Hi everyone,
So how do you perform off axis measurements of your dipole/open baffle ?
To get an idea I used to just quickly measure the driver response moving around it, in place, but nothing very precise and always influenced by the room.
And I also relied on what I could find online about each driver, official's spec, diy measurements, driver size etc.
Now I'd like to get a better idea of my power response.
I've read about the outdoor method, high off the ground on a rotating plate, but I don't see how I could do it easily.
Well certainly not now in winter but I have space so I could try this summer, but the "speaker" is not that easy to move around.
It's in multiple parts, all nudes (nupole?
), woofers on the floor and hanging drivers on a tripod, everything tied in its place and angled with zip ties (yes ).
Excluding the Woofers it's a MTW arrangement, this mostly for aesthetic reason, low mid below is just plain ugly, and I do like the sonic result as even bass male voices are always at ear level.
So I would need a huge rotating table and I'm a bit afraid of moving stuff around on high ground, see:
So in the mean time, yesterday I tried it a bit more precisely but still very quick (I'm impatient) with the mic in hand, at driver's height, maybe 60/70cm off, with a mark on the carpet for each angle. Easy as the carpet is round and the drivers are vertically aligned in the center (it was for the look only).
Here are the results for the horn (Azura AH-1100 w/ Faital pro HF108), in Var smoothing to make it viewable:
For the high mids (SB Audience NERO-6MRN150D) :
0-90 by 15 deg steps
0-180
The Low mids (NERO-12MWN400D):
0-90
0-180
And the woofers (SB Audience BIANCO-18SW450), mic on the floor here:
The pic at 70Hz is not present at my chair, it was surprising, maybe the walls?
Then the sums for the highs, measured:
0-90
I was surprised by the good summing even closed to the drivers
For the other sums I didn't save them all, so I rebuilt them with "arithmetic sum"
(comparing real measures of high mid-tweeter to their arithmetic sums gave me confidence the rest should be good enough)
So woofer to low mid:
low mid to high mid:
So still very jumpy with basically big holes at each crossover points.
While optimised for my listening point about 3 meters away in front, it is like this (no sub, Var/no smoothing):
It's not a surprise that off axis don't sum well but I still assumed better for the low end.
I'm about to retune each bandpass to make sure it smoother.
Of course being nudes the drivers don't give me much flexibility, but being PA their high sensitivity helps.
The low and high mids C-C is 37cm, so around 1Khz wavelength, I can’t really go higher than around 500HZ for the XO, and the 6" is more limited, cancellation is strong below 400Hz.
I was thinking about getting a regular driver here, with lower FS and higher Xmax.
The woofer to low mid C-C is 100cm, so 340Hz wavelength, and here I'm confortable with 150-200Hz XO, above the woofers don't sound right anyway (being clamshell maybe).
I could probably lower the high pass on the 12", it has good Xmax and I could boost it.
But before, am I plain wrong thinking about these two options?
A - don't care about off axis near field measurements, they don't represent anything and better summing would appear at longer distance anyway.
B - care, and retune everything to make it smoother
Thx!
So how do you perform off axis measurements of your dipole/open baffle ?
To get an idea I used to just quickly measure the driver response moving around it, in place, but nothing very precise and always influenced by the room.
And I also relied on what I could find online about each driver, official's spec, diy measurements, driver size etc.
Now I'd like to get a better idea of my power response.
I've read about the outdoor method, high off the ground on a rotating plate, but I don't see how I could do it easily.
Well certainly not now in winter but I have space so I could try this summer, but the "speaker" is not that easy to move around.
It's in multiple parts, all nudes (nupole?
), woofers on the floor and hanging drivers on a tripod, everything tied in its place and angled with zip ties (yes ).Excluding the Woofers it's a MTW arrangement, this mostly for aesthetic reason, low mid below is just plain ugly, and I do like the sonic result as even bass male voices are always at ear level.
So I would need a huge rotating table and I'm a bit afraid of moving stuff around on high ground, see:
So in the mean time, yesterday I tried it a bit more precisely but still very quick (I'm impatient) with the mic in hand, at driver's height, maybe 60/70cm off, with a mark on the carpet for each angle. Easy as the carpet is round and the drivers are vertically aligned in the center (it was for the look only).
Here are the results for the horn (Azura AH-1100 w/ Faital pro HF108), in Var smoothing to make it viewable:
For the high mids (SB Audience NERO-6MRN150D) :
0-90 by 15 deg steps
0-180
The Low mids (NERO-12MWN400D):
0-90
0-180
And the woofers (SB Audience BIANCO-18SW450), mic on the floor here:
The pic at 70Hz is not present at my chair, it was surprising, maybe the walls?
