Transcript
Pro Sound Web Live Chat With Jim Brown

March 27, 2001

(Editor’s Note: Jim has provided some additional information in this transcript not included during the chat session. This copy is indicated in italic type.)

Moderator: Welcome, everyone, to tonight’s live chat session with Jim Brown. Jim, could you kick things off by giving us a brief summary of your professional background?

Jim Brown: I started out in ham radio and amateur theater in Huntington, WV in the mid ‘50s. After high school, I got an EE at the University of Cincinnati, finishing in 1964. My college was coop, and I worked in several radio and TV stations, for a ham radio manufacturer and for a broadcast consultant.

Jim (continued): After that I worked briefly for Motorola (which brought me to Chicago) and then taught at DeVry for five years. Then some time in the civil rights movement, and from there to sound and video contracting. During that time period (the ‘70s), I started working part time in jazz clubs doing reinforcement and learning to mix. And as part of that, I did a lot of recording of what I was mixing, then brought it home to study.

Jim (continued): Next I moved into selling pro audio gear to broadcasters and studio folks, and did that for about eight years. From there I was recruited by a consulting firm (Kirkegaard) to help with the O'Hare United Terminal. I've worked full time as a consultant ever since (about 1985).

Jim Wilson: What DeVry Campus?

Jim Brown: The original one on Belmont Avenue near Pulaski in Chicago.

Charlie Hughes: Hi Jim. I know you do a lot of stereo & LCR (left-center-right) cluster designs. What are some of the things one needs to watch out for when designing this type of a system? How much of the room can you expect to receive reasonable stereo coverage?

Jim: Hi, Charlie. Great question. The biggest thing is keeping a handle on arrive times from the various clusters, and keeping them within a range that the ear doesn't have trouble with. I aim for 20 milliseconds (ms) between adjacent clusters, but will sometimes let it go to 25 or 28 ms in a very big room.

Jim (continued): To do that, you have to have pretty good ceiling height, and not too much width. And you have to spend a lot of time with your EASE model studying it. The other thing to watch for is relative levels between the clusters. I try to hit the far side of the room about 6 dB hotter than the near side. This helps to compensate a bit for precedence. So if you're on the far left side of the audience, you'll hear the right cluster about 6 dB hotter than the left one.

[The next day, realizing I didn't answer part of Charlie's question: In most rooms where I do stereo, I'll get good stereo over at least 80-90% of the seats, and decent mono in the rest.]

Tom Young: Hello Jim and Charlie. So how does this (arrival times from clusters) relate to maximum width that the LCR (left-center-right) can be from one another?

Jim: Hi Tom. That depends a lot on cluster height, as well as how wide the room is. Luckily, EASE will do all the number crunching for us, but we have to know to ask! In a small hall, I start with L/R clusters about 35 ft. apart. My intent is not to create an image that is a lot larger than life, but only to keep the life in it that would otherwise be lost due to comb filtering from summed mics to a mono cluster.

Jim (continued): I may end up wider than that 35 ft, depending on what the geometry of the room is and what else has to happen around the proscenium. And I try to minimize the size of the front clusters, doing as much as possible with delays.

Charlie Hughes: So the maximum time difference will define the maximum cluster spacing? Anything else to consider when locating the clusters?

Jim: Right. Other things -- well, with most systems, we aren't alone. There are other disciplines trying to do their part to make a show happen (like lighting, a theatrical set), or architecture to make a venue work if it's a permanent install. And the configuration we need to make a cluster work to cover a room will sometimes dictate where it needs to go.

Jim (continued): For example, if a room is wider in the front, that tends to pull the clusters wider. I try to use the minimum number of components in a system to minimize the interaction between them. Coming wider will sometimes allow me to cover more with less. Also, it helps in a wide room to go to three clusters if there is money for it. That helps the time difference issue, since we're generally panning left to center, and center to right.

Gary Whisman: Hi Jim: The room is 48 ft. wide and 16ft. high at the sides, 22 ft. at the center. Will LCR work? And where would you put the subs?

