I don't think that true in general, but it is for a line source. For a piston the near field is like a columnated (sp?) beam, it goes straight out like a flashlight. There is no fall-off, except that along the way there are several peaks and dips. The higher the frequency the more of these until at LFs they just disappear.
That is an interesting observation. There's been an increase in research in low frequency perception the past few years over concerns regarding health and low frequency noise pollution from wind farms and I've chanced to read some of it. Some of it has noted that at your mentioned 100Hz, the ear shifts from being dominated by sound pressure and inner hair cells to particle velocity and the outer hair cells as frequency decreases with particle velocity becoming the dominant mode roughly an octave above your mark of 20Hz (with ten times to voltage output on the nerve to boot, though there does seem to be a blanking phenomenon that occurs when much higher frequency content comes back into play). Perhaps that has some relation to the tactile sensation you mention?
I wonder about that as, again, with the more recent studies there have been findings that, in fact, the audible signal is the more important in perceiving low frequencies. There was one study in particular I found that is especially interesting in that it compares low frequency audibility between hearing test subjects who would both hear and feel the bass and deaf test subjects who would be limited to pure tactile perception. The results? The hearing subjects heard 20Hz bass 30dB lower than the deaf testers could feel it and could still hear 3Hz roughly ten dB before it could be felt. (Measured at SPL only.) When tactile perception was looked at, the hearing subjects still "felt" the bass a few dB before the deaf subjects. I've attached a graph of the results to better visualize it.
So, low bass is definitely heard, not merely felt, but it does seem to be "felt" when heard just a little perhaps through some sort of synaesthetic bleed over between the senses. Given how many times I've "felt" a bit of low bass when listening to music on my IEMs, I'm not surprised at that at all.
Attachments
It was some open ear headphones in my case. Just turned down the oscillator frequency gradually and the tone got deeper and deeper and deeper, never sounded perceptually like it switched or transitioned to any type of harmonic to me. Slowly swept the oscillator up and down in frequency a few times listening for any evidence of mode changes, but detected none. Oh well, I thought at the time, I guess it is possible to hear below 20Hz. Much later I heard Sean Olive say the old 20Hz to 20kHz range is a very rough approximation. He said young children may hear up to 30kHz or possibly higher, but by age 25 it was mostly limited to about 16kHz. That's on the high end of course. People still argue about that as well.
Earl, you should know this and I am sure that you do so I am puzzled. The FR and the CSD are indeed derived from the same impulse response, but they contain different information. The reason is that the FR is the impulse response with the window starting at t=0. For each subsequent slice of the CSD, this window slides over the impulse response with a fixed time interval. It thus shows the decay of the original impulse response.
For good measure I include two plots where the same impulse response is processed first into the FR, and then into the CSD.
Frequency response
Cumulative decay
There is no predicting the CSD plot from the FR. With the same FR, the CSD plot could stop on a dime or go on forever. Only correct processing from the impulse response will subtract the cumulative decay information from it. This information gets lost in the FR because it only looks at the impulse response with the window starting at t=0.
I would have loved to post this immediately after reading yours, but I was travelling, sorry for the delay.
Nonsense. You can't retrieve information that was outside the window of the first FFT.
They are somewhat related, but the larger the diaphragm, in relation with the stiffness of material, there will be certain things that are not predictable. Just last week I sent Peter Larsen some data, hoping that his software could simulate such design in some way. Yes, I did have to show him some Klippel scans to at least draw his interest. There are lots of things I hope would make trying out some new concepts easier.
ha, If they're not made out of mental I'll have to get a refund.
Dr. "Shwiensgruber" heartily recommends them in the comments section. The secret, I'm sure, is in the mental material.
Review: boarseman KR25D - AudioBudget
Last edited:
I think that you need to think this through better, because it is a simple fact that I can generate the CSD from the t=0 response alone. This means that all the information in the CSD is contained in the t=0 result. This should be obvious since the impulse response contains all information and the CSD is just showing ever smaller portions of it. At no time does the CSD use any data that is not in the original t=0 response. Windowing, as you imply, is not the answer here.
@ post 746
Vac, what you show is rather amazing.
Firstly, the FR is 1/3 smoothed, whereas the waterfall is unsmoothened.
Such heavy smoothing makes narrow band dips and peaks disappear.
Secondly, the waterfall O dB line stops providing usefull FR info below 1000 Hz, which info is in the FR plot, running to 20 Hz.
Basically the 0 dB line in the Waterfall should be identical to the FR plot, although the graphic presentation of the FR usually looks a bit smoother, because of somewhout different scaling.
And now you want us to believe both plots are derived from one and the very same impulse without touching the windowing markers in ARTA, in other words with the markers in exactly the same spot?
Eelco
Vac, what you show is rather amazing.
Firstly, the FR is 1/3 smoothed, whereas the waterfall is unsmoothened.
Such heavy smoothing makes narrow band dips and peaks disappear.
Secondly, the waterfall O dB line stops providing usefull FR info below 1000 Hz, which info is in the FR plot, running to 20 Hz.
Basically the 0 dB line in the Waterfall should be identical to the FR plot, although the graphic presentation of the FR usually looks a bit smoother, because of somewhout different scaling.
And now you want us to believe both plots are derived from one and the very same impulse without touching the windowing markers in ARTA, in other words with the markers in exactly the same spot?
Eelco
What you say is true. But then we have to figure out ways to look at how much of t=0 still remains beyond t=0, do you know of a better way to show it? How will it assist in showing deficiencies more clearly? How will it more clearly show differences related with audible differences?
Last edited:
You are exactly correct here, it is HD and/or IMD and it is audible. I've observed the same
thing with different speakers. No, I did not do a double blind experiment. Your speakers
and mine are "well behaved" with gradual onset of non-linearity allowing us to push them
hard so that we can hear the gradual onset of HD.
Earl, made his assertions based on specific contrived experiments.
An untrained listener will hear more bass in the distorted speaker and MIGHT like it more - I don't.
I also believe that the observations will change when a better reference is available
in A/B comparisons.
I started a thread one time: "When Bad is Good".
Audio and especially human perception are complex.
Last edited:
I should point out that it is not just me that believe this, it is also professed by JBL's entire engineering staff and The THX people, just to name a few. I just happen to be the only one willing (or crazy enough) to come to the forums to express those beliefs.
And any results that are reported that are not blind can certainly not be held up as evidence against tests done correctly.
Forums should be a place for brainstorming and finding ways to improve on existing technology. Let’s focus on ideas rather than trying to discredit people. Everyone has different experiences, and we know that audio systems are far from perfect. How can we move to improve on existing methods should be much more productive.
That's JBL's problem too, they don't use "mental material". In all seriousness though a lot depends on personal preferences and what one is willing to put up with or adjust to. Linear distortion and dispersion are generally recognized to be more important that other types of distortion.
Electronics and amps are for the most part transparent.
Last edited:
Member
Joined 2009
Paid Member
subjectively, I've thought of near-field as being too close to use multi-way speakers due to poor integration of the sound. As soon as I'm far enough that a multi-way speaker sounds coherent then I'm no longer in the near-field. For computer speakers, single driver full-range is the way to go.