AVS Forum banner
1 - 18 of 18 Posts

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #1 ·
I recently finished full-range measurements (using Room EQ Wizard) on my 5 channel HT and I decided to take some near-field measurements. The graphs produced all show a rather steep roll-off above roughly 14KHz. I expect to see a gradual roll-off in the treble at my listening seat 12 feet away, but all of the measurements below were taken just 3 feet from the left speaker. My speakers use a Scan Speak silk dome tweeter that is supposed to have response out to at least 25KHz.




I've checked the output of my processor by connecting it to REW and the response is flat. I also moved one of my speakers to another system I have in a bedroom and the response rolls off in the treble when connected to that system too.


It seems like the problem is with the speakers, but it is just odd that I would see the same roll-off in every one of my five channels.


Thoughts?
 

·
Registered
Joined
·
1,363 Posts
Sounds like it's a problem with your mic or SPL meter that you are taking the measurements with.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #3 ·

Quote:
Originally Posted by 11001011 /forum/post/15491576


Sounds like it's a problem with your mic or SPL meter that you are taking the measurements with.

I thought that too, but when I use a different mic I get very similar results.


This measurement was taken of the right speaker with the Behringer ECM8000 (the previous measurements were with a DBX brand mic).
 

·
Registered
Joined
·
794 Posts
Mic should be placed at tweeter height. I would also explore distance of mic from speaker. I own Dunlavy speakers and the recommended distance for proper measurement of high frequency response is 9 feet. If I test the speaker response closer than 9 feet, I get a severe roll off at 14 kHz.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #5 ·
shumi -


Thanks for the info. That is odd isn't it? I can understand that a crossover might not integrate the drivers properly at too close a distance, but 14K is well above any crossover.


BTW - Here is a measurement taken on the tweeter axis @ 3 feet from the speaker.



You can see that the crossover was designed to blend the mid with the tweet on the woofer axis. On the tweeter axis, there is cancelation at the crossover point. Either way the high treble is still rolled off.


I will play with distances even though that makes no sense to me.
 

·
Registered
Joined
·
820 Posts
I don't think either your crossover or tweeter are defective. Neither would cause such a 'problem' as you are seeing at those frequencies. Your wording of the tweeter's performance is interesting. You say it is "supposed to have response out to at least 25KHz.". Well, there's nothing in your graphs which dispute that statement. Did you mean to say that it's supposed to have flat response to 25kHz? Can you tell us which model number Scan Speak tweeter this is? The graphs look like normal high end rolloff to me.


How do the speakers sound? Is this a new phenomenon, or did you only notice it once you measured?
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #7 ·
Aaron -


The tweeter is a Scan Speak 2905/9500 that, according to the SS data sheet, has flat response to 30KHz (I was being conservative with the 25K figure).

http://www.tymphany.com/files/produc...905_950000.pdf


I can only hear up to roughly 17KHz, so this is more academic than a serious audible problem. I'd just like to figure out if my measurements are off, or if the tweeters are malfunctioning. Can a tweeter even malfunction this way?


FWIW - I have measured several other speakers in my room and in the near-field measurements have always shown response that extends out to 20KHz.
 

·
Registered
Joined
·
820 Posts

Quote:
Originally Posted by hifisponge /forum/post/15493197


Aaron -


The tweeter is a Scan Speak 2905/9500 that, according to the SS data sheet, has flat response to 30KHz (I was being conservative with the 25K figure).

http://www.tymphany.com/files/produc...905_950000.pdf


I can only hear up to roughly 17KHz, so this is more academic than a serious audible problem. I'd just like to figure out if my measurements are off, or if the tweeters are malfunctioning. Can a tweeter even malfunction this way?


FWIW - I have measured several other speakers in my room and in the near-field measurements have always shown response that extends out to 20KHz.

