Murphy Corner Line Array (MCLA)

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John,
Thanks for posting the design. I find it very intriguing. There are some old and new ideas presented.

I've been thinking a lot lately about corner-loaded speakers. Klipsch "cornered" :rolleyes: this market once upon a time, but you don't see it anymore ,and I'm not sure why this is. (Of course, I don't have my speakers in the corners, either)

I guess the thing I'm left wondering after reading your writeup is that you continually describe the "corner reflections" as mirrored images. This makes sense to a point; however as you know, not all frequencies behave the same directionally (bass = omnidirectional, treble = directional, midrange = transition). So I'm just sort of wondering how you factor this frequency discrepancy into the design to keep the overall FR "flat" (or x-curve or whatever your desired FR needs to be)

Again, it's a very interesting design; I only wish I could hear it. Reading about it and the theory behind it doesn't take the place of an in-person experience.

Ahh, the classic K-horns.

Speaker enthusiasts traditionally say that bass is omnidirectional and that the treble is directional. In the context of a typically sized speaker enclosure this makes sense because the baffle is usually about as wide as middle frequency wavelengths where 1 kHz has a wavelength of about 1 foot. At higher frequencies the baffle is an obstruction which limits rearwards radiation and imposes half-space loading. At longer wavelengths (lower frequencies) the baffle is small with respect to the wavelength and the bass radiates omnidirectionally in spite of the baffle obstruction.

In general, the directional characteristics of a sound source depend on the dimensions of the source and the dimensions of nearby obstructing and reflecting surfaces. Here's a simple example that clearly contradicts the speaker enthusiasts rule-of-thumb that bass is omnidirectional. If a 1 foot (12inch) woofer beams above about 2 kHz a 100 foot woofer would beam ((radiate in only a narrow angle) above 20 Hz. That is, a 100 foot diameter woofer would produce highly directional bass. From this extreme example we see that bass frequencies are only omnidirectional because of the (relatively) small size of our usual speaker enclosures compared to bass wavelengths.

Today line arrays are finding increasing application in large concert venues because they provide a degree of low frequency pattern control due to their long length. This tends to improve the sound of these large rooms where low frequency reverberation times can be unusually long and troublesome. Long line arrays can limit the low frequency radiation to a narrow vertical angle (compared to smaller sources) and direct less energy into the room and more on the concertgoers. This holds down the tendency of these large venues to ring obnoxiously in the bass range.

In-room the MCLA system has wide frequency dispersion throughout the frequency spectrum. It does narrow slightly and smoothly in the top octave but offsetting this is the fact that the listening angle is automatically restricted to less than 45 degrees by virtue of the corner placement. Because all frequencies are handled by the same driver there are no abrupt changes in directivity such as occur when crossing between a larger woofer and a small tweeter.

Build some and tell us what you think.

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John,

Any chance on posting some frequence responses of the speakers using the exact same eq outdoors or as close to anechoic as possible, to compare to those taken in the room.

I would like to see that measurement also...but with cold weather arriving and my other projects heating up it may take a while to get to that.

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

And yet again, all the heads turn to the JVC 3x5s. That is the obvious choice i presume, same Fs, freq extension. But much less xmax, but then much cheaper, so u can use as many as u want!

I think alot of people will come in to say that, its one of those situation where without a tweeter line, it sounds perfectly good and extends high enough etc. etc., Yet, when a tweeter line IS added, you'd notice whats missing. Hence I've been living with my JVC3x5 tweeterless arrays happily, making sure that I DON'T get to hear one with a tweeter line

I am sure these MCLA sounds fantastic, but I am also confident that you can do just as well with cheaper drivers like the JVCs, or peerless, TB, Aura, etc. etc. 2-3" fullrangers with some calculated comprimises - e.g. distortion OR SPL, and ovals are A PAIN IN THE *** to drill, when u have a 100 ovals to cut.....

