We can always rely on Soundminded for an upbeat assesment of the state of the industry
.I wouldn't vouch for the value of all products as most high-end products charge a great deal for the snake oil proportion, but I have seen a great deal of important research on a constant basis. It is amazing what we have learned since, say, Vilchur's day.
David
Gedlee,
I am now 61 and I know my ears are not young. I can only hear tones upto about 15kHz.
Recently as an experiment with some very cheap orthodynamic tweeters, I added a little ribbon tweeter to a Mid Bass driver that I was running without it's crossover.
I crossed in the tweeter using two passive single pole filters (LR2) @ 7kHz. Turning the treble only amplifier ON OFF made an enormous difference to what I was hearing. Even listening to FM radio which is bandlimited to ~15kHz sounded completely wrong without the ribbon operating.
You are putting in and taking out everything above 7 kHz - of course you will hear that. That is not what we are talking about.
The point was that changing the design to replace frequencies above 7 kHz with a different driver was not going to make much of a difference, especially if the OP was over 50.
The point was that changing the design to replace frequencies above 7 kHz with a different driver was not going to make much of a difference, especially if the OP was over 50.
It might have an important impact if the OP is often listening off axis, as even a biradial design will start to seriously beam above 7khz with a 2" driver, especially in the vertical plan (if the slot is in the horizontal one, as with the P-audio horns he is using).
On the other hand a TH4001 or A290 would present a sufficiently large directivity up high to avoid using a tweeter (there will be lobes, but that is a fairly acceptable trade off to do), and will also allow for a lower crossover point with the mid woofers...
On the other hand a TH4001 or A290 would present a sufficiently large directivity up high to avoid using a tweeter (there will be lobes, but that is a fairly acceptable trade off to do), and will also allow for a lower crossover point with the mid woofers...
Last edited:
You didn't look at the reference I pointed to in post #40.
http://www.jblpro.com/pages/pub/components/2380a.pdf
It shows that the d.i. of the horn is quite flat to the 12.5 k 1/3rd Octave. Thats pretty good and shows no real evidence of beaming from the 2" throat driver. The reason is that it is a 90 x 40 horn with a diffraction slot less than an inch wide feeding the wide horizontal angle, so no beaming there. In the vertical angle the throat directivity is essentially the same as the narrow 40 degree coverage of the horn, so no beaming there either. The equalized off axis curves they show fully support that. (as they would have to if the d.i. curve is valid.)
So if maintaining a flat power response above 10k is a criterion then these curves show no need of the complication of a 7k crossover to another driver. In my experience you will both certainly mess up the power response and also, more importantly, mess up all the near off axis curves. If you don't time align to some degree you will even mess up the on axis curve. As long as the driver is good to a high enough frequency (with or without EQ) the crossoverless case will be better in practice.
Of course the higher order modes will give you headaches or cancer or something.
For the rest of you debating whether it is audible or not. Why doesn't somebody go into Cool Edit Pro and simulate some response errors in pink noise above 10k? If you're still bickering about it when I get home I will take a stab at it.
David
Here are some samples of pink noise with a response error added. Simply, a parametric EQ was added to give a peak with a Q of 2 and a rise of +6 for three versions and +3 for the final sample. In every case I alternated 1 second of flat, then one second of peaked.
Peak frequency was 10k, 16k and 20k. Remember that any peak has width, so a peak at 16kHz has some effect at 10k and 12k. This is realistic because a "peaky" tweeter will have a width to that peak. Clearly, if the peak gets narrower and lower there is probably a threshold where you just can't hear it.
As always, I can't attach MP3 files so I have falsely renamed these as .zip files. Save them to your desktop, change the extension back to MP3 (don't unzip) and play them on your computer into decent headphones.
Enjoy.
Peak frequency was 10k, 16k and 20k. Remember that any peak has width, so a peak at 16kHz has some effect at 10k and 12k. This is realistic because a "peaky" tweeter will have a width to that peak. Clearly, if the peak gets narrower and lower there is probably a threshold where you just can't hear it.
As always, I can't attach MP3 files so I have falsely renamed these as .zip files. Save them to your desktop, change the extension back to MP3 (don't unzip) and play them on your computer into decent headphones.
Enjoy.
SpeakerDave, you are taking an extreme position that no one ever suggested.
If there is sufficient energy in the top octave, then yes you will detect it (even if you have a moderate hearing impairment). The issue was whether reproduction in this top octave was especially critical or not. IOW, do you spend your money fixing other problems first. The question regards priorities. IMO, focusing on the rest of the spectrum is the bigger priority. The top octave is usually the least of your worries.
If there is sufficient energy in the top octave, then yes you will detect it (even if you have a moderate hearing impairment). The issue was whether reproduction in this top octave was especially critical or not. IOW, do you spend your money fixing other problems first. The question regards priorities. IMO, focusing on the rest of the spectrum is the bigger priority. The top octave is usually the least of your worries.
That may not be your definition of beaming but it certainly is mine. When you consider that many musical instruments can be heard with little or no tonal variation at any horizontal or vertical angle and at some frequencies many loudspeakers show little variation too, a driver that only covers 1/36 of what a spherical radiator at high frequencies would produce is a fairly narrowly focused beam in my book.
"We can always rely on Soundminded for an upbeat assesment of the state of the industry"
Look on the bright side. The speakers that have broken the million dollar a pair price barrier are in many ways no better than far less expensive speakers so if you are a designer of less ambitious products, they haven't really raised the bar much higher. There's still plenty of room for improvement.
I was never taking a position in the arguement about the relative importance of the top Octave, although I have stated a strong oppinion about the benefits (or lack there of) of high crossover points. I'd rather have response roll off a few dB by 15kHz than extend to 20k+ but suffer through crossover problems at 7k.
Anyhow, listen to the samples.
David
The question wasn't whether the directivity was wide or narrow, but rather whether it became significantly narrower above 7k, to the point of warranting crossover to another unit.
I would argue that the particular horn has very little variation within its advertised 90 x 40 beamwidth. i.e. "little or no tonal variation".
The problem with that idea is that while there will be early room reflections from many angles in the general direction of the speakers at other frequencies, you won't be getting any at high frequencies. This is spectral reflection distortion. There is no way to beat the listening room short of creating an anechoic chamber. You are far better off trying to exploit it instead, you have a much better chance of success. IMO Roy Allison was headed in the right direction. Even a 5 db falloff at 15 khz 60 degrees off axis wasn't good enough for him. That's why he designed AR LST and his own speakers the way he did.
I am over 50 and can hardly perceive pure tones over 10khz. But with the following program, i discovered that the low passing of a program over 10khz is quite audible to my ears.
Audio Musings by Sean Olive: Harman's "How to Listen" Listener Training Software Now Available as Beta
Audio Musings by Sean Olive: Harman's "How to Listen" Listener Training Software Now Available as Beta
All that debate over whether response problems above 10 k are audible and only one person has downloaded the examples? Come on guys.
David
I don't think that a compressed file would tell the truth ,as we are talking about those kind of frequencies that characterize the sound of an instrument by timber,amplitude and other important factors. We should bring to our ears what the real sound is , then argue if it's natural , audible in some little refinements etc.
Indeed , what we ear from our 'system' is a series of distortions , which are caused by :
intermodulation of complex sounds ; beats and amplitude modulation caused by nearby harmonics( we can't hear a 20 KHz , moreover a 22 KHz , but the difference between them , 2 KHz , is perfectly audible ! );so our hearing is biased by the level of the treble , which changes individually ,as a treble signal played at low and at high volume , is perceived as higher in frequency when played louder .
Indeed , what we ear from our 'system' is a series of distortions , which are caused by :
intermodulation of complex sounds ; beats and amplitude modulation caused by nearby harmonics( we can't hear a 20 KHz , moreover a 22 KHz , but the difference between them , 2 KHz , is perfectly audible ! );so our hearing is biased by the level of the treble , which changes individually ,as a treble signal played at low and at high volume , is perceived as higher in frequency when played louder .
Once again, the notion of beats from supersonic tones is a canard. You would have a hard time proving that it existed in music to any real degree. If it did then the audible difference tones would have already been generated and would be recorded and reproduced by an accurate system (so no need to record the inaudible supersonic components). If your system is converting supersonic tones into audible beats, then that indicates huge distortion.
MP3 files aren't worthy of showing the effect of a response error? Then I guess you won't hear anything if you try them.
David
MP3 files aren't worthy of showing the effect of a response error? Then I guess you won't hear anything if you try them.
David
No , I was meaning the beats produced by normal instruments like drums , violins ,guitars . And indeed I was thinking of the addition of a supertweeter to the existing system , not to evaluate the contribution to sonic of the existing treble unit . If digital audio is somehow truncated at 22 KHz by LPF ( well , that's 16 /44 standard , a little old ... ) you won't perceive nothing with the addition of a stw , but in reality , it does .
I haven't tried it but it logically adds something .
And ,no , I' haven't the PC attached to my system
) you won't perceive nothing with the addition of a stw , but in reality , it does .I haven't tried it but it logically adds something .
And ,no , I' haven't the PC attached to my system
Well, there's no arguing with that!
Beats in musical instruments are perceptual without being real tones. For example (I tune pianos occaisionally) a piano with a string at 440 and one at 441 will have a 1 cycle per second beat. You will perceive the amplitude modulation, but you aren't hearing 1Hz. A system that captures 440 to 441 (a modest bandwidth!) would fully capture the effect.
David
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.