To be clear, I respect both listening and measurements.
When there are limited measurements, I take great notes of listening impressions.
When there limited listening impressions (few listened), I take note of measurements.
I'm also cognisant of the science of perception- for instance, that sightedlistening impressions are biased.
I’m also aware that measurements are incomplete. (?can they ever be complete)
When there extensive measurements and extensive listening are in agreement, then I have a belief that we are on the right path.
I have have never heard the SB12MNRX2, nor the NE149W.
So this is how I would read data sheets for both (I have used the MS Paint to resize the NE149W-4 (bottom picture) to match the scale of the SB12MNXR2-4 (top picture) as closely as possible...)
Analysis:
1. Frequency response:
1a) ON-AXIS: The SB12MNRX2 has straight and smooth response from ~400Hz to 5KHz, apart from the mild dip at 1500Hz.
The NE149W-04 has a straight and smooth frequency response from 200Hz to 5KHz, apart from the dip at 1700Hz.
--> practical implication- These are the ideal region of use for these drivers.
b) OFF-AXIS: looking and compare the axis responses to the on axis.
There are 2: 30 and 60°.
On the SB12MNRX2 they diverge above 6KHz (peak at 7Khz on axis, but dip at 6KHz 60° off-axis (red trace)
On the NE149W-4 there is a peak at ~6 KHz on axis, but a dip at ~5Khz Hz at the 60° off-axis.
—> This suggests the regions where chaotic (less predictable) breakup occurs.
Practical effect- ideally, would want to cross them below 6KHz and 5.3KHz, ideally.
(based on driver drivers sizes and directivity, the directivity would start to narrow earlier than this)
2. Impedance response:
a) Impedance curve flatness
The SB12MNRX2 impedance is 15 ohms @20KHz)
The NE149W is 8.6 ohms @ 20KHz)
—> the NE149W would have lower inductance; and thus potential for lower distortion in the 3rd and 5th harmonics.
a) Impedance curve smoothness
both curves are relatively smooth, apart from the mild blip at the frequency where is a dip. It is hard to see for untrained eyes. It depends on the scale used.
let's resize-
Can we (be convinced) that there is an impedance blip at around 1500Hz / 1700Hz yet?
It has been shown that frequency peaks are more audible that frequency dips. Of course, if things being equal it would be ideal that the dip is not there.
—> depending on baffle size (Amy cause a relative boost around 1KH; and our crossover, this may be managed with our crossover.
3. Driver size, excursion.
a) The SB12MNRX2 has a smaller cone, which may allow for wider dispersion.
b) The NE149W-04 has a larger cone, and thus will tend to have lower distortion/play cleaner at higher SPL.
—> maximum SPL is rarely limited by the midrange, and usually limited by the woofer due to air displacement requires at low frequencies.
At 90.50 or 89.2dB / 2.83V /1m; either be be suitable for a single 12” woofer like the SLS-830669 in a 60-75L box. After baffle step losses, the sensitivity of an single 830669 box will have lower sensitivity 89.2dB/2.83/1m
Caveats:
If the datasheets are taken under equivalent conditions eg. IEC baffle, that how I would I interpret the datasheets.
Unfortunately there are some clues that the measurements are taken under different conditions- the SB datasheet shows a ragged response under ~300Hz.
This is typically know in measurements is typically taken in a IEC baffle without a rear enclosure for the driver, and thus cancellations between the rear wave and front wave cause interference in the response.
-> Implications- to study the sub-300Hz response for this driver, you would need to simulate this in the cabinet of your own choosing.
The Peerless datasheet shows a smoother response below 300Hz, which would suggest that the rear wave it not interfering with the front wave. Peerless may also using an IEC (or larger) baffle , but must have a large enclosure (>100L) enclosing the rear of the driver to provide a closer approximation of the bass response in a "infinite" baffle
-> Implications, you would need to simulate this in a cabinet of your choosing.
Both drivers wild be suitable for a midrange. On the same cabinet/baffle, the NE149W would have lower distortion, but the SB12MNRX2 would have lower directivity (wide dispersion)
Experienced loudspeaker designers can pickup this information at a glance, others still learning what to make on this limited data.
