Good evening dorks. I picked up a pair of these speakers because I wanted a small accurate desktop monitor for editing dialog. They are good sounding speakers, but I did want something fairly accurate, so I took quasi anechoic measurements of the speakers this evening.
The measurements are good to about 200hz, where I gated. I also took measurements in the nearfield, merging them with the farfield gated polars, getting a reasonably accurate picture of the speaker's response from 0-120 degrees, in 10 degree increments. If you don't know what gated measurements are, basically it is a clever way of ignoring parts of the impulse response which occur after some set duration, enabling one to remove room effects from the response. This distinguishes these measurements from others seen on the internet - these are quite accurate above 200hz, and reasonably representative below, as if they were taken in an anechoic chamber.
Measurement notes:
ARTA using MLS sequences for the upper frequency range and sine sweeps for the nearfield woofer measurement
Speaker was set to '0db' for the LF trim and '-2db' for the HF trim.
Nearfield woofer measurements spliced in around 200hz, and I think I may have set it a bit too high.
Mic is a manufacturer calibrated emm-6 about 2' from the tweeter at tweeter level
I didn't do vertical responses because I am lazy
Ignore the absolute level- absolute SPL is meaningless here
Importing the measurements into VituixCAD, a great freeware speaker design package, I get the following:
Well, considering the treble is already attenuated by 2db, this is not very flat. Note the vertical scale of the measurements; nevertheless I am seeing a pronounced non resonances in the 2-4k region. This is bad; on the other hand, the polar response is excellent. The green line on the top graph is an average of the responses from 0 out to 15 or 30 degrees - basically a reasonable hifi listening window. You can hardly see this line because it matches the direct response (0 deg) out to 10K! This would be exceptional for a conventional hifi speaker; very few baffle mounted tweeters have dispersion this wide that high up.
Also in the 'good news' category is the directivity sonogram seen below. This is not normalized, and shows clear directivity control down to 500hz or maybe even a bit below. This is the stuff that can't be corrected with DSP - this is the fundamental acoustic design of the speaker, and it is outstanding, especially for a small speaker. You might get a polar response this smooth with a conventional 2 way, using a high crossover frequency and a LR2 slope, but you won't get it as tight.
We can also see that in the area of confidence (say, 300hz up) that there are no spatial discontinuities - there's no holes or peaks in the response as you move up and down at a given frequency. This is nice, because it means that if the direct response was EQ'd flat, you wouldn't make anything worse off axis.
So what can we do? Well, with dsp we can do some correction of the speaker. Generally DSP correction is discussed in the context of low frequency, to help combat room effects below the Schroeder frequency. I am more interested in making the speaker more linear in the 200-20K region. Using Room EQ wizard, we can do this pretty easily.
Notes on correction:
I set REW to only correct above 300hz - it's pretty rare that there are sharp resonances in speakers below this range.
I used the 'generic' preset, which gives you 20 bands of PEQ. If you're using an equalizer with fewer bands, I'd focus on the low Q corrections, but then again I'm not exactly an expert.
Here is the response before and after:
As you can see, it's making a ton of corrections, but the minimalist in me would suggest the most important is the boost from 300 to 1K, and the taming of the peak around 3K.
Anyway, here's my review - good sounding speakers but not what I would expect for a reference monitor. When I use them for that purpose I'll be using REAPER, and I'll throw compensating EQ into the master bus. I've heard that works. I'm not sure if this EQ makes the speaker sound better; I haven't listened yet, but bringing the mids and taming that peak would be what I would do if I was designing this speaker. Will take further measurements as sanity check.
The measurements are good to about 200hz, where I gated. I also took measurements in the nearfield, merging them with the farfield gated polars, getting a reasonably accurate picture of the speaker's response from 0-120 degrees, in 10 degree increments. If you don't know what gated measurements are, basically it is a clever way of ignoring parts of the impulse response which occur after some set duration, enabling one to remove room effects from the response. This distinguishes these measurements from others seen on the internet - these are quite accurate above 200hz, and reasonably representative below, as if they were taken in an anechoic chamber.
Measurement notes:
ARTA using MLS sequences for the upper frequency range and sine sweeps for the nearfield woofer measurement
Speaker was set to '0db' for the LF trim and '-2db' for the HF trim.
Nearfield woofer measurements spliced in around 200hz, and I think I may have set it a bit too high.
Mic is a manufacturer calibrated emm-6 about 2' from the tweeter at tweeter level
I didn't do vertical responses because I am lazy
Ignore the absolute level- absolute SPL is meaningless here
Importing the measurements into VituixCAD, a great freeware speaker design package, I get the following:
Well, considering the treble is already attenuated by 2db, this is not very flat. Note the vertical scale of the measurements; nevertheless I am seeing a pronounced non resonances in the 2-4k region. This is bad; on the other hand, the polar response is excellent. The green line on the top graph is an average of the responses from 0 out to 15 or 30 degrees - basically a reasonable hifi listening window. You can hardly see this line because it matches the direct response (0 deg) out to 10K! This would be exceptional for a conventional hifi speaker; very few baffle mounted tweeters have dispersion this wide that high up.
Also in the 'good news' category is the directivity sonogram seen below. This is not normalized, and shows clear directivity control down to 500hz or maybe even a bit below. This is the stuff that can't be corrected with DSP - this is the fundamental acoustic design of the speaker, and it is outstanding, especially for a small speaker. You might get a polar response this smooth with a conventional 2 way, using a high crossover frequency and a LR2 slope, but you won't get it as tight.
We can also see that in the area of confidence (say, 300hz up) that there are no spatial discontinuities - there's no holes or peaks in the response as you move up and down at a given frequency. This is nice, because it means that if the direct response was EQ'd flat, you wouldn't make anything worse off axis.
So what can we do? Well, with dsp we can do some correction of the speaker. Generally DSP correction is discussed in the context of low frequency, to help combat room effects below the Schroeder frequency. I am more interested in making the speaker more linear in the 200-20K region. Using Room EQ wizard, we can do this pretty easily.
Notes on correction:
I set REW to only correct above 300hz - it's pretty rare that there are sharp resonances in speakers below this range.
I used the 'generic' preset, which gives you 20 bands of PEQ. If you're using an equalizer with fewer bands, I'd focus on the low Q corrections, but then again I'm not exactly an expert.
Here is the response before and after:
As you can see, it's making a ton of corrections, but the minimalist in me would suggest the most important is the boost from 300 to 1K, and the taming of the peak around 3K.
Anyway, here's my review - good sounding speakers but not what I would expect for a reference monitor. When I use them for that purpose I'll be using REAPER, and I'll throw compensating EQ into the master bus. I've heard that works. I'm not sure if this EQ makes the speaker sound better; I haven't listened yet, but bringing the mids and taming that peak would be what I would do if I was designing this speaker. Will take further measurements as sanity check.
