Putting the Science Back into Loudspeakers

[OzMikeH]

to continue the vision analogy. a high power camera flash can produce pulses of light, some as narrow as 1/10,000th of a second. we see the flash very easily. We all know we no longer see flickering at 20Hz.

So why believe it's impossible for a person to detect certain acoustic effects that are clearly outside our normal timescales of detection?
You can't hear a 30kHz tone but you may still be able to detect an effect that is of similar duration such as a phase offset. You may not even be conscious of the effect.
Slam a door or trigger a strobe, which is more likely to startle a person?

the thing with the crossover is they roll off. you have multiple drivers producing the same frequencies, even multiple parts of the same driver when you add a whizzer cone. you have sound coming from 2 places at once with a slight distance difference and therefor phase offset. it's been made blurry. I can see how it's possible even if I know the timescales are microscopic.

Phase accurate: why do we accept such terrible speakers? No idea but 2 dimensional TVs are still being sold.

 

[Pano]

a high power camera flash can produce pulses of light, some as narrow as 1/10,000th of a second. we see the flash very easily. We all know we no longer see flickering at 20Hz.

A 1/1,000,000 has been done for years, I think Edgerton did it. You can see it. And at 20Hz you can see a flicker, for sure. If you are close to the light source, you need to be above ~70Hz.

Anyhow, you can sure see a flash that is much, much more brief than what the eye sees as continuous . I agree with you Mike. That's due to "persistence of vision." The chemicals in the retina decay slowly. But - do we have "persistence of hearing?" It's a good question.

The reason you can see an ultra brief flash of light has to do with the slow chemistry of the eye. Can/could the ear work in a similar manner?

 

[CeramicMan]

So why believe it's impossible for a person to detect certain acoustic effects that are clearly outside our normal timescales of detection?
You can't hear a 30kHz tone but you may still be able to detect an effect that is of similar duration such as a phase offset.

Well maybe we're just not asking the right questions?

For example, what might usually be involved in a stereo localization test? Multiple speakers arranged in an orb-shape surrounding a person? (Such as the picture in: http://www.aip.org/pt/nov99/locsound.html ) What about the sounds produced? Natural wildlife sounds?

If a test uses broad-spectrum pulses then the ears get a large amount of information all in one tiny fraction of a second. All of the necessary diffraction and shadowing information is given straight away.

I've heard of tests suggesting the existence of "virtual lobing", where one ear hears one frequency while the other ear is fed a slightly different frequency. A beat frequency is sensed despite the lack of physical interference between the sounds.

If a test uses long-lasting sinusoidal waves, similarly, the brain can statistically work out the coordinates of the sound source because the 'pulses' may repeat over a 1000 times in one second.

Perhaps it would be a good idea to test that "post processing" system by limiting both the time and frequency information that the brain receives and finding the point at which it starts to break down? For example, try to localize short "blips", low frequencies, or focus on peculiar combinations that form phase differences between our ears such as 360 degrees?

Rather than scientific flattery, why not ask how bad our hearing might be in some situations? See at what point human processing starts to make mistakes? After all, common sense suggests that if we want to fool the human ear we should start with the most obvious weak points.

 

[Pano]

Multiple speakers arranged in an orb-shape surrounding a person? (Such as the picture in: http://www.aip.org/pt/nov99/locsound.html )

Wow! Great link! Thanks. Reminds me of a "sonic tree" at Radio France that had over 200 individually addressable speakers. Crazy sound.

There seems to be a lot of good info in the link - I want to stay up and read it over a few times, but it's time for bed. ZZZzzzzzz

 

[phase_accurate]

Just a sidenote: I will most probably attaend an AES meeting in August with guest speaker John W. who will be demonstrating some of his constructions.

Regards

Charles

 

[DCPreamp]

to continue the vision analogy. a high power camera flash can produce pulses of light, some as narrow as 1/10,000th of a second. we see the flash very easily. We all know we no longer see flickering at 20Hz.