So whilst measurements can show pros/cons, and one cannot place utmost important value on just frequency response and impedance data, unfortunately the vast majority of manufacturers only show us these 2 (apart from T/S parameters)
Looking forward to hearing feedback on how others read FRD and Z data.
In my view; they are both suitable.
But after putting them into cabinets, and doing the crossover to have perfectly equal frequency responses in the on axis; the 2 other difference, in my view, is the differences in directivity
(ie. focus vs spaciousness) or differences in distortion (richness or cleanliness, particularly at lifelike SPLs)
But what sounds “better” to one person might sound “better” to another.
When there are limited measurements, I take great notes of listening impressions.
When there limited listening impressions (few listened), I take note of measurements.
I'm also cognisant of the science of perception- for instance, that sighted
I’m also aware that measurements are incomplete. (?can they ever be complete)
When there extensive measurements and extensive listening are in agreement, then I have a belief that we are on the right path.
I have have never heard the SB12MNRX2, nor the NE149W.
So this is how I would read data sheets for both (I have used the MS Paint to resize the NE149W-4 (bottom picture) to match the scale of the SB12MNXR2-4 (top picture) as closely as possible...)
Analysis:
1. Frequency response:
1a) ON-AXIS: The SB12MNRX2 has straight and smooth response from ~400Hz to 5KHz, apart from the mild dip at 1500Hz.
The NE149W-04 has a straight and smooth frequency response from 200Hz to 5KHz, apart from the dip at 1700Hz.
--> practical implication- These are the ideal region of use for these drivers.
b) OFF-AXIS: looking and compare the axis responses to the on axis.
There are 2: 30 and 60°.
On the SB12MNRX2 they diverge above 6KHz (peak at 7Khz on axis, but dip at 6KHz 60° off-axis (red trace)
On the NE149W-4 there is a peak at ~6 KHz on axis, but a dip at ~5Khz Hz at the 60° off-axis.
—> This suggests the regions where chaotic (less predictable) breakup occurs.
Practical effect- ideally, would want to cross them below 6KHz and 5.3KHz, ideally.
(based on driver drivers sizes and directivity, the directivity would start to narrow earlier than this)
2. Impedance response:
a) Impedance curve flatness
The SB12MNRX2 impedance is 15 ohms @20KHz)
The NE149W is 8.6 ohms @ 20KHz)
—> the NE149W would have lower inductance; and thus potential for lower distortion in the 3rd and 5th harmonics.
a) Impedance curve smoothness
both curves are relatively smooth, apart from the mild blip at the frequency where is a dip. It is hard to see for untrained eyes. It depends on the scale used.
let's resize-
Can we (be convinced) that there is an impedance blip at around 1500Hz / 1700Hz yet?
It has been shown that frequency peaks are more audible that frequency dips. Of course, if things being equal it would be ideal that the dip is not there.
—> depending on baffle size (Amy cause a relative boost around 1KH; and our crossover, this may be managed with our crossover.
3. Driver size, excursion.
a) The SB12MNRX2 has a smaller cone, which may allow for wider dispersion.
b) The NE149W-04 has a larger cone, and thus will tend to have lower distortion/play cleaner at higher SPL.
—> maximum SPL is rarely limited by the midrange, and usually limited by the woofer due to air displacement requires at low frequencies.
At 90.50 or 89.2dB / 2.83V /1m; either be be suitable for a single 12” woofer like the SLS-830669 in a 60-75L box. After baffle step losses, the sensitivity of an single 830669 box will have lower sensitivity 89.2dB/2.83/1m
Caveats:
If the datasheets are taken under equivalent conditions eg. IEC baffle, that how I would I interpret the datasheets.
Unfortunately there are some clues that the measurements are taken under different conditions- the SB datasheet shows a ragged response under ~300Hz.
This is typically know in measurements is typically taken in a IEC baffle without a rear enclosure for the driver, and thus cancellations between the rear wave and front wave cause interference in the response.
-> Implications- to study the sub-300Hz response for this driver, you would need to simulate this in the cabinet of your own choosing.
