Head-Fi.org › Forums › Equipment Forums › Sound Science › Learning more about the science of sound
New Posts  All Forums:Forum Nav:

Learning more about the science of sound - Page 9

post #121 of 395

Are you an audio engineer? You seem to be omniscient like Mr Winer

post #122 of 395
Hi. Let me toss a thought in here that almost came up a few pages ago.

We talk about quality gear as being able to reproduce recorded material accurately, but what does that really mean? A recording is a chain of voltages (or bits describing a chain of voltages), but we don't hear voltage, we hear sound pressure. How can any given sequence of sound pressures be considered faithful to a string of voltages?

Especially consider this question with respect to modern recordings, nearly all of which do not closely resemble the original microphone recordings made while the album was being tracked. The string of voltages never existed as sound pressures, it was always voltages. Objectively, how then do you decide if a produced sound is "accurate?"

We can talk about the quality of equipment that exchanges volts with other equipment and we can certainly be objective about the fidelity of those processes. But saying that an accurate voltage transfer results in high sound quality doesn't follow logically. Electrical potential isn't sound pressure.

If similarity to reality was what was most desirable, there would be no point in multi-tracking, reverb, compression, or any of the other processes used in making a good sounding recording. All the best records would be a live performance captured by an accurate microphone with zero processing.

But that isn't reality. Reality is people prefer to hear sounds that never existed naturally. What sounds good is completely reliant on how a system produces sound. In most cases there is no original material to compare against. Every time any system plays a recording it creates its own interpretation of that string of voltages as a string of sound pressures. There cannot be a right or wrong about it because there is no basis for comparison. Every one of these interpretations is just as "correct."

The only yardstick we have with any actual meaning is whether a person enjoys what they hear.
post #123 of 395
Quote:
Originally Posted by RevMen View Post

Hi. Let me toss a thought in here that almost came up a few pages ago.

We talk about quality gear as being able to reproduce recorded material accurately, but what does that really mean?

Look up high-fidelity sound reproduction. It is well defined. Flat frequency response, low distortion, noise ...

 

Quote:
A recording is a chain of voltages (or bits describing a chain of voltages), but we don't hear voltage, we hear sound pressure. How can any given sequence of sound pressures be considered faithful to a string of voltages?

input -> ideal DAC --> ideal amp --> non-ideal transducer (loudspeaker) -> ideal transducer (mic) -> ideal pre-amp -> ideal ADC -> output

 

Compare input and output. If the loudspeaker reproduces the signal faithfully the output will be close to the input.

 

 

Quote:
Especially consider this question with respect to modern recordings, nearly all of which do not closely resemble the original microphone recordings made while the album was being tracked. The string of voltages never existed as sound pressures, it was always voltages. Objectively, how then do you decide if a produced sound is "accurate?"

The same way. It doesn't matter how a track is made. It could be recorded, but it could also be generated by a computer program.

Measurement systems, for example, use generate sine sweeps, noise or MLS signals.

 

 

Quote:
We can talk about the quality of equipment that exchanges volts with other equipment and we can certainly be objective about the fidelity of those processes. But saying that an accurate voltage transfer results in high sound quality doesn't follow logically. Electrical potential isn't sound pressure.

You mean conversion of an electric signal into sound? Yes, an accurate transfer does result in high sound quality and I do not see how this is not logical.

(If I apply the logic in your message then saying that an accurate conversion from bits into an electric signal results in high sound quality doesn't follow logically either? Since bits are not an electric signal.)

 

 

Quote:
If similarity to reality was what was most desirable, there would be no point in multi-tracking, reverb, compression, or any of the other processes used in making a good sounding recording. All the best records would be a live performance captured by an accurate microphone with zero processing.

What would you do if the singer messed up during the performance? What if after the recording the artist says he wants a bit more reverb on his voice only? ...

You see, the flexibility added by multi-tracking is what matters. Again, it doesn't matter if the tracks sound completely different to what was recorded, it wouldn't even matter if the output is just random noise.

 

 

Quote:
But that isn't reality. Reality is people prefer to hear sounds that never existed naturally.

Why are you excluding proper jazz, classical .. recordings?

 

 

Quote:
What sounds good is completely reliant on how a system produces sound. In most cases there is no original material to compare against. Every time any system plays a recording it creates its own interpretation of that string of voltages as a string of sound pressures. There cannot be a right or wrong about it because there is no basis for comparison. Every one of these interpretations is just as "correct."

The only yardstick we have with any actual meaning is whether a person enjoys what they hear.

Whether a classical recording sounds good depends on how accurate (see above) the system reproduces sound. You can compare how instruments sound in such recordings to how they sound in actually.

If a violin on the recording sounds like a viola then the system is inaccurate. And no, it's not just a different interpretation that is just as "correct", it's wrong.

 

Only a few genres require, for example, an overblown subwoofer that drowns out the rest of the frequencies and I have no problem with people enjoying that. What I do have a problem with is calling it accurate, or demeaning objectivity in general.


Edited by xnor - 4/17/13 at 5:11am
post #124 of 395

How do tube amplifiers colour the output? The unique sound that they possess, is it due to certain frequencies being deviated from flatness or is there a different way in which they colour the sound?

post #125 of 395
Quote:
Originally Posted by uchihaitachi View Post

How do tube amplifiers colour the output? The unique sound that they possess, is it due to certain frequencies being deviated from flatness or is there a different way in which they colour the sound?

