Film-Tech Forum ARCHIVE: The X-curve for theatres

This is topic The X-curve for theatres in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=005321

Posted by Sumit Guha (Member # 1887) on 09-14-2003, 10:24 AM:

I would like an explanation of the X-curve that is used to check the frequency response of loudspeakers in cinemas. If a loudspeaker with flat frequency response is used, why does the curve show a roll-off from 2Khz onwards?

Posted by Gordon McLeod (Member # 33) on 09-14-2003, 10:39 AM:

First I would highly recomend that you read Ioan Allens articles on the development of the x curve
They started with research by Bell Labs in the late 20's of what the psycoaccoustic properties were of reproduced sound in large rooms
By using large concert halls and transmitting orchestras to them they used audiences to determin what electro accoustic response the mean audience group felt was natural in a large room
It was discovered that the brain in a large room expects a certain attinuation of high frequency sound and the x curve eventually derived from that
The large the room the more rolloff the smaller the room less roll off

Posted by Manny Knowles (Member # 1171) on 09-14-2003, 12:59 PM:

So, what's considered a "small" room?

Are there any commercial theatres that are so small they don't require any HF rolloff?

Most of my theatres are in the 200 seat range. I have a few that are 100 seats. Our "big" houses are about 300 seats or so.

Posted by Gordon McLeod (Member # 33) on 09-14-2003, 03:13 PM:

The 150 seaters I tend to leave a bit brighter and those over a 1000 a little duller otherwise I follow the curve

Posted by Christopher Seo (Member # 7) on 09-15-2003, 06:57 PM:

Gordon,

Could you point me to the articles you're referring to? The Dolby manuals reference papers, but they (or the ones I could find, anyway) are all either on the development of optical stereo or noise reduction. I haven't read anything that specifically and thoroughly addresses the research that went into the X-curve. Thanks.

Posted by Gordon McLeod (Member # 33) on 09-15-2003, 09:20 PM:

There is a booklete put togather by DOlby by Ioan Allen on cinema design and it historys the x curve
Also there was several article in SMPTE over the issue around the time Quintaphonic came out

Posted by John Pytlak (Member # 331) on 09-16-2003, 10:13 AM:

Links about the X-Curve:

http://www.robertmargouleff.com/publications/surround_0100b.htm

http://www.genelec.com/ht/pdf/aes23rd.pdf

http://www.genelec.com/ht/pdf/aes114th.pdf

http://www.realspace.com/SigTech/aes_90sf.html

http://www.thx.com/mod/techLib/pdf/guidelines.pdf

http://www.americanmusical.com/Manuals/JBL%20Pro/CinemaSoundSystem_manual.pdf

Posted by Ray Derrick (Member # 1001) on 09-16-2003, 10:54 PM:

The standard X-curve that we all know is optimised for a 500 seat theater. For smaller or larger rooms a correction factor is made to the slope of the curve above 2kHz. For example, for a 1,500 seat room you increase the slope of the curve by 1dB at 4kHz and 2dB at 8kHz. For a 30 seat room you reduce the slope of the curve by the same factors.

Some RTA's provide X-30, X-500 and X-1500 versions of the curve, otherwise you need to interpolate the amount of increase or decrease at the other frequencies above 2kHz.

I have edited this post to add, for those interested, the rolloff slopes specified in the standard for various theater sizes (the number after the X being the number of seats):

X-30 2dB/octave 2kHz to 10kHz, 4dB/octave above 10kHz
X-150 2.5dB/octave 2kHz to 10kHz, 5dB/octave above 10kHz
X-500 3dB/octave 2kHz to 10kHz, 6dB/octave above 10kHz
X-1000 3.5dB/octave 2kHz to 10kHz, 7dB/octave above 10kHz
X-1500 4dB/octave 2kHz to 10kHz, 8dB/octave above 10kHz
X-2000 4.5dB/octave 2kHz to 10kHz, 9dB/octave above 10kHz

[ 09-16-2003, 11:56 PM: Message edited by: Ray Derrick ]

Posted by Manny Knowles (Member # 1171) on 09-17-2003, 10:32 AM:

Thanks for all the info.

Can a half-dB really make that much of a difference? (I guess I'll just have to go listen and find out!) [thumbsup]

Posted by Antonio Marcheselli (Member # 373) on 09-17-2003, 06:17 PM:

Ray,

Should't I lower the highs in larger rooms?