So why believe it's impossible for a person to detect certain acoustic effects that are clearly outside our normal timescales of detection?
You can't hear a 30kHz tone but you may still be able to detect an effect that is of similar duration such as a phase offset. You may not even be conscious of the effect.
Slam a door or trigger a strobe, which is more likely to startle a person?

the thing with the crossover is they roll off. you have multiple drivers producing the same frequencies, even multiple parts of the same driver when you add a whizzer cone. you have sound coming from 2 places at once with a slight distance difference and therefor phase offset. it's been made blurry. I can see how it's possible even if I know the timescales are microscopic.

Phase accurate: why do we accept such terrible speakers? No idea but 2 dimensional TVs are still being sold.

I’m pretty sure part of why we can see a flash of 1/10,000th of a second is not just because of persistence-of-vision, but also has to do with the amplitude of the light pulse. View the same 1/10,000 pulse with an LED limited to 100 mcd and see if it’s detectable. Plus, and I really don’t know much about camera flashes, but is there a thermal time-constant, and therefore decaying light output, to the cooling of the xenon bulb, or it the light output really -3dB at 1/10,000th of a second?

With the flash analogy, it would be like listening to a 30KHz tone at 170dB. I’ll bet you could hear it, or at least perceive it, albeit with some gnarly side-effects.

 

[graaf]

Just a sidenote: I will most probably attaend an AES meeting in August with guest speaker John W. who will be demonstrating some of his constructions.

Regards

Charles

hello again!

what about this demostration? How was it?
Have You attended the meeting?
Has anybody?

best regards,
graaf

 

[thoriated]

Re. 'Putting the Science Back into Speakers', very interesting comment by the author regarding that ceramic magnet speakers cannot reach the 16 bit resolution of a CD.

This is not to say that a ceramic magnet speaker cannot resolve the full dynamic range of many recordings on CD which use significantly less than the full 16 bit resolution on average.

I post this just to be fair to ceramic magnet speakers, not because I like their sound very much. In fact, I just replaced the ceramic magnet woofers in a pair of my home built speakers with an alnico version of the same driver (the HF driver already was alnico) with a resultant dramatic improvement of the inner detail and soundstaging of good recordings.

 

[gedlee]

An interesting read. Not quite as "scientific" as I would like to see, but in general I can agree with his conclusions, even if I don't agree with all the data that he used to get there. His points about loudspeaker diffraction being a masker for artifacts is one that I have been saying for many many years now and a key reason why my loudspeakers are received so well. The room is also a diffraction source, which he fails to mention. I too could not hear MP3 and WMA artifacts until I built the Summas and then I found that I was bothered by these codecs. I only use lossless encoding now.

Everyone should read this paper, but don't assume that everything said is absolutely correct - at least the methodology for doing the subjective tests is not sufficient to draw conclusive results.

His psychoacoustics is a bit coarse and not exactly clear, but generally factual.

His points about magnets are a bit weak, and they ignore the fact that any good driver should have a shorting ring which dramatically reduces flux modulation. All of the drivers that I use have shorting rings - and ceramic magnets.

Graaf, thanks for the link.

 

[phase_accurate]

what about this demostration? How was it? Have You attended the meeting? Has anybody?

The demo was fun and I could indeed get some insight into his constructions. Technically speaking most of the info could already be found within his papers - but it was funny to hear him life since he is very humorous and he makes funny comparisons like:

"The grill on a modern Rolls Royce doesn't have any function at all whereas in earlier days it had the function of the radiator but it is still there because people immediately recognise it as a Rolls. Speakers are most always rectangular and people easily recognise them as speakers - so we made a speaker that LOOKS rectangular but that actually isn't - just because people can recognise them as speakers"

It was his Legend Line of speakers that he was talking about in this case. They are actually built from cylindrical elements and they are omnidirectional by the use of deflectors. In their corners they have four poles over which a sock is pulled and that gives them the look of a fairly ordinary rectangular box.
He was taking two models of these with him but mostly used the smaller model for demoing. Although this has respectable performance for its size I assume that it would have been more fun if more demos had been done with the larger legend.

One thing that I missed was the Cabar studio monitor. But maybe there will be an opportunity to hear this one also one day.

