Sorry if I haven't been paying enough attention, but are we talking 105dB in the throat of the horn or a meter in front? I know that the SPL in the throat of most horns is loud enough for very significant THD and 3% 2rd harmonic is not all that rare. Similar amounts of IM should also be likely.
Regards,
David S.
Regards,
David S.
Yesterday I measured the throat SPL in the Maltese horn at 149.1 dBA when it was producing 110 dB at 2 meters.
That is about as loud as the horn can go and still sound clean.
The recordings and test details of the VHF generated sub harmonic tones including pictures are in #179 #182.
The 104 and 103.9 dB tones were converging at 22.5 inches from the horn mouths.
Although the results are rather amazing (unbelievable to some
) one can apply the inverse square law and find that the effects probably have little contribution at normal listening volumes and distance.Nonetheless, if my damaged hearing can hear the effect at a loud volume, those blessed with "golden ears" may hear the effect at normal volumes with speakers capable of ultrasonic range.
Art
Last edited:
Hi Art,
I finally got a chance to listen to your samples, which are fascinating.
As I understand you are using two narrow angle horns, with an HF tone being fed to each. This means no IM at the drivers because any mixing would be outside the speakers on the way to their common receiving point at the microphone. Levels are about 104 at 2 feet.
The most telling one is where your hand is waved in front of the mic and the difference tone dissappears.
I would agree that the hand would shadow tones up near 20k but not a 1k difference.
This means, I think, that the nonlinearity (that must be present) is on your side of the hand position. In other words, the intermod has not yet occured by the time the sound hits your hand. If it had then the 1k, or 100Hz difference would already exist and would diffract around your hand.
I think that this also means that the intermod is being caused by your microphone. I see that this is a high quality B&K, but I don't see any way around the conclusion that the intermod products must be made near the mic position and there is no reason why the air would be nonlinear near than your hand and linear farther than your hand.
The fact that you hear it in person might just mean that your ear also forms a similar amount of intermod. i.e. the mic hears it and you hear it because both mechanisms cause a similar amount of intermod and create the difference tones.
I would return to my previous comment that you should try the mic nearer and farther away. If intermod percentage goes up as you get nearer (and level at the mic increases), then it is mic distortion phenomonon. If it were an air nonlinearity issue, then as you got closer the the distortion wouldn't change or might be less (?) from less distance through the nonlinear medium.
Regards,
David
I finally got a chance to listen to your samples, which are fascinating.
As I understand you are using two narrow angle horns, with an HF tone being fed to each. This means no IM at the drivers because any mixing would be outside the speakers on the way to their common receiving point at the microphone. Levels are about 104 at 2 feet.
The most telling one is where your hand is waved in front of the mic and the difference tone dissappears.
I would agree that the hand would shadow tones up near 20k but not a 1k difference.
This means, I think, that the nonlinearity (that must be present) is on your side of the hand position. In other words, the intermod has not yet occured by the time the sound hits your hand. If it had then the 1k, or 100Hz difference would already exist and would diffract around your hand.
I think that this also means that the intermod is being caused by your microphone. I see that this is a high quality B&K, but I don't see any way around the conclusion that the intermod products must be made near the mic position and there is no reason why the air would be nonlinear near than your hand and linear farther than your hand.
The fact that you hear it in person might just mean that your ear also forms a similar amount of intermod. i.e. the mic hears it and you hear it because both mechanisms cause a similar amount of intermod and create the difference tones.
I would return to my previous comment that you should try the mic nearer and farther away. If intermod percentage goes up as you get nearer (and level at the mic increases), then it is mic distortion phenomonon. If it were an air nonlinearity issue, then as you got closer the the distortion wouldn't change or might be less (?) from less distance through the nonlinear medium.
Regards,
David
OK, I've done the test with 18KHz in one tweeter, 20Khz in the other. Playback and recording were 96Khz sampling rate to keep things clean.
Results:
Results:
- With one or the other tweeter playing a single tone, distortion is low. I just see H2.
- With both tweeters playing - 18K in one, 20K in the other, there is a definite 2K tone showing up at about 69dB below the fundamentals.
- moving the mic back to twice the distance of the first recording brings the 2K tone to about 72dB below the fundamentals. Or 3dB less.
The narrow dispersion Maltese horns I used concentrate the VHF in a much smaller convergence zone, where the microphone was located.
Pano's tweeters, being much wider dispersion, make the effect work over a wider area, but at a reduced level.
Check out the link in my next post for technical details plentiful enough to satisfy most any propeller head :^).
Art
Last edited:
Glad you got a chance to listen to the two short recordings. There are several minutes of recordings, but those two give one a good impression of what I could hear in the overlap region of the horn.
The microphone was placed on axis in the convergence zone, the beat tone was at maximum there.
The tone was audible in front of the microphone and for a ways behind the microphone, varying in level with the inverse distance law.
A 60 year old friend happened to stop by on a return road trip (he also has hearing loss) and was able to hear the tones with his “good ear”.
This effect is not something new, or of my own creation, you can read more in:
Sound from ultrasound - Wikipedia, the free encyclopedia
Dr. F. Joseph Pompei of the Massachusetts Institute of Technology published a paper in 1998 (105th AES Conv, Preprint 4853, 1998) describing a working device that reduced audible distortion essentially to that of a traditional loudspeaker.
According to the Wikki article, Pompei says the demodulation process is extremely lossy, with a minimum loss in the order of 60 dB from the ultrasonic SPL to the audible wave.
This is far greater than the 32.4 dB loss in my experiment, but at distance, air loss would rapidly reduce level at 50 kHz where his system seems to be operating.
As I mentioned before, I had tried this experiment about 10 years ago, but the transducers were not capable of the level needed, and I did not fully understand the concept, so did not spend much time figuring out why it did not work.
A JBL 2402 driver loaded with a 2405 diaphragm results in a rising HF response, peaking at some point above 16 kHz . I say “some point” because I only tested one using a hand held octave RTA some time around 1981, the driver’s slope went up at around 6-12 dB per octave, the (last) 16 K LED was at the top of the response curve. At the time, I was a bit flabbergasted, a normal 2402 “bullet” rolls off in the high end.
These drivers, painted an olive drab, were used for ultrasonic traffic signal control in rural areas.
One driver emitted an ultrasonic tone, the other , across the road, wired as a microphone received it.
When a vehicle dove by, it would break the sound beam, causing the lights to change in their favor.
The devices were mounted low enough that any typical vehicle would trigger the light change relay, the rural environment led to wholesale theft of the drivers. My “inside information” came from a PA company owner who “liberated” dozens of these drivers for VHF extension in his system.
Electrotec also used a pair of JBL 2402 driver loaded with a 2405 diaphragm in their PA cabinets from around 10K up.
Those drivers would work well for this “sonic holography”, although it seems that response to 50 kHz would be desirable, the 20 kHz range still can cause hearing damage at the levels needed to create the sub harmonic tone.
I do still seem to have more tinitis than usual after spending much of a day mixing the two tone VHF recordings.
My experimentation regarding this phenomenon is pretty much done, others have preceded me in explanations and working devices, I have no desire to duplicate their efforts.
A good designer like yourself, with connections to a major transducer company, may be able to exceed those pioneer’s results.
At any rate, neither my ears or the B&K microphone were distorting during the recording of VHF tones
.Art
Last edited:
Even wide dispersion tweeters are pretty directional at 18-20 kHz, unless they are about 1/4" diameter, in which case they are not capable of enough SPL for the beat tone to be audible.
"My Radio Shack meter said 92dB at 3.5" for the first measurement (both tones playing). "
Were the tweeters toed in to the microphone?
What was the level of each tone individually at the microphone location?
Two non-correlated tones add about 3 dB to the measurement level, 89 dB at only 3.5" is not much level, though it could be far higher because of the RS "A" scale rolloff.
The RS meter on a dBA setting will be rolled off that high, but your measurement mic could give a relative idea of SPL.
Most likely the close distances you used preclude your ear from getting in to the convergence zone.
Seems the RS meter does not adhere to the A or C scale weighting in the HF, it should be 9 or 10 dB down at 20K.
I'll compare my old RS analog SPL meter to Smaart and the Terasonde Audio Toolbox response on "A" scale next time I do some HF testing.
With all the ringing in my ears the recent compression driver and VHF tests have caused, that may be some time from now, must give the ears a rest...
Attachments
At any rate, neither my ears or the B&K microphone were distorting during the recording of VHF tones
Art
I'm not sure you understand what Earl and I have been saying (over and over). The fact that you hear and record the difference tones are proof that something was distorting. Those difference tones can only occur via nonlinearity in the system. They are not beats, they are intermod products.
Tis the nature of nonlinear systems that harmonics are generated for single tones and new products at other mathematical frequencies are created for multiple tone stimulus.
Note that the order of the nonlinearity is important. For HF two tone tests you will get adjacent sidebands from 3rd order nonlinearity and sub tones from second order. It goes like this: (1) x 20k - (1) x 19k = 1k There will also be a 39k component. The 1 and the 1 add to 2 for 2nd order nonlinearity required. 19kHz and 20kHz give 18 and 21 from third order nonlinearity because (2) X 19K - (1) X 20K = 18K, and (2) x 20k - (1) X 19k = 21K The 2 and the 1 add to 3 and are evidence of 3rd order nonlinearity.
Not proof but at least a good way to remember it.
I remember the bullet tweeters on traffic signals. That was a little before my time at JBL, but they still talked about that as a lucrative contract. Imagine doubling your business due to widespread theft of your product!
David S.
If I mix the 18K and 20K in software, there is only* the two tones in the mixed result. If I mix them in the air, there is the the 2K spike. Is that the air, or microphone? How can I tell?
I did move the mic to double the distance and saw a 3dB relative drop in the 2K spike. The mic? Or something else?
*Below -110dB there were a lot of artifacts, I guess not even a software mixer is perfect.
I did move the mic to double the distance and saw a 3dB relative drop in the 2K spike. The mic? Or something else?
*Below -110dB there were a lot of artifacts, I guess not even a software mixer is perfect.
The nature of the air non-linearity is explained quite well here:
Sound from ultrasound - Wikipedia, the free encyclopedia
As Earl and David have been saying (over and over:^)." The fact that you hear and record the difference tones are proof that something was distorting".
The somewhat surprising detail is that the distortion is in the air itself (not our ears or a 150 dB low differential distortion capable microphone recording a 104 dB event) non-linear enough to cause the audible intermodulation products at a much lower SPL than I would have thought possible.
Several companies have exploited the audible intermodulation products to make various commercial devices for targeted communication in a single point in space.
I was a bit surprised to find the basics of that technology can be duplicated with standard audio gear.
Art
Looking back at SpeakerDaves post 191, (took me this long to set up and check this out) I set up two tone generators and wye'd them with a 4.7K Ohm resistor on each posative lead and tied the grounds together and used this to single endedly drive one channel of a Crown D-45 amp in turn driving a Jbl 2451 compression driver.
Generator one was monitored on channel one of my scope, gen two- ch two and the summed signal on ch 3, the amp output on ch4.
Channel three and four show the expected "motor boating" of the amplitude of the single sinewave. With gen one set at 2kHz and gen two set at 2.2kHz the sound is modulated as one would expect. Gen one at 20,000 Hz and gen two at 22,000 Hz ch 3 and 4 show amplitude modulation, no audible sound beyond the noise floor of the set up, as you would expect in a linear system.
Just my 2 cents.
Barry.
Generator one was monitored on channel one of my scope, gen two- ch two and the summed signal on ch 3, the amp output on ch4.
Channel three and four show the expected "motor boating" of the amplitude of the single sinewave. With gen one set at 2kHz and gen two set at 2.2kHz the sound is modulated as one would expect. Gen one at 20,000 Hz and gen two at 22,000 Hz ch 3 and 4 show amplitude modulation, no audible sound beyond the noise floor of the set up, as you would expect in a linear system.
Just my 2 cents.
Barry.
Barry,
That is exactly what I found in my tests of single drivers.
The sub tone audible air IM distortion only happens in the convergence zone of two drivers each driven with a discrete tone, if the discrete tones are above audibility, the IM tone is more obviously audible.
If you want to hear the audible IM effect created by air, simply put each tone in to a pair of drivers, the more "beamy" they are at 20 kHz, the easier it will be to hear. I set the drivers in my tests to 104 dB output.
Be forwarned, high SPL 20 kHz can still cause hearing damage even though you don't hear it.
OSHA sets a limit of one hour exposure to 105 dBA "slow" per day, a lot more SPL than I like to expose my ears to.
Art
Last edited:
Me too, as I know how these devices work. The Ultrasonic levels they use are very high, thus we expect the non-linearity. My levels were much lower, yours a bit above.
Perhaps it's as Earl says, the non-linearity is progressive. I certainly don't know how much at what levels.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.