There are many factors, but in general, phase coherency is at a lower magnitude of influence.
FR and CSD are giving different information and one cannot be gleaned from the other. There are perfectly flat speakers with horrible FR and good decay, and the reverse is also true.
Please, those of you in doubt, forget this red herring. You cannot guess correctly the CSD from the FR. But you can know exactly the FR from the CSD, because it is the graph at t=0.
Hello Vac,
Would you care to explain what is meant by:
"There are perfectly flat speakers with horrible FR"
Basically you are challenging the Fourier Transform with the full version of your statement - including the part of good decay-above.
Now that is quite a position.
Care to clarify & substantiate the above? I am bit puzzled to put it mildly.
Eelco
Would you care to explain what is meant by:
"There are perfectly flat speakers with horrible FR"
Basically you are challenging the Fourier Transform with the full version of your statement - including the part of good decay-above.
Now that is quite a position.
Care to clarify & substantiate the above? I am bit puzzled to put it mildly.
Eelco
An impulse contains energy at all frequencies and if it's perfect, it will have an infinite bandwidth. A sound card isn't going to produce this anyway as it is band limited at the top end.
Anyway, it isn't practical to use a pulse for measurement even though it should be possible.. you'll get more signal to work with sweeping slowly.
Just a mention of a crude experiment with a surprising result.
I had a cheap 15 inch poly coned woofer years ago. the thing was crap. Couldnt handle any power and always sounded, well for lack of better explanation, slow and ponderous. Meaning the opposite of lively and dynamic. Sorry but thats the best wording I can come up with.
Anyway at the time I was looking at the effects that well damped surrounds have on sound quality in the bass. This particular cone had a surround that was designed to give a cleaner midrange I believe. It was a material that was "lossy" , not a springy material.
Soo I Coated the surround on both sides with regular old clear silicone calk. This gave a more springy nature to the suspension of that cone.
Now I cant say for sure what all was changed in that driver but to this day I have never heard such a transformation to sound from a driver mod. The sound went from dead and lifeless to alive and dynamic.
Yes I am painfully aware of how small changes in FR can alter sound, and I tried to set up to maintain same overall curves etc. But I gotta say, in the end there seems to be something that the FR and distortion measurements arent telling when it comes to the character of bass and mid bass sound.
I had a cheap 15 inch poly coned woofer years ago. the thing was crap. Couldnt handle any power and always sounded, well for lack of better explanation, slow and ponderous. Meaning the opposite of lively and dynamic. Sorry but thats the best wording I can come up with.
Anyway at the time I was looking at the effects that well damped surrounds have on sound quality in the bass. This particular cone had a surround that was designed to give a cleaner midrange I believe. It was a material that was "lossy" , not a springy material.
Soo I Coated the surround on both sides with regular old clear silicone calk. This gave a more springy nature to the suspension of that cone.
Now I cant say for sure what all was changed in that driver but to this day I have never heard such a transformation to sound from a driver mod. The sound went from dead and lifeless to alive and dynamic.
Yes I am painfully aware of how small changes in FR can alter sound, and I tried to set up to maintain same overall curves etc. But I gotta say, in the end there seems to be something that the FR and distortion measurements arent telling when it comes to the character of bass and mid bass sound.
I'm sorry, but this is not correct. The CSD derives directly from the impulse response, the exact same impulse response whose Fourier Transform is the frequency response. There cannot be any information in one that is not in the other.
The example that you state could not exist in the real world because it would be impossible. The FR at t=0 is based on everything that follows after t=0, all the exact same information that is in the CSD.
As I said there is one single situation that can occur where one could not guess the CSD from the FR and that is a dual close resonance, like in a room. When a room has two modes close by and is excited by a tone between them, the steady state sound field will be at the excitation frequency, but the room sound will decay at the two frequencies of the modes. Hence the FR would show a shallow peak at the driven frequency, but the decay - the CSD - would show two separate sharp peaks. But in a loudspeaker this would be a very uncommon situation, although I have seen it.
CSD makes it easier to understand how the speaker decays at different frequencies, normally it is possible to find relationship with sound coloration. I look at CSDs of drivers and horns, and notice that the horns I measured seem to decay slower than direct radiating drivers, but they have a more uniform decay over the spectrum.
It is the same response with one caveat, if the speaker gets overloaded by the large impulse then the response might differ. This is why people don;t actually use an impulse, but swept sine, or noise, because spreading the energy over time is a lot easier on the DUT. Hence the impulse response is simply a convenient way of obtaining the necessary information, but not actually the way that we derive it.
All audio testing based on the time/frequency relationships of the FFT are band limited. AT the low end it is limited by the length of the time sample and at the high end it is limited by the sample rate. We cannot know anything about the system outside of these limitations.
Yes I thought someone might say that. I cannot show solid test regime to validate BUT after spending many hours tailoring and comparing I came away thinking that suspension change had great effect on bass sound.The driver was used from about 300 hz down.