Bye
A

Posted by Ray Derrick (Member # 1001) on 09-17-2003, 06:24 PM:

Yes, Antonio, "increase the slope of the curve" means lower the highs.

Posted by Richard Fowler (Member # 893) on 09-17-2003, 06:26 PM:

To a trained ear a 1/2 db is noticeable. I supplied dailies for a Disney movie where the production manager insisted on Flat reproduction for the screening room. I EQ for his spec and for the proper rolloff similar to the X-30 curve...after a 3 minute preview with the FLAT horns piercing the room, the sound mixer and production manager went for the X-30. For a Paramount movie, the mixer was able to hear a 1/2 db change between projectors and with my test equipment, he was exactly right on the level

Posted by Ray Derrick (Member # 1001) on 09-17-2003, 06:51 PM:

Also we are talking 1/2dB "per octave" not 1/2dB across the board. Audio rolloffs are often expressed as so many "dB per octave" which means for each doubling in frequency, the attenuation increases by the number of dB.

So at the first octave, say 4kHz (if you are starting at 2kHz), the attenuation is 1/2 dB, but at the next octave (8kHz) it increases to 1dB, and so on up the spectrum. So the effect is more pronounced the higher you go up the spectrum. In the case of the X-curve(s), the slope doubles above 10kHz, so the effect is even more pronounced above that frequency.

Of course "dB per octave" also implies that the in-between frequencies are proportionally attenuated. So for a 3dB/octave slope starting at 2kHz, the attenuation is 1dB at 2.5kHz, 2dB at 3.15kHz, 3dB at 4kHz (1 octave above 2kHz), 4dB at 5kHz, 5dB at 6.3kHz, 6dB at 8kHz (2 octaves above 2kHz) and so on.

Posted by Antonio Marcheselli (Member # 373) on 09-18-2003, 07:06 AM:

Ray

Sorry, I didn't note the "slope"!
Very interesting this thread!

I've read some of the John's links. But finally, the eq curve is attenuated to highs for a psicological matter or for a physical matter? It is not clear to me...

Bye!
A

Posted by Ray Derrick (Member # 1001) on 09-18-2003, 09:54 AM:

It is really not a matter of either. Look at it this way, it is all to do with how you measure it.

In essence, the size of the room has very little effect on the high frequency response. It is the effect the room has on your measurements that varies.

The various X curves are compensating for an imperfect system of measurement, and little else.

Neither your ears, nor the sound changes (to any great extent) with the room size, what changes is the effect the room has on the accuracy of your measurements. It's as simple as that.

I should add that there are some losses at high frequencies due to distance though the air and varying humidity but such loses are fairly insignificant in all but the very largest auditoriums.

Posted by Antonio Marcheselli (Member # 373) on 09-18-2003, 10:58 AM:

Ray,

Are you talking about the reflections that comes out of phase in a larger room instead of a smaller one?

Bye
A

Posted by Ray Derrick (Member # 1001) on 09-18-2003, 08:58 PM:

Antonio, as the size of a room increases the amount of reverberation and colouration due to reflections usually increases. Reflections and reverberation have reduced high frequency energy because high frequencies are absorbed by most surfaces more easily.

If you stand in a large cinema, your ear discriminates between the direct sound from the speakers and the reflected sound and reverberation, at least in the upper part of the frequency spectrum. A microphone placed in this same space cannot discriminate between direct sound and reflected sound, it can only average all of the sound that it receives and send this to an analyser. So in a larger room the reflected sound and reverberation form a greater percentage of the sound that a microphone picks up and hence the average will be deficient in high frequency energy, even though the direct sound may be flat and sound flat to a listener.

So the X-curve essentially compensates for the inability of the microphone to duplicate what the the human ear (or rather human brain) does by nature.

Now this is a slightly over-simplified explanation but it will give you some idea of why the X-curve is necessary.

Posted by Manny Knowles (Member # 1171) on 09-18-2003, 09:11 PM:

What happens to this theory when the room isn't that reflective because it has really good wall treatment?

Posted by Gordon McLeod (Member # 33) on 09-18-2003, 09:38 PM:

I usually choose the slope based on the RT60 calculation

Posted by Ray Derrick (Member # 1001) on 09-18-2003, 09:53 PM:

Firstly if you read the literature you will find there are alternate theories about all of this, including more recent ideas that there are psychoacoustic factors involved, such as that a person expects a large room to sound duller than a small room. I think the jury is still out on this idea despite some interesting research.

What is clear is that the process is complex and to my mind no one theory provides a totally satisfactory answer. However the explanation I gave above is the oldest theory and to my mind is still the easiest to grasp.

Posted by William Hooper (Member # 71) on 09-19-2003, 12:47 AM:

quote:
What happens to this theory when the room isn't that reflective because it has really good wall treatment?

The ear still expects the same thing. As Ray said:

quote:
If you stand in a large cinema, your ear discriminates between the direct sound from the speakers and the reflected sound and reverberation, at least in the upper part of the frequency spectrum. A microphone placed in this same space cannot discriminate between direct sound and reflected sound,

An illustration is if you've gone to a concert in a large auditorium, then listened to a bootleg tape made at that concert with a mike in the auditorium near where you were. It sounds very inferior to the concert you heard, because part of your perception when listening live was discriminating to minimize the banging & booming reflections. The X-curve, as I understand Ray to be describing it, mimicks the way that perception works by attenuating the high end (what would be the banging part).

It would seem that in larger rooms where the highs would decay faster you could push more of the highs, but this is still counter to what you hear in that environment. Listening to a show in a big arena is more open & with fewer highs than if you listened to it in a small, club-like space. The small one is more bangy & immediate, the big room is big & with fewer highs. That characteristic is what the X-curve mimicks.

Posted by Ray Derrick (Member # 1001) on 09-19-2003, 01:39 AM:

In answer to Manny's question, I doubt if it would make much difference at all unless you were comparing a heavily treated room to an excessively live one. In my experience the speakers in "live" cinemas do tend to sound a little brighter than "dead" ones when equalised to the same X-curve, which is what you might expect if the theory is correct. But this impression may be caused by other factors such as increased hf content in the reverb of live rooms.

Many years ago when single mic analysers were the norm, I did some interesting experiments with microphone types when equalising studio audio control rooms and later, cinemas. I found that I got better results at hf end of the spectrum when using a cardiod pattern mic pointing towards the sound source, than with an omni pattern mic. Of course the omni was still better for equalising the mid and low end. This lead me to conclude that it was because the cardiod mic was not being influenced so much by reflections and reverb. Whether or not that was a valid conclusion was largely immaterial, the results spoke for themselves.

Posted by Gordon McLeod (Member # 33) on 09-20-2003, 09:02 PM:

I have done several very large rooms that had ajustable accoustics via flown in resonators and damping materials
The bottom line was the reaction to the film sixing staff present to the playback of their source material was the more reverberant the room the less High frequencies they wanted

Posted by Ray Derrick (Member # 1001) on 09-22-2003, 11:00 PM:

So Gordy, I guess your experience is further evidence that the simplest (and in this case the oldest) theory is the right one. Somebody mention Arcom's Razor?

Posted by Edward Jurich (Member # 1787) on 09-24-2003, 09:36 AM:

I recall that when magnetic sound came out in the 50's, the audio had a much brighter high end than optical. In fact, one of the reasons for doing magnetic was for "high fidelity" sound. And this was in large ( 1500 to 3000 seat ) theaters in Chicago. Seems like in this case any curve was thrown out the window, the high end was quite bright. I think those old studies were made with live concert music in mind and not necessarly movie sound. Seems like you'd want to try and make your sound as good as possible under any conditions. I know that when I went to those magenetic presentations in the 50's, the sound sounded outstanding compared to normal mono optical of the time.

Posted by Mark Gulbrandsen (Member # 72) on 09-24-2003, 09:55 AM:

"I supplied dailies for a Disney movie where the production manager insisted on Flat reproduction for the screening room."
_________________________________________________________________
I was on way too many location dailies jobs where the sound mixer, these were some big time guys, didn't have a clue about proper room equalization or the X curve. It seems they spend most of their lives wearing a pair of headphones. After setting things flat at their request and they heard the sizzling bacon high end they would usually want to tune it by ear. When I checked it after they left it was almost always near the X-30 curve as well, however it sometimes was an odd curve that was derived from the fact that it was done by ear and I suppose they were attempting to approximate the sound of the headphones they used.
Mark

Posted by Mark Seaton (Member # 1398) on 09-25-2003, 12:24 PM:

Hi all,

Coming of lurk mode for a moment here...

