What uncontroversial audible differences cannot be measured?

[semente]

9 hours ago, sdolezalek said:

 

   4) timbre accuracy (which to me is more than just frequency response) as it encompasses accuracy of frequency as well as of harmonics, decay and room reflections in the original

 

 

'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.

 

[semente]

9 hours ago, fas42 said:

 

To a small extent, yes. Simulations of amplifiers using, say LTspice, my choice, typically assume everything around it is perfect, including power supplies - which leads one to a fairy land of performance numbers. By adding a real world mains supply, and typical component power supply made up of Spice parts with all their parasitic behaviours, I was able to see the high treble distortion of a decent amplifier design increase by the order of a 100 times - why? Because the linearity of the amplifier was compomised by the design's feedback having to deal with the non-ideal voltage rails; the complex, and real world, electrical environment ensured that the amplifier performed nowhere near as well as assuming ideal operating circumstances.

 

Thus we may conclude that it wasn't a decent amplifier design, since it didn't take into account the PSU issue you have mentioned.

[Jud]

8 hours ago, esldude said:

Then AJ has given you a good answer.

 

I tend not to think of binaural.  For whatever reason I have not found it to work for me.  It just sounds like bad stereo.  I really hate that Chesky now does only binaural recordings. When they did blumlein recordings I could enjoy their work.  Now in binaural.....well for me just forget about it.

 

Brings up another issue I've very briefly mentioned in the thread, and have a great deal of interest in: Presumably to the Chesky folks (and I'm assuming to buyers, or they wouldn't be marketing them exclusively), these recordings have a more natural sound.  Maybe it's primarily for headphone users?  But for those who prefer this sound from speakers, in contrast to you disliking it, is it because of genetics, because of what you are accustomed to listening to and you now hear as "right," some combination of these...?

[Jud]

2 hours ago, semente said:

 

Thus we may conclude that it wasn't a decent amplifier design, since it didn't take into account the PSU issue you have mentioned.

 

The problem is, there are a lot of "indecent" (  ) designs around.

 

Or if you want to look at this another way: Let's take software engineering.  Between education, best practices, various tools, and quality assurance testing, most of the "indecent designs" (egregious errors) are going to be caught before a lot of software products make it out the door.  But then there are always "bugs" found by users.  Why?  Because the variety of usage situations (associated hardware and software, ways people try to use the software you may not have anticipated) is nearly infinite. 

 

So yes, with software, and with audio hardware, you can anticipate what you can think of.  But there will always be situations you haven't thought of.  (One version of this is "Whenever you think you've designed something foolproof, along comes a better fool.")

[AJ Soundfield]

5 hours ago, STC said:

 I was expecting direct answers to the questions raised

Now that is funny.

Frank has been teaching himself, creating and grading his own tests since Kindergarten. A+ all the way ;-).

You're going to see 18k posts stating the same circle over and over and over and...

I thought the YouTube stuff and the "Phillips HTIB" uber system were dead giveaways to the prank, but alas...

[mav52]

20 hours ago, Jud said:

 

Lexicon then went on to do great things with Oppo. 

Jud the only great things Lexicon did regarding to OPPO was to copy the OPPO BDP-83 ($500) player and release the Lexicon BD-30 ($3500). Lexicon changed the outside but its identical to OPPO on the inside and hint, they used the same internals of the OPPO. Not really a great thing.

[STC]

44 minutes ago, AJ Soundfield said:

Now that is funny.

Frank has been teaching himself, creating and grading his own tests since Kindergarten. A+ all the way ;-).

You're going to see 18k posts stating the same circle over and over and over and...

I thought the YouTube stuff and the "Phillips HTIB" uber system were dead giveaways to the prank, but alas...

 

The only reason I paid attention to this thread because it reminded me of someone who did a wonderful job with a JVC DVD player and some other cheap stuff together with 8 or so cheap subwoofer. He died bringing his secret to his grave. 

[mmerrill99]

1 hour ago, Jud said:

 

The problem is, there are a lot of "indecent" (  ) designs around.

 

Or if you want to look at this another way: Let's take software engineering.  Between education, best practices, various tools, and quality assurance testing, most of the "indecent designs" (egregious errors) are going to be caught before a lot of software products make it out the door.  But then there are always "bugs" found by users.  Why?  Because the variety of usage situations (associated hardware and software, ways people try to use the software you may not have anticipated) is nearly infinite. 

 

So yes, with software, and with audio hardware, you can anticipate what you can think of.  But there will always be situations you haven't thought of.  (One version of this is "Whenever you think you've designed something foolproof, along comes a better fool.")

Funny, when I made the exact same point, I got an ad-hom attack from Mansr - let's see if you are similarly treated?

 

"You have no clue about software, do you?"

[mav52]

20 hours ago, Jud said:

Long, long ago, Bell Labs/Western Electric came to the conclusion a perfect loudspeaker would be a sphere of energy.  (Something like the Hill Plasmatronics, but full range and, oh yes, not poisoning people who breathed near them.)  You'd just need 8 of these, one in each of the corners of the room, radiation patterns adjusted continuously for room acoustics and instrument radiation patterns....

Jud, having worked for Western Electric, Bell Labs, AT&T and Lucent , can you provide the information on this perfect loudspeaker is a sphere of energy. I just don't recall it in my days working for them..  I do know from the pictures that Western Electric in 1926 introduced its first hifi speaker  and that's 88 years ago and its not a sphere its a horn.  Thanks

[AJ Soundfield]

2 hours ago, Jud said:

 

The problem is, there are a lot of "indecent" (  ) designs around.

Where?

[AJ Soundfield]

11 minutes ago, mav52 said:

 can you provide the information on this perfect loudspeaker is a sphere of energy.

I'm not familiar with that either.

Here is something more recent that doesn't require AES membership to view:

http://www.linkwitzlab.com/Recording/acoustics-hearing.htm

(pay close attention to pg 26-36)

[mansr]

2 hours ago, Jud said:

The problem is, there are a lot of "indecent" (  ) designs around.

 

Or if you want to look at this another way: Let's take software engineering.  Between education, best practices, various tools, and quality assurance testing, most of the "indecent designs" (egregious errors) are going to be caught before a lot of software products make it out the door.  But then there are always "bugs" found by users.  Why?  Because the variety of usage situations (associated hardware and software, ways people try to use the software you may not have anticipated) is nearly infinite. 

 

So yes, with software, and with audio hardware, you can anticipate what you can think of.  But there will always be situations you haven't thought of.  (One version of this is "Whenever you think you've designed something foolproof, along comes a better fool.")

I don't think the software analogy is a good one. An amp has an input and output. As long as the input voltage and load impedance are within specified limits, the output should be predictable. A handful of test signals are enough to fully characterise its behaviour. Software, on the other hand, has a near infinite variety of inputs and it's not even well defined what the output ought to be in many cases. Consider also the internal complexity. While an amp has perhaps a few hundred parts (a simple one needs only a handful), an operating system has tens of millions of lines of code. It is not humanly possible to form a mental model of the entire system at one time, so we have to use high-level abstractions for all but the small part we're currently working on. When a module violates the rules of these abstractions, obscure bugs occur.

[fas42]

43 minutes ago, STC said:

 

The only reason I paid attention to this thread because it reminded me of someone who did a wonderful job with a JVC DVD player and some other cheap stuff together with 8 or so cheap subwoofer. He died bringing his secret to his grave. 

 

Various people have got similar results - I'm always on the lookout for others who "get it", and those that do usually ascribe their success to some special ingredient - using an amplifier with a particular design philosophy; or horn speakers that are particularly well sorted; or going overboard on a grounding strategy. For me, these are just different roads to the same destination: achieving a sufficiently competent playback quality; from then on the ear/brain takes over, and constructs a powerful illusion which makes the listening a delight.

 

IOW, I refuse to have a hangup about needing special speakers, room setups, the adding of magic sauce - it's the attention to detail, worrying about the most minor aspects, that gets me the results.

 

There will certainly be other people after me who will chance upon their version of the answers - what matters is making at least some people aware that a very special quality of sound can be achieved, if they are prepared to pursue it.

[mansr]

5 minutes ago, fas42 said:

Various people have got similar results - I'm always on the lookout for others who "get it", and those that do usually ascribe their success to some special ingredient - using an amplifier with a particular design philosophy; or horns speaker that are particularly well sorted; or going overboard on a grounding strategy.

Many of those cases are probably nothing but dumb luck being wrongly attributed to whatever part of the system is slightly unusual.

[AJ Soundfield]

2 minutes ago, mansr said:

Many of those cases are probably nothing but dumb luck being wrongly attributed to whatever part of the system is slightly unusual.

Or grading ones own self administered "tests" 

[fas42]

13 minutes ago, mansr said:

I don't think the software analogy is a good one. An amp has an input and output. As long as the input voltage and load impedance are within specified limits, the output should be predictable. A handful of test signals are enough to fully characterise its behaviour. Software, on the other hand, has a near infinite variety of inputs and it's not even well defined what the output ought to be in many cases. Consider also the internal complexity. While an amp has perhaps a few hundred parts (a simple one needs only a handful), an operating system has tens of millions of lines of code. It is not humanly possible to form a mental model of the entire system at one time, so we have to use high-level abstractions for all but the small part we're currently working on. When a module violates the rules of these abstractions, obscure bugs occur.

