What uncontroversial audible differences cannot be measured?

[mansr]

17 minutes ago, Jud said:

Approaching this from a practical angle:

 

You are at a live or studio performance.  The microphones respond to the compressions and rarefactions.  They produce (edit: electrical) waves of higher amplitude corresponding to the compressions, and lower amplitude corresponding to the rarefactions.

Not quite so. Compressions produce a signal of one polarity and rarefactions the negative. Their magnitudes determine the amplitude of the signal.

Quote

From there you ought to be able to measure, in fact predict, any processing that alters phase.

Again, phase relative to what?

 

I thought the "absolute phase" thing was about a recording/playback system such that a compression of the air results in the speaker drivers moving forward. (In other words, the correct term is simply "polarity.") Some say reversing the polarity of the speakers produces an audible change with asymmetric waveforms (with plain sine waves it obviously can't make any difference). I haven't done any tests to determine whether or not I can hear a difference. Even if I could, I have no intention of rewiring my system per track.

[AJ Soundfield]

My favorite Absolute Polarity aka "Witch Effects" article, by Dr Shanefield 

http://www.enjoythemusic.com/magazine/bas/1208/

 

[Jud]

Perhaps interesting, from a paper on using evolutionary algorithms to design a simple multiplier circuit in an FPGA:

 

Quote

Firstly, note that all these designs require only 7 cells, hence they are more efficient than the designs
produced by human designers! It is interesting to note that in all circuits the multiplier has been
synthesised as two separate sub-circuits. In multiplier A one sub-circuit produces P1 and P3 and the
other produces P2 and P4. While in circuit B, one sub-circuit produces P2 only and the other produces
P1, P3 and P4. Incidentally note also that the sub-circuit in B which produces P2 is equivalent to the
circuit in the conventional cellular multiplier in figure 11, as the two inversions on the AND gate
outputs cancel out at the ex-OR gate. In multiplier C, one sub-circuit produces P2 and P3 while the
other sub-circuit produces P1 and P4. One is tempted to ask: How many intrinsically different 7-cell
multiplier circuits are there? Also: Is 7 cells the theoretical minimum? This division of the multiplier
into sub-circuits is interesting and quite subtle. Take for example the way in which P3 is calculated in
circuit A. What have A2,B2 and A1,B1 got to do with P3? This is highly non-intuitive. Also note that
P2 is calculated within circuit A in an identical way to the conventional design but then P4 uses this
circuit in an extraordinary way! It is clear that evolution is coming up with very novel ways of building
these circuits....

 

There have also been efforts to produce analog hardware circuits using evolutionary algorithms, which have also resulted in some circuits that outperformed human designs, occasionally in ways quite difficult for humans to duplicate (because, for example, the circuits relied on aspects of individual component operation that hadn't been seen by human designers as notable, so they couldn't be adapted to use other components).

 

While this is admittedly not relevant to unmeasurable quantities, it seems to me it does speak to our ability (or lack of same) to evaluate what measurements are important.

[Jud]

3 minutes ago, mansr said:

Not quite so. Compressions produce a signal of one polarity and rarefactions the negative. Their magnitudes determine the amplitude of the signal.

Again, phase relative to what?

 

I thought the "absolute phase" thing was about a recording/playback system such that a compression of the air results in the speaker drivers moving forward. (In other words, the correct term is simply "polarity.") Some say reversing the polarity of the speakers produces an audible change with asymmetric waveforms (with plain sine waves it obviously can't make any difference). I haven't done any tests to determine whether or not I can hear a difference. Even if I could, I have no intention of rewiring my system per track.

 

Not necessary.  Some software players have switches for this. 

[mansr]

3 minutes ago, Jud said:

Not necessary.  Some software players have switches for this. 

True. I could also simply invert the files. Nevertheless, it's not something I fret over, and I have no intention of starting.

[sdolezalek]

10 hours ago, semente said:

 

'Timbre accuracy' is a property of the recording.

Harmonics, decay, spatial cues, detail, all that.

If the system is "transparent" (faithful, accurate to the recorded signal) then it will reproduce the properties of the recording.

Distortions can be added to the signal (frequency response aberrations, harmonic distortion, phase distortion, noise, off-axis frequency aberrations, cone and dome resonances, etc. to "increment" aspects like "spaciousness" (decay/reverb), "air", "bloom", "speed", "presence", "detail", etc. 

 

Audio review (particularly speakers) generally highlight these aspects as if they were qualities when in fact all the equipment is doing is adding a form of distortion...

It makes no sense to me.