Jim: Hi Gary. That room sounds pretty similar to one I'm currently having installed near Baltimore. I designed it for L/R stereo. It's a contemporary congregation, but they're not real loud. Ralph Heinz is building some custom boxes for the project. I think it will work fine. In that room, the sub is going at the peak of the roof at centerline, but spaced away from the peak to create a bass array working against the roof deck. That system is L/R. The loudspeakers are about two-thirds of the way from centerline and they are flanking the platform. That's all there is in that church - it's not very big.

Tom Young: What is the primary purpose of the center cluster in LCR ?

Jim: Several things. First, to create a strong center image. Second, to provide a good voice channel. But my center channels are full range, full power elements - they are not intended to be only a wimpy speech or vocal channel. I see an LCR system as an integrated thing, not mono surrounded by stereo.

Charlie Hughes: So for a LCR design, is the center truly center only? Or do you put some left source in the center cluster covering the right, and right source in the center cluster covering the left, to get more imaging at the side of the room?

Jim: In my view of the world (grin), a center channel (I'm looking for the right phrase here) should be a co-equal player with the other two channels. If I'm mixing for it, I would typically try to mic the choir with three mics and give one to each cluster exclusively. If that was too wide, I'd let my ears tell me to pan the side mics a bit to the center.

Jim (continued): But that's a mixing decision. I see the pan pot as panning from L to C with no R, and from C to R with no L. When it is in the center, I want nothing in left and right. And I don't see that as a mixing decision, and I think you can guess that I don't like the Allen and Heath LCR pan arrangement. But I do like the implementation that most others do.

Mike M: What is the maximum delay between left and right that you would tolerate in an LCR system?

Jim: Hi Mike. If you have an LCR system, the delay between left and right only comes into play when you are playing stereo tracks. And for me, that's a secondary function of most reinforcement systems. So the delay I'm most concerned with is between L and C, and between C and R.

Jim (continued): Of course, it is never a good idea to try to go a lot wider than life, so I would never spread just for the sake of spreading (except on a theatrical production for effect). If we can keep the arrivals within a reasonable range, we give the mixer and other creative people more options for how they mic and mix a show. So the best answer I can give you is the 20-25 ms for L to C, and from C to R. With L/R, we're pretty much stuck with 25 ms as the max. And understand, these max times generally only get bad at the front corners in typical rooms.

Mike H: So you delay the L and R in relation to C?

Jim: No. The times I'm talking about here are the arrival time differences at any listener's ears from the clusters. In other words, if we take TEF or Smaart to a seat and do an ETC, what's the difference between the arrival times from L, C, and R? Those times can easily be computed in EASE, and we can either plot them as a color map, or look at them in greater detail with the Probe function.

Mike H: So then you are talking about the delay from physical spacing, not inserting a delay?

Jim: Exactly.. If you think about it, these times are all the result of doing three-dimensional trig for the path of sound from the various clusters. That's part of what EASE is doing for us.

Mike M: Is LCR feasible for a semicircular audience and thrust stage, and how do you deal with the geometry of it?

Jim: Boy, that's a tough one. The nice thing about EASE is that's easy to do a quick model of the seating and try some loudspeakers in it to see what the trig works out to. It's also important to realize that it's not at all important to be purist about any part of what we do. Every room is different. Every use is different. It's important that we tackle every project with a fresh piece of paper and as many fresh ideas as can. For example, a popular technique in the sort of geometry you describe is some variation of alternating L and R.

stephen: How do you achieve the difference in levels, i.e. far side hotter? Basically by cross-firing the loudspeakers?

Jim: Hi Stephen. Yes, the level variations are achieved mostly with loudspeaker directivity, and also by arraying things. And, of course, by cross firing a main pair. Vance Breshears has developed a nice technique of adding an element to one cluster to work on delay for another cluster. That helps him get sound across the room. This is an interesting variation on what I saw Jim Gundlach and Pete Tappan do in the ‘80s at Ordway. But I think Vance's implementation is slicker.