The Peerless datasheet shows a smoother response below 300Hz, which would suggest that the rear wave it not interfering with the front wave. Peerless may also using an IEC (or larger) baffle , but must have a large enclosure (>100L) enclosing the rear of the driver to provide a closer approximation of the bass response in a "infinite" baffle
-> Implications, you would need to simulate this in a cabinet of your choosing.
Both drivers wild be suitable for a midrange. On the same cabinet/baffle, the NE149W would have lower distortion, but the SB12MNRX2 would have lower directivity (wide dispersion)
Experienced loudspeaker designers can pickup this information at a glance, others still learning what to make on this limited data.
So whilst measurements can show pros/cons, and one cannot place utmost important value on just frequency response and impedance data, unfortunately the vast majority of manufacturers only show us these 2 (apart from T/S parameters)
Looking forward to hearing feedback on how others read FRD and Z data.
In my view; they are both suitable.
But after putting them into cabinets, and doing the crossover to have perfectly equal frequency responses in the on axis; the 2 other difference, in my view, is the differences in directivity
(ie. focus vs spaciousness) or differences in distortion (richness or cleanliness, particularly at lifelike SPLs)
But what sounds “better” to one person might sound “better” to another.
Last edited:
I have used the SB15NBAC30 in two different active 3-way speakers, and I was very pleased with the results. Actually in one case it was the CAC ceramic coated cone rather than the NBAC black aluminum cone, but they are very nearly interchangeable.
I think with this driver we would have to carefully consider if the sensitivity is enough to match up with a 90 dB/m/2.83V woofer. Assuming the sensitivity makes sense, I think this driver should be added to the list of candidates. It would definitely need a parallel notch filter in series with the driver.
j.
I think with this driver we would have to carefully consider if the sensitivity is enough to match up with a 90 dB/m/2.83V woofer. Assuming the sensitivity makes sense, I think this driver should be added to the list of candidates. It would definitely need a parallel notch filter in series with the driver.
j.
Awesome analysis and very helpful (not just to me, but for the casual follower of the thread.
How do you get this...I had a post, as did diyiggy and mbrennwa a few pages back regarding the sensitivity after baffle step losses. This not only matters for our midrange selection, but there was a recent comment regarding tweeters too.
Many of you are again recommending too expensive drivers again. This is a budget project and mids for about 75-100€/$ don't fit that plan i think. Try to keep it under 50€/$ please (or at least in that region). That NBCAC is a good mid, but about 85€ in europe, and so are some of the others.
Datasheet of the 830669 says sensitivity is 88.7 dB/2.83V
If you model in winISD (2pi) the sensitivity should be around this.
If you model in another, that takes into account baffle dimensions Eg. VuitixCAD2, it will be lower than this.
If you’re getting higher than the datasheet, either the datasheet is slightly wrong, or you’ve by chance, discovered free energy.
Joking aside, bass is determined by TSP parameters when modelling in sealed and bass reflex enclosures.
We’re not talking about tapped horns or horn loaded midranges etc.
In a midrange, it doesn’t not matter too much, just put it in a well designed sub-enclosure of several litres, with deadening and damping materials and plenty of space behind the driver.
If you model in winISD (2pi) the sensitivity should be around this.
If you model in another, that takes into account baffle dimensions Eg. VuitixCAD2, it will be lower than this.
If you’re getting higher than the datasheet, either the datasheet is slightly wrong, or you’ve by chance, discovered free energy.
Joking aside, bass is determined by TSP parameters when modelling in sealed and bass reflex enclosures.
We’re not talking about tapped horns or horn loaded midranges etc.
In a midrange, it doesn’t not matter too much, just put it in a well designed sub-enclosure of several litres, with deadening and damping materials and plenty of space behind the driver.
Yes, I will add it.
I had some subjective reasons why I left it off MY short list, but that list was not supposed to be anything official but rather a starting point.
- It was brought up as a good choice for the "more expensive" option with the SB34NRX. If that build happens, I think it would be cool to have different midranges so they are more differentiated than just the woofer.
- It is more than the NE149W. My subjective ranking is that if we went up to a $75 woofer, the NE149 is the pinnacle of value. (With the frequent 12% off sales at PE the NE149W would be $66 for US buyers. )
- The notch filter. I included the SIG150 just because the price is good, but otherwise left aluminum cones off my list (but again, that was MY own subjective list)
@waxx
Do you have any experience with Lavoce drivers?
