dynamics of the original performances.
Luckily, we're all really soooo poor at quantitatively figuring these nuances out, that we instead invent entire treatises that devote themselves to describing how using hemp fibers and liquified insect glue ... create the best sound ever.
GoatGuy
The microphones are composed of gossamer films that vibrate (one hopes) faithfully in accordance to the thousands of infinitesimal pressure changes that constitute sound. Yet, they have mass, and thus they are as much displaced as accelerated. This odd 2nd order behavior gives all microphones a remarkably different "sound", all other factors being equal.
Microphone signals - especially for quality ensemble recordings - are very low indeed. They need amplification of hundreds to thousands of times in amplitude. The nature of ALL amplifiers is that they're only modestly linear near that zero-amplitude point. Moreover, the 2nd order behavior of microphone diaphragm mass becomes greater (proportion of the signal) as the sound gets more quiet overall. Hence, what is being amplified changes dynamically per the coloration due to the microphone, and the coloration due to the zero-crossing characteristics of the mic preamps used. Its unavoidable;
Then you have all the work done by the recording engineers and artists to shape what is recorded in the raw, to what they want to represent on the final recording. For modern genre music, the modifications are so pervasive and massive, that whatever is on the track is literally only a harmonic approximation of whatever the performers played and were first recorded. Artificial echoes are added, pitch correction done, timing correction done (by repeating microslices of the recording), whole drum kits are substituted for the original drummer's kit. Things are compressed, companded, expanded, dynamics are squashed, harmonics are enhanced. Wow. But anyway, you get a recording. How on earth is one supposed to know it was INTENDED to sound?
For classical and "live" recorded music, its nearly as badly misrepresented. Its not like those overhead and surface-effect microphones are mixed together, and recorded straight to CD tracks, you know. No ... at best, they're very carefully preamplified, then transferred to a "multitrack" medium, whether in analog to tape, or digitally to a hard disk. From there - again - the recording engineers ply their trade, setting mix levels, adjusting compandors and compressors. They'll play up far-field mics with time-advances (making up for the distance, digitally), but with attack-gates and other things to make the enhanced sound natural. By the time its all mixed down ... does that recorded waveform really faithfully represent the sound of the original performance(s)? Nah... not by a long shot.
Then though, never mind the fact that it's all quite far-removed from the original "live" performance. It is made nice, and recorded, and I have a copy. Now, I play it ... and sure enough, the various nonlinearities of my system kick in, and further color the picture in their own (often irreducible) way. Devices such as digital-to-analog decoders might have "a sound", but really they tend not to. Or, when they do, it is because - again - they're introducing some kind(s) of signal transform functional "corrections" to fix something. Something perceived, something real, something mythical, something popular.
Plenty of words go into describing the exquisite engineering and exacting manufacture of wickedly cool little pieces of the chain - needles, D/A digitization techniques, aligned cryogenically treated copper crystals in cables, massive power supply isolation filtering, and more. The hand-made discrete op-amps, or the purely "Class A" (sounds good) amplification. Whatever. Truth is, like DF96 often says, "if there IS a noticeable sound difference, that means the device is acting as an effects box, too". Yep, it is true. Maybe some of the stuff is worthwhile, maybe not. Maybe some of the "Tube Sound" is undoing the compressive effect of real-world microphones and preamplifier compression. Maybe not.
Yet, I'm not done! Because following analog waveform reconstruction, the signal now passes through perhaps more preamplification (exerting its inevitable slightly non-linear first and second order transfer functions onto the signal), then the amplifier; Except for the physics of the microphone, and the correspondingly first, second and third order behavior of speaker cones in resonant boxes, the amplifier has its own effect, and the effect is very, very real.
A lot of "effect" of that amplifier can't be measured as "faithfulness to the incoming signal", because, well, that's what both linearity and feedback conspire to make so. Instead, it is the case that the very last output stage can have its own remarkable interaction with both the forward and reactance-caused reverse EMF of the speakers and speaker cables themselves.
On the one extreme, an amplifier with no "damping factor" just presents voltages to the output pins, and makes no adjustment for counter-force EMF coming back from the speaker. On the other, the amplifier is tightly trying to CONTROL that voltage to very tight coherence to the amplified input signal. Between, amplifiers do a bit of both. And "damping factor" is just one factor. Another is "stored energy", which is the effect that tube-amplifiers have that solid state simply do not (at least they don't without an output transformer!) ... The inductor stores substantial energy, which the output tubes are constantly modulating as a function of the voltage of the signal.
