audibility of ultrasonic forments

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I was doing some research (shudder) and ran across this article by Dick Sequerra

www.sequerra.com/common/data/ribbons.doc

Now I've had discussions over the years as to the audibility (or not) of ultrasonic sounds harmonically related to the fundamental and their influence on the quality of the reproduced (and live) sound from a stereo or musical instrument. I'm not sure I place much credence in the argument that trumpets/pianos/sax/ whatever instruments actually produce ultrasonic harmonics that influence the quality of what is heard, (James Boyk of CalTech is a BIG proponent of this position), but that's not really the thrust of my question.

Dick Sequerra (and others I've read) make a big deal over the ability of true ribbons to reproduce freqs. out to 100 Khz and more, and that this somehow explains their superior sound qualities. While I don't dismiss that ribbons/planars sound superior, i do question the logic used ....

If, as I understand it, we can't actually HEAR the harmonically related ultrasonics, but they are important in that they induce FORMENTS (i.e., difference frequencies within the audible range) that are produced from the ultrasonic harmonics, and that these must be produced by the speaker to make an accurate reproduction of the original signal. No problem if that is taken as a measure of quality (not sure if I believe it's necessary for accurate reproduction, though).

My issue is, why do the speakers need to produce these downconverted harmonically related sounds? Shouldn't they already be recorded in the electrical signal that is captured on the medium being used (not the ultrasonics, mind you, but the forments as espoused by Sequerra et.als)? This goes to the crux of tweeter capabilities, upsampling, high bit pcm, etc. arguments wrt cd's records, etc.

I mean, if these forments are present when they're being recorded, and they are indeed within the 20 - 20Khz audible range and influence the sq, why do we need capabilities beyond 20 Khz? The info has already been "decoded" during the recording (or listening) process, as I see it.

Any thoughts? Seems all this hype about the need beyond 20 Khz is driven more by marketing departments than any fundamental science.

John L.
 
I'm no expert, but I agree with you. From an "intuitive" standpoint I can't see any benefit from a tranducer doing something on its own, whether it's in the audible range of frequencies or on either side of it, that is not part of the input signal. In most schools of thought I know of this is called distortion!

Regarding Sequerra's paper, I read only subjective claims - he refers to no studies, measurements, data etc. to back them up. Hard for the inventor to be impartial about his invention.

As for ribbons being superior to other types of transducers, I humbly refer the reader to "Battle of the Non-Domes" at Zaph Audio.
 

Attachments

  • null.gif
    null.gif
    42.3 KB · Views: 322
ultrasonics

I've read several of David Griesingers papers... at least he does experiments and presents data to support his work... so I tend to lend more credence to that. Plus, it makes more sense to me as well. Not that he isn't trying to earn a living too...;)

John L.
 
I am pretty much convinced that better FR in the ultrasonic range improved the phase response in the audible range, and thus effects the shape of transient wave forms. These are audible, but may not be a great concern. Additionally, during the design of audio equipment, if this range is not accounted for during design evaluation, then adverse effects may occur. Much problems occur when the energy cannot be fully transferred to acoustic energy at the right time, and thus tends to remain somewhere in the production chain until damped out. I would not put designing to such extreme as first priority though, more accurate low end reproduction is much more appreciated by the listenters.
 
"I am pretty much convinced that better FR in the ultrasonic range improved the phase response in the audible range, and thus effects the shape of transient wave forms."

This is interesting - and I'm sure testable - so I would be a believer if I saw a test. As for effects on the shape of transient wave forms, as long as those effects create a more accurate reproduction of the input signal, then yes there could be a benefit - as long as it's audible.

I don't care about improvements if I can't hear them.
 
"I am pretty much convinced that better FR in the ultrasonic range improved the phase response in the audible range, and thus effects the shape of transient wave forms. These are audible, but may not be a great concern. Additionally, during the design of audio equipment, if this range is not accounted for during design evaluation, then adverse effects may occur."

I guess my real concern is at what price? If these hyper-response transducers are also hyper-expensive, how do we evaluate the cost/benefit ratio?

I've got planar drivers in my latest build that measure pretty flat beyond 50 Khz, but I'm not sure that means much (except to dogs, bats, etc.). Griesingers' point about the highly directional aspect of ultrasonic signals seems relevant, since one is unlikely to be physically located to take advantage of any enhancement such performance might provide.