Thanks for the information. Indeed, the on-axis response does certainly extend to 30k. Are you measuring the response directly on the tweeter axis, at the height of the tweeter? Moreover, do you know what topology crossover is used in the speaker? Different order crossovers will have different sizes and directions of output lobes. I would try doing a close mic measurement of the tweeter (6" or less away) and see what you get. If you were making measurements off the tweeter axis, you could (and should) very easily get the rolloff which you're seeing, as shown by the two lower response graphs on the spec sheet.


I've never known a tweeter to malfunction in that way. However, I also can't hear much beyond 17k, and my test instruments (consisting of an SPL meter and a sine wave generator) are probably less reliable than your measurements. So were a tweeter actually to lose the top octave of it's response, I'd probably never be able to tell.



Your best bet may be to search around the many DIY speaker forums and find other people who have used this tweeter, to see if they found similar results.
 

·
Registered
Joined
·
2,691 Posts
I'm not familair with the measurement method you're using. I can't tell whether it's basically a room measurement (even at 3 feet), or whether it's a quasi-anechoic response above 200 Hz or so. If it is a standard anechoic FFT measurement, and if you are measuring on the tweeter axis at about 3 feet (which is close to one meter), then the response at 20 khz should be higher than shown. If you get closer than 3 feet, the response will tilt downward, but earlier than shown. If this is more of a room measurement with a long sample time and smoothing, then the room itself will absorb some of the high frequencies. Still, I would not expect the steep roll off shown. I doubt very much that all of your tweeters are defective. And there's little a passive crossover designer can do (or would want to do) to produce that kind of roll off--a small series inductor or a parallel cap will cause a roll off, but at a much more gradual rate. Finally, I doubt that your measuring system can resolve any information above 20 khz, so under no circustances should you expect to see flat reaponse to 25 or 30 khz. So, putting all that together, I obviously don't have a clue what's going on. It sure looks like an artifact of your measuring system. Are you sure other speakers were flat to 20 kHz?
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #10 ·

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


I'm not familair with the measurement method you're using. I can't tell whether it's basically a room measurement (even at 3 feet), or whether it's a quasi-anechoic response above 200 Hz or so. If it is a standard anechoic FFT measurement, and if you are measuring on the tweeter axis at about 3 feet (which is close to one meter), then the response at 20 khz should be higher than shown.

I'm using Room EQ Wizard with the default impulse response windows of 125ms left and 500ms right. Even if I decrease the right window (the time measured after the signal has stopped), it has no effect on the HF response.


The 4 different plots in the first graph of this thread are at different mic positions. One on the woofer axis, one on the tweeter axis, one 6" to the left and one 6" to the right. I could find no axis in which the HF extended to 20KHz in some flatish way.

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


If you get closer than 3 feet, the response will tilt downward, but earlier than shown.

Can you rephrase this? I don't follow.


Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


If this is more of a room measurement with a long sample time and smoothing, then the room itself will absorb some of the high frequencies. Still, I would not expect the steep roll off shown.

This is an in-room measurement. I would expect to see some roll-off if the mic were further away from the speaker, but not at three feet.

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


I doubt very much that all of your tweeters are defective.

I tend to agree, but I sure wish I could figure out what is causing the roll-off in the reading.

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


And there's little a passive crossover designer can do (or would want to do) to produce that kind of roll off--a small series inductor or a parallel cap will cause a roll off, but at a much more gradual rate.

I didn't figure a crossover would affect the top of a tweeter's response, but I didn't want to rule anything out.

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


Finally, I doubt that your measuring system can resolve any information above 20 khz, so under no circustances should you expect to see flat reaponse to 25 or 30 khz.

I'm certain that with a calibration file I can get flat response from my mic up to 24KHz, and I expect to see flatish treble response up to that point, at least on some axis. I haven't found that axis yet though.

Quote:
Originally Posted by Dennis Murphy /forum/post/15495183


So, putting all that together, I obviously don't have a clue what's going on. It sure looks like an artifact of your measuring system. Are you sure other speakers were flat to 20 kHz?

Here is a reading of a small Monitor Audio Radius speaker I use in a bedroom system. This is using the same test equipment that I used to measure the speakers shown in the top graph in this thread.




Here's Home Theater Mag's measurement of that same speaker.



So I am getting response up to 20KHz from my equipment, but it does look like it may not be flat, since the spike at 20Khz in the HTMag graph is higher than the one in mine. Thoughts on the differences between the two?
 

·
Registered
Joined
·
3,509 Posts
Subscribed. I am interested in this. I haven't used the impulse within REQW yet, have you tried the old-school sine sweep from 20khz down to 20hz? Your response will be more comb filtered which you can artificially remove by decreasing the resolution to 1/12 octave.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #12 ·
I did a little research tonight, and I'm starting to think the roll-off is being caused by my somewhat cheap mic (even though it is not spec'd to roll-off).


I came across a thread in the HTShack forum that tests the Behringer ECM8000, which is very similar to the mic I use (a DBX RTA-M). I believe both mics use the same mic element, but I was lead to believe that the DBX was more accurate (better compensation in the mic circuitry). The FR tests of the ECM8000 mic show roll-off before 20KHz. Not as early as mine, but just the fact that there is any roll-off raises reasonable doubt about my mic.

http://www.hometheatershack.com/foru...lots-pics.html
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #13 ·

Quote:
Originally Posted by Jonomega /forum/post/15504685


Subscribed. I am interested in this. I haven't used the impulse within REQW yet, have you tried the old-school sine sweep from 20khz down to 20hz? Your response will be more comb filtered which you can artificially remove by decreasing the resolution to 1/12 octave.

I have been using the log sweep function of REW (the big "measure" button), which gathers the impulse response at the same time. You just click on the impulse response tab to view it.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #15 ·

Quote:
Originally Posted by WojtekO /forum/post/15505042


Out of curiosity, what did you use to produce such a nice graph?

The audio analysis tool is called Room EQ Wizard and you can export any of the graphs as .jpg files.


REW is freeware and can be found here:
http://www.hometheatershack.com/roomeq/
 

·
Premium Member
Joined
·
8,710 Posts
I suspect you have an issue with your measurement equipment or the way you are using it. Be certain your mic is pointed directly at the tweeter, btw. It is unlikely all your speakers would have the same fault.


Note: The uptick in the HF response of the HTMag plot is to be expected from any speaker not designed as a near field monitor. Since HF's roll off quicker over distance, these are typically boosted to provide a more flat response at whatever distance the designer thought would be the typical listening distance.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #17 ·

Quote:
Originally Posted by Dennis Erskine /forum/post/15505610


I suspect you have an issue with your measurement equipment or the way you are using it. Be certain your mic is pointed directly at the tweeter, btw. It is unlikely all your speakers would have the same fault.


Note: The uptick in the HF response of the HTMag plot is to be expected from any speaker not designed as a near field monitor. Since HF's roll off quicker over distance, these are typically boosted to provide a more flat response at whatever distance the designer thought would be the typical listening distance.

If I point the mic at the tweeter, I get this:




And I believe that the peak in the HTMag graph is showing the "tin can" resonance of the metal dome tweeter rather than being an intentional lift in the treble by the speaker designer.
 

·
Registered
Joined
·
7,466 Posts
Discussion Starter · #18 ·
It's taken me a while to get to the bottom of this problem, but after testing every component in the chain, it turned out to be a bad mic calibration file and improper mic orientation. I was using a generic calibration file that I got from another forum and I needed to point the mic at the speaker instead of having the tip up. Measuring with the tip up is good for capturing the room sound but bad for capturing high frequency response.



The yellow line is the response with the correct mic cal file and horizontal mic position, the blue line is with the mic in the vertical position.
 
1 - 18 of 18 Posts
Top