The datasheet looks a bit too smoothed to my liking.
This 4” which may be competitive with the SB12MNRX2
32€ in Europe from THLP
and $45 from PE.
Is this price OK?
Do I even REALLY need to model? 2pi minus 4 to 6dB gives me a good enough estimate at this stage of driver selection, and then if we are cutting it really close do more precise modelling with both the woofer and midrange?
By the way, in my post I did ask for help checking my logic...mainly because I'd like to understand this more precisely. And I was hoping you would see that post and comment, at least if I was quite wrong in some aspect. Take a look at post 507.
Now I'm sure we've seen all kinds of wacky frequency responses:
Here two ~5.5" drivers that could be considered as a midrange.
1. The first one is measured in-box, centered, dimensions 34.3 x 19.8 cm (LxW) with a round-over of ~3/4"
Frequency responses:
Impedance response:
2. The second one is measured in a box 41.4 x 22.2 cm (L x W) ; with a round-over radius of 10mm
2.83V drive:
Frequency responses:
Impedance response:
Anyone want to guess what these graphs may (or may not show), and comment for suitability in our budget 3 way?
Here two ~5.5" drivers that could be considered as a midrange.
1. The first one is measured in-box, centered, dimensions 34.3 x 19.8 cm (LxW) with a round-over of ~3/4"
Frequency responses:
Impedance response:
2. The second one is measured in a box 41.4 x 22.2 cm (L x W) ; with a round-over radius of 10mm
2.83V drive:
Frequency responses:
Impedance response:
Anyone want to guess what these graphs may (or may not show), and comment for suitability in our budget 3 way?
There was a 3 way project done by Klang&Ton and a single SLS12 SPL in a finished loudspeaker was measured around 87 dB/2.83V/1m.
I miss a few things in this summary. The (rising) acoustic impedance of a cone also influences output. You can't just look at the electric impedance curve and judge there have to be cone resonances. Also: good soft cones (paper, mainly PP and the B&W FST of course) are breaking up by design, to reduce the radiating area with rising frequency. Mainly to keep directivity right. As long as those breakups aren't culminating in severe resonances a cone could do a better job than one not breaking up.
This (the 5FE125) still seems a very wise advice to me. Easy round basket, nice price, proper motor design.
Last edited:
No experience, and i hear very mixed sounds about them, the main complaint is inconsistancy of the specs on different copies of the same model. That's why i don't consider them. There are better EU brands who make the same style like Faital, Beyma, B&C, 18Sound, ... and the first two are actually quiet affordable, the the others mostly not (but not extreme expensive neighter).
To be usefull about the german mag. that measured the SLS-12, we need to know what is the measured frequency or if it is averadged on its filtered audio band ?
Dibirama measured one at 91.5/2.83V IEC baffle I assume. Doesn't matter much as IEC the baffle is almost invisible at the low frequencies of interrest.
I simmed (in a post above) with 37 cm baffle in Vituix I had 4 dB loss at 100 Hz. At looking at twice it goes to the 6/-7 dB towards 40 Hz as @mbrennwa inputed (it was with a driver 12PR320 weak already in its very low end cause the unit is a pro midwoofer not a sub or hifi woofer with low Fs).
So with the room modes or according where they are in the low end, 87 dB is certainly a good approximation. And that's here I'm a little lost cause we have to understand at 250 hz cut off we are more at the middle of the baffle step I assume. So 3Pi at cut off. Rigth in the middle between full and half space.
I measured a loss of 3 dB max for 37 cm baffle for the mid unit between 1 K hz and 250 Hz. So I am lost in the - 6 dB to choose the mid efficienty ?!
My two cents logic says it is 97 dB + 3db needed at 250 Hz and -6dB if the ref is the low end (should it be, I assume we all are talking of flat on axis spl power response up to 1 Khz).
So What we should look is the 40 to 50 Hz or at least the -3 dB of the woofer measured 4 Pi, then take +6 dB for the choice of the midrange or a little less due to the gain of the filter ? (The filter loss however we often see in Rew or Vituix in the midrange from the filter curves window is 2 Pi I surmise).