In fact, there are commercial implementations of just such directionality to advantage...

http://www.atcsd.com/site/content/view/34/47/

Since there most likely aren't any non-electronic instruments with step risetime waveforms, and few electronic ones anyone would want to listen to, why pay exhorbitant prices for performance not needed?

John L.
 
sdclc126 said:
"I am pretty much convinced that better FR in the ultrasonic range improved the phase response in the audible range, and thus effects the shape of transient wave forms."

This is interesting - and I'm sure testable - so I would be a believer if I saw a test. As for effects on the shape of transient wave forms, as long as those effects create a more accurate reproduction of the input signal, then yes there could be a benefit - as long as it's audible.

I don't care about improvements if I can't hear them.

I quite agree that if one cannot hear the difference of extended range, then it means nothing in audio application. I have used drivers that go up to 40KHz, and it does make a difference to me. Also at a friends location, he added a supper tweeter that went so high that I could only measure the rising portion up to 48K, and he claims his system sounded like the 20KHz system I brought along until he added his super tweeter. I have also listened to various systems that have added the ELAC super tweeter. All these implementations resulted in a more precise imaging of instruments during listening.

auplater said:
"I am pretty much convinced that better FR in the ultrasonic range improved the phase response in the audible range, and thus effects the shape of transient wave forms. These are audible, but may not be a great concern. Additionally, during the design of audio equipment, if this range is not accounted for during design evaluation, then adverse effects may occur."

I guess my real concern is at what price? If these hyper-response transducers are also hyper-expensive, how do we evaluate the cost/benefit ratio?

I've got planar drivers in my latest build that measure pretty flat beyond 50 Khz, but I'm not sure that means much (except to dogs, bats, etc.). Griesingers' point about the highly directional aspect of ultrasonic signals seems relevant, since one is unlikely to be physically located to take advantage of any enhancement such performance might provide.

In fact, there are commercial implementations of just such directionality to advantage...

http://www.atcsd.com/site/content/view/34/47/

Since there most likely aren't any non-electronic instruments with step risetime waveforms, and few electronic ones anyone would want to listen to, why pay exhorbitant prices for performance not needed?

John L.

I have used super tweeters that go to 40KHz but really do not cost much if you get them in quantities. The issue is lots of people don't care about that little difference, hence sales volume cannot get to a good size, so they have to sell them at a high price until they have volume. So it's a chicken and egg problem. So if someone want's to get 10000 per year, of super tweeters that go to 40KHz, give me a call.:D

The HSS thing may be good for non-hifi applications where you only need to listen to mid frequencies. Maybe just using ultrasonic harmonics to excite the audible range?
 
But hang on. Where is this ultra high frequency coming from? What's the limit in the recording process? Microphones?

And on play back we're sampling at 44Khz. That means the maximum frequency/slew rate is a triangle wave at 22Khz off the CD, then a brick wall to anything higher thanks to the filter after the D to A converter.

Sure, it's nice to have a flat phase response to the limit of audibility, but then what HF components are available are locked to the clock in the D to A convert.

The advent of 4 channel explored the upper regions of vinyl, which some studios took advantage and produced some very clean recordings. But play them once with a standard cartridge and say goodbye to the top end.

Is sound on the latest HQ audio DVDs better in terms of phase and freq response? Unless we have a source, I cant see a reason for going much higher than audibility. Or am I missing something?

Then, all the top end distortion on a cd may keep the neighbours dog out, and drive mine insane.

Geoff.
 
Hey Geoff -

I completely forgot about that component - if it aint recorded, it aint gonna be reproduced. Most components in the audio playback chain have a frequency response near or identical to the range of human hearing 20 - 20kHz, so even if the recording picks up the ultrasonics (assuming they actually exist), chances are slim they will make it through to your speakers.

soongsc: "I have used drivers that go up to 40KHz, and it does make a difference to me." Was it a measured difference or a perceived difference?

"Also at a friends location, he added a supper tweeter that went so high that I could only measure the rising portion up to 48K, and he claims his system sounded like the 20KHz system I brought along until he added his super tweeter." With what input signal did you measure the supertweeter - a test tone, or music? If you measure with music and it does not produce 48k, the test tone means nothing. If the tweeter makes 48k with music input you then need to test the music input signal before the tweeter and see if that same information is there - now we're getting somewhere. If that information is NOT in the music input signal, your tweeter is producing its own artifact - to my knowledge such things have never been shown to improve musical reproduction. As for your friend's claim - well he makes my point - a claim is just that.

"I have also listened to various systems that have added the ELAC super tweeter. All these implementations resulted in a more precise imaging of instruments during listening."

Imaging, to my knowledge, is not a measurable parameter of loudspeakers - please correct me if I'm wrong - I've just never seen data on it.

I don't mean to be coming down on you - I just see no science here. Again - I'm just waiting for real data.
 
Administrator
Joined 2004
Paid Member
It's funny - even if I can hear 'em, I don't like 'em. Everything above about 15K sounds artificial to me. It might sound nice, but I just don't hear that stuff in nature.

Those frequencies exist in nature, but they always seem too pushed forward in Hi-Fi. So you can "hear" them. Like surround sound speakers that are pushed up too loud, so that you'll "know" they're there.

Everyone's ears are different.
 
Ultrasonic frequencies from natural sources as such don't sound, because we can't actually hear a tone. They also don't occur solo in nature, but as part of a sound or noise spectrum. If you force yourself to hear an artificial ultrasonic tone, it sounds alarming and distracting, as high power in the highest frequencies is a natural warning signal of something very quickly moving - very closely...

But we perceive ultrasonic frequencies (with the very low sensitivity of our eardrum to them), everyone with a healthy chochlea can try this by placing a transducer on the bone behind the ear.

What our eardrum makes out of it is basically intermodulation/convolution residuals that 'intertwine' with the rest of the original spectrum. The eardrum basically 'demodulates' parts of this ultrasonic content. This is what can also be done by any other transducer with limited bandwith, as can easily be tried with a regular dome tweeter, a 96kHz sound card and a software function generator!

I for one can clearly distinct the sound of a bird from the recording of the sound of a bird, for example. I can also hear well above 25kHz with my inner ear, but not really above 15kHz with my outer ear.

The more important question is: content. Is there any ultrasonic content in anything mankind has recorded in the last couple of decades? I think you need an especially prepared recording to try to perceive the difference... Unless we know how to do this - and have ultrasonic excitation in the recorded source to speak of - we won't find out for ourselves.

OTOH, newer SACD or DVDA recordings that have been produced with bandwith in mind are reported to sound different and more open. But as studio engineers reveal, this is achieved mainly by using studio production techniques (i.e. enriching the spectrum with harmonics of a voice or instrument).

Cheers,
Sebastian.
 
I don't think I'm out here to convince anyone, just to share some experience. But throughout the audio chain, we have a series of phase shifts at different stages, the only technically reasonable thing to do is to minimize it at each stage. Since speakers seem to be the most nonlinear part in the reporduction chain, it seem logical to start here. Mics may be the next to look into, but probably the mic builders would like to look into that. We are seeing higher bandwidth in other parts of the chain, but since the effects cannot be well demonstrated through lower bandwidth speakers, the trend is not catching on. Really a pity. That means we have to let MP3s win because most don't care about the difference?

Some people may think SACD and DVD Audio really don't sound better than current CDs or LPs, I just think we don't have the speakers and probably other parts of the playback chain good enough to reveal the potential of such standards, and the increased bandwidth might even be exciting unwanted harmonics of the low bandwidth systems. Some things to think about. In some of my measurements of super tweeters, the tope end roll-off seems well behaved, but when I measure drivers with a lower cutoff frequency, I sometimes see a rise in FR after roll-off indicating resonance in the ultrasonic region.
 
Sound repro is all about accepting compromise. 20 to 20K is beyond audible hearing through normal sound propagation, i.e. via air at sea level.

It's an industry standard laid down many years ago, which at the time, was a challenge to the manufactures. Today it's a stroll in the park.

The normal hearing range of a young person is from about 32Hz to 18K. Below 32Hz we tend to feel rather than hear. Above 18K gets filtered by the outer ear, which functions as a horn.

As we mature, the upper threshold of hearing diminishes.

I have been blessed with good hearing. At the age of 17, my upper limit was 17.8 Khz. At 25 it was still 17.6. At 40 I suspected trouble. A trip to the doctor, and a good flushing, restored it to 16K. Now at 53, its down to about 13K. Am I sad about the degrading of my second most treasured sense? No. Violins on CDs actually sound sweeter. I am not hearing the high distortion content available.

It has been a while since I looked at the specs of CD players. When I did last look, they were still not quoting IMD at 10Khz, for the same reason those figures were not published for magnetic tape. The figures would frighten the audiophile into the woods.

I think, after many years of listen to many systems, we are better of improving the performance in the region we can plainly hear, and where the musical qualities are.

Our biggest hurdle is size. We have had the technology for years. Those systems need a fair lump of floor space, and thus low WAF.

One of the most highly respected speaker systems ever made requires about a square metre of floor space, and stands about 1.8 metres high. It won't go to 20K, or reach down to 20. But where it does go, it does without much harmonic distortion. What I think we should be aiming for, is those qualities in a couple of cubic feet.
 
Geoff H said:
...

I think, after many years of listen to many systems, we are better of improving the performance in the region we can plainly hear, and where the musical qualities are.

Our biggest hurdle is size. We have had the technology for years. Those systems need a fair lump of floor space, and thus low WAF.

One of the most highly respected speaker systems ever made requires about a square metre of floor space, and stands about 1.8 metres high. It won't go to 20K, or reach down to 20. But where it does go, it does without much harmonic distortion. What I think we should be aiming for, is those qualities in a couple of cubic feet.


Good point. These are the places that show most significant improvement.
 
Whether there are any ultrasonic sources does seem to be the most important question to answer before exploring this, but I wanted to comment on the human perception issue.

In previous online arguments on this issue I've seen people state that 20-20k are the absolute limits of human perception. The argument was that our hearing is much like our vision where we can see only a limited range of frequencies, we can't see infrared or ultraviolet, can't take x-rays with our eyes etc.

Reading that I couldn't help but think of what a flawed analogy it was. For example when we step outside on a sunny day and notice how blindingly white a friend's t-shirt is, the optical brighteners in his laundry detergent may be playing a role. These compounds absorb UV light and emit blue light. (http://en.wikipedia.org/wiki/Optical_brightener) And the infrared? Well it certainly would be a different experience to stand out in the brilliant sunshine in Antarctica than miami. My point is only that while we may not directly see these bands they can still be a part of the overall experience of being somewhere.

Back to audio, lets take a live trumpet as an example. I've read that it produces sound into the 50k-80k range. We don't hear that directly, but it seems reasonable that it's going to interact with the environment in some way we do detect. It could certainly excite lower resonances that we do hear and who knows, it might tickle the hairs on the back of your neck. :D

I've mentioned here before that cymbals sound the most "wrong" to me when played back. When I'm playing music with a drummer the cymbals have a very complicated sound, it sounds like many different pitches and beats going on and a feeling of beat frequencies happening right at my ears. When played back it sounds like the familiar splash of white noise you hear on recordings.

In the low end anybody who's listened to headphones can attest to the difference between hearing the bass and feeling it. A kick in the chest may be only 1hz but you're still going to feel it.

So that's my bit, I can't say for sure that frequencies outside 20-20k would enrich the experience of listening to recorded music, but I'd love to see a serious exploration of the issue.
 
poptart said:
...

I've mentioned here before that cymbals sound the most "wrong" to me when played back. When I'm playing music with a drummer the cymbals have a very complicated sound, it sounds like many different pitches and beats going on and a feeling of beat frequencies happening right at my ears. When played back it sounds like the familiar splash of white noise you hear on recordings.


I quite agree that cymbals are very difficult to reproduce. This seems more of a matter of energy storage in the speaker system though, and more easily explored through Cumulative Spectral Decay characteristics.

Generally, the higher frequency that can be produced, the more linear the system is behaving at small signal reproduction. I would love to see a driver that does well from 500Hz to 50KHz. In the next few years, I think a good driver from 100Hz to 30KHz is possible. There is something funny going on around the 20KHz~50KHz region that I haven't quite figured out yet, but seems really complicated because everything seems to start resonating at this range.
 
Regarding the cymbals: There are two major concerns about cymbals and sharp 'S' sounds.

First, speakers can't reproduce them well and second, recording systems can't process them well.

The industry accounts for that. Microphone pickup techniques and EQ adjustments on the FOH and recording consoles were developed to get around the bandwith and power limitations of recording and playback. That's why high frequencies sound strange and artificial when compared to the real instrument. When done without these limitations in mind, a cymbal recording can sound as loud, cracking and deafening like for real - it depends on (low distortion) bandwith and sound pressure the system is capable of.

It's not really a limitation of your (or my) speakers, but the average playback system determines the production quality goals... :(
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.