Not sure of course what was really going on. Was the suspension now simply holding the VC in the gap much better than before? Was the more linear spring nature of the modified suspension actually causing a different sound in the bass? Combination? The driver had a simple motor BTW, no copper and no special pole geometry.
I have read somewhere( btw this was after the experience I had with this driver), cant remember where, that some resurch was done on this very thing. Comparing a number of commercial systems and their bass quality. and it was concluded that the systems with bass driver suspensions that were less lossy ( Qms > 3) and more linear spring in nature tended to rank high and that systems whos bass driver used a more lossy suspension/surround ( Qms < 3) did not.
Anyway I realize my "experience" is not exactly the most scientific but like said after much playing around I cant help thinking the suspension had a lot to do with its bass characture
Last edited:
There could be 100's of reasons for your perception, but let me just guess at one that I have seen before. If this is a "cheap" speaker with a small magnet then odds are it has very little linear excursion capability in order to get sufficient SPL output. This is extremely common for "musical instrument" loudspeakers because the net distortion is actually thought to be a good thing.
When you added stiffness to the surround, and I would guess, a lot of it, it raised the resonance much higher and resulted in a much lower excursion for any given piece of music. This would, in the above case, dramatically reduce the distortion.
When you added stiffness to the surround, and I would guess, a lot of it, it raised the resonance much higher and resulted in a much lower excursion for any given piece of music. This would, in the above case, dramatically reduce the distortion.
Yes that sounds likely. It was cheap speaker. Magnet was maybe 30 oz. The resonance before was around 25 hz, after mod was about 40 hz.
So I guess you dont think the mechanical properties of the suspension(lossy vs springy) had much effect. That the difference was mostly if not entirely simply whatever control was afforded by the raised resonance?
The only question in my head is that this sonic quality was obvious even at low volumes.
So I guess you dont think the mechanical properties of the suspension(lossy vs springy) had much effect. That the difference was mostly if not entirely simply whatever control was afforded by the raised resonance?
The only question in my head is that this sonic quality was obvious even at low volumes.
Again, I was only guessing, without an elaborate investigation into the situation that's all I can do. The mechanical properties of the surround do have a lot to do with sound quality, especially at the higher frequencies, but not so much at the lower ones.
Without an elaborate investigation into the situation its hard to know what the issue is. My point is that it is dangerous to jump to conclusions from simple experiments, especially when one is using their ears as a measurement device.
Without an elaborate investigation into the situation its hard to know what the issue is. My point is that it is dangerous to jump to conclusions from simple experiments, especially when one is using their ears as a measurement device.
Yes I agree totally with the conclusions from such "experiments". This one was just such a remarkable change that it had me playing with it for some time.
I do understand that surrounds have quite an impact at higher frequency's where lossy materials play a role in cone behavior. In this case however I was working that driver below 300 hz and in another setup I used it below 100hz. In both the difference was obvious compared to stock.
I do understand that surrounds have quite an impact at higher frequency's where lossy materials play a role in cone behavior. In this case however I was working that driver below 300 hz and in another setup I used it below 100hz. In both the difference was obvious compared to stock.
Do a Young’s modulus and Poisson’s ratio search of the various material and you probably have the answer.
Aliexpress.com : Buy 2017 New Boarseman KR25D In Ear Earphone Hifi Bass Alloy Tune Dynamic Earbuds Noise Cancelling For IPhone Android PC from Reliable in ear earphones suppliers on AK Audio store
I'm ordering these on 11/11 because I can save two dollars. I don't normally listen to headphones but hey why not, I deserve it.
Yes, that is useful. A lot of tweeters might have a gradual fr roll off that looks okay (approaching fs) but in reality they are full of resonances and this appears in the CSD and non linear harmonic distortion. We all know a tweeter crossed too low sounds bad.
A nice uniform decay indicates good energy release. It would be interesting to do a csd of a single driver vs the same driver in an array.
This website does decay plots for its test, here is one compression driver (actually looks pretty good especially dispersion).
Test Bench: A Redesigned DE990TN-8 1.4” Compression Driver from B&C Speakers
However it doesn't look like horns have faster or better csd's than tweeters so IDK
There is obviously more information in a csd since its showing you the change over time not just an average spl at a given frequency
Last edited:
It's just a different presentation of the same data. Learn what each tells you and how they are all related and you can get more out of them in terms of what you hear.
In days of yore (1970's about when silicone glue appeared, if I recall), people used the same silicone glue to coat surrounds but thinned it first (with vinegar??). I think mostly to improve the imperviousness of accordion cloth surrounds. Parts Express now sells some new-fangled paint goo for coating.
As others have said, hard to guess the effect. I stupidly did some coating not long ago on a 15-inch sub surround and it raised the resonance from about 22 to about 50 Hz. Also, surrounds are developed to have ideal spring performance and well-behaved damping. But new layers of coating, esp silicone glue, are likely to behave differently and so introduce distortion. Of course some folks may like the way the newly added higher-freq harmonics sound.
B.