 

Ah, but then what is the measurement for transparency?  Obviously there is not a single measurement that captures all of the above? So what is the series of measurements that add up to transparency and in what order should we consider them?  I do not doubt that speakers and room response have a much bigger effect on these qualities than do DACS, amplifiers, cables, etc.  but I believe the original post was not limited to only measuring those items.  

We once tried to measure where in a system the most aggregious distortions were and it immediatewly became clear that speaker distortions outdistance all others in a high quality system.  If we think the transient response of an amplifier, DAC or cable is poor and contains pre or post ringing, you should see the behaviour of the average paper conse.  And, yes, I agree that you can "distort" the input signal to try to affects or lessen those distortions (room response correction being only one).  But, I believe that most of those  "corrections" themselves involve some tradeoffs and that different earrs evaluate those tradeoffs differently.  

[Ralf11]

2 hours ago, mansr said:

Where did I say SPICE is perfectly accurate? If electronic components were as complex as some software, defects in amps wouldn't be talked about in terms of noise and distortion. You'd have resistors suddenly start acting as diodes if you played a particular sequence of music tracks for two hours or other utterly bizarre things.

 

He was just making a strawman argument to avoid responding to what you actually said.  Don't fall for it.

[mansr]

1 hour ago, mansr said:

The one I'm thinking of went haywire if presented a steady-state signal (such as a test tone) for more than a few seconds. This made the usual measurements almost impossible to perform. One might be tempted to term the feature a "defeat device."

Found it: https://www.stereophile.com/content/schiit-audio-ragnarok-integrated-amplifier-measurements

 

Schiit may tell good stories, but I for one will not be buying any of their products.

[Ralf11]

Here's a tidbit from optics:

 

https://www.osapublishing.org/oe/abstract.cfm?uri=oe-21-18-21119

 

[sdolezalek]

57 minutes ago, Jud said:

While this is admittedly not relevant to unmeasurable quantities, it seems to me it does speak to our ability (or lack of same) to evaluate what measurements are important.

 I think we are approaching the point where it may be useful to concede to Professor Scott that the differences we should care about can be measured, but what we lack is a "system" that tells us not only how important each measurement is, but, as there are compromises to be made in almost every system design, which set of compromises will lead to the optimum result.

Secondly, I don't believe we have well addressed the question you frequently come back to, which is differences in the listening endpoint (namely us), whether those are genetic, learned, age-based, or otherwise and the degree to which those differences would suggest prioritizing different measurements as more important.

[wgscott]

@Jud

 

In our case, the detector measures intensity, which is the absolute square of the waveform, so you can get the amplitude back by taking the square root, but the phase, having been multiplied by its complex conjugate, is unobtainable.  (You can infer stuff indirectly from interference, and there are tricks to re-establish the missing information, but the detector itself cannot measure it, because it measures intensities only).

 

i.e, we measure I, so we can get A, but not exp{ -iø}

 

I = A exp{ -iø} A exp{ +iø} = A^2

[mansr]

15 minutes ago, wgscott said:

In our case, the detector measures intensity, which is the absolute square of the waveform, so you can get the amplitude back by taking the square root, but the phase, having been multiplied by its complex conjugate, is unobtainable.  (You can infer stuff indirectly from interference, and there are tricks to re-establish the missing information, but the detector itself cannot measure it, because it measures intensities only).

In audio terms, what you have is an SPL meter, not a microphone.

[wgscott]

51 minutes ago, mansr said:

In audio terms, what you have is an SPL meter, not a microphone.

 

Thanks.  That makes sense.

 

[semente]

1 hour ago, sdolezalek said:

Ah, but then what is the measurement for transparency?  Obviously there is not a single measurement that captures all of the above? So what is the series of measurements that add up to transparency and in what order should we consider them?  I do not doubt that speakers and room response have a much bigger effect on these qualities than do DACS, amplifiers, cables, etc.  but I believe the original post was not limited to only measuring those items.  

We once tried to measure where in a system the most aggregious distortions were and it immediatewly became clear that speaker distortions outdistance all others in a high quality system.  If we think the transient response of an amplifier, DAC or cable is poor and contains pre or post ringing, you should see the behaviour of the average paper conse.  And, yes, I agree that you can "distort" the input signal to try to affects or lessen those distortions (room response correction being only one).  But, I believe that most of those  "corrections" themselves involve some tradeoffs and that different earrs evaluate those tradeoffs differently.  

 

A single measurement will only characterise a single parameter.

By "transparency" I mean accuracy and to measure it we have to use all the available measurements for a given category of equipment.