Then the sums for the highs, measured:
0-90
I was surprised by the good summing even closed to the drivers
For the other sums I didn't save them all, so I rebuilt them with "arithmetic sum"
(comparing real measures of high mid-tweeter to their arithmetic sums gave me confidence the rest should be good enough)
So woofer to low mid:
low mid to high mid:
So still very jumpy with basically big holes at each crossover points.
While optimised for my listening point about 3 meters away in front, it is like this (no sub, Var/no smoothing):
It's not a surprise that off axis don't sum well but I still assumed better for the low end.
I'm about to retune each bandpass to make sure it smoother.
Of course being nudes the drivers don't give me much flexibility, but being PA their high sensitivity helps.
The low and high mids C-C is 37cm, so around 1Khz wavelength, I can’t really go higher than around 500HZ for the XO, and the 6" is more limited, cancellation is strong below 400Hz.
I was thinking about getting a regular driver here, with lower FS and higher Xmax.
The woofer to low mid C-C is 100cm, so 340Hz wavelength, and here I'm confortable with 150-200Hz XO, above the woofers don't sound right anyway (being clamshell maybe).
I could probably lower the high pass on the 12", it has good Xmax and I could boost it.
But before, am I plain wrong thinking about these two options?
A - don't care about off axis near field measurements, they don't represent anything and better summing would appear at longer distance anyway.
B - care, and retune everything to make it smoother
Thx!
Kudos on your design and implementation.
I think you should care about off axis. Having a smooth, downward sloping power response is almost as important as being flat and smooth on axis. To get both, you need to start with a speaker architecture capable of both. Yours is indeed that, I believe, but it would likely benefit from a rear facing tweeter.
Nude drivers maintain a dipole response as high as possible for each driver, suffering reduced low end response as a consequence. As long as you stay within each driver's linear and dipole ranges, as you seem to have done, you should be good.
You might consider doing simulations, as I show in a current thread, to tune the design for best combined power response plus axial response. ABEC is likely too much for you unless you already have some experience with it, but a decent job can be done fairly easily using Vituix diffraction models. As your measurements improve, you can replace the diffraction models with the measurements and get closer to reality. As you already have measurements, you can simply plug them into Vituix to see power response. 30 degree steps is fairly coarse for looking at power response but its a start.
Dipole Simulation Thread
I think you should care about off axis. Having a smooth, downward sloping power response is almost as important as being flat and smooth on axis. To get both, you need to start with a speaker architecture capable of both. Yours is indeed that, I believe, but it would likely benefit from a rear facing tweeter.
Nude drivers maintain a dipole response as high as possible for each driver, suffering reduced low end response as a consequence. As long as you stay within each driver's linear and dipole ranges, as you seem to have done, you should be good.
You might consider doing simulations, as I show in a current thread, to tune the design for best combined power response plus axial response. ABEC is likely too much for you unless you already have some experience with it, but a decent job can be done fairly easily using Vituix diffraction models. As your measurements improve, you can replace the diffraction models with the measurements and get closer to reality. As you already have measurements, you can simply plug them into Vituix to see power response. 30 degree steps is fairly coarse for looking at power response but its a start.
Dipole Simulation Thread
You are measuring your speaker in a small room. The room will have a strong influence on the smoothness and accuracy of your measurements in a bad way. The only way to really measure medium and low frequencies for a dipole or OB speaker is outside and away from lateral boundaries. There is no magic, you must eliminate reflections (from walls, buildings, etc.). If you can go up on a flat roof, that can work, or outside in a parking lot, etc. The farther away you can be the lower in frequency you can obtain accurate data.
When I am designing speakers I initially measure the midrange and tweeter indoors in a large room with a high ceiling at a distance of about 16 inches or 40cm with the driver elevated above the floor by at least 1m. I do a series of angles in front and behind the driver to see if there are any problems, see where the pattern is breaking down at higher frequencies, etc. Because these drivers will not operate very low in frequency I can gate out the roof reflections from the impulse and get quasi-anechoic data with decent resolution. But this works less well below 300-400 Hz and is not useful when you want to measure the system response after the speaker is built in order to confirm your crossover is working properly, etc. That is much more difficult.