To me this is just a good old full range line array stuffed in a coner, so as opposed to a "novel design" its more like a "novel placement/positioning" ? I mean, not many speakers designs here would ask you to do corner loading, other than subs. But as far as a "fullrange line array" is concerned, how much more designing is required other than finding a fullrange driver, and putting them into a correct size box? Maybe spend a little time on an EQ. But sounds like the drilling and cutting will be much much more effort!

A 3x5 should be fine. The Xmax may be adequate or not depending on the low end output the listener requires. I would encourage everyone to experiment if they have the time and inclination. You will almost certainly need to determine and setup your own custom EQ but that's not really hard to do. Don't expect the stunning bass capability of the MCLA with the ND90-8 if you use a less capable driver...but you may not need that level of output for your application. Remember I designed the MCLA to also use as the sound system for my band which is a much more demanding than the usual CD playback or home theater applications.

Regarding "missing highs" that some people report with line arrays: I still have my older 2-way line arrays (with 16 dome tweeters) which I listen to regularly as I have for 30+ years. I can't perform a quick A/B comparison because the old and new systems are in separate rooms on different floors. (apologies to Kathy Mattea) But I can definitely say that nothing appears to be missing in the new full range arrays...except for the crossover coloration. As a musician I definitely feel that I would hear if there were any missing harmonics. I can very easily hear the slightest "flanging" or "chorus effect" when I deliberately apply that type of effect to my guitar sound. Flanging is where you deliberately introduce a comb filter and then slowly sweep the notches around as a "sound effect". While I have definitely heard flanging with large scale concert sound systems when I walked the room while listening to pink noise, I have never heard it on either of my line arrays. In my opinion, there are no missing highs. The high end sounds constant as I walk around the room and as I change my listening height by sitting down or standing up. The MCLA is actually more consistent this way than my 2-way line arrays because I hear a very slight shift in response due to the crossover as I move laterally which changes the phase relationship between the woofers and tweeters. To me the full range array is more "ideal" with respect to shifting horizontal position.

Yes, it was a pain in the **** to cut 48 of those 3" holes. I can't imagine cutting ovals. This is why man invented NC machinery.

Yes, it is a full range line array stuffed in the corner. But I do believe the "image friendly" enclosure design is sufficiently novel that a US patent would have been possible. But that does not mean much these days...they still will patent things that have been in the public domain for seemingly forever. That's why I made such a fuss over the public disclosure. I want to make it easy for the patent examiner to deny anyone's patent application for a line array "stuffed in the corner". Someone could still obtain a patent and then have their lawyers hound anyone who was making the enclosures. The builders would then have to mount a lawsuit to overthrow the clearly bogus patent. That takes time and money. With the backing of the company I was working for I did manage to shut down harassment from attorneys for a patent holder trying to extort licensing fees for a "derived" crossover I was using in one of our products. I was rather arrogant in my response to their demand letter and explained to them how their patent was worthless because they had foolishly patented a circuit that was CLEARLY in the public domain long before their patent application was even filed. I included copies of the published papers detailing crossovers derived by subtraction and I never heard from them again. So I felt it important to make it very clear that this design is now in the public domain.

As you note, the design part for this closed box system was really quite simple. All that nasty crossover design stuff just vanishes...so does the nasty crossover sound coloration. I used to find myself making the case that a 2-way was always preferred over a 3-way (or higher) speaker because crossovers were a "necessary evil" that should be avoided whenever possible. Now I find myself making the case against using any crossover at all. While crossovers remain "evil", in my opinion, they are no longer "necessary" to achieve the highest quality sound reproduction.

Try this thought experiment...or maybe even try it for real. Imagine two identical wide range drivers (say the ND90) in separate closed boxes but with the drivers located close together. Now you add a crossover at, say, 2 or 3 kHz so that each driver operates over only a portion of the frequency range. Next you rig a switch to change between the crossed over pair and a single driver operating with no crossover. Then you conduct a listening test asking only if you can hear any difference. Do you think you could hear a difference? When I tried this test (around 30 years ago) I could always hear the difference. The crossover was always audible even in this "ideal" application of crossing between two identical wide range drivers in the middle of their operating range. In real 2-way systems we are always crossing between a larger driver and a smaller driver which just adds another layer of audibility as we introduce the "dispersion discontinuity" that attends the shift from a narrowing dispersion woofer to an ultra wide dispersion tweeter. All those audible effects completely go away when you eliminate the crossover. Try it yourself and tell us what you hear.