I certainly agree that most any room other than the case of near field monitors will sound overly bright if you see a flat line on an RTA. What everyone needs to remember is what exactly they are measuring with the RTA. An RTA takes an average of energy per band over a selected quantity of time (slow/fast). In terms of acoustics, this time is very long. So, you are in fact looking at the total energy in the room over a given amount of time, which includes both direct sound from the speaker, and a large amount of reflected energy from the room. Rooms generally have more absorbtion at higher frequencies, and if you are lucky, or have a good acoustics designer, they remain fairly constant in the mid-midbass range. Cinemas have the benefit of using generally similar acoustics and layouts where we can correlate the sustained energy in the room (as measured by the RTA) to the direct sound radiated by the loudspeakers.

In fact most would be quite suprised if they saw a measurement of a system voiced to an X-curve then measured by a system which can gate or window out late arriving energy. An MLS, TDS or Log-Swept-Sine type measurement all are capable of this in different manners. When doing any measurement, the most important bit to know is exactly WHAT you are measuring, and what your measurement does and does not tell you. In fact many system would likely measure much closer to flat when looking at the direct sound than when looking at the RTA.

Unfortunately you are only looking at a snapshot of the response of the speakers and room with the RTA, as most speakers change in response and distortion quite dramatically as power levels increase beyond about 1/4 of rated power. Combine these many factors, and we can easily account for the "art" which is often needed in best balancing a system.

Regards,

Posted by Mark Gulbrandsen (Member # 72) on 09-25-2003, 07:01 PM:

Mark,
Obviously you are speaking in terms of a one mic spectrum analyzer here. What exactly does going to four mics through a multiplexer accomplish(as in the R-2)?
Mark @ CLACO

Posted by Matt Zeiner (Member # 1914) on 09-26-2003, 09:17 AM:

quote:
Unfortunately you are only looking at a snapshot of the response of the speakers and room with the RTA, as most speakers change in response and distortion quite dramatically as power levels increase beyond about 1/4 of rated power. Combine these many factors, and we can easily account for the "art" which is often needed in best balancing a system.

Yes... and to complicate matters further, most amplifiers and all analog filters will respond differently to an instantaneous transient versus a steady-state of the same amplitude. Steady-state measurements can be spot on, but the system may still sound less than stellar. Ears are key.

Posted by Manny Knowles (Member # 1171) on 09-26-2003, 04:20 PM:

Let's make it "calibrated" ears and we have a deal! [Wink]

Posted by Matt Zeiner (Member # 1914) on 09-26-2003, 04:22 PM:

calibrated. deal?

Posted by Ray Derrick (Member # 1001) on 10-04-2003, 12:11 PM:

Mark said "Unfortunately you are only looking at a snapshot of the response of the speakers and room with the RTA, as most speakers change in response and distortion quite dramatically as power levels increase beyond about 1/4 of rated power".

It is indeed interesting that we spend so much time and energy worrying about frequency response and distortion in audio electronics when, given the physical nature of loudspeakers, logic would suggest that they would be the most significant contributors to distortion. However I don't recall ever seeing distortion figures quoted in any of the speaker manufacturers' data sheets. A quick check of the spec sheet for JBL's current 3 way cinema system makes only passing reference to distortion (something like "extremely low distortion"). It is almost as if the distortion is so insignificant that it is not worth quoting figures for.

So in your experience Mark, what degree of distortion are we really talking about here and where or how have you obtained such information on typical cinema loudspeaker systems? Also what degree of variation in frequency response with output power are we talking about?

Posted by Mark Seaton (Member # 1398) on 10-07-2003, 02:03 PM:

quote:
It is indeed interesting that we spend so much time and energy worrying about frequency response and distortion in audio electronics when, given the physical nature of loudspeakers, logic would suggest that they would be the most significant contributors to distortion.

A very astute and dead on observation Ray. As an extreme example, consider that if any electronic device had the characteristics of common loudspeakers with respect to distortion and the ability to preserve a waveshape, they would likely be deemed BROKEN! That said, remember that marketing departments want you to focus on improvements they can sell you now. There was a time when heads of major speaker manufacturers would boastfully state "there is no need for speaker response below 70Hz" where 6 months to a year later they would be announcing a new subwoofer which extended lower. Of course no one would ever need lower bass than this! [Roll Eyes]

quote:

However I don't recall ever seeing distortion figures quoted in any of the speaker manufacturers' data sheets. A quick check of the spec sheet for JBL's current 3 way cinema system makes only passing reference to distortion (something like "extremely low distortion"). It is almost as if the distortion is so insignificant that it is not worth quoting figures for.

Good marketing work on their part. In fact the opposite is true. Distortion is bad enough that they don't want you to see the numbers. We are so used to seeing miniscule distortion figures on electronics spec sheets that single and double digit distortion numbers would scare people.

What is more complicating is that direct correlation of distortion to what we hear is a complex and dynamic matter. That is not to say that distortion does not matter, but it is akin to saying two speakers with +/-3dB response over the same bandwidth will sound the same. 99 times out of 100 you need more information. Distortion figures give insight as to what happens at dynamic levels. Distortion is also not percieved the same at all frequencies. We are more sensitive to particular types of distortion, as well as distortion at different frequencies due to our hearing sensitivity and masking.

Making one example, we are very sensitive in the 2-6kHz region. Not suprisingly any distortion products falling in this region can be quite objectionable and will greatly increase the percieved loudness of the system. In loudspeakers the most distortion is found in the 2nd to 5th harmonics, particularly the 2nd and 3rd. This suggests that we want to keep distortion very low in the 400Hz to 3kHz range. It is a bit ironic that looking at music and soundtracks, this range commonly sees the greatest energy over a longer time span. If you've ever noticed that a more powerful speaker yeild much less ear fatigue when played at the same signal levels, this explains why.

quote:

So in your experience Mark, what degree of distortion are we really talking about here and where or how have you obtained such information on typical cinema loudspeaker systems? Also what degree of variation in frequency response with output power are we talking about?

Distortion is an interesting matter to examine. The large majority of distortion is tied to increasing excursion in drive units, with the remainder typically caused by extreme intensities of the air in the throat of mid & high frequency horns, or by port turbulence and choking off at low frequencies. All of these effects are limiting/compression effects, somewhat analagous to saturation of an analog tape, or overdriving of an amplifier. One important factor to remember is that once a speaker is overdriven to distortion, anything being produced at the same time by the device will also be distorted. So even if a loudspeaker may have maybe 1% distortion at relatively high levels in the middle of its range, if a signal at its low end drives it to distortion, that 1% distortion figure is gone as any signal present will be similarly distorted.

So far as the distortion levels we typically listen to, most direct radiating subwoofers I have measured reach 5-10% distortion at less than 50W of input, many times with as little as 10W. There certainly are some which will behave much better than others, but at realistic playback levels it is common to see the better direct radiators at a few percent distortion. This is typically in the 105-115dB @1m range. Consider that most cinemas are about 10m deep to the center, and that is 20dB lower than at 1m. I have measured our bass horns as well as a few others and typically distortion is 1/5-1/10th that of direct radiating subwoofers. Where we see the most difference between products is in the onset of distortion. It is not uncommon in many prosound applications where subs are driven at levels resulting in 15-25% distortion.

In main speakers we still have similar problems. At modest levels you will find better speakers with ~0.5% distortion over some bandwidth. Quite often speakers operate yeilding distortion in the 1-5% range. 1% distortion at say 110dB at 1m would be described by most listeners as fairly clean. Certain horns have some interesting distortion mechanisms which show up at moderate levels and higher due to high intensity in the throat, often referred to as "throat distortion."

We have, and you can EQ two different speakers to have almost identical frequency response yet you will still hear the signature of any distortion in the systems. A horn which "honks" badly will still retain that character even after EQing.

So far as level variations, this is mostly due to power compression. As you push more power through a voice coil or diaphram, the wire heats up. As commonly used metals heat up, the resistance increases, which reduces the amount of power delivered to the coil, and hence power compression. Note that many of todays very high power drivers achieve much of this power handling by withstanding much higher temperatures before glues melt and things deform. JBL did a study a few years back which actually proved that at some very high power level you can push more power into some drivers, and because of the heating and resistance rise, you actually get LESS output. To put this in real terms, if the resistance doubles, power halves, resulting in 3dB of power compression.

In the ideal when you increase the signal/power by XdB, you expect to see the speakers get XdB louder. Above about 1/4 rated power for many speakers, this no longer holds true. What you see in measuring a speaker is that this effect is frequency dependent, depending on the design. This effect is seen both from heat as well as from overdriving of a speaker in terms of excursion. Subwoofers are easy examples as they are less complicated devices operating at long wavelengths, but a good example can often be seen in vented boxes where measuring the response at increasing power levels shows significant changes in the shape of the low end roll off. To varying degrees similar effects occur at midbass and midrange frequencies as well.

This got quite lengthy, but I hope someone finds it of interest.

Enough for one post... [Wink]

Posted by Steve Guttag (Member # 268) on 10-13-2003, 12:00 AM:

Ray,

There was actually a published article on the JBL 5674 3-way speakers and their distortion components (based on order) to show the superiority of the 3-way system over "antiquated" 2-way designs.

I'm pretty sure it was published in the SMPTE (and possibly other trade magazines)...I'm sure John P can provide the links are methods to obtain the article...he is so good at such fact-finding.

Steve

Posted by John Pytlak (Member # 331) on 10-13-2003, 08:49 AM:

quote:
I'm sure John P can provide the links are methods to obtain the article...he is so good at such fact-finding.

Steve must think I am just a librarian! [Wink]

The paper is "Improvements in Motion Picture Sound: The Academy's New Three-Way Loudspeaker Systems" by John Eargle, M. Mayfield, and David Gray, SMPTE Journal, July 1997, Volume 106, page 464.

Here are some links:

http://www.jblpro.com/pub/cinema/cinedsgn.pdf

http://www.aes.org/journal/sample_issue/JAES_V50_11_PG954.pdf

http://www.aes.org/sections/la/meetingnotices/sept_97.pdf

http://www.smpte.org/members_only/library/indexes/ (SMPTE Members only)

http://www.doctorproaudio.com/doctor/bookstore.htm

Posted by Kevin Wale (Member # 1851) on 10-14-2003, 07:35 PM:

I'm a big time "use your ears" believer.

Every mic is different. Every analyzer is different.

Something else that people tend to overlook is speaker placement. Move the cabinet an inch and it will change the curve. I have also wondered about speaker bases commonly in use today. Most I have seen lately are metal. Thin metal at that. When you tap on them, they resonate. Instant distortion in any frequency range that is sympathetic to the metal's resonation.

The screen it'self by virtue of even being there is automatic distortion, holes or no holes. So, what distortions are acceptable and able to be worked around and which ones are damaging to the presentation? Obviously the screen is such a low amount of distortion that 1 in a million people might be able to hear the difference with the screen up and with it down on the same settings.

My next question then comes to this. What speakers currently yield the least amount of distortion? If the JBL's have what has been said is a significant amount of distortion(despite thier claims that it is very low), what are the options?

Back on the x curve...

Obviously different parts of a movie are at different volume levals. When I record or mix live sound, I use compressors, limiters, and parametric eq to make things sound as consistant as possible at the various volume levals. I think it's something like at every 10 dbs certain midrange frequencies begin doubling to the human ear's perception. An eq curve that works at 70 dbs ceases to be effective at 90 dbs.

So far, changing the curve automatically(morphing between) isn't possible in real time. Or is it? I know theoretically how it could be done, but I certainly haven't the knowledge or technical skill to even begin. Volume sensitive switching is being employed in lighting, so why not sound?

I don't know if any of that makes sense... I'm just thinking out loud more than anything but it seems to me that the problems are not only needing more mics to measure but also more measurements at different volumes. The main reason biggere auditoriums need the xcurve modified is because they require more volume to fill the room. More volume means the eq is percieved differently by the ear(espeically midrange and hense spoken dialogue). That in itself implys that the movie's different volumes would idealy need a different curve too. Correct?

Posted by Gordon McLeod (Member # 33) on 10-14-2003, 07:43 PM:

It doesn't need more volume to fill a large room just more power or higher efficiency of speaker
Eargle and several others have published many papers on the issue of the Theile Small parameters on direct radiator designed speaker cabinets and also the design of crossovers for them