There is another funny one that just jumped up: When he had finished his speech he was given a small cardboard box by the Swiss AES section vice president. He then asked if this was something that one isn't supposed to take on a plane in one's hand luggage. The vice president said yes. John W. then answered: "Well - that doesn't matter ! I don't fly anymore anyway - since I am always considered a terrorist by the security staff !"

Edit: forgot the link:

http://www.swissaes.org/nl/aes_newsletter_110.pdf

Regards

Charles

 

[graaf]

An interesting read.
(...)
Graaf, thanks for the link.

You are welcome

You have said sometime ago that

the point of my being here is to educate people so that audio won't continue on in its "dark ages"

I think it is also Mr Watkinson's "mission impossible" - "dragon slaying" as He calls it

some more entertainment with food for thought (speaker-related or general audio-related) selected from Mr Watkinson’s writings for Resolution Magazine
(http://www.resolutionmag.com/slaying.html ):

CROSSOVERS

Like all industries, aviation has its dragons to be slain, top of the list being the popular view that Bernouilli’s theorem explains where lift comes from. To hold this view, you have to neglect the fact that Bernouilli himself made it clear that his theorem doesn’t apply to that case. However, aviation is a money-where-mouth-is discipline, whereas audio isn’t. Thus on the whole there are fewer aircraft that don’t fly than loudspeakers that don’t reproduce anything like the original sound. A lot of this has to do with the crossover. (...)
in order to get a pair of complementary signals it is necessary to use a subtraction stage in the crossover. This is fundamentally impossible in a passive crossover because passive circuitry can’t subtract. Thus passive crossovers are not crossovers at all, but a pair of filters, one high pass and one low pass, whose turnover frequencies happen to be similar. The outputs of these crossovers cannot and do not sum to the original waveform. (...)
With the speakers set up normally, assess the quality from the normal listening position, then go just outside the open door of the listening room and listen again. In most cases the sound quality will be noticeably coloured. It shouldn’t be. A good loudspeaker will have a sufficiently accurate response off-axis that it produces an uncoloured sound field that will be acceptable in the next room. Listening from the next room, a very goud loudspeaker playing an accurate recording might make you unsure whether the source was a speaker or the instruments themselves. Very few loudspeakers can pass that test. It is important to test speakers for realism with suitable recordings. Any instrument that has a spectrum straddling the crossover frequency is a good candidate. The cello is one such instrument. Female speech is also a good signal for testing. (...)
Don’t use passive crossovers. They are fundamentalny incapable of producing waveforms that sum to the original.

GETTING THE BASS RIGHT

Once a linear-phase woofer is available, it becomes possible to make some meaningful experiments in bass management. It is soon revealed that a separate subwoofer with all LF mixed into it is inferior to each loudspeaker emitting its own LF. This is the reverse of conventional wisdom, which suggests that subwoofers can be put anywhere. If you look at the surround sound research which showed subwoofers could be put anywhere, you will find that the subwoofers used were not phase linear and nor were the crossovers. In other words the timing of the sound from the subwoofer was subject to so much linear distortion that it would be wrong wherever the subwoofer was put.

THE SPEAKER CABLE SNAKE

Bi-wiring with lawn-mower cable from the electrical shop will always out-perform single wiring with audiophile cable and at lower cost. [BUT "Bi-wiring can improve quality, but only if the speaker is designed for it."!] I have been recommending lawn mower cable for a long time and am often asked what mysterious properties it has which ordinary cables lack. Finally, I am prepared to divulge this closely guarded secret: lawn mower cable has two cores whereas ordinary mains cable has three.

EQUIPMENT SPECIFICATIONS

Distortion is a factor that is seldom described by specifications. It comes in two flavours. Linear distortion is where there is no change to the spectrum, but some frequencies lag behind others in time. Harmonic distortion is where extra frequency components are found in the spectrum due to a non-linear transfer function. Linear distortion is relatively inaudible on continuous sounds, but it destroys realism and imaging information in transients. It’s typically not found in specifications. (...)
The audibility of the artefacts depends on the information capacity or resolution of the loudspeakers. Sometimes I wonder if I have become too cynical, but then along comes another wall of physics-defying propaganda. The current advertising for DAB digital radio makes a big play of the low-noise reception and slags off FM radio for buzzes and crackles. My FM radio doesn’t buzz or crackle, nor does it suffer from MPEG compression artefacts. My first exposure to a real live DAB radio in a real live domestic environment was a massive disappointment. Instead of the rich pure sound promised on the box, it sounded to me like it had a sock in it. Maybe they were relying on the old meaning of pure, which was a substance used in leather tanning whose primary ingredient was dog droppings.

GROUNDING

It is impossible to build a high-resolution audio system using unbalanced connections. The fact that high-end hi-fi systems frequently use phono connections says it all.
The use of audiophile phono cables with massively over engineered gold plated plugs and exotic wire makes about as much sense as gold-plating a cow pat. From a linguistic standpoint, the term audiophile is derived from the Greek philos, meaning love. However, in the conventional English usage, it can mean love of an unwholesome or perverted nature. I think there should be a register of audiophiles so we can keep our children away from them.

DESIGN

Our educational system does not teach how to design. Most of education is about analysis, taking things apart if you will, whereas design is synthesis. (...)
After a technology matures all that can be done is to make it cheaper. Regrettably an increasingly common driver for new design is the doubtful pursuit of novelty. For some reason the consumer is supposed to equate ‘new’ with ‘better’. Maybe when technology was evolving rapidly that equation held. Today with depressingly regularity ‘new’ means untested. (...)
As the company grows further it becomes too valuable to take risks and its actions become increasingly conservative. At this point the creative founders have become bored and leave (Usually with handsome pay-offs. Ed). The company is now run by accountants who have no knowledge of the technology. The only thing they understand is how to make it cheaper. That is the start of the decline, because making things cheaper isn’t design; it’s production (...)
In many cases the best designs came from people who couldn’t buy what they needed and had to make their own. Being your own customer is a powerful design tool as if it’s not right it gets changed. Today the gulf between designers and users has never been wider

and my personal favourite:
DOES HUMOUR BELONG IN PRO AUDIO?

One of the most powerful uses of humour is to find out whether people are genuine or not. People who know what they are talking about and are not seeking to mislead respond to teasing quite differently to those who don’t understand their topic or who have a hidden agenda. The latter often take themselves far too seriously. One of the functions of the traditional court jester was to tease visitors in order to establish their bona fide. I have been acting the court jester for a long time now and I have learned that the people who matter don’t mind and the people who mind don’t matter .

The room is also a diffraction source, which he fails to mention.

apparently He is of an opinion that room reflections are not a problem at all
here is what He writes about His Cabar studio monitor:

Use of omni-directional HF radiation in the Cabar contributes to cost saving by reducing or eliminating the need for acoustic treatment of the listening area

best regards,
graaf

 

[graaf]

It was his Legend Line of speakers that he was talking about in this case. They are actually built from cylindrical elements and they are omnidirectional by the use of deflectors.

sounds very much like Time Domain Yoshii 9 (cylindrical omni - but Yoshii 9 are one-way fullrange and don't have any deflectors)

One thing that I missed was the Cabar studio monitor. But maybe there will be an opportunity to hear this one also one day.

In fact I was hoping that this was the demonstrated model

There is another funny one that just jumped up: When he had finished his speech he was given a small cardboard box by the Swiss AES section vice president. He then asked if this was something that one isn't supposed to take on a plane in one's hand luggage. The vice president said yes. John W. then answered: "Well - that doesn't matter ! I don't fly anymore anyway - since I am always considered a terrorist by the security staff !"

what can I say? the guy is just great!

best!
graaf

 

[graaf]

ps.

more on Cabar which is "a single unit active stereo monitor loudspeaker system" (image attached):

Polar Response
The enclosure shapes are not just the whims of external design fancy, but combined with the transducer systems to produce an acoustic polar diagram that is a dipole at high frequencies and gradually shifts into an omni-directional pattern for the low frequencies. As a result oif carefully engineering, the polar diagram is very smooth and without any lobing: The reverberant sound field excited by a loudspeaker within a room is in effect the frequency response of the indirect emissions from the loudspeaker: If this indirect, or polar response is not clean and even, then the reverberation heard will be coloured and will add clutter to the overall imaging. A secondary benefit of a good polar response is the reduced requirement for acoustic treatment for the room: normal room reverberation in itself does not reduce the ability to listen accurately, so long as the reverberation is excited in a mimic of the natural trigger, in other words triggered by a smooth and even polar response.

