All kinds of good arts are actually perfectly scientific.
If we don't understand, that's because we haven't figured out the science yet, not because the science doesn't exist. 😉
If we don't understand, that's because we haven't figured out the science yet, not because the science doesn't exist. 😉
After looking at the said article, anybody who invokes Heisenberg in his argument about audio looses me instantly.
The point about the reflex box can be seen in Smalls expression for acoustic power output.
Looking at the QB5 II filter assisted alignments for instance, for an f3 of half that of the driver fs you can increase the acoustic output by around 90.5 times over that of the equivalent sealed box.
What else I wonder can give such a dramatic improvement in a loudspeaker system performance parameter such as this, and it is done by using simple low cost things like a cardboard tube and two capacitors, I ask what lengths he needs to go to to render the reflex box obsolete, its all to do with Heisenberg no doubt.
rcw
The point about the reflex box can be seen in Smalls expression for acoustic power output.
Looking at the QB5 II filter assisted alignments for instance, for an f3 of half that of the driver fs you can increase the acoustic output by around 90.5 times over that of the equivalent sealed box.
What else I wonder can give such a dramatic improvement in a loudspeaker system performance parameter such as this, and it is done by using simple low cost things like a cardboard tube and two capacitors, I ask what lengths he needs to go to to render the reflex box obsolete, its all to do with Heisenberg no doubt.
rcw
it's Your problem, on the other hand Linkwitz, for example, says: "The books by John Watkinson are highly recommended."
search Amazon with "John Watkinson"
Active electronics that give good bass extension and good impulse behaviour.
One example of that would be the Linkwitz transform circuitry.
A speakers bass roll-off is just a mechanical high pass filter. And you can add filters to other filters to make a new filter slopes, it does not matter if they are mechanical or electrical filters.
What is also very interesting is that Mr Watkinson demonstrates his speakers with a life musician as a reference.
Iaw the audience first listens to the real musician, witch gets recorded, then to the recording. In this way you get to compare the speaker with the real thing.
How many speaker manufacturers do that? I know only 1.
Iaw the audience first listens to the real musician, witch gets recorded, then to the recording. In this way you get to compare the speaker with the real thing.
How many speaker manufacturers do that? I know only 1.
It is a simple fact Graaf that Heisenberg uncertainty is about the physical concept of action, and that is position and time not frequency and time, and only applies to very small scale systems, and has nothing to do with the large scales that are found in audio.
His observations about magnets are also highly dubious, I would imagine that any thermal phonon motion would completely swamp any real effect of that sort, although it might theoretically exist in some theoretical construct such as a one atom thick sheet at absolute zero.
I have not seen the book and it might be free of this sort of nonsense, but can only go by the article and as others have observed, it is a bit of a curates egg, and I found the not so good parts very poor indeed.
rcw
His observations about magnets are also highly dubious, I would imagine that any thermal phonon motion would completely swamp any real effect of that sort, although it might theoretically exist in some theoretical construct such as a one atom thick sheet at absolute zero.
I have not seen the book and it might be free of this sort of nonsense, but can only go by the article and as others have observed, it is a bit of a curates egg, and I found the not so good parts very poor indeed.
rcw
Qb5 Alignments, cap assisted (I or II?) are certainly adding science into design. I only wish i understood the ESP article and how it leads to a series cap value. Better read it for the nth time 😀
it is a simple fact that You became prejudiced because of a sort of pupular rethoric used in the article
it is also a simple fact that JW is a respected author and an AES Fellow
speaking of AES:
Audio Engineering Society UK Loudspeaker Design: Tradition Versus Science
"Despite enormous progress in understanding how the human auditory system works, most present day loudspeakers cling to outmoded and discredited techniques that have not changed in decades.
The availability of advanced materials and design tools means that the task of advanced speaker design has never been easier, but the necessary steps simply are not taken.
This presentation will look at the criteria for accurate sound reproduction and will show that these criteria can be met. Demonstrations of some alternative loudspeaker designs will be given."
a comment from a UK forum (bolds mine):
"(...)
Randomly.......
A loudspeaker has to make your ears and brain think they are hearing the music correctly. Your brain does a reasonable amount of processing to what the ears give them. (...)
We are sensitive to phase. Not phase in a continuous tone but in transients- our survival mechanism. So you need a speaker linear phase in transients
(...)
Passive crossovers are rubbish. Test this by putting a signal into a x-over, recombine the output and then then measure. You wont get the same signal. So active is the answer.
(...)
Dont start with a resonance in the system to help its response and that means YOU bass reflex and for a slightly different reason transmission line. Start with a sealed box and fill it correctly. Understand adiabatic effects- you can read about these in the LS3/5A designs documents from the BBC.
You can get high power magnets and therefore high efficiency drivers so you dont need massive amps. And you dont need massive amps when you go active.
You dont need a perfect frequency response but you do need good off axis response- ie the same as the on axis as the ear is more sensitive to off axis response changes than wrinkles in the frequency response. To get a good off axis response you end up with an omni-directional speaker and that is what he was showing.
(...)
He gave a good tip to measure the "bandwidth" of your loudspeaker by taking an audio signal and data compressing it. When you can hear the compression you have reach the bandwidth limit of the speaker.
Throughout there were musical illustrations from CD and to show off his MI amps live vocals and Rhodes piano from a lady whose name I shamefully cant remember. John also played steel guitar through the combo. CD clips with bells to show transient skills, organ for bass, vocal jazz for vocal quality."
from:
AES Lecture, new loudspeaker designs? - General HIFI Discussion - HIFICRITIC FORUM - HIFICRITIC FORUM : hi fi audio systems forum
a twitter comment:
"just heard my first omnidirectional speaker system made by John Watkinson @AESorg (UK) meeting. I'm totally blown away at what is achievable"
from:Twitter / bongobeardy: just heard my first omnidi ...
ps.
another impressed attendant at the lecture shared his impressions at gearslutz forum:
http://www.gearslutz.com/board/stud...-work-matter-if-my-monitors-arent-up-par.html
Audio Engineering Society UK Loudspeaker Design: Tradition Versus Science
"Despite enormous progress in understanding how the human auditory system works, most present day loudspeakers cling to outmoded and discredited techniques that have not changed in decades.
The availability of advanced materials and design tools means that the task of advanced speaker design has never been easier, but the necessary steps simply are not taken.
This presentation will look at the criteria for accurate sound reproduction and will show that these criteria can be met. Demonstrations of some alternative loudspeaker designs will be given."
a comment from a UK forum (bolds mine):
"(...)
Randomly.......
A loudspeaker has to make your ears and brain think they are hearing the music correctly. Your brain does a reasonable amount of processing to what the ears give them. (...)
We are sensitive to phase. Not phase in a continuous tone but in transients- our survival mechanism. So you need a speaker linear phase in transients
(...)
Passive crossovers are rubbish. Test this by putting a signal into a x-over, recombine the output and then then measure. You wont get the same signal. So active is the answer.
(...)
Dont start with a resonance in the system to help its response and that means YOU bass reflex and for a slightly different reason transmission line. Start with a sealed box and fill it correctly. Understand adiabatic effects- you can read about these in the LS3/5A designs documents from the BBC.
You can get high power magnets and therefore high efficiency drivers so you dont need massive amps. And you dont need massive amps when you go active.
You dont need a perfect frequency response but you do need good off axis response- ie the same as the on axis as the ear is more sensitive to off axis response changes than wrinkles in the frequency response. To get a good off axis response you end up with an omni-directional speaker and that is what he was showing.
(...)
He gave a good tip to measure the "bandwidth" of your loudspeaker by taking an audio signal and data compressing it. When you can hear the compression you have reach the bandwidth limit of the speaker.
Throughout there were musical illustrations from CD and to show off his MI amps live vocals and Rhodes piano from a lady whose name I shamefully cant remember. John also played steel guitar through the combo. CD clips with bells to show transient skills, organ for bass, vocal jazz for vocal quality."
from:
AES Lecture, new loudspeaker designs? - General HIFI Discussion - HIFICRITIC FORUM - HIFICRITIC FORUM : hi fi audio systems forum
a twitter comment:
"just heard my first omnidirectional speaker system made by John Watkinson @AESorg (UK) meeting. I'm totally blown away at what is achievable"
from:Twitter / bongobeardy: just heard my first omnidi ...
ps.
another impressed attendant at the lecture shared his impressions at gearslutz forum:
http://www.gearslutz.com/board/stud...-work-matter-if-my-monitors-arent-up-par.html
Last edited:
Linkwitz describes the article as, "thought provoking", and goes on to elucidate his more benign thoughts, and only hints at his less benign ones, which I suspect are very similar to mine.
I am not likely to meet Watkinson at an AES convention however, so I have no need of being diplomatic.
rcw
I am not likely to meet Watkinson at an AES convention however, so I have no need of being diplomatic.
rcw
no need of being diplomatic at all - it is the article that is "thought provoking" - not the Author
Attachments
I have seen (and heard) him live as well at an AES meeting in Switzerland. He is by no way boring to listen to - quite the contrary to be exact.
He demonstrated two different models of his "Legend" speakers amongst other things.
Those rellay had the property to disappear.
The live perfomer was Lisa (or Liza ?)Marshall BTW.
There is no free lunch in speaker building and there will never be a perfect speaker either. So you will have to pick your poison: I.e. do I want better transient response (=closed box) or higher SPL/lower THD with a given driver ?
About the fith-order alignment: There was an article in JAES recently by Mr. Thiele where he showed how such a tunings could be realised in practice.
Regards
Charles
He demonstrated two different models of his "Legend" speakers amongst other things.
Those rellay had the property to disappear.
The live perfomer was Lisa (or Liza ?)Marshall BTW.
There is no free lunch in speaker building and there will never be a perfect speaker either. So you will have to pick your poison: I.e. do I want better transient response (=closed box) or higher SPL/lower THD with a given driver ?
About the fith-order alignment: There was an article in JAES recently by Mr. Thiele where he showed how such a tunings could be realised in practice.
Regards
Charles
Last edited:
Think of it like this:
Its more difficult to determine the frequency of staccato notes than it is for longer lasting notes.
If you want to tune an instrument, you have to compare a note to a reference note. The accuracy of the tuning is determined by how long you can compare the reference note to the note that you want tuned. The longer the comparison the better the frequencies can match.
For the comparison of transients its useless to measure for a few seconds because the transients do not last that long. You only have to compare them. And transients have all frequencies in them, so you don't know the frequency content very accurate.
An other way:
To measure the velocity of an object you need a distance and a time. The longer the distance the higher the accuracy of the (average) velocity. If you make the distance very short, accuracy goes down. A velocity measurement is always an avarage over a specific distance.
You can determine the place of an object but then you do not know the velocity, because you need 2 places for that measurement.
And one more:
Wind speed can be measured with a flag. If you ever see a big flag and a small flag together you will notice that the small flag moves much more than the big flag. The surface area (from witch the distance is calculated) of a small flag is smaller and the accuracy of the measurement is smaller to hence the bigger movement of the small flag.
Important thing to notice:
This is about how we get accurate measurements, not how we do maths from these measurements.
The thought that I had was that someone wrote an article about putting science back into a subject, and then starts by making a statement that is scientific nonsense about that subject, but is the sort of thing quite often used in marketing hype.
I then think I have reasonable cause to question the intentions of the writer.
rcw
I then think I have reasonable cause to question the intentions of the writer.
rcw
Instead of bantering about “science”, how about speculating about this concept; If a perfect measurement microphone exactly captures the acoustic pressure of one single point in space and converts it into a Voltage signal, why wouldn’t the ideal speaker be at least a close reciprocal of that?
How can science (the vast array of measurement tools available now) be used to understand where and why it falls short?
Best,
Tom
How can science (the vast array of measurement tools available now) be used to understand where and why it falls short?
Best,
Tom
I agree that in principle that is true.
The trouble is that real microphones have various pick up patterns and various mixes of microphones are used in real recordings, and in some cases tracks are directly coupled to the mix with no microphone at all.
So in effect the closest we ever come to that ideal situation is a single sound source being recorded by a single microphone, and in stereo perhaps the classical crossed pair of microphones.
Non of this applies to the vast majority of signals that anybody actually listens to and so it really doesn't matter very much to me.
But as a scientist I am very much aware of the difference between actual science and concepts being borrowed from science to make marketing hype more"scientific" sounding.
Quite often you hear people say such things as Albert Einstein said all things were relative in his relativity theory. All things might well be relative but that is not what that theory is about, all things might also be uncertain but Heisenberg has not anything to do with it.
Some might dismiss this as pedantic, but it is just clear definition that leads to clear thinking.
rcw
The trouble is that real microphones have various pick up patterns and various mixes of microphones are used in real recordings, and in some cases tracks are directly coupled to the mix with no microphone at all.
So in effect the closest we ever come to that ideal situation is a single sound source being recorded by a single microphone, and in stereo perhaps the classical crossed pair of microphones.
Non of this applies to the vast majority of signals that anybody actually listens to and so it really doesn't matter very much to me.
But as a scientist I am very much aware of the difference between actual science and concepts being borrowed from science to make marketing hype more"scientific" sounding.
Quite often you hear people say such things as Albert Einstein said all things were relative in his relativity theory. All things might well be relative but that is not what that theory is about, all things might also be uncertain but Heisenberg has not anything to do with it.
Some might dismiss this as pedantic, but it is just clear definition that leads to clear thinking.
rcw
I agree in principle, but we have to be aware the visual system can mislead us. Case in point: square waves from a linear-phase speaker look much better on an oscilloscope than the square waves from a speaker using an allpass crossover. The allpass function makes a complete mess out of the square wave.
However ... where the scope does not match hearing is display of nonlinear distortion. As a former Tektronix employee, we were aware that visually seeing sinewave distortion on a scope is very, very difficult. For trained observers (we're talking Tek here), 3% is about the limit, and that takes really good eyes. But 3% distortion for audio is quite bad, and clearly audible, particularly if high-order harmonics are present. Fifth-order (and higher) harmonics sound pretty bad, yet are invisible on a scope. (Scope electronics can have up to 1% nonlinear distortion, since greater linearity cannot be seen on the display.)
If we switch domains to a spectrum analyzer, then the analyzer discards time information, displaying only log-scale amplitude in the frequency domain. So a loudspeaker with very poor impulse could still look good on the spectrum analyzer.
So now we're in the soup of trying to match visual displays to audio perception. In visual terms, linear allpass distortion looks far worse than 10% nonlinear distortion, which is fairly subtle visually. But I think most of us here would agree that 10% distortion is completely unacceptable for quality reproduction, while there might be debate about audibility of linear allpass distortion.
Last edited:
- Status
- Not open for further replies.