 

The software analogy is quite reasonable - IME, a handful of test signals is not good enough to characterise the behaviour of, not the amplfier, but the system as assembled - everything I've read has confirmed that no-one has a good handle on how to test a complete audio rig, to come up with a set of numbers which fully describe its capability. So, I rely on my ears to give the feedback as to whether a setup is approaching the competence I seek.

[AJ Soundfield]

Reminds me of Norman Korens Loose Screws? story ;-)

(just below A list of books and publications I have for sale on page)

[STC]

58 minutes ago, STC said:

 

The only reason I paid attention to this thread because it reminded me of someone who did a wonderful job with a JVC DVD player and some other cheap stuff together with 8 or so cheap subwoofer. He died bringing his secret to his grave. 

 

And a PM from someone I respect spoke very highly of his system. I thought we had something in common since I also use cheap stuff like $8.99 DACs.  

[fas42]

5 minutes ago, mansr said:

Many of those cases are probably nothing but dumb luck being wrongly attributed to whatever part of the system is slightly unusual.

 

Luck is a huge part of it - it happened to me as a total surprise, I was just trying to get my system sounding better along the same lines as everyone else - and suddenly this quantum change in the subjective presentation popped up like a Jack-in-the-Box ... "What the hell's going on ??!!" ...

 

[esldude]

3 hours ago, Jud said:

 

Brings up another issue I've very briefly mentioned in the thread, and have a great deal of interest in: Presumably to the Chesky folks (and I'm assuming to buyers, or they wouldn't be marketing them exclusively), these recordings have a more natural sound.  Maybe it's primarily for headphone users?  But for those who prefer this sound from speakers, in contrast to you disliking it, is it because of genetics, because of what you are accustomed to listening to and you now hear as "right," some combination of these...?

Not really a matter of like or dislike.  It sounds okay on speakers. It is that due to odd ears or something binaural doesn't work for me. It sounds like poor recordings still suck in my head with weird overhead imaging.

[semente]

There's going to be a Reproduced Sound conference in November at the Nottingham Conference Centre.

 

http://reproducedsound.co.uk

 

The core-topic for this year’s conference is sound quality by design. Other topics of interest include:

• Acoustics of Rooms
• Live Sound
• Audio System Design
• Headphone Listening
• Loudspeakers and Microphones
• Intelligibility
• Measurement & Evaluation
• Cinema Sound
• Case Studies

&

• Dr Toole will be attending the conference to give this year’s Peter Barnett Memorial Award lecture after receiving the award from the Institute.

 

 

[jabbr]

1 hour ago, mansr said:

An amp has an input and output. As long as the input voltage and load impedance are within specified limits, the output should be predictable. A handful of test signals are enough to fully characterise its behaviour

 

To a large extent this should be true although as the number of individual components rise, with each of their own part to part tolerances and temp dependencies etc. the range of possible outputs can soar. 

 

More importantly is the failure to realize not only that the power supply is part of the circuit, but that the speakers are directly in the circuit, an integral part of the circuit, and so the crossovers etc., in fact the speaker cabinet is part of the circuit -- point being that "load impedance" is more complex than may first appear.

[mansr]

30 minutes ago, jabbr said:

To a large extent this should be true although as the number of individual components rise, with each of their own part to part tolerances and temp dependencies etc. the range of possible outputs can soar. 

 

More importantly is the failure to realize not only that the power supply is part of the circuit, but that the speakers are directly in the circuit, an integral part of the circuit, and so the crossovers etc., in fact the speaker cabinet is part of the circuit -- point being that "load impedance" is more complex than may first appear.

Sure, there are interactions between components and part to part variations. Nevertheless, the input to an amp is always a voltage. It is not possible to present it with, say, a cat instead. That makes it much simpler to verity correct operation than with software.

[Jud]

2 hours ago, mansr said:

I don't think the software analogy is a good one. An amp has an input and output. As long as the input voltage and load impedance are within specified limits, the output should be predictable. A handful of test signals are enough to fully characterise its behaviour. Software, on the other hand, has a near infinite variety of inputs and it's not even well defined what the output ought to be in many cases. Consider also the internal complexity. While an amp has perhaps a few hundred parts (a simple one needs only a handful), an operating system has tens of millions of lines of code. It is not humanly possible to form a mental model of the entire system at one time, so we have to use high-level abstractions for all but the small part we're currently working on. When a module violates the rules of these abstractions, obscure bugs occur.

 

So folks who design using SPICE or similar simulation software on amps that don't perform perfectly the first time tested or in the field are plainly incompetent?  And perhaps designers shouldn't even need SPICE, because they should be able to form a mental model of the whole?

[Jud]

2 hours ago, AJ Soundfield said:

Where?

 

Let's start with every DAC that is sensitive to changes in input noise or jitter levels.