Timing
All loudspeaker systems should be aperiodic, that is, without resonance. Whilst resonance may assist in extending the low frequency performance of very cost conscious designs, it comes with a penalty - the delaying of the low frequency rise and stop timing, which can only confuse the overall sound image. The Cabar system has no resonance and therefore no lag, no delay and no hangover. This combined with the very close spacing of the high and low frequency drivers means that the acoustic output of the system is truly time aligned in every sense, for the start, duration and termination of every signal.

Positioning
As with all dipole loudspeakers, the Cabar will perform best when exciting the natural reverberation of the listening room - to do this the Cabar should be some distance from the walls or close boundaries. Ideal positions are mounted on the penthouse top of an audio mixing or video editing console, suspended from the ceiling or free standing, supported on a suitably rigid stand. It should be noted that the Cabar, like most dipole loudspeakers, does not sound at its best in a non-reverberant room.

Watkinson's design priorities are very much like Yoshii's of Time Domain

I only wonder in what sense the Cabar is "dipole speaker"? Certainly not in a conventional meaning of the term, perhaps more like Swiss Stereolith: http://www.stereolith.ch/

best regards,
graaf

 

[phase_accurate]

Each of the wide-range drivers that you see on the picture have a "brother" on the back.


Regards

Charles

 

[graaf]

Each of the wide-range drivers that you see on the picture have a "brother" on the back.

I see - thanks for clarification
so these are "tweeters"?
are the "woofers" at the ends firing to the sides?

best!
graaf

 

[phase_accurate]

I think you are doing these drivers injustice by calling them tweeters.

It is difficult to recognise on your picture but the narrow shiny part between the red center tube and the black end-pieces is in fact acoustically transparent. The woofers are using the ends of the red tube as "box" and their sound is emanating through this acoustically transparent part. The electronics occupy the middle part of the tube.

All in all the device looks a bit like a stylish ghetto blaster.

Regards

Charles

 

[graaf]

I think you are doing these drivers injustice by calling them tweeters.

well, exactly in order not to do them injustice I have called them " "tweeters" " with ""
"tweeters" because they are high-passed

these are Bandors, aren't they?
the "woofers" are also Bandors?

It is difficult to recognise on your picture but the narrow shiny part between the red center tube and the black end-pieces is in fact acoustically transparent. The woofers are using the ends of the red tube as "box" and their sound is emanating through this acoustically transparent part. The electronics occupy the middle part of the tube.

thanks for detailed description

best!
graaf

 

[phase_accurate]

Don't know about the drivers in the Cabar but I asked him about the drivers in the flat speaker that you can see in the AES newsletter and he said that the only thing that it has in common with a Bandor 50 was the surround.

Regards

Charles

 

[graaf]

Don't know about the drivers in the Cabar but I asked him about the drivers in the flat speaker that you can see in the AES newsletter and he said that the only thing that it has in common with a Bandor 50 was the surround.

Regards

Charles

thanks for the answer

there is one thing I don't understand about Cabar:
JW writes about:

omni-directional HF radiation in the Cabar

therefore I thought it must have something in common with Stereolith

but elsewhere JW writes that the same Cabar

is a dipole at high frequencies and gradually shifts into an omni-directional pattern for the low frequencies

so - and omni or a dipole?
Charles - perhaps You can clarify? as someone who actually have seen the loudspeakers

best regards!
graaf

 

[graaf]

some interesting audio-related science:

"Sound-Source Recognition: A Theory and Computational Model": http://sound.media.mit.edu/Papers/kdm-phdthesis.pdf

what makes a loudspeaker a HiFi loudspeaker?

"Spatial Hearing with Simultaneous Sound Sources: A Psychophysical Investigation":
http://ses.library.usyd.edu.au/bitstream/2123/576/1/adt-NU20041221.13524302whole.pdf

everything You ever wanted to know about sound source localization and imaging

best,
graaf