The relevance of measurements to audible quality of sound

[Magnum Innominandum]

It looks like you prefer universal science based knowledge pre-chewed and ready to go.

Yes please.

You will find, among other things that +18dB at 20Hz is waaaayyy too much 'compensation'.

I will check out the true-fi software. It has a 10 day trial. But the price is steep, I will likely not buy it.

I'm, still a little stuck with the whole Amp thing. I think the Monoprice unit mentioned seems good.

It is listed as 10V output, which exceeds my 7V minimum requirement.

But can it run well from my Laptop? My laptop uses Realtek ALC269. Line voltage is 1.5v RMS max. It seems with only 6dB gain I will not see the full output. I don't really want to add another box for a DAC, cables and all that.

Magnum Innominandum

 

[Thomas savage]

<snip>

Yet so far no solid research has been cited and it seems limits are a matter of who you ask (everyone has a number they work with) and Amir kind of measures because he can measure and recommends on the concept of "better engineering", which kind of makes sense to me but seems just another belief (better engineering is desirable in the context of good sound).

Magnum Innominandum

Great engineering, achieving as close to ideal measured performance is the idea we are chasing . That’s not a belief, if you then conflate those objective ambitions with ‘good sound ‘ we then stray into value judgements and beliefs. This then leads to confusion wrt preference and you get in a mess.

Until it gets to the transducer I tend to think we are dealing with known and understood EE and there’s no big mystery , we can measure . The transducer itself and its workings are well known too but if there’s doubt we can conduct ‘controlled ‘ listening tests post transducer but how that will show any reliable conclusions Im not sure as the real ‘unknowns ‘ lay in the additives our psychological selfs impart on the end experience, a very faulty and complicated human experience it is too and understanding, recognising and controlling all those variables is impossible. We then rely on statistics to weed out trends and likelihood's ( I’m not keen on that)

So I stick to the known predictable bit, get that right and the rest is just choices and not worth trying to enforce some kind of universal truth on. The less human the better as it’s us that are the problem in this equation, not the failings in understanding performance of the equipment imo.

 

[Magnum Innominandum]

Great engineering, achieving as close to ideal measured performance is the idea we are chasing . That’s not a belief

Forgive me, but it sounds to me like a belief.

As in: "As long as the measurements are perfect the sound must be and the closer to perfect measurements are the better."

As others have remarked - there are limits to what is audible and going further seems pointless.

To come back to my car analogy, a care with a (say) 1,100 BHP engine, 0 to 62mph in under 3 seconds, 230mph max speed and the breaks and suspension to handle this sure sounds like a hell of a car, but do I need it for my commute? What good is 230mph if the limit on the interstate is 115mph?

More something does not, in my view, equate automatically to better. Hence my quest for reason and for the reasoning behind all this.

Magnum Innominandum

 

[solderdude]

I'm, still a little stuck with the whole Amp thing. I think the Monoprice unit mentioned seems good.
It is listed as 10V output, which exceeds my 7V minimum requirement.
But can it run well from my Laptop? My laptop uses Realtek ALC269. Line voltage is 1.5v RMS max. It seems with only 6dB gain I will not see the full output. I don't really want to add another box for a DAC, cables and all that.

Magnum Innominandum

It will give you 4.3V peak (8.6Vpp) voltage which is enough for normal to somewhat loud usage.
Indeed you will never reach the max output voltage but don't think you will be using it.
When you want to EQ passively the 6dB gain will be gone and so will be the possibility to play loud.
Strike or modify the Monoprice or start looking for something with say 20dB gain if you want to EQ passively and use the line-out of the laptop.

 

[Wombat]

There seems to be a group who wish to support their higher-level of individual positive subjective audible perception experience of higher sampling rates, compared to RBCD or next level above, with confirming evidence.

All I can say is keep looking and let me know when you find it. Substantial evidence, not anecdotal comment or internet waffle, is required.

I am still waiting to be convinced.

 

[Magnum Innominandum]

Strike or modify the Monoprice or start looking for something with say 20dB gain if you want to EQ passively and use the line-out of the laptop.