However, the actual amperage or current flow is quite removed (waveform wise) from the original signal! There are the speaker's dynamics, the stored energy of the inductor, the tightness of its magnetic couplings to the output, and the step-up or reverse-emf of the linear-motor aspect of practical loudspeakers. Oh... then there are those pesky crossovers, in the speaker cabinets.
What I'm trying to say here - is that it is REALLY straight forward to spec the steady-state behavior of each element, and say that it is 0.013% varying from linear (as an example). But by the time the real-world, first-, second-, and third-order differential response of the whole system is taken into play, the signal can be quite different - at the speaker cone - than envisioned by the data encoded in that primary recording.
And then, to add insult to injury ... the damned speaker cones also have masses, and thus have all sorts of resonances, eigenvalues and nodes which they compress and rarefact the air, every moment during the recording. In essence, what we actually hear is a "reasonable approximation" of the spectral dynamics of the recording, which is itself a "reasonable artistic interpretation" of the spectral dynamics of the original performances.
Luckily, we're all really soooo poor at quantitatively figuring these nuances out, that we instead invent entire treatises that devote themselves to describing how using hemp fibers and liquified insect glue ... create the best sound ever.
GoatGuy
Yes, that's the starting point. Is the reproduction of a solo piano, spanish guitar, violin convincing? Going by the standard at the recent audio show, asked "Are we there yet??" - the answer is no! At least, in the sense that you can't buy a few pieces of kit, plug them together, and out pops 'authentic' sound - 99.99% of the time, 😛.
However, it mostly certainly can be made to happen, with some effort. And, when it gets the mimicing of single, acoustic instruments right - really, really right - then everything else falls into line too. Including, yes, even really messed up, highly artificial recordings. How can one say that? Because, at one level they become easy to listen to, all the threads of sound within become straightforward to listen to, it all makes acoustic sense; and, at another level, frequently even the most manufactured recordings have snippets or clips of sound which are real world noises added in - when these sound 'right' it also confirms the correctness of the reproduction ...
Some would say it's flawed because it's not accurate. In the strictest sense, non-hifi. But there is also the issue of measurable distortion that isn't audible. So while some amp can be demonstrated by measurement not to be as accurate as a lower distortion amp, it may not matter. The higher distortion may simply be inaudible. You have to be careful not to be swayed by what isn't audible.
As to whether and amp with a little euphonic distortion is inherently flawed, that's another subject. 🙂
A key word here is 'convincing' - systems, no matter how theoretically "deficient", are capable of creating sound which can consistently fool one - as in, if you're at the other end of the house, or outside in the garden, they tick all the boxes. At the best level, such a system will maintain the illusion even when you're standing literally only inches away from the speakers ... and, do it with a really, really nasty recording ...
Something working like that is good enough for me ... 🙂
There have been a fair few SS amps that go up to 50Khz or more for no good reason. These do cause harsh sound when the tweeter does not have a low pass filter of say 22Khz between it and the amplifier. If the input to the tweeter contains any artifact in the signal much above the figure in the data sheet it will drive the cone in the tweeter faster than the outer edge can travel. It will produce ripples that run through the cone that will mix with the wanted signal and lower harmonics will be produced that go on to be heard as harshness or distortion well below the rated power of the amplifier. By there nature tube amplifiers have iron transformers and the heavy cores of these will kill the ultrasonic parts of the signal and take the harshness out before the tweeter ever gets to receive the signal.
Pano, unless a person at some stage has experienced sound that's clean enough for the 'effect' to happen - and more particularly, had it pointed out to them what's going on - they may always have trouble with the concept. I'm sure that hearing good sound would go right over the heads of many people ... because it sounds, well, normal! Especially at lower volumes, it doesn't call attention to itself, and then you raise the volume, and then you raise it again - and there's no flashing lights or fireworks going off, acoustically!! So, what's the fuss, some people might say ...
I note that group muso's don't tune into clean sound, it has to start to compress, to subtlely distort - then they say, "Gee, now that's good - it sounds loud!", 😀 ...
I note that group muso's don't tune into clean sound, it has to start to compress, to subtlely distort - then they say, "Gee, now that's good - it sounds loud!", 😀 ...
- Status
- Not open for further replies.