Luckily at low frequencies below 200-300Hz most small and medium size sources are operating well below the dipole peak and you can rest assured that the on-axis and off-axis responses will be similar (e.g. it has near constant-directivity), so I don't actually bother to measure down there. This approach is valid for a plain open baffle, BUT NOT for a U- or H-frame because those will or may have resonances and different response shapes on and off axis at lower frequencies and you should measure or at least model the data.
When I am designing speakers I initially measure the midrange and tweeter indoors in a large room with a high ceiling at a distance of about 16 inches or 40cm with the driver elevated above the floor by at least 1m. I do a series of angles in front and behind the driver to see if there are any problems, see where the pattern is breaking down at higher frequencies, etc. Because these drivers will not operate very low in frequency I can gate out the roof reflections from the impulse and get quasi-anechoic data with decent resolution. But this works less well below 300-400 Hz and is not useful when you want to measure the system response after the speaker is built in order to confirm your crossover is working properly, etc. That is much more difficult.
Luckily at low frequencies below 200-300Hz most small and medium size sources are operating well below the dipole peak and you can rest assured that the on-axis and off-axis responses will be similar (e.g. it has near constant-directivity), so I don't actually bother to measure down there. This approach is valid for a plain open baffle, BUT NOT for a U- or H-frame because those will or may have resonances and different response shapes on and off axis at lower frequencies and you should measure or at least model the data.
Looks very good to me, naturally it is difficult to know what is causing each dip and peak.
C - keep mic position fixed and rotate the speaker. Make a turntable for it using two sheets of leftover plywood/mdf. Then you will get same boundary reflection artefacts in every measurements
D - use different right window gating and smoothing settings, simple to do after measurements to study various issues.
Measure room response at listening position and use dsp only with low Q values and individual driver gain (spl). Trust your subjective impression!
C - keep mic position fixed and rotate the speaker. Make a turntable for it using two sheets of leftover plywood/mdf. Then you will get same boundary reflection artefacts in every measurements
D - use different right window gating and smoothing settings, simple to do after measurements to study various issues.
Measure room response at listening position and use dsp only with low Q values and individual driver gain (spl). Trust your subjective impression!
Thanks nc535, I am definitely interested by Vituix, had an eye on it and need to study it more to see if I could manage (need to find a pc too as my old macbook is too slow for heavy stuff).
I agree, well I feel like I miss rear highs too. I constantly change my house target and that’s probably a proof that I’m not perfectly happy yet.
I am actually studying how to do some tests with a rear firing horn to get an idea. Had an amt there years ago but it was a completely different room.
So changing for regular domes is still on the table, especially if I change the mids, for back to back, or vertical offset etc (Rudolf’s thread is an awesome help).
And I will make sure to use 10deg steps in the future, for now 15/30 were easier on readibility, but Vituix would make all this a lot more understandable!
Thanks for the tips CharlieLaub, for sure getting a clear idea of what’s coming from the crossover and what’s coming from the room is the goal.
So I think I will have to go through these outside measurements, and will start to plan for it, good thing is I’m in the middle of nowhere.
There is a beauty in nude drivers as they’re easy to disassemble and test alone.
Well still have to readjust the cord/zip ties but it’s still much easier than building specific baffles!
And I need to study more the gating part as I never fully mastered it, I wasn't able to see big differences when I tried.
Great idea with « C »! This way I could still vary the mic distance and better grab what’s going on.
It should be easy to build too, awesome, thanks. I should have thought of it when looking at the round carpets.
I try to keep low Q on mids but they’re so sensitive their responses need severe EQ (not the horns).
I might have taken a wrong path with them, but it was useful to reduce amps power too.
It’s so easy to get trapped in the visual aspect of measurements, correct too much when in real it’s not that different or even worst.
Got work to do thanks guys!
Once I realized that I could very easily (even indoors if you are careful) investigate each nude driver on its own for its on and off axis responses I found the process of identifying good drivers for potential dipole loudspeakers to be relatively easy. More difficult is to make measurements on the fully assembled speaker in order to get good data for doing crossover development. That requires a few tricks and thinking...