In the end I spent days measuring and tweaking the EQ to get it "just right". I am hoping that those who precisely clone my MCLA will not really need further EQ tweaking; or at least that it will be minimal. For me the hard part was building the enclosures. But I know that many DIYers excel at woodworking and I look forward to seeing some really good looking implementations of the MCLA. I guess I'm getting long winded here on this lazy Sunday afternoon.

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

On the max output issue, I meant a lower x-max mini would not be able to produce 30-50 hz as well, meaning the upper freqs would need to be attenuated more if you wanted flat 30-20k hz response.

Your design concept is pretty interesting, and i think it illustrates some things we all sort of suspect, how reflections affect performance. I never thought of attacking the problem the way you did. I'm pretty intrigued, but I'm just trying to think of a more cost effective way of implementing this.

I am aware you can't provide support if I use different drivers, but I think it's really cool you made this a public domain project, and I hope people build it and give their impressions.

I do have a question about the corner coupling. How long does each wall have to be to effectively a reflector. Where I would probably mount something like this, I have 4 ft of wall then an opening for entry into an adjoing room. Would this be acceptable for reflection redirection and reinforcement?

Yes, the driver must have an adequate combination of Xmax (linear displacement) and S(D) (piston area). The product of these two parameters is V(D), the peak volume displacement of the driver. The V(D) establishes the maximum acoustic output of the driver. Based on its 4mm Xmax and 2.5" piston diameter a single ND90-8 has 0.773 cubic inches of linear volume displacement. A single MCLA with 24 ND90s has 18.5 cubic inches of displacement capability. Two arrays can displace 37 cubic inches. Two arrays with corner reflections and just their first two floor/ceiling reflections (effectively 980 drivers!) effectively provide 757 cubic inches of displacement. So the ND90-8 might even be considered overkill considering the contribution from the reflected images when located in a room as specified.

Certainly there is room to experiment with even smaller drivers is search of a lower cost implementation. As a product designer it is the high-value products that ring my bell. I enjoyed reading Roger Russell's description of using a $.69 driver to prototype his full range line array. http://www.roger-russell.com/columns/columns.htm

I have always liked the challenge of designing a product as inexpensively as possible consistent with the required performance. Take the simple example of a power amplifier; let's say a stereo power amp rated at 100 Watts per channel into 8 Ohms. Such amps can be had for a price ranging from around $200 to $20,000. From a design perspective, building an excellent power amp with the parts budget implied by the $20,000 price tag is a no brainer...even a trivial task. It certainly would not be a challenge to a good design engineer. It is the low end of the budget where the design challenge becomes mush more interesting to me personally. That's one reason I'm interested in designing an analog EQ for the MCLA...it could be built for much less than the $300 Behringer digital EQ I'm currently using and specifying for the MCLA. I don't need for a product to be expensive in order to accept that it has excellent audio performance. For example, I really enjoyed having my ($100) TrueRTA software recommended as a replacement for a bench-top spectrum analyzer costing $80,000 in a published article describing a standardized communications industry test procedure.

Back to jmc's questions...The 4 foot wall span before the doorway may have a small impact on the bass but it would not stop me from trying it. At worst you may need an EQ adjustment. It may depend most on how you listening area is located within the room. I assume you have a standard ceiling parallel to the floor.

Regards,

John

[Paul Carmody]

Re: Murphy Corner Line Array (MCLA)

John,
Thanks for posting the design. I find it very intriguing. There are some old and new ideas presented.