Ok. Will adding a DAC solve this? It seems to me that the difference between 2V from a DAC and the 1.5V from my laptop is small.

Magnum Innominandum

 

[solderdude]

No indeed it won't, that difference is only 3dB.

When you want to use passive EQ you need to find a (portable ?) amp that has a high or configurable gain.
It's the reason why I designed my own portable amp so I can configure it and apply targetted EQ in an analog signal. Alas unless you find someone willing to build you one it is of no use to you.

 

[Thomas savage]

Forgive me, but it sounds to me like a belief.

As in: "As long as the measurements are perfect the sound must be and the closer to perfect measurements are the better."

As others have remarked - there are limits to what is audible and going further seems pointless.

To come back to my car analogy, a care with a (say) 1,100 BHP engine, 0 to 62mph in under 3 seconds, 230mph max speed and the breaks and suspension to handle this sure sounds like a hell of a car, but do I need it for my commute? What good is 230mph if the limit on the interstate is 115mph?

More something does not, in my view, equate automatically to better. Hence my quest for reason and for the reasoning behind all this.

Magnum Innominandum

The reason is often academic imo, but your still conflating a human value judgment with objective measured performance and that’s likely why it all sounds like a ‘ belief ‘ . To me it’s just a signal that should not be added to or subtracted from and then needs to be conveyed through a transducer in the most ideal way, why? Because we can so why not? We end up with accurate ( note I did not say ‘good’ or ‘perfect’) sound , great. Why do otherwise at that stage of reproduction?

Unlike your car analgy there’s no associated danger or vast costs in observing near ideal engineering in terms of audio reproduction. I know what your driving at, unusable performance but they are not remotely the same scenario .

At some point you have to satisfy your own mechanism of discerning the reliability and trustworthyness of the data, is it good data. You could make a argument that knowledge relays on belief, our reality relies on belief but there’s no constructive end to that line of thought and the real mechanisms that underpin our experience lay beyond our control.

Who am I..

 

[Blumlein 88]

Yes please.


snip...................

Magnum Innominandum

College level texts on hearing or senses might be a good place for you to look. Lots of basic info in one place. It may not be pre-chewed enough. You'll find spending some effort earns large dividends. Some is not directly related to things like THD or reflections etc. But enough is so you'll be able to put 2 and 2 together if you'll do the work.

https://www.amazon.com/Introduction..._rd_t=40701&psc=1&refRID=WQEATP0TZ07J3FYDKPMY

This is a good example of one that is worthwhile. You'll learn basics of how hearing works, theories describing how our ears/brain process sound, masking and enough to go a long way toward filtering out what matters and doesn't.

https://www.amazon.com/Fundamentals..._rd_t=40701&psc=1&refRID=2AE2H5DH68RBW971MH4P

This one is more enlightening I think, though not quite as easy as the Moore text.

Now do you have to become educated in the psychophysics of hearing just to buy audio gear? No you don't. Many of the various limits of audibility are nailed down with in reason. It is worthwhile to note it isn't like "wow, you can really hear that!" and you reduce it by fraction of a percent and the effect is suddenly inaudible. If some of the suggested limits are off of maximal limits of the best hearing humans alive, its not like you'll have a massive degradation nor that removing edge case spurious signals will make it all sound like you've never heard it before. You might barely detect .1 or .2 % distortion with test signals in just the right frequency range. It is doubtful anyone would find that level of distortion audible with music.

Basic frequency response of up to 20,000 hz within .1 db of flat, distortion below .1%, and noise levels low enough they aren't obvious won't leave much on the table if anything with music. Audiophiles aren't advancing the psychophysics of hearing. They just let bias convince them they are.

Various other guidelines people are telling you are backed up by research somewhere. They unfortunately aren't in a wikipedia of audible hearing limits.

If you don't want to learn a bit more, or don't have the time, you'll have to decide what kind of info is worth trusting. You might choose wrongly. One guideline that helps plenty with some of the more dismissible claims is people really can't hear above 20 khz. People really have to have it LOUD to hear anything above 15 khz. Whenever someone is selling you on an idea involving ultrasonic signals there is usually something wrong with their claims.

