Murphy Corner Line Array (MCLA)

[time]

Re: Murphy Corner Line Array (MCLA)

I'm interested in your MCLA design. I like the theory behind it. I do have questions though.

You make this statement on your site: "A subwoofer is not required and would likely even degrade the performance."

Why would a sub(or pair) degrade the performance? I suppose I can think of two reasons on my own. Forgive my novice termonlogy.

Of coarse, there is the added crossover. With part of theory behind the array being 'no crossover', I can see where adding a sub would pose a problem.

There is also placement issues. My sub pair will be quite large and should be placed beside or behind the mains. I may have a way to do this in my room, but most will not.

I guess in simple terms I'm questioning the bass capabilities of the speakers. I wouldn't want to build the design and find I'm not happy with the lows and not be able to add a sub. After adding sub to my current system, I heard things in some of my music I didn't know was there.

[richidoo]

Re: Murphy Corner Line Array (MCLA)

Thanks for the thorough explanation, John. That makes sense. There is recorded depth and then there is the additional sense of depth added as room distortion which can be pleasant or not depending on the setup. The recorded depth should come across clearer if room distortion is lowered as the corner arrays should do. I have listened to music in pseudo anechoic space, and I loved being able to hear the recorded spatial cues so clearly without added room reflections. But like any other distortion we get used to front wall reflections an grow to prefer them until we unlearn the habit with something new.

I bought a pair of the N90 drivers, just to see what they look like and hear them in a small box, as you did. They are very high construction quality. I am looking forward to getting started on the project.
Thanks again
Rich

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John, I have been reading your speaker design articles on your website for a long time, and recently stumbled onto your new MCLA design. Thanks for sharing it. I tried to find it again a couple days later, not so easy. But I got it bookmarked now. A friend of mine brags about his pipedreams all the time, so I want to give your array a try. I am accustomed to in room speakers which have good imaging and depth, with tweeters spaced 5+ feet from the front wall.

Referencing Curt's post above, can you comment on the soundstage depth of the MCLA? The mirror effect that you describe has me thinking that maybe the front wall kind of disappears due to the reflections, and there might be decent depth perception. I will be using these primarily for stereo music listening. Do the images seem to be natural size, or stretched tall? Pipedream and Maggies claim to have a naturally full size (large) presentation.

Thanks
Rich
my 1st post on PE forum

Hi Rich, welcome to the MCLA thread at Techtalk!

I've been thinking about your question because I am reluctant to get very deep into subjective descriptions of the "sound" of the system. Most of my own listening is centered on popular rock/country/jazz recordings where the perception of depth arises primarily from reverb and delay settings applied when the recording was mixed. On these recordings the sense of depth seems to be conveyed properly. In my opinion, the MCLAs provide a full sense of width across the soundstage as well as appropriate perceived depth.

Concerning soundstage depth vs. speaker placement away from the front wall I would just note that as the speaker is moved away from the front wall it becomes progressively more separated in space from its front wall reflection. The speaker moves forward 5' but the reflected image moves back 5'. Moving the speakers off the front wall will change the system impulse response such that the front wall reflection and all of its higher order reflections will all be delayed about 10 ms (for the case of speakers 5' from the front wall) compared to the direct sound and the other reflections. This small rearrangement of the playback system impulse response is likely to be audible but not necessarily better or worse; rather, it seems likely that it would just be randomly "different". In one room it might sound "better" while in another room it sounds "worse". From a "system" point of view we want to optimize the total system impulse response given a direct sound impulse and numerous early arrival image impulses which then repeat with progressively closer spacing and evolve into reverberation before decaying. We cannot make the reflected impulses "go away" (as with anechoic testing) but we do have some freedom in aligning the timing of the reflection impulses. We would achieve the highest listening resolution if we could align all of the impulses at any one time. Where that is not possible the next best thing with the earliest arrivals is to align the impulses so that they are clustered as tightly as possible. If we can squeeze all the impulses into a single "fusion window" (about 20ms or so) our human brains will perceive them as a single acoustic event. In the case of speakers placed along a wide front wall I would probably want to start with the speakers on or near the wall in order to move the reflection as close as possible to the source and tighten the timing between the two impulses. The ideal case would be flush mounting the speaker in the front wall so that the source and the image are not just close but virtually merge into one sound source. This happens again when you move the speaker into the front corner; you then merge the side wall reflected images with the sound source and front wall image. With narrower front walls corner placement works out nicely for either point sources or line arrays.

Getting back to Rich's question about the image sounding "tall"... I wonder if perhaps the notion that line arrays provide a "tall" image does not arise from the fact that the image is always perceived to be exactly at the listeners height such that when the listener stands the image moves upward to track the height of the listener's ears. When you sit, you hear the image track down with you. This is a readily noticeable difference from the way point source speakers behave. In any event, once I am happily seated (or standing) I do not perceive the performers as being "tall". But it is a little unusual at first to hear how the performer seems to sit when you sit and stand when you stand.

Regards,

John

[Humboldt_G]

Re: Murphy Corner Line Array (MCLA)

I am really intrested in this design. I noticed that you said the room MUST have corners to put into, well I only have one corner and the other side is just a wall to the rear.Should I look at another design?

I am truly an amateur when it comes to DIY speakers.

I really like the idea of full range speakers for the simplicity of it. I know you have strict eq settings to achieve the sound you have but I dont have an Eq. Where can I get an eq for this without breaking my bank? drivers for your original design are going to cost over $600 so I would like to keep costs down.

-Coop

[audiorasp]

Re: Murphy Corner Line Array (MCLA)

Thanks for the reply John.

I just finished cutting all of the mdf and I'm ready to start cutting the holes in the front baffles. I've decided to go with 25 drivers per side with 5 groups of 5 wired in series-parallel. This should give a calculated impedance of 8 ohms.

Geo

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

Is there any specific reason or advantage for mounting the ND90-8 drivers on the inside of the front baffle?

Geo

I elected to mount the drivers from the rear because I thought they looked better in the simulations. Rear mounting also provides the drivers with a bit of protection. The flat portion of the rubber surround serves as the gasket. I checked a few drivers with the stethoscope (chest piece removed) to listen for air leaks and heard none.

Regards,

John

[richidoo]

Re: Murphy Corner Line Array (MCLA)

Hello John,
Of course, pushing the speakers back into the corners will most likely result in the loss of soundstage depth <snip>
C

John, I have been reading your speaker design articles on your website for a long time, and recently stumbled onto your new MCLA design. Thanks for sharing it. I tried to find it again a couple days later, not so easy. But I got it bookmarked now. A friend of mine brags about his pipedreams all the time, so I want to give your array a try. I am accustomed to in room speakers which have good imaging and depth, with tweeters spaced 5+ feet from the front wall.

Referencing Curt's post above, can you comment on the soundstage depth of the MCLA? The mirror effect that you describe has me thinking that maybe the front wall kind of disappears due to the reflections, and there might be decent depth perception. I will be using these primarily for stereo music listening. Do the images seem to be natural size, or stretched tall? Pipedream and Maggies claim to have a naturally full size (large) presentation.

Thanks
Rich
my 1st post on PE forum

[audiorasp]

Re: Murphy Corner Line Array (MCLA)

Is there any specific reason or advantage for mounting the ND90-8 drivers on the inside of the front baffle?

Geo

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John,

I downloaded your updated PDF file for the enclosure construction reflecting the 1/4" roundover change. Unfortunately, the file shows the roundover on the interior face of the front baffle and the text of the note seems to suggest the same interior face.

I have corrected the enclosure drawing pdf for this error reported by Marvin and will be releasing a revised total project pdf shortly. Thanks for catching this error Marvin.

Here is a link to the revised enclosure pdf:

http://www.trueaudio.com/array/downl...nclosure_2.pdf

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John,

I really like your unobtrusive concept. I will consider it for a future build. I may need to hear them before committing to them, though.

I downloaded your updated PDF file for the enclosure construction reflecting the 1/4" roundover change. Unfortunately, the file shows the roundover on the interior face of the front baffle and the text of the note seems to suggest the same interior face.

4. The Front Baffle is to have 24x 3.0" diameter holes
on 3.5" centers along the center line and equally
spaced from each end. Each 3" hole will have a
radius (.25") on the front edge (narrow side) of the
baffle.

That disagrees with your comment in the quote above..."the radius would be on the outside of the baffle toward the listener"....which I think is where you intend the radius to be. The drawing text is confusing because the "narrow side" of the baffle is the inside due to the angle cuts on the edges.

Perhaps you can answer a couple of questions for my own application:
My listening area has a seven foot ceiling. Would you simply remove 4 drivers from each enclosure?

How would this affect the wiring?

Would your initial EQ settings change with a different number of drivers?

Also, is it critical that no furniture obstruct direct line of sight (or hearing) between all drivers and the listener?

Thanks for sharing your design,
Marvin

Marvin, THANK YOU for spotting this error. The note will be changed to indicate that the radius is on the WIDE side of the baffle, that is, the exterior to the enclosure. I have a redlined version of the pdf (lots of typos...) and will make this correction as I release the next revision of the pdf.

Your custom design for a 7' ceiling would work well with 20 drivers just as you propose. As you say, remove 4 drivers from the enclosure and reduce the total length by 3.5 x 4 = 14". So the total exterior length of the enclosure would drop from 85.5" to 71.5".

You also have an option for using 21 drivers (3 x 7 = 21) but this would result in either 3/7 of 8 Ohms or 7/3 times 8 Ohms net impedance. That is, either 3.4 Ohms or 18.7 Ohms. The higher impedance is impractical but the lower impedance of 3.4 Ohms is certainly usable with many amps. A 21 driver array would extend 75" and would more nearly approximate a complete floor-to-ceiling line array.

For 20 drivers (5x4 = 20) you would use either a 4/5 or a 5/4 wiring array. That is, 4 paralleled strings of 5 in series or 5 paralleled strings of 4 in series. The result would be either 6.4 Ohms or 10 Ohms nominal system impedance. I'd take the 6.4 Ohm case with 5 paralleled strings of 4 in series.

For 21 drivers you could achieve the 3.4 Ohm nominal impedance by using 7 paralleled strings of 3 drivers in series.

Changing the length of the radiating line just this much may have a small affect in the bass range but the published EQ for the 24 driver system should still be a good starting point.

It is certainly preferred that you have a clear line of sight to the arrays for all listeners without major furniture items directly in the listening path.

Regards,

John

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

John,

Over the last week or so, a friend and I have been looking at implementing a 28 driver pair in his livingroom. The room is about 5400 cubic feet. I have the cabinet under control, but I am not clear about how difficult the equalization portion of the project will be. While we could probably use your settings, ultimately it would be necessary to optimize the arrays for the room that they are in.

What additional tools do you recommend in addition to the equalizer, to adequately adjust the response? I noted that you did quite a bit of data taking. Approximately how long does it take and how complicated is it to make the necessary measurements and adjustments?

As a side note, I was curios about other options available for equalization that may not require user measurement. I contacted Chris Kyriakakis at Audyssey to ask if their MultiEQ product could automatically flatten the array output. I thought that for only a little more money than a Ultracurve and a calibrated microphone, an AVR could be used for both array equalization and power. Mr. Kyriakakis felt, from my description of the project, that their MultiEQ product would not provide adequate resolution. However, he was confident that anything running MultiEQ XT would. Unfortunately, the XT product is only available in higher end receivers and their Professional products, placing the use of MultiEQ XT well outside what I think is a reasonable cost for the equalization of a DIY project. I suspect that products from Trinnov would be able to accomplish the same task.

Dana

I use my own measurement software TrueRTA which I sell through my web site at trueaudio.com. But whatever measurement method you use I strongly recommend that you perform a spatial average measurement where you make multiple measurements and then average them to get a response representative of the systems performance in-room.

I usually perform from 10 to 20 measurements in the listening area with small changes in mic position between measurements. Then I average these measurements and finally copy the response to a separate memory and apply a small amount of smoothing, say 1/6th octave. This is the response you want to force to be flat using the 1/3 octave equalizer.

I use separate EQ setting for each speaker and I measure the speakers one at a time. I usually measure the average response for a speaker and then make EQ adjustments at the 1/3 octave EQ. Then I repeat this cycle a few times until the response is as flat as I can get it with the graphic EQ. Then I start introducing bands of parametric EQ to address any remaining peaks or dips. Following each adjustment I repeat the measurement cycle until I'm satisfied with the response.

Next, I carefully record these EQ settings in a spread sheet where I can then apply a house curve (such as the cinema industry standard X-curve) and see the revised EQ needed to achieve the house curve. After saving the flat response to memory in the EQ I then readjust the EQ for each house curve and save the house curves with a descriptive name. It probably takes me about 15 minutes to collect my data, average the responses and smooth the final response. Then I print that average response from TrueRTA and print the spreadsheet with the EQ settings used to achieve that response. This gives me a starting point for the next EQ adjustment cycle.

The three curves I start with are:

1: FLAT

2: SMALL ROOM X-CURVE

3: X-CURVE

In addition to these basic curves I also keep a sequence of "concert" curves in which I progressively scoop out the midrange using one band of parametric EQ. To create one of these concert settings I start with the X-curve EQ and then set a band of the parametric EQ to -2 dB at 1099 Hz at a bandwidth of 5 octaves. I also have saved concert curves with -3, -4, -6, -8 and -10 dB of mid range attenuation. So I guess I have a total of nine responses saved to the EQ memories and available for instant recall.

The flat response is only used for playback of very select material of the highest quality. My everyday response is the x-curve. I occasionally dial up -2 to -3 dB of midrange cut on material that sounds to prominent in the midrange. The higher degrees of mid cut are primarily for playing the system high sound pressure levels where the midrange would otherwise be too loud. Say for those occasions where you want to show off the bass capability of the system or perhaps you are having a loud party and your friends want to dance to loud music with lots of bass. On those occasions it is a shame to be stuck with a playback system that only offers one quasi-flat response. These midrange scoop responses should only be used by exception however. Critical listening or typically is performed in a quiet room using the x-curve response.

I see no reason why the MultiEQ XT would not be useful provided the EQ is sufficiently detailed and that it will average a sufficient number of responses. You also want to be able to switch the x-curve in and out as appropriate. I don't know if it will provide this kind of flexibility.

One reader wrote me to suggest a very inexpensive ($125.00) bare board digital EQ which might appeal to some builders. Here is a link to the MiniDSP:

http://www.minidsp.com/products

A product like the MiniDSP is appealing because has a USB interface which allows you to make all your EQ adjustments at the PC (where you can save and load files) and download then settings to the EQ unit. This would allow for easy distribution of EQ files without everyone having to manually enter all the settings as we have to do with the project's Behringer EQ.

Regards,

John

[marvin]

Re: Murphy Corner Line Array (MCLA)

Geo,

That 0.5" radius was in error. I just uploaded revised enclosure plans changing that to a 0.25" radius. There is only 0.5" between driver cutouts so the .5 radius would remove far too much material between drivers. The radius would be on the outside of the baffle toward the listener. I did not use any gaskets in my prototype. But a gasket would be a nice refinement and would minimize any air leak noises.

Here is a link to the revised enclosure plans:

http://www.trueaudio.com/array/downl...nclosure_2.pdf

Happy Thanksgiving,

John

John,

I really like your unobtrusive concept. I will consider it for a future build. I may need to hear them before committing to them, though.

I downloaded your updated PDF file for the enclosure construction reflecting the 1/4" roundover change. Unfortunately, the file shows the roundover on the interior face of the front baffle and the text of the note seems to suggest the same interior face.

4. The Front Baffle is to have 24x 3.0" diameter holes
on 3.5" centers along the center line and equally
spaced from each end. Each 3" hole will have a
radius (.25") on the front edge (narrow side) of the
baffle.

That disagrees with your comment in the quote above..."the radius would be on the outside of the baffle toward the listener"....which I think is where you intend the radius to be. The drawing text is confusing because the "narrow side" of the baffle is the inside due to the angle cuts on the edges.

Perhaps you can answer a couple of questions for my own application:
My listening area has a seven foot ceiling. Would you simply remove 4 drivers from each enclosure?

How would this affect the wiring?

Would your initial EQ settings change with a different number of drivers?

Also, is it critical that no furniture obstruct direct line of sight (or hearing) between all drivers and the listener?

Thanks for sharing your design,
Marvin

[danakruse]

Re: Murphy Corner Line Array (MCLA)

John,

Over the last week or so, a friend and I have been looking at implementing a 28 driver pair in his livingroom. The room is about 5400 cubic feet. I have the cabinet under control, but I am not clear about how difficult the equalization portion of the project will be. While we could probably use your settings, ultimately it would be necessary to optimize the arrays for the room that they are in.

What additional tools do you recommend in addition to the equalizer, to adequately adjust the response? I noted that you did quite a bit of data taking. Approximately how long does it take and how complicated is it to make the necessary measurements and adjustments?

As a side note, I was curios about other options available for equalization that may not require user measurement. I contacted Chris Kyriakakis at Audyssey to ask if their MultiEQ product could automatically flatten the array output. I thought that for only a little more money than a Ultracurve and a calibrated microphone, an AVR could be used for both array equalization and power. Mr. Kyriakakis felt, from my description of the project, that their MultiEQ product would not provide adequate resolution. However, he was confident that anything running MultiEQ XT would. Unfortunately, the XT product is only available in higher end receivers and their Professional products, placing the use of MultiEQ XT well outside what I think is a reasonable cost for the equalization of a DIY project. I suspect that products from Trinnov would be able to accomplish the same task.

Dana

[johnlmurphy]

Re: Murphy Corner Line Array (MCLA)

Hi John,
Have you tried the MCLA as a free standing P.A. speaker. What would they sound like in a small club? Or a backyard deck/pool party.

I've not tested them out of the corner but I would expect a serious loss of bass output. If they could be properly placed in the front corners of a small club I bet they would sound great. They could be EQ'd to be flat in free space but would not be usable as low as they can be operated in-room. In-room I'm currently operating them with -3 dB frequency of 25 Hz but without the benefit of corner loading they would probably still have strong output down to around 60 Hz. The usable low frequency limit just depends on the required SPL level. A nice benefit of having a highly flexible digital EQ in the system is that you can use different bass cutoff frequencies for different applications in order to employ the speakers optimally for each application.

Regards,

John

[IVI]

Re: Murphy Corner Line Array (MCLA)

Hi John,
Have you tried the MCLA as a free standing P.A. speaker. What would they sound like in a small club? Or a backyard deck/pool party.