For speakers you can get anechoic frequency response on- and off-axis (polar), listening window, waterfall, THD+N, impulse response, step response, sensitivity, impedance. A (comparative) room response plot is also very informative.

JA even tests amplifiers using a dummy load to assess how well they can cope with real world loads.

 

[AJ Soundfield]

39 minutes ago, semente said:

By "transparency" I mean accuracy

 

Quote

JA even tests amplifiers using a dummy load to assess how well they can cope with real world loads.

Quite appropriate for those magic, audiophile targeted amps/special effects processors, especially those with high output impedance. However also quite limited by its static nature. A far better way is using a PowerCube (with AP loadbox).

Though he probably doesn't want to confuse the "mindless acceptance of voodoo science" crowd further 

[christopher3393]

On 6/6/2017 at 11:17 AM, wgscott said:

Just to clarify, I fully understand that there are differences that cannot be reduced to measurement, like an emotionally-charged vs. bland rendition -- you can measure these differences, presumably, but there is little if any indication how to interpret such measurements in terms of how emotionally compelling a rendition might be.  (The differences are objectively measurable, even if the interpretation of them in this manner is not possible.)

 

Bill, I'm curious about how you evaluate your gear. You have a TEAC NT-503. Given your approach, how did you go about evaluating it? Particularly regarding sound quality, not feature set. What specs did you look for? Did you find any reviews that included independent measurements? Did you do any measuring yourself? And how about listening? Any blind testing? What were you looking for? Finally, what role did sighted testing have and how did you evaluate the emotional factor for yourself, how did you decide you liked the sound, what in particular did you like about the sound, and did that help you decide to keep the unit? Do you think that emotional response to SQ had anything to do with that decision?

[semente]

9 minutes ago, AJ Soundfield said:

 

Quite appropriate for those magic, audiophile targeted amps/special effects processors, especially those with high output impedance. However also quite limited by its static nature. A far better way is using a PowerCube (with AP loadbox).

 

Perhaps AudioGraph won't lend him one...

[AJ Soundfield]

27 minutes ago, christopher3393 said:

 

Bill, I'm curious about how you evaluate your gear. You have a TEAC NT-503. Given your approach, how did you go about evaluating it? Particularly regarding sound quality, not feature set. What specs did you look for? Did you find any reviews that included independent measurements? Did you do any measuring yourself? And how about listening? Any blind testing? What were you looking for?

He had Frank listen to it via YouTube.

For "artifacts"

[mansr]

Just now, AJ Soundfield said:

He had Frank listen to it via YouTube

Well, if AQ can demonstrate the improved sound of their HDMI cables on YouTube...

[AJ Soundfield]

Just now, mansr said:

Well, if AQ can demonstrate the improved sound of their HDMI cables on YouTube...

it's best to assume that any perception in 100% real

[semente]

[jabbr]

13 minutes ago, AJ Soundfield said:

it's best to assume that any perception in 100% real

In the very same sense, it's best to assume that any statement is 100% true... how are we defining "100% true" btw?

[AJ Soundfield]

 

Audiophilia explained

 

 

[wgscott]

1 hour ago, christopher3393 said:

 

Bill, I'm curious about how you evaluate your gear. You have a TEAC NT-503. Given your approach, how did you go about evaluating it? Particularly regarding sound quality, not feature set. What specs did you look for? Did you find any reviews that included independent measurements? Did you do any measuring yourself? And how about listening? Any blind testing? What were you looking for? Finally, what role did sighted testing have and how did you evaluate the emotional factor for yourself, how did you decide you liked the sound, what in particular did you like about the sound, and did that help you decide to keep the unit? Do you think that emotional response to SQ had anything to do with that decision?

 

The Teac NT-503 was new and very hard to find when I got it. So I looked at the measurements (Archimago) for the 501 (along with lots of subjective reviews) and made the assumption that the 503 would behave similarly or perhaps marginally better.  Then I ordered it on Amazon without listening to it, figuring I could return it if (a) it was defective or (b) sounded worse than what I already had.  

 

If I was spending $10k on speakers, where audible differences are likely to really matter, I would be more careful.

 

I can honestly say there was no emotional factor involved with getting this, (unlike my Peachtree Nova, or my bikes). Maybe because TEAC is mere consumer-level rather than boutique audiophile equipment. I'm about as emotionally involved with it as I am with my toaster, whereas if my bikes got stolen I would be suicidal with grief.

[wgscott]

1 hour ago, AJ Soundfield said:

He had Frank listen to it via YouTube.

For "artifacts"

 

Blizzard thought it was shouty; it was good enough for me (I knew I remembered you from somewhere!)