Avoid crossover in the voice-range?

[Dr.db]

Hello,

I´m wondering which crossover-frequency is in generall worse; 800hz or 1200hz?

Both are right in the human voice-range (300-3500hz) and usually speaker-companies try to avoid crossing over in there due to phase anomaly.
But if you´re dealing with horndrivers, there is no way to avoid it.
So if you have a 2"-driver with a suitable horn that could reach 500hz and a cone-driver underneath like 2123 or similar that could easily play up to 1600hz, which crossover has the advantage?

In the big old 4-way monitors like 4355 JBL used a 2"-driver but a very small horn and high crossover ob 1200hz.
Why didn´t JBL use a much bigger horn, the horndriver could have handled a much lower crossover :confused:


Regards,
Olaf

[ivica]

Hello,

I´m wondering which crossover-frequency is in generall worse; 800hz or 1200hz?

Both are right in the human voice-range (300-3500hz) and usually speaker-companies try to avoid crossing over in there due to phase anomaly.
But if you´re dealing with horn drivers, there is no way to avoid it.
So if you have a 2"-driver with a suitable horn that could reach 500hz and a cone-driver underneath like 2123 or similar that could easily play up to 1600hz, which crossover has the advantage?

In the big old 4-way monitors like 4355 JBL used a 2"-driver but a very small horn and high crossover ob 1200hz.
Why didn´t JBL use a much bigger horn, the horndriver could have handled a much lower crossover :confused:


Regards,
Olaf

HI Olaf,

You have to be aware of the situation that the diaphragm displacement is inverese proportional to the Frequency-squared, so as 800Hz : 1200Hz =1.5, so
the diaphragm amplitude movements is 2.25 TIMES greater at 800Hz then at 1200Hz, so the distortion would be (I can guess about the same larger) but if it is accepted that in standard sound program sound energy is almost inverse proportional to the frequency, so at the mentioned frequency it would be about 1.5 TIMES more energy at 800Hz then round 1200Hz, so multiplying these 1.5*2,25 = 3.375 times , (I would expect distortion rise).

Proper horn loading (large and LONG) horn would help just in order to rise driver efficiency, so less diaphragm displacement, but as You have mentioned that larger driver sounding membrane (cone vs CD diaphragm) would help (mentioned mid-bass vs CD driver) too.

Long horn would produce some other not so pleasant sound reproduction (it can be "seen" if You compare JBL 4333 vs 4350/55 speakers, as in 4333 the horn 2312 is used (about 12" long) but in 435x it is used 2311 (nearly 4.5" long) (internal drivers horn has to be assumed too...but...)
From my taste I prefer compression driver (CD) to operate at 1200Hz then at 800Hz. (the sound is less harsh), especially I prefer 4" CD diaphragm vs 1.75" diaphragm....

Not to mention horn controlled sound dispersion...
I believe that some of our Forum members can give better explanations...

regards
ivica

H

[1audiohack]

Hi Olaf;

I have taken to crossing my large format horns at or about 850Hz. I don’t have any solid reason for doing so.

I always measure the native (unfiltered) response and make decision based on several factors to get started but often end up about there.

I have a friend who is working around the issue of cutting into the vocal range and has built some immense “Smith” horns and they do OK down to 300Hz, if played about conversation level but if you get loud they squawk.

What are you working on?

Barry.

[TimVG]

After testing, and listening to, many different speaker models from various sources, my preference went and goes out to speakers that remain relatively flat (or well behaved at least) through the crossover range, both on- and off-axis.

The actual crossover range didn't seems to matter as much once the above statement was true. Models with higher order crossover designs did better in the off-axis department from what I've seen.

Just my two cents.

Best wishes

Tim

[Dr.db]

Thanks a lot for all your helpful comments! :)

I believe increased cone-movement and therefor increased distortion is one of the major factors to prefer an 1200hz over an 800hz crossover.
My concern was, that the human ear could be more sensitive around 1200hz than 800hz and because of that the lower crossover would be preferred!?

Regards,
Olaf

[bubbleboy76]

I prefered 650Hz LR 48dB/oct over 800Hz 24dB/oct when I did active crossover on JBL 4365.
My feeling was that the compression-driver is better than the woofer, in that region, in that speaker.

[toddalin]

Hello,


Both are right in the human voice-range (300-3500hz)


Regards,
Olaf

This seems to be a general consensus and there in lies the problem. The fundamentals of human voice go down to ~80 Hz for a bass and a soprano tops out at ~1,400 Hz.

When I was designing the Kick-A-- system, I wanted to set my crossover points at ~80 Hz and ~2,800 Hz to cover the vocal range and first octave overtones all in the 10" woofer.

But an "honest" 80 Hz (<3 dB roll-off) typically requires a foam edge woofer in a larger cabinet, and this is what I been avoiding. Also, the 2251 gets a bit louder after ~2,500 Hz. So, I'm thinking of setting my crossover points to about 200-250 Hz and ~2,400 Hz for the 2251J.

[Ian Mackenzie]

Hello,

I´m wondering which crossover-frequency is in generall worse; 800hz or 1200hz?

Both are right in the human voice-range (300-3500hz) and usually speaker-companies try to avoid crossing over in there due to phase anomaly.
But if you´re dealing with horndrivers, there is no way to avoid it.
So if you have a 2"-driver with a suitable horn that could reach 500hz and a cone-driver underneath like 2123 or similar that could easily play up to 1600hz, which crossover has the advantage?

In the big old 4-way monitors like 4355 JBL used a 2"-driver but a very small horn and high crossover ob 1200hz.
Why didn´t JBL use a much bigger horn, the horndriver could have handled a much lower crossover :confused:


Regards,
Olaf

Hi Olaf,

You have asked several questions

There are a number of factors around your first question 800 versus 1200 Hertz.

Looking at older vintage horns (Not wave guides) it was about allowing an overlap of both the low and HF drivers , the moth size to achieve correct loading and the extension of the woofer

Take for example the older Altec horns 500 and 800 hz versions)

Compromising the size of horns since the 50s in domestic systems has pushed up the lower crossover point

Not so with the OMA 300 hertz conical mid horns that are very highly regarded.

https://oswaldsmillaudio.com/ac1

So it was more of a mechanical issue.

Then scientific studies suggested 2000 hertz and above was the no go zone in terms of human hearing sensitivity to certain anomalies

At the time Jbl development in bi radial horns occurred and more research was done on horn and woofer directivity at the crossover point.

Jbl set out to design monitors that matched the directivity of both the woofer and the horn at the crossover point and maintain a relatively smooth directivity index with frequency.

The Jbl 4435 is an example of this style of thinking.

The crossover point is 1000 hertz

The bi radial horn was a breakthrough at the time

More recently Jbl designed the M2 wave guide and due to its size it will crossover lower and has an even wider dispersion to match the lower crossover point and a uniform directivity index with frequency.

Outside Jbl guys like Earl Geddes have researched this topic

The thinking behind the 4350 in the early 70s was to purely limit the bandwidth of each driver to it optimum operational range. That was a smart idea at the time and overcame some basic limitations of the earlier two way monitors

The exponential flare only requires a small mouth for the 1200 hz range in the 2311 horn.

My suggestion is try 1000 hertz on your Smith Horns with the 2123

The 2397 has something like 140 degreeks dispersion at the lower range so you want to match what the 2123 is doing.

For FWIF in cone systems l prefer a crossover point of 250-3000 hertz but that is not an uncommon design approach with cones. Small cones like 4 inch Aerogel drivers are excellent for this application but they will never produce the levels of a large horn .

[Dr.db]

Thanks a lot for all your helpful answers! :)
A great diversity of experience and knowledge.

First of all, this is not to tweak the crossover of my existing DIY-build, it is for designing a new DIY-system in the future.
I understand the history of horns shrinking in size due to comercial demands.
But if you have a big horn like a yuichi a-290 or JBL 2350 with a 2"-driver that is easily capable of 800hz AND you have a good 10"-midcone that could easily reach to 1200hz on the other hand, where do you actually set the crossover?
Is the human ear more sensitive to anomalys like phase-shifts at 800hz or 1200hz ?