^I found that room response measurement was best way to match levels. All kind of nearfield and ground plane are difficult, because they miss the full 3D sound power, which is what we hear (below Schröder) in a room. If xo is below 100Hz it is rather easy, but my xo is 150-200Hz.
https://www.acousticfields.com/schroeder-frequency/
https://www.akutek.info/Papers/MS_Schroeder_Revisited.pdf
https://www.acousticfields.com/schroeder-frequency/
https://www.akutek.info/Papers/MS_Schroeder_Revisited.pdf
That perfectly represents how I feel about my setup. Still undecided if I want to keep the subs or not, so I keep playing with both settings.
With electro they certainly bing some fun, but with acoustic most of the time I turn them off and leave the woofers free (4x18sw450).
It’s not much cleaner on measurements, it just feels better, maybe the room modes less excited.
But I will continue to optimise maybe I can get best of both worlds.
I know I’ll never get that, but I still miss the almost suffocating chest thump I had in my car with 2x15’’ right behind me.
I spent some time with OB designs, and first completely gave up the idea of producing whole-room "deep bass" with OB. And then, I gave up the idea to produce "deep" bass altogether. Personally and by "deep", I mean something around 20Hz ... 25Hz.
As human beings we are not hearing like a microphone. Instead, we are subject to the ISO 266:2003 loudness curves:
http://www.sengpielaudio.com/ISO226LoudnessCurves.gif
Therefore, even with such a quite serious and impressive setup, we all can simply forget about enjoing a subjectively "loud" listening at 20Hz: The brave-hearted 105dB @ 20Hz of this very monster array do correspond to less than 60 mickey Phon referenced at 1kHz. This is a quite annoying, but true fact, indeed.
For a more moderate low-end bass goal things get rapidly better: This same array might produce 112dB@30Hz and 117dB@40Hz . These values translate, for the 30Hz case into about 80 Phone @ 1kHz, and for the 40Hz case into about 100 phone @ 1kHz.
So, even with this huge system, "loud" listening stops at 40Hz. This might not seem this impressive ... And this also goes for any other approach, not even OB. Therefore and in a slightly kidding mode, I encourage for the bottom 20Hz...40Hz octave instead to better take a shaker and attach it anywhere onto your body. You might find a decent location to do so for best subjective results.
As human beings we are not hearing like a microphone. Instead, we are subject to the ISO 266:2003 loudness curves:
http://www.sengpielaudio.com/ISO226LoudnessCurves.gif
Therefore, even with such a quite serious and impressive setup, we all can simply forget about enjoing a subjectively "loud" listening at 20Hz: The brave-hearted 105dB @ 20Hz of this very monster array do correspond to less than 60 mickey Phon referenced at 1kHz. This is a quite annoying, but true fact, indeed.
For a more moderate low-end bass goal things get rapidly better: This same array might produce 112dB@30Hz and 117dB@40Hz . These values translate, for the 30Hz case into about 80 Phone @ 1kHz, and for the 40Hz case into about 100 phone @ 1kHz.
So, even with this huge system, "loud" listening stops at 40Hz. This might not seem this impressive ... And this also goes for any other approach, not even OB. Therefore and in a slightly kidding mode, I encourage for the bottom 20Hz...40Hz octave instead to better take a shaker and attach it anywhere onto your body. You might find a decent location to do so for best subjective results.
Erratum
Sorry, these above and wrong data would rather fit a closed box with a 12dB/oct slope. OB instead do have a 6dB/oct slope. And I also took a wrong graph for theirs extrapolation. So the (now hopefully!) correct values for this OB system are as follows:
Open Baffle (-6dB/Oct):
20Hz / 105.0dB / 50 Phone (approx values)
25Hz / 106.9dB / 65 Phone
32Hz / 109.1dB / 79 Phone
40Hz / 111.0dB / 89 Phone
50Hz / 112.9dB / 100 Phone
Same principle, but for Closed Box (-12dB/Oct)
20Hz / 105.0dB / 50 Phone (approx values)
25Hz / 108.8dB / 70 Phone
32Hz / 113.1dB / 85 Phone
40Hz / 117.0dB / 100 Phone
Hope, this better fits reality now.
These values also illustrate a main difference between OB and CB in terms of bass resproduction: Starting at a fixed value (105dB@20Hz), it gets clear, how more bass potential in terms of Phone a CB has compared to the OB. A OB might have other vertues than subjectively sounding loud in the lower bass, however.
Sorry, these above and wrong data would rather fit a closed box with a 12dB/oct slope. OB instead do have a 6dB/oct slope. And I also took a wrong graph for theirs extrapolation. So the (now hopefully!) correct values for this OB system are as follows:
Open Baffle (-6dB/Oct):
20Hz / 105.0dB / 50 Phone (approx values)
25Hz / 106.9dB / 65 Phone
32Hz / 109.1dB / 79 Phone
40Hz / 111.0dB / 89 Phone
50Hz / 112.9dB / 100 Phone
Same principle, but for Closed Box (-12dB/Oct)
20Hz / 105.0dB / 50 Phone (approx values)
25Hz / 108.8dB / 70 Phone
32Hz / 113.1dB / 85 Phone
40Hz / 117.0dB / 100 Phone
Hope, this better fits reality now.
These values also illustrate a main difference between OB and CB in terms of bass resproduction: Starting at a fixed value (105dB@20Hz), it gets clear, how more bass potential in terms of Phone a CB has compared to the OB. A OB might have other vertues than subjectively sounding loud in the lower bass, however.
Daihedz, dipole loss is -6dB/oct, but with real drivers only above driver's Fs. Below it you must add the sloping that the driver shows in IEC or infinite baffle!
+1 on what Juhazi says above
For OB woofers that will be used below 100Hz, you really should include the droop of the driver's free-air response. This happens well above Fs, actually. The OB loading is essentially like the driver is in free air, so the MFG TS parameter values (e.g. Qts) can be taken as the Q value of the response before any EQ is applied. You can estimate the SPL loss from the response droop by putting the driver in any box modeler for a closed box, but set the box volume parameter as high as possible, e.g. 99999 liters or whatever. You will see that these losses are not insignificant.
Here is a quick online calculator web page that you can use for this estimation.
http://www.mh-audio.nl/Calculators/CBCV.html
I put in the values:
Fs=30 Hz
Qes=0.35
Qts=0.4
Vas=180 liters
Vdr=0.2 liters
Vb=999999 liters (just ignore the warning that is should be 0-1000 liters)
This produces a curve that is:
-2.6 dB @ 70 Hz
-4.2 dB @ 50 Hz
-5.7 dB @ 40 Hz
-8.0 dB @ 30 Hz
-12 dB @ 20 Hz
below the passband level. Higher Qts drivers will have less of these losses above Fs. Lower Q drivers (I used 0.4, which is higher than any pro driver) will have even higher losses. Use the calculator and plug in the correct values for your intended driver. Keep in mind, like Juhazi mentioned, that these losses are in addition to the dipole losses (the 6dB downslope from dipole cancellation).
For OB woofers that will be used below 100Hz, you really should include the droop of the driver's free-air response. This happens well above Fs, actually. The OB loading is essentially like the driver is in free air, so the MFG TS parameter values (e.g. Qts) can be taken as the Q value of the response before any EQ is applied. You can estimate the SPL loss from the response droop by putting the driver in any box modeler for a closed box, but set the box volume parameter as high as possible, e.g. 99999 liters or whatever. You will see that these losses are not insignificant.
Here is a quick online calculator web page that you can use for this estimation.
http://www.mh-audio.nl/Calculators/CBCV.html
I put in the values:
Fs=30 Hz
Qes=0.35
Qts=0.4
Vas=180 liters
Vdr=0.2 liters
Vb=999999 liters (just ignore the warning that is should be 0-1000 liters)
This produces a curve that is:
-2.6 dB @ 70 Hz
-4.2 dB @ 50 Hz
-5.7 dB @ 40 Hz
-8.0 dB @ 30 Hz
-12 dB @ 20 Hz
below the passband level. Higher Qts drivers will have less of these losses above Fs. Lower Q drivers (I used 0.4, which is higher than any pro driver) will have even higher losses. Use the calculator and plug in the correct values for your intended driver. Keep in mind, like Juhazi mentioned, that these losses are in addition to the dipole losses (the 6dB downslope from dipole cancellation).
It’s not that simple… In an H-baffle, the air loading will lower Fs and increase Qts, usually as much as 25%.
And the room... Below 1st mode there is some air loading, but still dipoles loose to sealed. BR loose too, because the response drops so fast.
Measurements in the listening room are the real test!
Measurements in the listening room are the real test!
Last edited:
Sure, but still even wtih 25% higher Qts if the builder starts off with a pro audio driver with Qts=0.3 you will just be right back at my value of Qts=0.4, or possibly even less.
The link is to a very interesting fellow's site. https://www.trans-fi.com/diy-speaker-blog
He takes you through his idea's evolution. A strange amalgamation of OB and isobarik loading.
I have never tried this so I have never heard it but it is a novel approach and worth a read.
He makes the unusual in concept and in its low cost turntable mat which I was surprised to see one of the STEREOPHILE writers uses. He also offers a litz tonearm wire at a very low cost. Just to give him some credibility.
I p[ersonally think his no longer available turntable to be the best, conceptually, I have ever seen. Again, not heard.
He takes you through his idea's evolution. A strange amalgamation of OB and isobarik loading.
I have never tried this so I have never heard it but it is a novel approach and worth a read.
He makes the unusual in concept and in its low cost turntable mat which I was surprised to see one of the STEREOPHILE writers uses. He also offers a litz tonearm wire at a very low cost. Just to give him some credibility.
I p[ersonally think his no longer available turntable to be the best, conceptually, I have ever seen. Again, not heard.
Based on the sheer number of systems shown, this person seems to make a new speaker every couple of weeks. Lots of enthusiasm. Many of his designs include some serious flaws. You can tell just by looking at them, what drivers are used, how large the drivers are, etc. You can't escape physics.
Also, very very light on measurements (essentially none) and I did not see much details in terms of crossover design, etc. Anyone can slap together some drivers in whatever configuration and use some DSP to "make it work" by using their ear (the worst way to design any type of loudspeaker except a simple fullranger). It's a recipe for what you see here, never satisfied and continually trying something new.
Although a couple of the systems with nude drivers on a "stick" above a compound dipole woofer setup have some merit I see little mention of how the crossovers were designed and implemented. Oh, I see now, this quote kind of sums it up:
Guy seemingly has no clue how to do anything properly. It's just random stuff that he tries based on whatever...
This should serve more as an example of how NOT to design and build speakers!
Great thread. Thanks for the link. I have seen his work before as he used the PRV 5MR450 for a while - a driver I really like. I think from his bizziliion experiments which arrived at a COTS mini monitor from Focal on top of isobaric woofers in bafflesss config, says something about a good 2 way to convey soundstage and image and leave the bass duties to a large OB woofer.
Thank you Juhazi and CharlieLaub for your objections. You are right.
I have some corrected and extended data, computed as CharlieLaub suggested, same driver data, same online program. The results also show the consequences of StigErik's comments about the H-Baffle and the discussion about Qts 0.3 vs. 0.4.
And apropos loudness assessments: I found a nice link for IEC 226:2003 data analysis
https://chart-studio.plotly.com/~intmarine/1/#/
I have some corrected and extended data, computed as CharlieLaub suggested, same driver data, same online program. The results also show the consequences of StigErik's comments about the H-Baffle and the discussion about Qts 0.3 vs. 0.4.
And apropos loudness assessments: I found a nice link for IEC 226:2003 data analysis
https://chart-studio.plotly.com/~intmarine/1/#/
Last edited:
Mr Laub,
I was mainly pointing folks to his low bass idea.
I am a horn guy so all of that other stuff did not pertain to me. Well, I guess NONE of it does but I had never seen anything like his idea for the bass and, yes, of course, it looks impossible to work well, but then there is always the bumblebee syndrome.
When it comes to bass I am very partial to the RYTHMIK system. I have four of them and being a music listener and not a sound effects guy I find these things load the room far better than the Edgar Titans I had two of at one time and they certianly have much lower perceived distortion. The room loading takes place immediately which is what fast bass is all about.
Small and simple. I think they are a bargain.
No question, as Mr. Geddes has said for years - the key to good bass reproductiuon is using as many drivers as you can stand. In my case it is four - I would not have a place for any more than that.
The cones never flap wildly - mine are all sealed - REW says they are down one dB at 10 Hz in my room. 95 dB.
Not to say I do not doubt LOW bass dipoles would be quite a fine experience. But about as practical as thinking one can power the nation's grid with solar panels.
I was mainly pointing folks to his low bass idea.
I am a horn guy so all of that other stuff did not pertain to me. Well, I guess NONE of it does but I had never seen anything like his idea for the bass and, yes, of course, it looks impossible to work well, but then there is always the bumblebee syndrome.
When it comes to bass I am very partial to the RYTHMIK system. I have four of them and being a music listener and not a sound effects guy I find these things load the room far better than the Edgar Titans I had two of at one time and they certianly have much lower perceived distortion. The room loading takes place immediately which is what fast bass is all about.
Small and simple. I think they are a bargain.
No question, as Mr. Geddes has said for years - the key to good bass reproductiuon is using as many drivers as you can stand. In my case it is four - I would not have a place for any more than that.
The cones never flap wildly - mine are all sealed - REW says they are down one dB at 10 Hz in my room. 95 dB.
Not to say I do not doubt LOW bass dipoles would be quite a fine experience. But about as practical as thinking one can power the nation's grid with solar panels.
For dipole bass, it could be a good idea to move away from the "cone speaker" paradigm altogether. Linear motion is already very limiting in boxed speakers, with ~1% efficiency. Horn speakers have more leverage, basically constraining the air in a narrower column so, on a per-wavelength basis, the speaker is able to compress it more. With dipoles, you kind-of want a LOT of swept volume but somehow improving upon the 1:1 ratio of the voice coil moving about like crazy.
One idea that comes to my mind is this:
1) An angular motor that works on the same basic principle as DC motors, except that the entire range of motion is limited to a few degrees or so, and therefore there is no need for brushes or field coils. Instead, the connection is hard-wired and expected to survive a bit of twisting for the life of the speaker.
2) A fan-like wing is attached to the coil / armature, along with a counter-weight on the other side.
3) To prevent torsional reaction forces, they could be operated in pairs, like insect wings.
There would certainly be a couple of engineering puzzles to solve, I'm almost inclined to do this myself! It's just a matter of finding some free time. Thoughts, anyone?
One idea that comes to my mind is this:
1) An angular motor that works on the same basic principle as DC motors, except that the entire range of motion is limited to a few degrees or so, and therefore there is no need for brushes or field coils. Instead, the connection is hard-wired and expected to survive a bit of twisting for the life of the speaker.
2) A fan-like wing is attached to the coil / armature, along with a counter-weight on the other side.
3) To prevent torsional reaction forces, they could be operated in pairs, like insect wings.
There would certainly be a couple of engineering puzzles to solve, I'm almost inclined to do this myself! It's just a matter of finding some free time. Thoughts, anyone?