Regarding gating, in your measurement software it is usually possible to take the measured impulse and "zoom in" the y-axis but leaving the x-axis (time axis) alone. You should see the decay of the direct sound and then a small "copy" of the initial part of the impulse peak but at a much lower amplitude. Unless you "blow up" the impulse data it might be easy to miss. This is the first reflection from some nearby room boundary, coming after a few milliseconds. Gating is the process of setting the value for the impulse signal just before this reflection begins, and for all times after it. This is done with a "windowing function" that may or may not smoothly reduce the signal level down to zero near the gating time. Once you remove the part of the signal that contains the echoes or reflections you are left with only the "direct" sound that traveled in a straight line from the speaker to the microphone. Since this is sort of like making a measurement in a reflection-free space, it's called "quasi-anechoic". But this process limits the remaining non-zero impulse in time (and number of samples) and the consequence is that the low frequency response for periods longer than the length of the impulse are unknown. Also, the resolution is then limited to "bins" of some minimum frequency and this smooths the frequency data over all frequencies. You can think of the data as more of a column chart than a smooth and continuous line. Inside each "box" of the column chart the data has the same value over some range of frequency due to the effects of smoothing and binning. So measuring outside and high above the ground to reduce reflections and allow for longer gated data length provides higher resolution data without glossing over any FR issues that might be smoothed out otherwise.
Playing a bit with the IR window I still have hard time to find the first reflection.
Do you have an order of magnitude (ball park) for it for a typical 4 way/medium size room?
Since I wasn’t sure so I played with the step response window based on the maximum loss I could accept.
Meaning finding a compromise between removing enough of the room effect, and keeping a curve still usable.
Even if I lose a lot of definition on the low end, it gives a curve that I can still use to define XO and EQ.
Ex on the woofer, with measurement from last week end, it’s in between 20-50ms:
For the high mid it’s more 5-20ms:
Am I on the right path?
Do you have an order of magnitude (ball park) for it for a typical 4 way/medium size room?
Since I wasn’t sure so I played with the step response window based on the maximum loss I could accept.
Meaning finding a compromise between removing enough of the room effect, and keeping a curve still usable.
Even if I lose a lot of definition on the low end, it gives a curve that I can still use to define XO and EQ.
Ex on the woofer, with measurement from last week end, it’s in between 20-50ms:
For the high mid it’s more 5-20ms:
Am I on the right path?
Looks like your gating settings are odd. Letft window means time before the impulse peak (t=0), also type of curve can be defined.
Right window sets the ending of analysis, after the impulse peak.
Reference point can me moved to if first (direct) impulse if it is lower in amplitude than the peak (where REW automatically sets t=0) This is common situation with dipoles at angles 70-120deg.
Directly from impulse or step you can see reflections, they make peaks (higher amplitude) in sound signal level. Floor is typically first and strongest, at 4-6ms after t=0
Please study REW Help files!
Right window sets the ending of analysis, after the impulse peak.
Reference point can me moved to if first (direct) impulse if it is lower in amplitude than the peak (where REW automatically sets t=0) This is common situation with dipoles at angles 70-120deg.
Directly from impulse or step you can see reflections, they make peaks (higher amplitude) in sound signal level. Floor is typically first and strongest, at 4-6ms after t=0
Please study REW Help files!
Last edited:
Ok I’ll perform some new measurements next week end.
So far I wasn't able to find the reflection points on the impulse, maybe on the ETC but nothing sure.
I should also measure stuff around (walls/floor etc to get an idea of where to look at).
On your exemple at 6ms you still have 10hz informations, even if a lot smoothed out, how did you keep this?
If I adjust to such a small time window I quickly lose stuff on the low end, can it be related to the sweeps settings, sample rate or length?
So far I wasn't able to find the reflection points on the impulse, maybe on the ETC but nothing sure.
I should also measure stuff around (walls/floor etc to get an idea of where to look at).
On your exemple at 6ms you still have 10hz informations, even if a lot smoothed out, how did you keep this?
If I adjust to such a small time window I quickly lose stuff on the low end, can it be related to the sweeps settings, sample rate or length?
On your exemple at 6ms you still have 10hz informations, even if a lot smoothed out, how did you keep this?
Low freq must be ignored, you can see when response looks smoothed. In 6ms gated below 1kHz. The dip at 500Hz is floor/ground reflection.
If gating slope of Right is changed to some more shallow filter, one can use a bit longer time, but again low freq is unreliable showing also reflections and room modes.
It is important to study impulse response, how strong the reflection peaks are and at what time delay they appear.
With room/listening spot response, I use 120-500ms right window and 1/2 or 1/3 smoothing. This is to show subjective freq balance/tonality.
REW is by default settings optimized to analyze room modes and reflections when measured at listening spot or around it multiple spots. Also good to match main and subwoofer responses. That's why it's defaut Right gating is 500ms.
Low freq must be ignored, you can see when response looks smoothed. In 6ms gated below 1kHz. The dip at 500Hz is floor/ground reflection.
If gating slope of Right is changed to some more shallow filter, one can use a bit longer time, but again low freq is unreliable showing also reflections and room modes.
It is important to study impulse response, how strong the reflection peaks are and at what time delay they appear.
With room/listening spot response, I use 120-500ms right window and 1/2 or 1/3 smoothing. This is to show subjective freq balance/tonality.
REW is by default settings optimized to analyze room modes and reflections when measured at listening spot or around it multiple spots. Also good to match main and subwoofer responses. That's why it's defaut Right gating is 500ms.
To give you or anyone a rough idea of when you can expect the first reflection of the direct sound to appear in the impulse just calculate the round trip time of flight (TOF) from the source (driver) to any boundary and compare this to the direct sound TOF. The TOF is just the distance in meters divided by the speed of sound, which is around 344 meters / second, with the result in seconds. If you prefer to use feet, c= 1100ft/sec. 1 millisecond is one one-thousandth of a second.
Example: you measure the driver elevated 1m off of the floor in the middle of a 4x5m room. The microphone is also at 1m high and 1m away. The closest surface is the floor, then the ceiling (assuming the room is taller than 2m) and then the walls. The floor bounce occurs midway between driver and mic and with this particular geometry the TOF happens to be SQRT( 5.0 ) meters or approximately 2.23m / 344m/s = 6.5 msec. The TOF of the direct sound is 1m / 344m/sec = 2.9 msec. The floor bounce reflection will appear 6.5 - 2.9 = 3.6 msec after the initial impulse peak from the direct sound.
As you can probably see, it is not realistic to get 20msec of data in a domestic indoor space without it being contaminated with one or more reflections. To get better resolution you have to find a very large indoor space and elevate the source, use a ground plane measurement technique, or go outside away from boundaries.
Example: you measure the driver elevated 1m off of the floor in the middle of a 4x5m room. The microphone is also at 1m high and 1m away. The closest surface is the floor, then the ceiling (assuming the room is taller than 2m) and then the walls. The floor bounce occurs midway between driver and mic and with this particular geometry the TOF happens to be SQRT( 5.0 ) meters or approximately 2.23m / 344m/s = 6.5 msec. The TOF of the direct sound is 1m / 344m/sec = 2.9 msec. The floor bounce reflection will appear 6.5 - 2.9 = 3.6 msec after the initial impulse peak from the direct sound.
As you can probably see, it is not realistic to get 20msec of data in a domestic indoor space without it being contaminated with one or more reflections. To get better resolution you have to find a very large indoor space and elevate the source, use a ground plane measurement technique, or go outside away from boundaries.
Since I was asked to show some old iterations, I've been searching for old pics to show some of the previous steps (took the time to clean up my photo folders, what a mess after all these years, probably 50% of the weight is from doublets/triplets/quadruplets).
I know I have some old rew measurements somewhere and it would be great to see them per layout, but no way I have the time.
Just too many and I'm not really well organised, plus I ended up not naming or not even saving most of them and more tuning "in live" now.
Too many files, too big and never re-opened anyway.
I think I found about half of them, missing a few and mostly layout variations of the same drivers like wmmt/wmtm/wwmt etc
But at least few pics would show the mistakes, misconceptions/misunderstanding I had, and the general evolution.
If I count a new driver or a driver new placement as a change, it’s probably the 30th version.
If I remember well, after few years in car audio I planned to build a home system, played a bit with oem stuff and was ready to build a box.
But then I saw the Reflector Audio.
This triggered something, I absolutely loved the look and wanted to mimic it, within in my constrains.
Started with this:
I think the stand was a weight holder, something from the gym. 2or 3 way, the horn was fine but bass was anemic, even with the first 18'', plus room was too small anyway.
getting better in bass with better coupling, and tpl gets in the loop.
clamshell bass was a big step, bigger room too, then I could focus on mids, 1 mid, then 2 in mmt or mtm.
spend quite some time there, the clamshell bass couldn’t go high, maybe 300hz max then it started to get blurry.
So I started with 2x5’', too low for them.
Then 2x6", too low again but great with the tpl, then 2x7’' and it was ok but less good on highs, the tpl150 is just too tall.
Had fun hanging the amps too, and treating the room.
Then moving back in France, I could restart almost from scratch as I only kept the tv stands.
I wanted to try the sb waveguide, too high a price on the tpl at the time.
Re-tested few layouts to be sure, but ended with basically the same.
Big 12’’ for low mid to avoid the previous issue with mtm and driver size not optimum, + I always thought one driver sounded better, even if measured not as well.
Kept this mtw for probably 2-3 years, really fun, and loud.
Tried a horn again, it was smoother off axis and allowd me to experiment again with tubes, but I was still missing the back side.
So last is with the Kartesian mid120, very detailed driver it’s impressive, sometime too much.
And good spl too, less than the 6'' but not by much as it is smoother to eq, the dipole peak on the 6'' was damn big
I tried 2 mid120 in mtm but again, not as clean sounding.
I think the same in 5'' would be perfect.
Still some work to do, I need to perform the full measurements outside, finetune.
I could fabricate a front plate for the tpl75, something optimized.
Or integrate it in a very small wood baffle that holds the mid too, experiment adding few cm for the mids but mostly for a cleaner/smoother look.
Then redo all hanging system with cords, zipties are great, solid and easy to adjust, but cords would look cleaner.
I also experiment with passive crossovers, or hybrids for now.
Low mid to high mid is passive, with great result.
High mid to tweet is hybrid, its break up is just too hard but I'll try again.
In the mean time I also wanted to check my own objectivity, I mean checking that I'm not too much biased about OB.
So I started a 3 way classic closed, to ultimately compare.
I don’t know what to expect as my room is heavily treated for dipole, but it should be fun.
Will do a thread one day it's just POC for now, I only have 1 and only just tested the impedance:
I know I have some old rew measurements somewhere and it would be great to see them per layout, but no way I have the time.
Just too many and I'm not really well organised, plus I ended up not naming or not even saving most of them and more tuning "in live" now.
Too many files, too big and never re-opened anyway.
I think I found about half of them, missing a few and mostly layout variations of the same drivers like wmmt/wmtm/wwmt etc
But at least few pics would show the mistakes, misconceptions/misunderstanding I had, and the general evolution.
If I count a new driver or a driver new placement as a change, it’s probably the 30th version.
If I remember well, after few years in car audio I planned to build a home system, played a bit with oem stuff and was ready to build a box.
But then I saw the Reflector Audio.
This triggered something, I absolutely loved the look and wanted to mimic it, within in my constrains.
Started with this:
I think the stand was a weight holder, something from the gym. 2or 3 way, the horn was fine but bass was anemic, even with the first 18'', plus room was too small anyway.
getting better in bass with better coupling, and tpl gets in the loop.
clamshell bass was a big step, bigger room too, then I could focus on mids, 1 mid, then 2 in mmt or mtm.
spend quite some time there, the clamshell bass couldn’t go high, maybe 300hz max then it started to get blurry.
So I started with 2x5’', too low for them.
Then 2x6", too low again but great with the tpl, then 2x7’' and it was ok but less good on highs, the tpl150 is just too tall.
Had fun hanging the amps too, and treating the room.
Then moving back in France, I could restart almost from scratch as I only kept the tv stands.
I wanted to try the sb waveguide, too high a price on the tpl at the time.
Re-tested few layouts to be sure, but ended with basically the same.
Big 12’’ for low mid to avoid the previous issue with mtm and driver size not optimum, + I always thought one driver sounded better, even if measured not as well.
Kept this mtw for probably 2-3 years, really fun, and loud.
Tried a horn again, it was smoother off axis and allowd me to experiment again with tubes, but I was still missing the back side.
So last is with the Kartesian mid120, very detailed driver it’s impressive, sometime too much.
And good spl too, less than the 6'' but not by much as it is smoother to eq, the dipole peak on the 6'' was damn big
I tried 2 mid120 in mtm but again, not as clean sounding.
I think the same in 5'' would be perfect.
Still some work to do, I need to perform the full measurements outside, finetune.
I could fabricate a front plate for the tpl75, something optimized.
Or integrate it in a very small wood baffle that holds the mid too, experiment adding few cm for the mids but mostly for a cleaner/smoother look.
Then redo all hanging system with cords, zipties are great, solid and easy to adjust, but cords would look cleaner.
I also experiment with passive crossovers, or hybrids for now.
Low mid to high mid is passive, with great result.
High mid to tweet is hybrid, its break up is just too hard but I'll try again.
In the mean time I also wanted to check my own objectivity, I mean checking that I'm not too much biased about OB.
So I started a 3 way classic closed, to ultimately compare.
I don’t know what to expect as my room is heavily treated for dipole, but it should be fun.
Will do a thread one day it's just POC for now, I only have 1 and only just tested the impedance:
Last edited: