Vocal range(80-1100Hz ) crossover design philosophy ..... Steep Xover slope at 1.2-1.4Khz is acceptable compromise in order to keep human voices on one driver.
The ISO 226 constant loudness curves(2003) provide one source of hearing sensitivity vs. frequency. From the curve we can see that on average humans are most sensitive to tones at about 3500 Hz, because these tones require the least gain to reach the threshold of hearing. The hearing threshold difference between 700Hz and 1400Hz is modest in the ISO226 tests. Legendary 2-way Altec speakers used a “modest gain” frequency 700-800Hz Xover between the woofer and controlled directivity horn. Modern technology 3-way “Vocal Range(80-1100Hz )” speakers use a 15”-18” woofer Xover at 80-100Hz, plus a 8”-12” midbass Xover at the 1.2-1.6Khz hearing threshold sensitivity cliff point to a controlled directivity tweeter/horn.
some 8" midbass drivers produce a 90 degree polar response ~1100Hz
some 10" midbass drivers produce a 90 degree polar response ~1300Hz
some 12" midbass drivers produce a 60 degree polar response ~1600Hz
The ISO 226 constant loudness curves(2003) provide one source of hearing sensitivity vs. frequency. From the curve we can see that on average humans are most sensitive to tones at about 3500 Hz, because these tones require the least gain to reach the threshold of hearing. The hearing threshold difference between 700Hz and 1400Hz is modest in the ISO226 tests. Legendary 2-way Altec speakers used a “modest gain” frequency 700-800Hz Xover between the woofer and controlled directivity horn. Modern technology 3-way “Vocal Range(80-1100Hz )” speakers use a 15”-18” woofer Xover at 80-100Hz, plus a 8”-12” midbass Xover at the 1.2-1.6Khz hearing threshold sensitivity cliff point to a controlled directivity tweeter/horn.
some 8" midbass drivers produce a 90 degree polar response ~1100Hz
some 10" midbass drivers produce a 90 degree polar response ~1300Hz
some 12" midbass drivers produce a 60 degree polar response ~1600Hz
Attachments
It's really a misconception that the human voice only goes up to 1100Hz. It is true that the fundamentals only go up that high, but we all produce those. What determines how we sound as individuals is the harmonic structure above this and the way our 'singers formant' resonates. You're going to want at least the 2nd and 3rd harmonics to be covered by the same driver, so up to around 3300Hz. With the 'buzz' or 'ring' of the formant usually going up to around 4kHz. 3.5kHz would be a nice compromise.
Single early reflection is not capable to destroy sound reproduction in my experience. Destruction is reality if early reflections are allowed increase early decay time (EDT) too much. Therefore also cardioid-family is able to sound accurate and clear, if room acoustics is designed for high resolution reproduction.
Strong directivity is (just) a way to skip acoustic work, but is trade off. Axial flutter echo is powerful enemy also for strong unidirectional directivity. Strong directivity only above midrange has another trade off against better imaging; high resolution step from low to high frequencies does not sound very natural and focuses listener to HF bling-bling. DI shouldn't increase more than about 12 dB from the lowest to highest octave. Bass directivity is only way to maintain that relation with unidirectional speakers.
This is so true! I came to that conslusion simply by turning the tweeter
off and listening without it for a while.
X/O within 3-5 kHz is good and easy way to combine 3”-5.25" mid and ordinary dome or ribbon tweeter. 'Most sensitive band of human hearing' means that I can’t compromise trend of power response (and DI if possible) within sensitive range. Vertical lobing and phase distortion of single LR2 at midrange are obsolete at least for me. Flat waveguide enables lower X/O if mid is 'too large' but that's all. Deep waveguides and horns are different story.
Why not use a good quality full range from 150hz to 8khz? Tweeter from 8khz up and woofer below 150hz. That's a good 3way design
If I remember, s's and t's upper harmonics resonate around 8khz.
If you are not ghosting in a tweet with 1 cap, or not using a 6db network, 24db@5khz works well (but a 6.5" doesn't have enough dispersion to get up there), and even 750hz@24db is better than 2khz@24db.
So I agree, 1-4khz is not a great place to cross.
I'll take that over a 12db or 18db slope any day of the week.
If you are not ghosting in a tweet with 1 cap, or not using a 6db network, 24db@5khz works well (but a 6.5" doesn't have enough dispersion to get up there), and even 750hz@24db is better than 2khz@24db.
So I agree, 1-4khz is not a great place to cross.
I'll take that over a 12db or 18db slope any day of the week.
DI of 3"..5.25" midrange driver jumps to the sky above 3-4 kHz. That's also main reason why higher X/O sounds more accurate and clear (if low-passed steep enough), and lower X/O at "sensitive range" sounds bright but inaccurate, death and nonprecense.
Unfortunately none that I know of. Considering how difficult it is to tune a resistance box, I doubt you'll ever get meaningful results anyway. 'Everything' matters: the shape of the baffle, driver size, exact internal dimensions of the box, the location of the box openings, the path length to the openings, the characteristics and distribution of the acoustic material in the box, etc.
It's nice to have you here, Kimmo!
In my view early reflections, EDT and reverb are subject to personal preference. I believe the scientific literature is with me on this, if you consider the often ambiguous results of preference tests. Some people like near anechoic conditions, while others prefer the a lot of room contribution. I find myself somewhere in the middle, although I've learned to appreciate both extremes. The problem is that there is no absolute reference for stereo sound reproduction, so some pursue hearing every single detail in the recording plus superbly sharp imaging, while others find that to sound too two-dimensional and artificial. They are more concerned with the music sounding 'real'. You simply need reflections for that.
Some reflections are simply bad. Strong axial reflections, for example. You'd rather not have a very strong front-wall boundary interaction or reflection and you don't want a strong and early rear-wall reflection. They can cause audible coloration and degrade imaging. A cardioid deals with the boundary interaction - an important feat in my opinion, because loudspeakers are usually placed much closer to boundaries than they should be placed for optimal performance. In general, I'd advice to put some absorbing material on the rear wall.
I think smooth directivity is beneficial down to somewhere in the transition region, but I doubt a step in directivity matters much once you are well into the modal range. In the bass-range, our hearing uses a much too large time-envelope to distinguish between direct and reflected sound. I can see no other reason to opt for cardioid bass, other than the better room-interaction, exciting a greater number of standing waves, yet less effectively so.
Allow me to actually go on-topic for a bit 😛. In my experience it doesn't matter much where you cross. A good crossover is practically inaudible. In order to accomplish that, you have to; 1) make sure the drivers add up coherently at and around the crossover frequency, for a wide frontal lobe; and 2) match directivities at the crossover point.
Thanks! Been here occasionally since 2005 but not as publisher.
Shortly off-topic about cardioid basses:
Cardioid is very adaptable bass application. It's not perfect but quite immune to speaker location, different room sizes and building materials. Super-cardioid with 10-14 dB back attenuation is more adaptable than standard cardioid, and all cardioid types are more consistent than monopole and dipole. I’ve played with resistance enclosures and monopole+dipole -combinations for a decade, room sizes from 6 m2 to 80 m2, including 16…20 m2 concrete bunkers. This experience may explain why I have severe difficulties to buy arguments against cardioid’s benefits at bass range.
Leaking box is ineffective at lower bass, but Sd >2000 cm2 per speaker with 1 kW power amp is quite ok. Resistance box is full range application with active shelving LP.
Monopole+dipole is difficult to design as powerful and extending to midrange at the same time, but real closed monopole offers some advantages for the low end. Polar pattern is adjustable as bi-amped.
I can live without cardioid bass, but the highest quality requirements should be switched off if optimal speaker placement is not possible.
Last edited:
In my experience you get the best bass with multiple subwoofers. I have experience configuring true cardioid subwoofers, but at this point I don't have much experience with them in a home environment. Do you have experience with the multiple subwoofer approach?
Only with two, which is not real multiple like four.
I prefer quite high lowest X/O (270...400 Hz) with 3-way due to risk of timing issues of low crossed boxed woofers. Good experiences with closed 2-way, but trade off may be somewhere else. Our living room (and wife) may accept 3 subs.
I disagree with the two comments above.
To the first comment, the science shows (Blauert) how very early reflections will mess up the perception of image - maybe you like a poor image, but most won't. I certainly don't. Imaging is key to me. It is true that early reflections do increase the feeling of spaciousness, but at the cost of imaging. You can achieve both with a lively room and narrow directivity speakers however.
I published an AES paper years ago on the modal response of monopoles, dipoles and cardioids in small rooms. In terms of response there was no winner in all cases. Everything depends on how they are setup and where. The dipole and cardioid suffered from the lowest efficiency of course, but did NOT result in few modes excited or a smoother response. Only multiple subs does that and in my experience monopoles work best.
Last edited:
Interesting combination of "small rooms" and "all cases" 😉 In my real life house and apartments stereo monopole has been looser in every single case compared to any type of cardioid. Dipole has been competitive in axially very short rooms or if front wall and corners (behind speakers) are not too massive and reflective. Some difficult cases dipole really sucks. Monopole and dipole need more careful positioning or/and EQ. Super-cardioid is plug & play (positioning for good sound stage and imaging if music is not relevant 😀) if listener is not very close to back wall. This is quite valid from 80m2 complex house to 16m2 concrete/brick bunker, but don't expect to get totally flat response and hear all bass note bursts exactly correct. It's just "not too bad", easy enough to forget and just listen music.
I have nothing against multiple monopoles if subwoofers are generally possible and acceptable approach.
Last edited:
As a layman seems there is a lot of arguing over preferences as there are relatively few absolutes.
So should we as diy'ers be designing specific to the listening room( if we cant/dont want to modify the room) so as to not compromise like commercial speakers.Seems that is a msasive advantage to this community!
So should we as diy'ers be designing specific to the listening room( if we cant/dont want to modify the room) so as to not compromise like commercial speakers.Seems that is a msasive advantage to this community!
That's the point of DIY, at least for me. Commercial speakers sound awful in my home. They may have a flat response in a controlled environment, but here they need serious equalisation.
Other designers' speakers are no good either, for the same reason. So I make my own.
My current SEAS L22/Morel EM1308/SEAS 22TAF/G xo at 850 and 4000, has a flat in room response. They make commercial speakers (I have some KEFs here) sound nasty.
There's a lot of talk about technical stuff - I just wonder if anyone listens to music anymore?
Other designers' speakers are no good either, for the same reason. So I make my own.
My current SEAS L22/Morel EM1308/SEAS 22TAF/G xo at 850 and 4000, has a flat in room response. They make commercial speakers (I have some KEFs here) sound nasty.
There's a lot of talk about technical stuff - I just wonder if anyone listens to music anymore?
Male bass voice fundamental ~80Hz E2
Female Soprano voice fundamental 1.1Khz C6
Measurement of harmonic content from professional Soprano singing voice shows harmonics are -30db to -40db SPL below the 1.1Khz C6 fundamental.
This data seems to support that it is "acceptable" to use steep slope Xovers designs ~1.2Khz - 1.6Khz as a "trade-off" to keep the fundamental vocal range on one driver.
Attachments
- Status
- Not open for further replies.