|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.180.4.250
In Reply to: Re: Much simpler answer posted by andy19191 on February 26, 2007 at 23:58:10:
I disagree with this opinion. Good audio tends to bring out the magic of the original performance. I have NEVER deliberately adjusted an audio design to sound different from the original live performance.
Follow Ups:
> Good audio tends to bring out the magic of the original performance.Is good audio different to neutral audio in ways that can be measured?
There is no such thing as neutral audio, because there has been no perfect component that passes audio perfectly. Mid fi usually leaves out information and often imprints a signature that is usually associated with solid state electronics. However, the better that you make solid state electronics perform, the less its presence in the audio chain will compromise the sound. Tubes are not perfect either, but they are usually more listenable than typical solid state.
> There is no such thing as neutral audio,Neutral is that which inaudibly modifies the signal which is straightforward to determine in an audibility experiment.
> because there has been no perfect component that passes audio
> perfectly.There is no requirement for perfection only that the imperfection is inaudible.
> Mid fi usually leaves out information and often imprints a signature
> that is usually associated with solid state electronics.Can this signature be measured?
> However, the better that you make solid state electronics perform,
> the less its presence in the audio chain will compromise the sound.This will only be the true if the presence of solid state electronics is audible in the first place. Is it?
> Tubes are not perfect either, but they are usually more listenable
> than typical solid state.What does listenable mean and can it be measured?
You have not responded to the question asked which was is good audio different to neutral audio in ways that can be measured?
The reason for banging on about measurement is that if something cannot be measured it cannot be addressed by the application of scientific knowledge. If this is the case with your claims then that is fine and we would be talking at cross purposes if I were to expect the scientific method to apply. Clark, for example, has stated that the scientific method does not apply to audio which makes his position clear. No problems.
"Neutral is that which inaudibly modifies the signal which is straightforward to determine in an audibility experiment"
Your original question is meaningless, because there is no such definition as 'neutral audio'. How do you measure that?
Personally, I can measure to 1 million to one, with THD and IM distortion. This appears to be about 10 times better than 'Stereophile' typically measures. This is about -120db. I also can separate individual harmonics and noise from the measurement, and accurately measure harmonics to 100KHz. There is even newer and better test equipment out there, but I cannot personally afford it. However I doubt that an even lower measurement would gain much. After all, sooner or later, an extremely low value in distortion, by itself, must be meaningless, even to the 'objectives' of this world.
I will give an example however of how I once changed the sound of a power amplifier that even surprised me.
It was 1973, and I was hanging out at Mark Levinson's house. We were making a new phono pre-preamp together, to be called the JC-1, and Mark also made some discrete line level modules for the Grateful Dead, for use in their new sound system.
Mark had retired his Marantz 9 tube power amps and bought a Phase Linear 700 power amp that went well with his new DW electrostatic loudspeakers.
Both Mark and I thought this combination pretty good, and well balanced. However, Mark's mother (did I mention that Mark lived with his family?) complained that the new setup hurt her ears, so Mark and I discussed the problem.
I told Mark: "We could change the bias and cut out the protection circuitry."
I knew that the Phase Linear 700 ran next to empty in bias, because I could just put my hand on a heatsink after some time in idle and it would be just barely perceptably warm over the ambient.
Well, we did what I suggested on his workbench. The bias was slightly increased to reduce the xover distortion to where I could just feel a slight difference in the heatsink temperature. (you must remember that by 1973, xover distortion had been proven to be inaudible in DB tests) and the protection circuits were snipped out.
Well what happened? After we got the unit running again, Mark asked me: "Where did the highs go! Did you change something that would affect the frequency response? " I said: "No, Mark, I just removed some of the xover distortion in the amp.
Well, the next time I came to visit, Mark had an electrostatic tweeter mounted on top of his big electrostatic speakers. Subsequently, the electrostatic speaker came out with a MK2 version with a piezo tweeter (ick) on top. Apparently, the electrostatic speaker manufacturer had exaggerated the actual high frequency performance of his speaker in his literature, but this was originally taken care of by the addition of higher order harmonics from the bigger amps that were necessary to drive the speaker properly.
So much for the inaudibility of xover distortion.
Now I want to ask you engineers out there: Why do Motorola piezoelectric tweeters sound the way that they do? What can I MEASURE to show how they sound, and why doesn't everyone use them?
Thanks for the length of the reply but you still seem to have avoided directly answering the question can you measure what you are reporting hearing. Nonetheless, the gist of your response would seem to be that you cannot measure what you are reporting hearing. Is this fair?
I remember the Dayton-Wrights well having heard Dr. John W. Cooledge's pair back in '76. While his MK Is were a tad shy on the top, I was taken by their incredible coherency and low level resolution capabilities. Thirty years later, I'm still a 'stat freak.Indeed the later MK II / MK III's were a mixed bag (no pun intended - remember the outer diaphragm used to contain the SF2 gas?). More efficient, somewhat more reliable, but had that damn tweeter. I worked for a dealer who picked up the line at the time of the MK IIIs. He drove them first with a Dunlap-Clarke amp (Dan D'Agostino was the rep for both D-W and D-C at the time) and later with Nelson Pass' first effort, the 800A and finally a Stasis 2. Some folks disconnected the piezo. The store used an identical driver for use as a - burglar alarm!
SF6 gas :)
rw
Well, I am a chemist :)
it bears repeating. Dr. Cooledge not only wrote for TAS from day one, but he has been a member of the Atlanta Symphony Chorus for about thirty years as a baritone. One of the Dayton-Wright panels required replacement. He had to first cut the outer diaphragm to gain access to the interior where the panels are located. After replacing the panel and resealing the outer diaphragm, he had to replace the SF6 gas. So...As you know as a chemist, SF6 is heavier than air. Exactly the opposite of helium in that regard. You can put it in a cup where it remains colorless, but you can feel it. JWC inhaled a bit and then sung a line from an opera. It sounded like Lurch at the Met! That was hilarious. :)
Cool! I have heard that people have tried that before but personally I have never played with SF6. I am glad most electrostats don't need it. I guess the panels were moisture sensitive or did it simply reduce the chance of arcing?
Electrical insulator to prevent arcing.
...if something cannot be measured it cannot be addressed by the application of scientific knowledge.Perhaps now you are beginning to understand the underlying challenge and limitations of current metrics. In the real world, final voicing of virtually every kind of audio component is done by human ear, not THD plots.
> > ...if something cannot be measured it cannot be addressed by the
> > application of scientific knowledge.
>
> Perhaps now you are beginning to understand the underlying challenge
> and limitations of current metrics.I was simply stating how to determine what is and what is not in the scientific domain. If John considers his observations to be unmeasurable then his statements are not in conflict with anything I have signed up for.
> In the real world, final voicing of virtually every kind of audio
> component is done by human ear, not THD plots.Listening under controlled conditions is a perfectly valid form of measurement. However, if a sound is loud enough to be heard under such conditions it will be straightforward to measure using more reliable and repeatable techniques.
I was simply stating how to determine what is and what is not in the scientific domain. >For stating the obvious.
Listening under controlled conditions is a perfectly valid form of measurement.
Right. We've known that for decades.
However, if a sound is loud enough to be heard under such conditions it will be straightforward to measure using more reliable and repeatable techniques.
Using what metrics? Which combination fully characterize the entire envelope of performance characteristics of any audio component? Every audio designer on the planet will surely benefit from your unique knowledge in this regard. Make it good!
> > However, if a sound is loud enough to be heard under such conditions
> > it will be straightforward to measure using more reliable and
> > repeatable techniques.
>
> Using what metrics?Metrics? Using a microphone plus normal kit.
> Which combination fully characterize the entire envelope of
> performance characteristics of any audio component?What has that, whatever it means, got to do with measuring sound?
> Every audio designer on the planet will surely benefit from your
> unique knowledge in this regard. Make it good!
What you propose Andy is what has been done for the last 70 years. Why do you think it will change due to your suggestion?The question is which measurements are meaningful and correlate with the sound. It is clear that the basic metrics like THD and IM do not.
So, what Estat wants to know is WHAT are you going to measure and how are you going to arrive at that metric and how does it correlate with listening? Assuming of course you accept that different gear actually sounds different, which at the moment it doesn't sound like you accept this basic tennet of the discussion. If you don't think that then there is no point in developing a metric to correlate to something that doesn't exist!
First of all, let's agree on the definition of "metric" in the proper context. My dictionary says this:A standard of measurement.
Anyone disagree? So what metrics do we have for measuring audio gear? THD. IM. Frequency response. Phase shift.
Metrics? Using a microphone plus normal kit.
Which measurement(s) will the microphone be capturing? Not sure what you mean by "kit". Is that a metric? What does it measure? What is the unit of measure?
> Which measurement(s) will the microphone be capturing?Usually they measure a variation in pressure.
> Not sure what you mean by "kit".
A microphone on its own is not much use without supporting kit such as a power supply, preamp, A/D and a storage device.
> Is that a metric?
Is what a metric?
> What does it measure?
Assuming it refers to the microphone, it usually measures a variation in pressure.
> What is the unit of measure?
Whatever unit for pressure you want. Pascal is quite a good one.
I am not sure much clarification has occurred although I do still remain intrigued.
Andy19191919191:However, if a sound is loud enough to be heard under such conditions it will be straightforward to measure using more reliable and repeatable techniques.E-Stat: Using what metrics? Which combination fully characterize the entire envelope of performance characteristics of any audio component?
Andy19191919191:Metrics? Using a microphone plus normal kit.
E-Stat: First of all, let's agree on the definition of "metric" in the proper context. My dictionary says this: A standard of measurement...
So what metrics do we have for measuring audio gear? THD. IM. Frequency response. Phase shift. Soooooooo.
Which measurement(s) will the microphone be capturing?Andy19191919191: Usually they measure a variation in pressure...Whatever unit for pressure you want. Pascal is quite a good one.
_________________________________________________________________________________
So, the measurement of pressure alone is capable of fully characterizing the complete set of all performance criteria of high resolution audio components. Huh?
> So, the measurement of pressure alone is capable of fully
> characterizing the complete set of all performance criteria of high
> resolution audio components. Huh????
I am still find it hard to work out what you are going on about. Are you discussing measuring the difference between two sounds that you can hear or are you discussing something else? If it is something else can you please let me know what it is.
You fail to grasp the concept that the conventional measurements on which you base your qualitative judgement (through your education, of course) are completely useless. Incapable of rendering meaningful correlations to that which is heard.If you haven't already, I recommend you buy the cheapest pro amp on the market like a Behringer and enjoy ordinary audio quality.
"There is no such thing as neutral audio"Yikes, if you restricted your comments to loudspeakers and turntables, you might have an arguable point.
"because there has been no perfect component that passes audio perfectly."
err....Perfection and neutrality, of course no audio component is perfect in the absolute sense, but however does relate to neutrality? Afterall, neutrality means the lack of own signature rather than absolute playback perfection.
"Mid fi usually leaves out information and often imprints a signature..."
How? the job of the components is simple, to playback the software, the measurements can give good indicator of the performance. If you claim that the measurements i.e. controlled listening tests and meters, do not tell the full picture, you are back to where you started? How do you know a component leaves out information.
"However, the better that you make solid state electronics perform, the less its presence in the audio chain will compromise the sound"
The questions stated previously also apply here, how do you know this without the use of some objective metrics. Of particular concern and relevance are the commments "Mid-fi leaves out information"
Music making the painting, recording it the photograph
"the measurements can give good indicator of the performance"
This post is made possible by the generous support of people like you and our sponsors: