I’ve completed the measurements for the Satori MW16P-4 mid-woofer. This is the new mid-woofer “flagship” driver from SB Acostics and is clearly targeted to compete with the very best mid-woofer drivers from e.g. ScanSpeak Revelator and Illuminator series and perhaps drivers from AudioTechnology and Accuton etc.
My tested drivers are the 4 Ohm version of the Satori mid-woofer.
For detailed measurements see: SB Acoustics “Satori” MW16P-4 Measurements!
How does it stack up against the competition?
If we first look at the price for this driver, it costs 288$ which isn’t exactly the cheapest driver out there. Here's a price comparison of similar drivers that compete in the same class:
The Satori mid-woofer places it self somewhere in the middle of the price range, but to be honest it is also a slightly larger mid-woofer than the rest of the drivers on list. There aren’t many true 6.5” mid-woofers on the market today and on the list the Accuton driver is the closest match.
If we increase the size to about 7”, which is the most common size today the comparison fails a bit since with this driver size we seldom can use them up to about 3kHz, as in the case of the Satori.
If we look at the technical aspect of the Satori mid-woofer and to sum-up its key performance features, we have the following:
- A low resonance frequency of 29Hz out of the box without any burn-in.
- A largelinear excursion capability of +/- 6mm.
- A very low inductance.
- A very stiff cone without any severe cone break-ups.
- An extreme linear and extended frequency response up to 15kHz, on-axis.
- A low loss rubber surround with very low mechanical damping.
- A low mechanical noise from surround at higher excursions
- A high Qms and low Rms together with the properties of the surround, provides the possibility of a fast and dynamic bass response.
- A very low distortion profile throughout the mid-woofers usable range.
- A usable frequency range up to 3kHz and it allows the use of shallow sloped filters.
- A high sensitivity of approximately 90dB 2.83V/1m
- A low bass response from a relative small enclosure of 18-22 liters (34-35Hz @ -6dB)
- A good consistency between the two tested driver unit samples.
Are there any cons? Yes, it has a modest spike of odd-order distortion and an elevated second-order distortion at the cone edge resonance.
I think this is an excellent top performer and worth its price compared to the competitors, but I haven’t yet subjectively listened to it in an actual loudspeaker design and I can’t tell if it is also high performer audibly, but it certainly has the potential to be a top sonic performer!
Right now I’m testing different tweeters to match the Satori and eventually there will be a “Satori Two – Monitor” design.
Merry Christmas & a Happy new Year!
/Göran
Attachments
I missed this when you posted it, very nice performance from the Satori there, it's a shame about that small resonance, but as the driver is so linear anyway, the peak in the 3rd order doesn't reach up that high and the 2nd order peak should be inaudible.
The absolute level of 2nd order products tends not to worry me in general with respect to driver performance and I prefer to use the third order as a guide to the quality of the drivers motor. Certainly having nominal levels closer to 0.01% rather than 0.1% places it among the best. In this regard it reminds me of the B&W FST, having similarly low levels of 3rd harmonic throughout its useful range, that then varies only by a small amount when pushed.
It absolute terms the Satori is probably the best driver you've measured so far, it will be interesting to see what your subjective impressions of it are.
The absolute level of 2nd order products tends not to worry me in general with respect to driver performance and I prefer to use the third order as a guide to the quality of the drivers motor. Certainly having nominal levels closer to 0.01% rather than 0.1% places it among the best. In this regard it reminds me of the B&W FST, having similarly low levels of 3rd harmonic throughout its useful range, that then varies only by a small amount when pushed.
It absolute terms the Satori is probably the best driver you've measured so far, it will be interesting to see what your subjective impressions of it are.
What i think is interesting is that the distortion profile does not change much over level.
I am working at the moment on power amps and found when the profile does not change much over level the amp sound better.
It seems to be that the ear is more sensitive to unmusical changes then in the absolute value of the distortion when it is under a certain percentage and low order.
I also like very much the low inductance over frequency of the Satori.
Inductive modulation is a very unpleasant type of distortion. It sound like "crrkk, crrk, crrk" in tact with the music. I know it because i can isolate that distortion on my Klippel.
I am working at the moment on power amps and found when the profile does not change much over level the amp sound better.
It seems to be that the ear is more sensitive to unmusical changes then in the absolute value of the distortion when it is under a certain percentage and low order.
I also like very much the low inductance over frequency of the Satori.
Inductive modulation is a very unpleasant type of distortion. It sound like "crrkk, crrk, crrk" in tact with the music. I know it because i can isolate that distortion on my Klippel.
Originally I think SB wanted to make a statement with the Satori, very high performance but with a relatively attractive price, then the value of neo went up
Actually the price increase was *after* neo went up in price and then down again (though not as low as before).
Personally I think they listened to Scan's marginal "Illuminator" line as their closest competitor and realized they had a better product - and therefor increased the price. (..might have had something to do with the new Wilson's as well.)
Yes, it has an incredible measurement performance. I have high hopes for this mid-woofer and I'm currently started some simulation and tweeter selection for a "Satori Two - Monitor" design.
Regards
/Göran
Thanks for the measurements. It is absolutely brilliant that we have people reporting measurements.
I have a few questions regarding the measurements:
1. I see you are using STEPS for the measurements. Can you post your measurement setup in STEPS? I'm just curious to find out what people are using and compare it to what I use. Maybe I'll learn something.
2. Your mic position is close to the woofer, but when measuring indoors, STEPS will integrate the output over the integration time. Reflections may have an impact on the distortion result. It may be useful to look at the frequency response to see if it is smooth. I would be really happy if you could post comparable measurements taken outdoors.
3. Everyone is excited at the low distortion level. But remember that is at 1m. I find 2m to be a much better distance to judge the distortion. I'm not saying measure 2m away, just normalize for 2m. That will raise the output by 6db over the 1m level.
I typically listen at 8-10 ft away. I find small bookshelf two-ways to not have enough dynamic capability. I don't know what common average SPLs are, but let's assume 75 db at 1kHz as an average level. Good classical recordings can have more than 20db dynamic range. That means 95db on peaks. Normalize 95db at 2m back to 1m and you get 101 db. Now look at the distortion and it may be a different picture.
Lastly, low-order distortion is more difficult to hear. Keep an eye on the 5th and 6th-12th order distortion. This will not be masked as much by the fundamental, whereas the lower orders will be masked.
I have a few questions regarding the measurements:
1. I see you are using STEPS for the measurements. Can you post your measurement setup in STEPS? I'm just curious to find out what people are using and compare it to what I use. Maybe I'll learn something.
2. Your mic position is close to the woofer, but when measuring indoors, STEPS will integrate the output over the integration time. Reflections may have an impact on the distortion result. It may be useful to look at the frequency response to see if it is smooth. I would be really happy if you could post comparable measurements taken outdoors.
3. Everyone is excited at the low distortion level. But remember that is at 1m. I find 2m to be a much better distance to judge the distortion. I'm not saying measure 2m away, just normalize for 2m. That will raise the output by 6db over the 1m level.
I typically listen at 8-10 ft away. I find small bookshelf two-ways to not have enough dynamic capability. I don't know what common average SPLs are, but let's assume 75 db at 1kHz as an average level. Good classical recordings can have more than 20db dynamic range. That means 95db on peaks. Normalize 95db at 2m back to 1m and you get 101 db. Now look at the distortion and it may be a different picture.
Lastly, low-order distortion is more difficult to hear. Keep an eye on the 5th and 6th-12th order distortion. This will not be masked as much by the fundamental, whereas the lower orders will be masked.
Last edited:
I’m also eager to have them up and running for a listening test, but I need some custom made enclosure for them and the plan is to build the new enclosures late spring or early summer.
I have some cross-over simulations ready to be tested. The first run of test will use a couple of different ScanSpeak tweeters on a stepped baffle together with a LR2 cross-over.
/Göran
I normally use ARTA Steps for distortion measurements and HolmImpulse for frequency measurements. There isn’t really much to setup in Steps besides having the sound card calibrated and setting the output level and frequency range for the measurement. I usually use 200-10000Hz for mid-woofers and 500-10000Hz for tweeters with a step of 1/12 octave. If I look for some specific narrow banded issue, I increase the step to1/24 octave or higher.
I usually do a lot more measurements than I publish on the website. The published measurements are a selection of measurements and are made under similar conditions so the results can be easily compared between the different driver units presented on the website. The close mic setup is used to minimize the room influence. Measuring outside is not an option when you live in a crowded city with a lot of ambient noise and with a current outside temperature of -5deg Celsius (23deg F).
I usually do distortion measurements for an output of 85, 90 and 95dB @ 1m, sometimes higher. This is basically the standard way most papers, magazines and websites publish distortion measurements. The data I publish is too give people a general idea of how different drivers distortion profile looks like, but I strongly recommend anyone to do their own distortion measurements if they have a specific area/setup they want to explore. My measurements don’t replace such measurements, by any means.
I publish only the 2, 3 and 5 order distortion on my measurements simply because I don’t want to clutter the picture with too many graphs, but all other higher distortion are there in the measurement and if there are any anomalies I report it.
In the distortion graphs I usually look for cone break-ups, rising distortion of tweeters lower frequencies and mainly the 3 and 5 odd-order distortion profile a couple of octaves below and above the intended cross-over frequency. Spikes of 2 order distortion could also be interesting, but is usually no problem if they are narrow banded.
I think most modern quality made drivers have a reasonable good distortion profiles, some are better like the Satori and some are worse. The distortion measurements could be a good guidance to see the driver’s limitations and to help deciding suitable cross-over frequencies and driver matching as well as dealing with cone break-ups etc.
As always there are hundreds of parameters and design choices to weight against each other to make a successful design. Distortion is one of them.
Regards
/Göran
Agree with what you say.
I find that cone breakup is more easily seen the on and off axis curves and a cumulative spectral decay plot. Regarding distortion, you can display the 6th to 12th harmonic in STEPS.
We don't have to do things because magazines do it that way. Besides, if you look at the soundstage network measurements (a superb resource!), they do report distortion 95db at 2m.
I was asking about what you use for the integration time, transient time, intra-burst pause and delay.
Don't get me wrong, I really appreciate what you do. Measurements, and especially good ones, are rare to come by. I'm here to learn and discuss our hobby.
I find that cone breakup is more easily seen the on and off axis curves and a cumulative spectral decay plot. Regarding distortion, you can display the 6th to 12th harmonic in STEPS.
We don't have to do things because magazines do it that way
. Besides, if you look at the soundstage network measurements (a superb resource!), they do report distortion 95db at 2m.I was asking about what you use for the integration time, transient time, intra-burst pause and delay.
Don't get me wrong, I really appreciate what you do. Measurements, and especially good ones, are rare to come by. I'm here to learn and discuss our hobby.
We don't have to do things because magazines do it that way
If I had my own anechoic chamber like they do, I would probably do similar distortion measurements!
However, I do 2m system frequency measurements of the finished loudspeaker design, but these measurements are only valid down to aproximately 400Hz in my room. I wish I could do 2.5-3m anechoic, but that would require a chamber, a larger room or an outdoor measurement.
I use low latency ASIO drivers for my sound card and default values seems to be working fine with my sound card and setup. I/O delay = distance/delay between the mic and driver unit.
Here is an explanation from a German STEPS guide:
Min. integration time (ms):
STEPS determines the frequency response of the portion of the signal, according to the "I/O delay" and "Transient Time" begins by integrating the Sine wave signal in the time domain. This time is called "Integration Time".
The time required depends on the lowest desired frequency. When a signal with the lowest frequency F [Hz] is to be measured, then the integration time must be a minimum of 1 / F
ARTA and STEPS still use a special filtration of Signal by applying a windowing to Kaiser. This requires complete a minimum of 5 cycles (250 ms at 20 Hz). If a faster measurement is wanted it can only be achieved by increasing the lower frequency.
Note: For distortion measurements you should double the Integration time that should be used (at 20 Hz, e.g. 500 ms).
Transient time (ms):
The measurement of the sinusoidal signal has to be done in the steady state. When the steady state is reached depends on the resonant behavior of the system or from the acoustic reverberation.
For room measurements the "Transient Time" should be at least 1/5 of the reverberation time. Values between 100 ms and 200 ms are common for normal rooms. For outdoor measurements, the "Transient Time" will be set from 50 ms to 100 ms.
I/O delay (ms):
Due to the distance between the microphone and speaker a delay always exists. To get a sensible phase response, this delay must be taken into account.
Intra burst pause (ms):
After a measurement, the system has to fade out before you can start with the next measurement. This break is called "intra burst pause". As a rough value for the duration of the "Intra burst pause", 1/5 the reverberation time will be accepted.
Sorry for the bad German to English translation!
I do not
Regards
/Göran
Goran, you don't need an anechoic chamber. Just measure at 1m but raise the SPL by 6db over target. So, if you want to measure 95db at 2m, raise it to 101db at 1m. I'm assuming you are doing something similar when you position your mic close to the woofer for the 1m distortion measurement.
Regarding the I/O delay, I set it to zero and then take a sample measurement. Then go to time record and see at what time the first impulse is received. Then use that as the delay. Anyway, I experimented with each of the four parameters and the I/O delay did not have a big impact on the FR or distortion.
Thanks for the translation! The STEPS documentation is quite poor. No guidance is given on what could be typical values. I tried looking online, but didn't find much guidance.
One thing that concerns me is the "Note" in guidance you mention above. They are suggesting multiplying the time by 10 (20 Hz or 50ms becomes 500 ms integration time). If you do this in a living room, all the reflections are going to get into the distortion result.
Regarding the I/O delay, I set it to zero and then take a sample measurement. Then go to time record and see at what time the first impulse is received. Then use that as the delay. Anyway, I experimented with each of the four parameters and the I/O delay did not have a big impact on the FR or distortion.
Thanks for the translation! The STEPS documentation is quite poor. No guidance is given on what could be typical values. I tried looking online, but didn't find much guidance.
One thing that concerns me is the "Note" in guidance you mention above. They are suggesting multiplying the time by 10 (20 Hz or 50ms becomes 500 ms integration time). If you do this in a living room, all the reflections are going to get into the distortion result.
Yes, of course, but I was referring on how they did their measurements, which looks like they actually have the mic at a 2m distance. By using the mic this way you get a better summation of the drivers for the distortion measurements.
I think the best way here is to test different settings and explore the results. As I earlier said, the default settings works fine for me in my setup, but it would be great if someone dug in to this deeply and test different settings more thoroughly and write a "best practice" guide.
Perhaps that´s something to pick up in a new thread?
Anyone tempted?
Regards
/Göran
the suspension system seems very advanced probably the most linear response as I have seen the kmx curve of this and its super linear didnt expect that it has such good and I believe it has phenomenal impact on the sound as the similar driver is being used in the Wilson Audio Alexia which seems to use this satori driver in it..
I think you should update your blog section on audioexcite because lot of people is reading it and is following you. Keep us informed about your work please, its better to have the same point of information on your website than searching here and there for some information.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.