In the scenario of the loudspeakers purchased in the shops, should we assume the real sensivity seen for a standalone amp is between -4 and - 6 dB for real in relatiuon to the numbers of the datashheet ?
I asked several time at Diyaudio and I was said to don't stand by the room gain (underrstanding if you peaks and deep this is the room modes).
So flat from say 50/60 Hz is asking a -6dB dB midrange if it is not padded down by resistor, so ideally -4 dB if you want some resistor attenuation for the voice coil power handling in front of the miid unit ?
What about the 5" SB NRX in spite of the 4" (50cm² Sd only) ?
Peereless has also a woven fiber midwoof @wolf_teeth liked a lot in one of his design.
The Monacor talked also is a good path as well.
I am not worried when we chose to move forward, the bass driver (price) and maxi size of the cabinet, people would like to take the train.
Dibirama measured one at 91.5/2.83V IEC baffle I assume. Doesn't matter much as IEC the baffle is almost invisible at the low frequencies of interrest.
I simmed (in a post above) with 37 cm baffle in Vituix I had 4 dB loss at 100 Hz. At looking at twice it goes to the 6/-7 dB towards 40 Hz as @mbrennwa inputed (it was with a driver 12PR320 weak already in its very low end cause the unit is a pro midwoofer not a sub or hifi woofer with low Fs).
So with the room modes or according where they are in the low end, 87 dB is certainly a good approximation. And that's here I'm a little lost cause we have to understand at 250 hz cut off we are more at the middle of the baffle step I assume. So 3Pi at cut off. Rigth in the middle between full and half space.
I measured a loss of 3 dB max for 37 cm baffle for the mid unit between 1 K hz and 250 Hz. So I am lost in the - 6 dB to choose the mid efficienty ?!
My two cents logic says it is 97 dB + 3db needed at 250 Hz and -6dB if the ref is the low end (should it be, I assume we all are talking of flat on axis spl power response up to 1 Khz).
So What we should look is the 40 to 50 Hz or at least the -3 dB of the woofer measured 4 Pi, then take +6 dB for the choice of the midrange or a little less due to the gain of the filter ? (The filter loss however we often see in Rew or Vituix in the midrange from the filter curves window is 2 Pi I surmise).
In the scenario of the loudspeakers purchased in the shops, should we assume the real sensivity seen for a standalone amp is between -4 and - 6 dB for real in relatiuon to the numbers of the datashheet ?
I asked several time at Diyaudio and I was said to don't stand by the room gain (underrstanding if you peaks and deep this is the room modes).
So flat from say 50/60 Hz is asking a -6dB dB midrange if it is not padded down by resistor, so ideally -4 dB if you want some resistor attenuation for the voice coil power handling in front of the miid unit ?
What about the 5" SB NRX in spite of the 4" (50cm² Sd only) ?
Peereless has also a woven fiber midwoof @wolf_teeth liked a lot in one of his design.
The Monacor talked also is a good path as well.
I am not worried when we chose to move forward, the bass driver (price) and maxi size of the cabinet, people would like to take the train.
Last edited:
I can't find the build, so I'm speculating. Would they have gone with 89dB if they could?
Were they limited by the mid choice, or was it padded down to get an 87dB speaker sensitivity?
Did they comment on room placement or tonal balance?
Edit "My two cents logic says it is 97 dB + 3db needed at 250 Hz and -6dB if the ref is the low end (should it be, I assume we all are talking of flat on axis spl power response up to 1 Khz)."
I was meaning of course 87 dB + 3 dB (for the midwoof, in ordder to be flat at the cut off and having few resistor attenuation to protect from thermal compression most of the voice coils of cheap mifwoofers). But 87 dB is certainly high enough as imputed Mbrennwa and the mids inputed till now.
I was meaning of course 87 dB + 3 dB (for the midwoof, in ordder to be flat at the cut off and having few resistor attenuation to protect from thermal compression most of the voice coils of cheap mifwoofers). But 87 dB is certainly high enough as imputed Mbrennwa and the mids inputed till now.
Last edited: