I think we can use measurements to judge whether one electronic component is more faithfull in amplifying or reproducing sound than another. I agree that we are not sure that we catch all that is important for listening but we can measure the difference between what goes in an amp and what comes out to such minute differences, with signals that are even more complex than music, that it is unlikely (but not proven) that we miss something that is important for reproduction.
It is possible that a listening panel prefers an amp that by our imstruments would be the less faithfull. You can't argue with preference. But that doesn't invalidate the measurements.
Edit: We don't really know what shapes our preference. Often, people prefer a certain amp because they feel it gives them the best audible illusion of the real thing. That's not something we can measure I think, although some designers have found the trick; how to 'shape' the amplifier transfer function so that it sounds great to many ears. But that then deliberately involves deviating from the most faithfull, the most transparent function. And they do it by looking at the measurements. I wish I knew what they were looking at 😉.
jan didden
Last edited:
Yes there seems to be acoustical pre-amplifiers resident in the outer cochlear hair cells.
The prevailing hypothesis is that this amplification stems from somatic electromotility of the outer hair cells attributable to the motor protein prestin. Thus outer hair cells contract and elongate in synchrony with the sound-evoked receptor potential.
Calling this a passive receptor is mistaken - it's a bit of a stretch to try to say that the active part only goes on between the ears! The ear is patently not a passive transducer & trying to defend that statement is futile - maybe it would be worthwhile getting some more up-to-date information than you have cited I should use - hearing-aid designers 😉
Well, have it your way, but believe me; hearing aid designers are lightyears ahead in this. It's a multi-billion $ market with almost exponential growth - the money they can throw to research there is almost limitless.
jan didden
Again, you're using your model of the ear as a passive transducer & you fail to grasp the distinct possibility that it is an active one which would then render your measurement system as misguided & unless you know the performance & characteristics of this active system, rendered useless for the task at hand.
Last edited:
What do you mean, have it my way - it's either correct or not that the ear is a passive transducer. Are you saying that this is what hearing aid designers & you are saying?
Can you cite a modern paper from these hearing aid designers stating this?
Last edited:
I think there is a fundamental difference between MyFi and HiFi, but I do agree with the point that e.g. some SETs produce less than average measurements but an above average sound experience. Problem here is that the measurement set applied to amplifiers, notably distortion, does not correspond one on one with subjective quality. I am the first to admit that we must still be missing out on something to measure. But I remain confident that one day we will find out.
A well designed horn can be more linear than other ways of mounting speakers, for the simple reason that a given SPL can be obtained with less cone movement. This lowers distortion and other non-linearities, so it actually proves my point. Also because of high directivity, horns may provide more direct sound over indirect sound as compared to other enclosure types, leading to less artifacts caused by the listening environment such as comb filtering.
As I mentioned, we do already have a fairly detailed understanding of the mechanism of hearing, but much of it is not directly relevant for the goal of faithful reproduction of sound, because whatever happens, it happens to both live and reproduced sound. From industry research I know that for loudspeakers, the perceived quality is directly correlated to the straightness of the frequency respons, so this creates a measurable design goal. From other research we know that the hearing system is, to a high degree, unphased by phase shift. Therefore, a linear phase response may be irrelevant, and therefore a 24 dB cross-over does not need to stand in the way of a reliable reproduction of music, in spite of the 360 degree phase shift that occurs.
Now back to the topic of this thread: "how better is a turntable compared to a CD". I don't exclude the possibility that at some stage, we will develop new ways of measuring, which will demonstrate that vinyl has a hitherto unmeasurable advantage over metalized polycarbonate, but I doubt it. CD has better stats for the moment, and to my ears sounds better too (I don't listen under laboratory conditions in a dust free environment).
As I said, maybe it's the model of the ear that seems to being used - a passive transducer. Maybe this could explain why the one to one correlation of measurement with subjective quality is difficult? It always seems that the brain is cited as the reason for something which sounds better - maybe that's only part of the picture, maybe the ear plays a much bigger role than the passive role it seems to have here?
I disagree based on the responses from here 🙂
Is it just flailing about? Citing that live & replayed sound is listened to by the same ears is immaterial - we don't fully know how those ears work yet. If you believe that your replay system can exactly reproduce the soundfield of the original performance then you would have a point but this is impossible for the reasons stated already. In the absence of that we need to know what to focus on in the sonic reproduction, no?
Or maybe new models that focus our measurements more accurately? But if you want to ignore this possibility & just get on with arguing how superior CD measures then go ahead. Pity our ears aren't linear systems 🙂
Last edited:
I do have a problem with the position that states "only a detailed understanding of the functioning of both the ear & hearing will allow us to say what are the important measurements & what can be ignored." Science and the progress of human knowledge in general doesn't work that way. It's these little steps that lead to a better understanding, and big steps.
With that detailed understanding of hearing, the sounds can then all be computer synthesized, and that will be the Hot 100 in 2111. Musicians will be obsolete.
And some "hold-outs" will still be spinning vinyl thru tube amps.
I've rethought something I posted earlier about common standards. "More lifelike" specifically. What does that mean? The violin sounding as it was played in the studio when recorded? Sounding like it's being played in the room where it's being reproduced? Or Carnegie Hall? Seems to me there's not just one valid answer there.
So unless there's some "meeting of standards" with this issue or that, there can never be a consensus. Which isn't necessarily a bad thing.
With that detailed understanding of hearing, the sounds can then all be computer synthesized, and that will be the Hot 100 in 2111. Musicians will be obsolete.
And some "hold-outs" will still be spinning vinyl thru tube amps.
I've rethought something I posted earlier about common standards. "More lifelike" specifically. What does that mean? The violin sounding as it was played in the studio when recorded? Sounding like it's being played in the room where it's being reproduced? Or Carnegie Hall? Seems to me there's not just one valid answer there.
So unless there's some "meeting of standards" with this issue or that, there can never be a consensus. Which isn't necessarily a bad thing.
Yes, that's something of a red herring IF one's goal is to accurately reproduce the signals coming from the microphones (or master tape). To date, there is NO experimental evidence using listening tests that the D/A and A/D conversions are audible to human ears. One does not need to know how the ear-brain works to make that experimental determination.
Now, this has nothing to do with perceived realism. The major compromise of reducing a 3D sound field to two 1D signals is overwhelmingly the limitation. That aside, mikes could be lousy, the placement could be poor, the EQ and compression applied at the studio/mastering end can be damaging... if someone likes the output of an effects box better than the direct signal from the mikes or master tape, one can't argue that. The research problem is then not "accurate reproduction," but rather "what processing effect do I like better." And of course, "why in the world do I need an effects box, can't the producers do a better job of getting rid of peaky mikes, overly enthusiastic compression, and EQ to compensate for their lousy monitors?"
Are you contending that all D/A & A/D conversion sounds the same or just that there have been no experimental listening tests done? Is D/A & A/D conversion the same as when it was first conceived & implemented in audio?
My point is that the ear is not a linear passive transducer but rather seems to be an active non-linear device. This is the target device for your reproduced audio signals. By not understanding the non-linearity you seem to be ignoring a fundamental issue & focus on issues like the measured linearity of the audio signal. This could well be a mistake.
Of course one doesn't need to know psychoacoustics to do tests but again are you not just flailing around in the dark, following the well trodden path hoping for some revelation. I already stated the analogy that it would seem in audio we are at the stage that art was before perspective was discovered. Of course things in the real world don't shrink when they are further away but that is one of the many ways that the eye perceives distance. If we were fixated on linearity in art we would never have realism in art, just naive art, we would never have Cinema or TV or the many ways the eye is fooled into a seeing something that approaches realism visually. Your statement about the problem of 3D to 1D sound-fields is exactly what I'm saying.
You seem to have mixed this up with the idea of "liking a sound" as opposed to what I'm talking about which is being aware of the important factors that the ear actively handles & the brain finds realistic in audio signals & paying attention to getting this correct in the reproduced audio signals.
So what I'm trying to introduce is step back from this focus on linearity & measurement & dismissal of everything else. Other factors & non-linearity could be much more cogent to the hearing mechanism - it's worth considering that this might be why we cannot draw a one to one correlation between measurements & what we hear
My point is that the ear is not a linear passive transducer but rather seems to be an active non-linear device. This is the target device for your reproduced audio signals. By not understanding the non-linearity you seem to be ignoring a fundamental issue & focus on issues like the measured linearity of the audio signal. This could well be a mistake.
Of course one doesn't need to know psychoacoustics to do tests but again are you not just flailing around in the dark, following the well trodden path hoping for some revelation. I already stated the analogy that it would seem in audio we are at the stage that art was before perspective was discovered. Of course things in the real world don't shrink when they are further away but that is one of the many ways that the eye perceives distance. If we were fixated on linearity in art we would never have realism in art, just naive art, we would never have Cinema or TV or the many ways the eye is fooled into a seeing something that approaches realism visually. Your statement about the problem of 3D to 1D sound-fields is exactly what I'm saying.
You seem to have mixed this up with the idea of "liking a sound" as opposed to what I'm talking about which is being aware of the important factors that the ear actively handles & the brain finds realistic in audio signals & paying attention to getting this correct in the reproduced audio signals.
So what I'm trying to introduce is step back from this focus on linearity & measurement & dismissal of everything else. Other factors & non-linearity could be much more cogent to the hearing mechanism - it's worth considering that this might be why we cannot draw a one to one correlation between measurements & what we hear
Last edited:
No. For example, an 8-bit conversion will be noisier than a 16 bit, audibly so. Audibly distorted as well, if the signal levels are low and there's no dither.
It's a pretty simple task isn't it? And the results are measurable not mysterious.
There may be a bit of fiddling around the margins with rounding, dither and so on, and you may get in a fret about how many parts per trillion of jitter you allow, but most of it's academic. I have a 22 year old CD player (with 8x oversampling DAC!) and I can't hear any difference between it and my supposedly superior Arcam machine (I don't know what DAC it's got). I think I once perceived 24 bit from my M Audio PC card as superior to 16 bit, but I probably imagined it.
Let's face it, if you're in middle age you can't hear the top end anyway, and you though you all talk as though the differences are like night and day, you all know that in a blind test you wouldn't really be able to tell the difference, because there is no audible difference. Luckily for you, you'll never have to put it to the test.
OK, CopperTop, so you are saying that D/A & A/D conversion sounds the same now as it did when the CD was first conceived - perfect sound forever, eh - brilliant, I need to get me one of them there new fangled machines - what'ya call them a Arcane CeeDee player was it or a SoMe CeeDee player? 🙂 Make way, I'm joining the linear brigade - they hear right & are so much smarter than the rest of us & know that one CeeDee contarption sounds the same as the next - everything else is just Booga-Boo
Last edited:
No. For example, if I set the levels such that I use only the four lowest bits and no dither on one recording, and the four lowest bits and dither on a second recording, they can be distinguished by ear.
I only said that my old CD *player* sounded the same as a newer, superior one, and I accepted that there had been some fiddling around the margins with dither and so on. But it is marginal.
Where does this mystique of A/D and D/A come from? It genuinely feels as though some of you lot think that two ADCs which generated the same digits at their outputs would somehow sound different. Someone here commented the other day that the trouble with cheap DACs is that they are encapsulated in plastic.
Get a grip! We're talking about a system for sampling a wiggly line, nothing else. If there's a discrepancy, then two ADCs will give give you different numbers which you can then analyse to your heart's content. If the input signal is a known signal, you can see what the error is. DACs are just the other way round, and you can measure the error they generate. It's all backed up with mathematical theory, so no mystery.
Edit: Just why would you think that my 22 year old CD player should sound different from a modern 'state of the art' one? Can you suggest what you would be left with if you could subtract their two audio outputs? Would you be able to hear it? Or would its level be one percent of the Brownian motion of air molecules on your ear drums?
Last edited:
- Status
- Not open for further replies.