I've been thinking a lot lately about corner-loaded speakers. Klipsch "cornered" :rolleyes: this market once upon a time, but you don't see it anymore ,and I'm not sure why this is. (Of course, I don't have my speakers in the corners, either)

I guess the thing I'm left wondering after reading your writeup is that you continually describe the "corner reflections" as mirrored images. This makes sense to a point; however as you know, not all frequencies behave the same directionally (bass = omnidirectional, treble = directional, midrange = transition). So I'm just sort of wondering how you factor this frequency discrepancy into the design to keep the overall FR "flat" (or x-curve or whatever your desired FR needs to be)

Again, it's a very interesting design; I only wish I could hear it. Reading about it and the theory behind it doesn't take the place of an in-person experience.

[justinc]

Re: Murphy Corner Line Array (MCLA)

John,

Any chance on posting some frequence responses of the speakers using the exact same eq outdoors or as close to anechoic as possible, to compare to those taken in the room.

[DDF]

Re: Murphy Corner Line Array (MCLA)

John,
Thanks for sharing your design. The benefits are articulated well in your article but I had a few questions about the trade offs. These aren't "gotcha" questions by any means since the research tends to vary on these effects. Since you've lived with this arrangement so long, hoping you have some experiences to share in these regards.

Some psycho acoustic research indicates that short delay lateral reflection colour the timbre and lead to spreading of the individual image sizes. Have you noticed anything to this effect with the MCLA? For example are image sizes the same as traditional speakers out in the room? Does the timbre sound similar to a box out in teh room that measures about the same, anechoic wise?

Other questions are in regards to imaging. One of the benefits of having the cabinets a few feet out from the walls is that this positioning can enhance the perception of width and depth. How do you find this trade off working for the tight corner design? Also do you miss any sense of spaciousness? Cabinets 5 or more feet from the boundaries can enhance this perception.

Thanks again,

Dave

[blamus]

Re: Murphy Corner Line Array (MCLA)

I'm just trying to think of a more cost effective way of implementing this

And yet again, all the heads turn to the JVC 3x5s. That is the obvious choice i presume, same Fs, freq extension. But much less xmax, but then much cheaper, so u can use as many as u want!

I think alot of people will come in to say that, its one of those situation where without a tweeter line, it sounds perfectly good and extends high enough etc. etc., Yet, when a tweeter line IS added, you'd notice whats missing. Hence I've been living with my JVC3x5 tweeterless arrays happily, making sure that I DON'T get to hear one with a tweeter line

I am sure these MCLA sounds fantastic, but I am also confident that you can do just as well with cheaper drivers like the JVCs, or peerless, TB, Aura, etc. etc. 2-3" fullrangers with some calculated comprimises - e.g. distortion OR SPL, and ovals are A PAIN IN THE *** to drill, when u have a 100 ovals to cut.....

To me this is just a good old full range line array stuffed in a coner, so as opposed to a "novel design" its more like a "novel placement/positioning" ? I mean, not many speakers designs here would ask you to do corner loading, other than subs. But as far as a "fullrange line array" is concerned, how much more designing is required other than finding a fullrange driver, and putting them into a correct size box? Maybe spend a little time on an EQ. But sounds like the drilling and cutting will be much much more effort!

[jmc]

Re: Murphy Corner Line Array (MCLA)

On the max output issue, I meant a lower x-max mini would not be able to produce 30-50 hz as well, meaning the upper freqs would need to be attenuated more if you wanted flat 30-20k hz response.

Your design concept is pretty interesting, and i think it illustrates some things we all sort of suspect, how reflections affect performance. I never thought of attacking the problem the way you did. I'm pretty intrigued, but I'm just trying to think of a more cost effective way of implementing this.

I am aware you can't provide support if I use different drivers, but I think it's really cool you made this a public domain project, and I hope people build it and give their impressions.

I do have a question about the corner coupling. How long does each wall have to be to effectively a reflector. Where I would probably mount something like this, I have 4 ft of wall then an opening for entry into an adjoing room. Would this be acceptable for reflection redirection and reinforcement?

