It's named "mensink" (my last name).
Yeah, something like that. My solution however would be to extent CD-behavior to well below the Schroeder-frequency and use (multiple) separate bass-speakers for the low frequencies.
Unfortunately Earl's polar app doesn't show 360°, so any speaker that performs good in the frontal half comes out as a "winner". Hypercardioid, anyone?
I'm wondering how the Nathan 10ca can have such a high DI at low frequencies?
Attachments
Having listened to some JBL stuff, i can't say I have encountered anything I would personally want to own, but lots of their reseaserch certainly trigger lots of ideas.
I think Earl is plainly addressing his design trade off consideration. There are different design constrains. For example, size and form factor, and placement are usually the important design factors for an average home, that sort of limits what you can do because low frequency directivity.
Fuzzy smoothed polar plots tell nothing about preservation of phase relationships across radiating area. 500Hz to about 5kHz. In case of crossover point about 1kHz, behavior of drivers 1/2 octave or so about this point have impact on clarity. Apparent smoothness of frequency response alone just doesn't yield much beyond Toole's observations of listener preference for flatter response and wider coverage.
Conical waveguide with big radius opening may radiate nicely, but 10-12" driver at 1kHz-1.5kHz has diffraction issues much greater than 5.25"-6" driver, leading to non uniform phase behavior off axis, leading to room reflections that yield poor cues to perceptual system.
Griesinger's work applied to small listening spaces instead of concert halls has many details to be worked out.
Two divergent drivers in same bandwidth are perfect emulation of bad diffraction problem.
Conical waveguide with big radius opening may radiate nicely, but 10-12" driver at 1kHz-1.5kHz has diffraction issues much greater than 5.25"-6" driver, leading to non uniform phase behavior off axis, leading to room reflections that yield poor cues to perceptual system.
Griesinger's work applied to small listening spaces instead of concert halls has many details to be worked out.
Two divergent drivers in same bandwidth are perfect emulation of bad diffraction problem.
Markus
If you don't understand then it must be your fault.
Well, if the signal is already corrupted, then the receiver isn't necessarily faulty
By the way, what's up with the strange DI plot at low frequencies for some of the newer data sets in your polar app?
Which city?
Last edited:
Montreux
To an American everywhere in Switzerland would be "in the neighborhood". Isn't the whole country smaller than Michigan?
I think that I said that the DI below the 200-300 Hz point is based on data taken near field and as such is not always correct for the technique that I use to calculate DI. I have had trouble getting a 0 dB DI at LFs to match with the DI above 200-300 Hz and sometimes the software does not produce such a good match. Above 200-300 Hz the shape of the DI is very good, but its actual value could be off by the same amount that the LF DI is off from 0. I don't use the DI data so much for its absolute value as I do to insure that there are no DI aberrations through the crossover. The absolute value of the DI is not important in that case, only the shape matters.
To an American everywhere in Switzerland would be "in the neighborhood". Isn't the whole country smaller than Michigan?
I think that I said that the DI below the 200-300 Hz point is based on data taken near field and as such is not always correct for the technique that I use to calculate DI. I have had trouble getting a 0 dB DI at LFs to match with the DI above 200-300 Hz and sometimes the software does not produce such a good match. Above 200-300 Hz the shape of the DI is very good, but its actual value could be off by the same amount that the LF DI is off from 0. I don't use the DI data so much for its absolute value as I do to insure that there are no DI aberrations through the crossover. The absolute value of the DI is not important in that case, only the shape matters.
Last edited:
The country is probably smaller than any US midsize SUV. Nevertheless there are 4 different languages in Switzerland.
Err, the shape matters? You show a DI that simply doesn't reflect reality.
Contract Out The Building
You can 'farm' all that out.
Custom Speaker Cabinet Boxes Manufacturing | A List Woodworks
Speaker Cabinet Manufacturers
That leaves you with only marketing & sales, plus maybe some leads for additional consulting assignments as well. If your products are as flawless as you claim, then that effort should be a 'cake-walk' for you.
Regards,
WHG
You can 'farm' all that out.
Custom Speaker Cabinet Boxes Manufacturing | A List Woodworks
Speaker Cabinet Manufacturers
That leaves you with only marketing & sales, plus maybe some leads for additional consulting assignments as well. If your products are as flawless as you claim, then that effort should be a 'cake-walk' for you.
Regards,
WHG
You show a DI of 8dB and higher above 200Hz for the Nathan 10ca. How is that possible for a speaker that has a rather omnipolar radiation pattern in that region?
There is absolutely no evidence out there that the phase response of reflections has any impact on perception. In fact there is little connection between frequency response and reflections other than broad response variation of reflections impacting their apparent level.
Read the Soren Bech papers where he simulated a speaker in a typical room by creating all the reflections of a typical room in an anechoic chamber. In the first generation of his tests he used a generic polar curve to define the level of each reflection. In the second study he used the particular true frequency response based on the angle of response originating each reflection, some of which had large holes due to crossover effects. At angles where big response holes occurred the reflection simply needed to be raised an appropriate amount to match the previous study values.
The phase shift of a reflection has many cycles of rotation compared to the preceeding direct response. The notion that we can detect some added differential shift in later reflections is pretty far fetched.
We need to be careful that we don't let audiophile idealism create requirements that are unjustified. We do not hear all the individual reflections that come at us from the room's surfaces. To think that a loudspeaker needs perfect response at every angle so that the reflections are equally perfect, is not supported by any study. Only the strongest and earliest reflections are part of perceived frequency response, but in an aggregate manner, not individually.
David S.
Although I am a proponent of constant or at least smoothly changing directivity, I do agree with David.
I'm not sure if I interpret correctly.
The off-axis response of a beaming driver is still minimum phase (as long as it doesn't break up heavily). You can see the membrane as a collection of an infinite number of point sources. Their responses add up and the negative interference you get off-axis is what causes driver directivity. Our hearing sums the responses of all point sources, so we hear it as a single source. So all you really have to do, is look at the response curves. As long as they are nice and smooth off-axis, there's not much to worry about.
Barleywater said:
I'm not sure if I interpret correctly.
The off-axis response of a beaming driver is still minimum phase (as long as it doesn't break up heavily). You can see the membrane as a collection of an infinite number of point sources. Their responses add up and the negative interference you get off-axis is what causes driver directivity. Our hearing sums the responses of all point sources, so we hear it as a single source. So all you really have to do, is look at the response curves. As long as they are nice and smooth off-axis, there's not much to worry about.
I agree. More important is probably delay, angle and level of reflections.
One could even argue that spectral similarity of direct sound and reflection is detrimental because it gets harder for our hearing to make the distinction between reflection and second sound source. The latter creating a new phantom source in combination with the direct sound (when delay of the reflection is small and level is high).
Not much data to support that.
Early, strong reflections are the issue, its not some absolute. The earlier they are and the stronger they are the more that they affect the image so "minimizing" them is the goal. Phase has nothing to do with the problem. Beyond 10 ms or so they are all good, although vertical ones do not appear to do much good or bad at longer delays.
However, tests by Olive suggest that the frequency response of the reverberant field is a factor and many of the comments here would imply that it is not a factor. The reverberant field needs to be smooth and comparable to the direct field or it will "color" the total response.