In general, it's non-flat response, added harmonic distortion and noise.  The degree varies from design to design, with some tube designs becoming indistinguishable from solid-state designs, to others that deliberately exaggerate some tubes tendency toward non-linearity. Part of the problem is that tubes operate at fairly high impedances, and have output impedances from hundreds to thousands of ohms.  In order to drive a low impedance load like a speaker or headphone, an output transformer is required, which has its own non-linearities, and response impact.  Some designs exploit the potential nonlinearities of tubes and transformers by operating the tube outside of its designed parameters, for example, starving it for plate voltage, or biasing the tube to an unusual point in its operating curve, which result in the addition of all sorts of harmonic distortion.  The noise content is in large part due to the fact that a tube runs very hot, which by nature increases residual noise in high gain circuits.  

 

You can build a pretty transparent tube amp, but there's no point to that as people seem to want a tube amp to color the sound.  So most tube amps are built to add "warmth", which is the above mentioned non-flat response and additional distortion components.  If you believe an amplifier should act like a "straight wire with gain", you'd be moving away from that desire.

 

But if you ask, "Why to people love tube amps", the answer would only partially relate to the above, and mostly encompass a huge amount of psychology.


Edited by jaddie - 4/30/13 at 6:41am
post #126 of 395

Then is the added distortion in general actually 'pleasant' to most ears? Or is it the added distortion that makes people think that wow there is indeed a difference in sound and this combines with placebo to provide an amazing sonic experience. In a double blind test, would somebody actually prefer the added distortion sound (generally speaking) over a source that is true to the recording?

post #127 of 395

Added distortion is never preferred in and of itself, though some will argue in favor of it.  Distortion is not really what people like, and is actually far less audible than most think.  ABX testing reveals that distortion is largely inaudible until some rather startlingly high levels of it are reached.  However, frequency response changes, if over a sufficiently large range of the spectrum, are easily audible.  

 

DBT isn't designed to reveal a preference in one device over another, it's designed to reveal an audible difference by matching two choices (A and B) against an unknown (X) which is randomly assigned to be either A or B.  If a statistical result shows that A or B can be easily matched with reasonable reliability over multiple trials and multiple listeners, then the difference can be said to be clearly audible.  

 

Tube amps may or may not present a clearly audible difference, though, in ABX/DBT testing.  Preference is another story completely.  The experience of tube amps is not only sonic (which can be fairly minimal) but also tactile, visual, and highly biased by expectation.  Tube amps always have their tubes visible.  The heat can be felt, the glowing filaments can be seen, and with the complete knowledge that a tube amp is being used, the expectation of smoother, more musical sound is often perceived as real.  Remove the knowledge of what is being heard as well as the visual and tactile stimulus, and the preference will also disappear.  The general preference across general population is for undistorted sound.

post #128 of 395

I should correct myself, there is one instance where adding distortion in a carefully controlled way results in a preference for some listeners.  The process is found in the Aphex Aural Exciter, which among other things, adds even-order harmonic distortion.  I'm not sure that it's the distortion that people like, but rather the effect of high frequency emphasis that results.  It was actually quite nice on instruments with plucked strings, though once we moved to digital recording the need seems to have diminished.  Useful on analog recordings though.

post #129 of 395
Quote:
Originally Posted by jaddie View Post
The general preference across general population is for undistorted sound.

I was wondering if there is a source you base this conclusion on? (Asking out of genuine interest)

post #130 of 395

 

 

post #131 of 395

If you are exposed to a frequency that you cannot hear at high volumes. Can your hearing get damaged?

post #132 of 395
Quote:
Originally Posted by uchihaitachi View Post

If you are exposed to a frequency that you cannot hear at high volumes. Can your hearing get damaged?

Yeah, the sound pressure is still there.


This is why when listening to any sound tests at very high/very low frequencies you should never increase the volume to a level that would be uncomfortable at other frequencies.

post #133 of 395
Quote:
Originally Posted by SunshineReggae View Post

I was wondering if there is a source you base this conclusion on? (Asking out of genuine interest)

David Clark did some extensive ABX testing using his "chamber of horrors" distortion generator back in the 1980s.  He found two things: 1, we can't hear distortion as well as we thought.  He could introduce some rather high levels without reliable detection.  2, distortion was never preferred.  The data is in an old AES paper.

 

But, why would anyone doubt that less distortion is preferable?  We're trying to reproduce sound, why would adding distortion resulting in harmonics not otherwise present be a good thing?  The Aphex case is special, and not universally positively accepted.  


Edited by jaddie - 5/2/13 at 8:02am
post #134 of 395
Quote:
Originally Posted by jaddie View Post

 

But, why would anyone doubt that less distortion is preferable?

 

The myth of euphonic distortion.

post #135 of 395
Quote:

The most important finding was that none of the different patterns of nonlinearity sounded in any way preferable to the undistorted reference.

Quote:
Unless and until somebody comes up with a "magic" pattern of nonlinearity that truly enhances sound quality, I will believe euphonic distortion to be a fantasy. The only "good" nonlinear distortion is that of a nature and amplitude such that the human ear cannot detect it.

From a SP article titled "Euphonic Distortion: Naughty but Nice?".

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Learning more about the science of sound