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

Would it be possible do the project with the 4ohm version, since it could be series/parallel'd to a more friendly 6 ohms nominal? They also have a slightly lower Fs, and may be more friendly to doing a clone with passive bass correction down to say 45 hz.

Also, since you're using digital correction, couldn't any 3 inch driver that was smooth out to 15k+ eventually be eq'd to 30hz, at a loss of total output?

Also, no offense meant here, but isn't getting back to design premise for the Bose 901, being to not try to avoid reflections, but use them to a positive end?

You could create similar systems with excellent performance using any good driver in the 3 to 4 inch diameter range. The 4 Ohm version of the ND90 would be a good choice. The net impedance of the MCLA using the 8 Ohm driver is 5.33 Ohms. If you extend the array to 25 drivers then you can wire them in a 5 x 5 configuration to achieve the same impedance as a single driver.

While you could use "any" 3 inch driver to recreate MCLA the system's distortion performance can only be as good as the driver you use. This is most important in the low frequency range below 100 Hz where distortion always increases sharply as cone excursion increases. I selected the ND90 for its honest 4mm of Xmax which allows for very low distortion in the low end.

Here is a single array producing 94 dB SPL at 1 Watt at 1 meter:



At 50 Hz the 2nd harmonic is -41 dB down for 0.9% distortion. The 3rd harmonic increases to -37 dB or 1.4% distortion. The distortion stays nicely low even up to extreme sound levels. If the selected driver has higher distortion than the ND90 then an array of drivers will also have higher distortion.

I'm not sure what you mean by " at a loss of total output".

The MCLA is really nothing like 901's except that the 901's designers simply acknowledged "room reflections" with no apparent awareness of, or design for, the reflected images. Otherwise the designers would have concluded that the enclosures should to be placed in the corners with the drivers facing into the room. The 901's are quite distant from their reflected images when placed in a room as recommended by the manufacturer. This implies big time delays among the widely dispersed images. The MCLA is all about keeping the images in a tight group with minimal delays between adjacent images.

Regards,

John

[jmc]

Re: Murphy Corner Line Array (MCLA)

Would it be possible do the project with the 4ohm version, since it could be series/parallel'd to a more friendly 6 ohms nominal? They also have a slightly lower Fs, and may be more friendly to doing a clone with passive bass correction down to say 45 hz.

Also, since you're using digital correction, couldn't any 3 inch driver that was smooth out to 15k+ eventually be eq'd to 30hz, at a loss of total output?

Also, no offense meant here, but isn't getting back to design premise for the Bose 901, being to not try to avoid reflections, but use them to a positive end?

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

No off-axis plots. Corner mounted line arrays are nothing new, I designed my TLAH, for instance, seven years ago. Much of what Murphy states is true, but I'd want to see what it does at 30 and 45 degrees off-axis.

The frequency responses I have posted currently are not precisely "on-axis" but rather an average response over the listening area. Here it is:



I'd estimate that these measurements ranged over +/- 15 to 20 degrees or so. ajinfla points out that the listener can never be in a location more than 45 degrees off axis so 45 degree measurements are unnecessary. Bill does make a good point about documenting the off-axis response. I'll try to put together some 30 degree off-axis measurements as time allows.

As I was reviewing the design I thought the restriction of the listerner to being less than 45 degrees off axis provided a nice balance to the fact that the radiating piston is 2.5" in diameter. Over the decades speaker designers seem to have largely come to accept that the "wide dispersion" of the 1" dome tweeter is necessary for good sounding high frequency reproduction. Actually, the larger piston diameter doesn't bother me at all for a number of reasons. Even though I have designed with 1" domes over the years and elected to use them in my beloved 2-way reference line arrays I have always had a slight preference for the sound of highs from those studio monitors which employ tweeters with less wide high end dispersion...such as horns. On in-studio playback of drum tracks the cymbols in particular, just sound more realistic to me on those monitors with slightly narrower high frequency dispersion. There's no comparison like hearing a live drummer in the studio and then stepping into the control room for a quick playback of what was just recorded. The 2.5 " piston of the Dayton ND90 used in the MCLA seems to provide high frequency reproduction in-between a 1" dome and a high quality high frequency horn.

I'm aware of Bill's TLAH design. Here's a link:

http://www.billfitzmaurice.com/TLAH.html


The TLAH is a shorter (40") 2-way array with a somewhat wide baffle which appears to be angled slightly downward toward the listener and appears not parallel to the front wall.

The MCLA spans (approximately/ideally) the full distance from floor to ceiling, is parallel to both adjacent walls and the enclosure has been specifically shaped to join the source with its reflected images as tightly as possible. The fact that the MCLA enclosure was designed in light of the geometry it forms with its reflected images in the walls, ceiling and floor is the most unique aspect of the desgn.

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

I have been following this project over at www.audioroundtable.com and wanted to get feedback here.
His test data does look good when using the DEQ2496. It goes pretty flat out to 20kHz.
Any thoughts (or concerns)?

Hello Tech Talk!

This is John Murphy, the guy who operates TrueAudio.com. I am the designer of the MCLA and author of a couple of audio software applications and a book on loudspeakers. I'd like to tell you about my new line array loudspeakers and invite you to build them for yourself and your friends. The first pair are now serving as my primary monitors in my home music studio where they are getting a good workout and have already exceeded my extreme expectations. I am also planning on building a pair to serve as my front stereo mains for my home theater.

Before proceeding I first want to say thank you Jerrald for initiating this thread. I've been putting off starting an "official" discussion thread for the MCLA project untill I felt like my web pages were sufficiently well developed to warrant it and I think I now have enough information for builders to build the system and voice it. Please bear with me for just a moment as I attempt to orient you on the project.

The Murphy Corner Line Array (MCLA) is my first open speaker design project for the DIY community even though I've been studying and designing loudspeakers professionally since 1978. That's when I changed careers from serving as a space systems analyst for NORAD in Colorado Springs, CO to working as Chief Engineer for Ford Audio and Acoustics (now Ford Audio-Video) in Oklahoma City. Starting in 1978 I routinely designed several unique lousdpeakers each month (using a hand held calculator to plot simulated responses) in support of rather large and complex sound system installations throughout the region using primarily JBL transducers and Crown electronics. I'm not (just) trying to brag about my audio experience but rather to give my background as it relates to speaker design in order to give those less seasoned DIY speaker builders a bit of confidence that they won't be wasting their time and money on someone's first speaker project. This is a speaker system I have designed for use as my own primary audio monitoring system and I thought there might be others who would like to build it for their own enjoyment and/or profit. I have formally disclosed the "invention" of the corner line array in order force it into the public domain and to prevent anyone from trying to patent it or otherwise restrict DIY builders from making or selling the MCLA. As far as I am concerned everyone is free to build and/or sell the design.

OK so here is a link to the official project web site:

http://www.trueaudio.com/array


I have also compiled the documents for the entire project into a .pdf file located here:

http://www.trueaudio.com/array/downl...%20Project.pdf


I will consider THIS to be the official discussion thread for the MCLA project. Your comments, criticisms and questions are welcome.

<flame suit on>

Cheers!

John

Re: Murphy Corner Line Array (MCLA)

No off-axis plots. Corner mounted line arrays are nothing new, I designed my TLAH, for instance, seven years ago. Much of what Murphy states is true, but I'd want to see what it does at 30 and 45 degrees off-axis.

You want to see what a corner speaker does at 45 degrees off-axis?

[billfitzmaurice]

Re: Murphy Corner Line Array (MCLA)

I think it's this one:

http://audioroundtable.com/forum/ind...12100&start=0&

Any thoughts (or concerns)?

No off-axis plots. Corner mounted line arrays are nothing new, I designed my TLAH, for instance, seven years ago. Much of what Murphy states is true, but I'd want to see what it does at 30 and 45 degrees off-axis.