Recently we had a long discussion about the merits of Rated Frequency Response. Most said it was a worthless meaningless spec. I conceded it was of limited value, but could be used for a superficial quick check of speaker compatibility.
Recently I've been browsing some speakers for a potential project; at project that, so far, is no more than a gleam in my eye. The first thing I noticed is that for many speakers, it is nearly impossible to get the basic specs and frequency response curves for them. Goldwood for example, how or why would I every use a Goldwood speaker if I can't look at a frequency and impedance charts?
Next, how is 'RATED' frequency response measure? Do they all use a fixed sized closed box, or is it every man for himself?
Let's take two examples -
The reasonably well respected Dayton DC250-8 10" Classic Woofer -
http://www.partsexpress.com/pdf/295-315.pdf
It's RATED frequency response is 25-2,500 Hz , yet on the graph above the response starts rolling off at 100 hz. To get to the rated response, the curve is down about 12db on each end. I would rate it 80hz to 2500hz which is about 3db down.
Now take this example -
Aurum Cantus AC-250MKII 10" Carbon Fiber Sandwich Woofer-
http://www.partsexpress.com/pdf/296-434s.pdf
Now this is a much more expensive speaker, but that really shouldn't matter. The claimed freq response is 23-3,000 Hz and a look at the freq response chart would indicate that is pretty close to real. I would estimate it at about 23hz to 2700hz at the roughly -3db points.
Isn't there some standard that speaker manufacturers must adhere to? How can these rated claims so wildly deviate from the actual response?
The next point is, given how far off some of these specs are, how can I possibly trust a manufacturer like Goldwood who doesn't make their detailed specs and charts available?
There not making this easy; yet, one would assume that if the want to sell speakers, they would make it easy.
This bring up the last question. If I program the T/S parameters into a speaker modeling program, is that going to be enough to accurately model the speaker? Or, can there be, and are there, odd quirks and aberrations that exist but don't reveal themselves in an analysis of the T/S parameters?
OK, I'm done whining for now.
Steve/bluewizard
Recently I've been browsing some speakers for a potential project; at project that, so far, is no more than a gleam in my eye. The first thing I noticed is that for many speakers, it is nearly impossible to get the basic specs and frequency response curves for them. Goldwood for example, how or why would I every use a Goldwood speaker if I can't look at a frequency and impedance charts?
Next, how is 'RATED' frequency response measure? Do they all use a fixed sized closed box, or is it every man for himself?
Let's take two examples -
The reasonably well respected Dayton DC250-8 10" Classic Woofer -
http://www.partsexpress.com/pdf/295-315.pdf
It's RATED frequency response is 25-2,500 Hz , yet on the graph above the response starts rolling off at 100 hz. To get to the rated response, the curve is down about 12db on each end. I would rate it 80hz to 2500hz which is about 3db down.
Now take this example -
Aurum Cantus AC-250MKII 10" Carbon Fiber Sandwich Woofer-
http://www.partsexpress.com/pdf/296-434s.pdf
Now this is a much more expensive speaker, but that really shouldn't matter. The claimed freq response is 23-3,000 Hz and a look at the freq response chart would indicate that is pretty close to real. I would estimate it at about 23hz to 2700hz at the roughly -3db points.
Isn't there some standard that speaker manufacturers must adhere to? How can these rated claims so wildly deviate from the actual response?
The next point is, given how far off some of these specs are, how can I possibly trust a manufacturer like Goldwood who doesn't make their detailed specs and charts available?
There not making this easy; yet, one would assume that if the want to sell speakers, they would make it easy.
This bring up the last question. If I program the T/S parameters into a speaker modeling program, is that going to be enough to accurately model the speaker? Or, can there be, and are there, odd quirks and aberrations that exist but don't reveal themselves in an analysis of the T/S parameters?
OK, I'm done whining for now.
Steve/bluewizard
OK, forget all my ranting and raving and concentrate on this one question-
If I program the T/S parameters into a speaker modeling program, is that going to be enough to accurately model the speaker? Or, can there be, and are there, odd quirks and aberrations that exist but don't reveal themselves in an analysis of the T/S parameters?
In short, to what extent can I trust the T/S parameters to give me the whole picture when modeling a speaker?
Steve/bluewizard
If I program the T/S parameters into a speaker modeling program, is that going to be enough to accurately model the speaker? Or, can there be, and are there, odd quirks and aberrations that exist but don't reveal themselves in an analysis of the T/S parameters?
In short, to what extent can I trust the T/S parameters to give me the whole picture when modeling a speaker?
Steve/bluewizard
Good questions.
It's difficult for the manufacturer too.
For instance, the bass cutoff depends on the kind of box. This may change the system resonance and system Q.
The more common method then is to claim a low end cutoff at free air resonance.
The graph commonly shown is infinite baffle response, and the rolloff is caused by the driver Qts.
The high end may be from speaker 'break-up' and thus difficult to interpret.
It's difficult for the manufacturer too.
For instance, the bass cutoff depends on the kind of box. This may change the system resonance and system Q.
The more common method then is to claim a low end cutoff at free air resonance.
The graph commonly shown is infinite baffle response, and the rolloff is caused by the driver Qts.
The high end may be from speaker 'break-up' and thus difficult to interpret.
At the bass end, yes. At least for small signals.
No. That was my earlier point. It's a "spec" that means absolutely nothing.
If they don't give T-S parameters, run. And even good charts and specs will not be applicable unless you are using the drivers in the same size box, same size baffle, and measuring at precisely the same point in space. That's assuming perfect consistency in the drivers which is one hell of an assumption.
Speaker design is not easy and it's a participation sport. If you're willing to take the time and trouble to acquire and learn to use good measurement and CAD software (like Soundeasy, CLIO, Speaker Workshop, Praxis), you've got a shot at succeeding. If not, stick with published designs and kits.
Sy, so what you are saying in essence is that RATED frequency response is a "spec" without a spec.
In other words, a spec without an official or unofficial Standard to back it up; "Every man for himself".
Thanks for the replies.
Steve/bluewizard
In other words, a spec without an official or unofficial Standard to back it up; "Every man for himself".
Thanks for the replies.
Steve/bluewizard
Just like the wattage rating of amplifiers or the handling power of speakers. Starting to get it now?
BlueWizard said:
Unfortunately I have found measured T/S parameters to vary alot from spec so that may also be of limited use. Smaller and cheaper drivers seem to be be further from spec.
ocool_15 said:
Do keep in mind that different measurement conditions will yeild different results. T/S parameters are really curves. Then there are variations due to the weather. Maximum driver-to-driver differences are typically on the order of 20% in my direct experience.
Writing down T/S is trying to nail down a moving target
dave
- Status
- Not open for further replies.