 

[Wombat]

College level texts on hearing or senses might be a good place for you to look. Lots of basic info in one place. It may not be pre-chewed enough. You'll find spending some effort earns large dividends. Some is not directly related to things like THD or reflections etc. But enough is so you'll be able to put 2 and 2 together if you'll do the work.

https://www.amazon.com/Introduction..._rd_t=40701&psc=1&refRID=WQEATP0TZ07J3FYDKPMY

This is a good example of one that is worthwhile. You'll learn basics of how hearing works, theories describing how our ears/brain process sound, masking and enough to go a long way toward filtering out what matters and doesn't.

https://www.amazon.com/Fundamentals..._rd_t=40701&psc=1&refRID=2AE2H5DH68RBW971MH4P

This one is more enlightening I think, though not quite as easy as the Moore text.

Now do you have to become educated in the psychophysics of hearing just to buy audio gear? No you don't. Many of the various limits of audibility are nailed down with in reason. It is worthwhile to note it isn't like "wow, you can really hear that!" and you reduce it by fraction of a percent and the effect is suddenly inaudible. If some of the suggested limits are off of maximal limits of the best hearing humans alive, its not like you'll have a massive degradation nor that removing edge case spurious signals will make it all sound like you've never heard it before. You might barely detect .1 or .2 % distortion with test signals in just the right frequency range. It is doubtful anyone would find that level of distortion audible with music.

Basic frequency response of up to 20,000 hz within .1 db of flat, distortion below .1%, and noise levels low enough they aren't obvious won't leave much on the table if anything with music. Audiophiles aren't advancing the psychophysics of hearing. They just let bias convince them they are.

Various other guidelines people are telling you are backed up by research somewhere. They unfortunately aren't in a wikipedia of audible hearing limits.

If you don't want to learn a bit more, or don't have the time, you'll have to decide what kind of info is worth trusting. You might choose wrongly. One guideline that helps plenty with some of the more dismissible claims is people really can't hear above 20 khz. People really have to have it LOUD to hear anything above 15 khz. Whenever someone is selling you on an idea involving ultrasonic signals there is usually something wrong with their claims.

Pearls before swine.

 

[Magnum Innominandum]

Pearls before swine.

Keel Zeeba! Reep to Shred! Reep off da Head!

Right?

Magnum Innominandum

 

[solderdude]

Nah... I think Wombat is of the opinion that Blumlein has written a lot of words and gave links for self study to someone who, most likely, is not going to digest all the cited and linked stuff and just wants recommendations for an amp.

 

[Cosmik]

The answer to the original post is: if an audio system could be measured and modelled, and the noise & distortion it produced in response to real music could be isolated, synthesised and made available as a WAV file, would it sound like:

• silence
• a quiet hiss that scales with the input amplitude
• " at constant amplitude
• a quiet, muted version of the input signal
• a quiet but harsh version of the input signal
• a dying alien

etc. when compared to the original signal.

I recently did this for a hypothetical transfer function that produces the much sought-after pure second harmonic distortion - for a single tone. With music, of course, this becomes decidedly non-harmonic IMD and sounds diabolical i.e. the dying alien.

But if the result of the noise & distortion is 'silence' or a very quiet hiss, I find it hard to understand why anyone should worry about it.

Distortion that is related to phase and timing is, perhaps, more difficult to isolate and dismiss in this way. It is, however, something that can be eliminated through DSP if the speakers are not ported, and if we accept some latency in the system.

Therefore, I see nothing there that we have to write off as 'mysterious'. We can (if we put our minds to it) check that the noise & distortion is likely to be way below audibility, and eliminate phase & timing errors pre-emptively. Then we can get back to listening to the music

 

[andreasmaaan]

I recently did this for a hypothetical transfer function that produces the much sought-after pure second harmonic distortion - for a single tone. With music, of course, this becomes decidedly non-harmonic IMD and sounds diabolical i.e. the dying alien.

If you can be bothered, could you upload these files perhaps? Would be very interested in hearing them.