Jim (continued): Getting back to Mike Miles' question, I think it's important to realize that we don't necessarily have to achieve directional realism to benefit from stereo. The principal benefit for me is that we minimize the comb filtering that occurs when we sum two mics together electrically. They sound a lot better when we sum them acoustically in spaced loudspeakers.

Jim (continued): Two reasons why this is true. First, our head is in the way of our ears, and causes our left ear not to hear stuff coming from the right as well as what comes from the left. Second, the acoustic travel path is sufficiently unstable that it helps de-correlate the two mics that are being summed, which also reduces the combing.

Tom Young: On Vance's technique: So the odd element in an opposite cluster must be level-set as well as delayed to not interfere with the primary sound from that cluster, right?

Jim: Sort of. The thing I'm talking specifically about here is putting an element in the center cluster to send left over to the right side, and vice versa. You don't want too much delay coming across the room there, but you do want some. I haven't tuned a system yet that uses this (I've got a couple designed, but they haven't been built yet).

Jim (continued): But what I expect to happen is the instability of the acoustic path to de-correlate the summing. If you think about it, this is pretty much like what happens with the delayed loudspeakers we've been using for years. I have yet to hear any combing between them (that is, between a delayed loudspeaker and the loudspeaker which is providing the precedence for the delayed loudspeaker).

Charlie Hughes: This is what I was asking about earlier in putting left source in the center cluster to cover the right and vice versa. Does this seem to work well in increasing the stereo coverage area in the room?

Jim: Well, as I say, I've only done it so far inside the computer, but Vance has done it in hardware. It looks good to me in the computer, and my brain likes it.

stephen: The extra element that you add only has to primarily cover the front of the room where the time difference is greatest, right?

Jim: Yes, but mainly it's a way to get the left channel to the far right front corner, and vice versa, without making the front center too hot.

Jim (continued) Something I also do a lot is delayed mono to fill in a stereo system. This works when you don't try to add in too much mono.): For example, let's say the front clusters are just starting to fall off by about 3-6 dB at the back of the house. I'll sum left and right and feed it to a mono delay ring (perhaps under a balcony). And I'll bring them in at a level equal to the mains. Now, the listener in those seats hears the mains and the delays at roughly equal levels, so the combing from the L/R sum never gets real deep.

Jim (continued): Also, the front clusters are still providing precedence, so the folks in those seats still hear stereo. I've made this work well in a lot of places.

Tom Young: Back to your statement about source-position realism and also the issue of multichannel versus mono loudspeaker systems: I find that sound operators do not (should not) go crazy with panning and, in fact, the majority of inputs are panned center. But by carefully choosing what does get panned, you can achieve all that is necessary for a "spacial" soundfield. Comment ?

Jim: Yes, Tom, I agree for the most part. It depends on the sources and the program, of course, but the key to any mixing is not overdoing it. I think we all agree that moderation is the name of the game in getting good sound, and an overly wide image is a great example of overdoing it. I really see the systems that we design as a creative tool that mixers will need to learn to use.

Jim (continued): And a trip back through all of the old books on mic technique, and stereo technique, and a lot of trying different things, and listening to those things, is what is going to make stuff sound good. We all got to where we are mixing to mono, and stereo is simply another tool in our kit.

Tom Young: Well put, sir.

stephen: What about underbalcony "stereo" systems where you alternate every other speaker? Is that worth the effort to help "open up" the sound?

Jim: It depends a lot on how deep the underbalcony is. So far, I haven't done one of them that way, but I certainly might, depending on the venue. If, for example, I had to go a long way under the balcony so that the listener under there was dependent on those delays with little support from the clusters, yes, I would be more inclined to do so.

Jim (continued) But that takes a lot of loudspeakers, because you have to crossfire them in pairs, so you need more loudspeakers, more power amps, more wiring, more DSP, etc. I designed one church with full LCR in a balcony, but they were so thin on money they chopped the building itself to pieces, so that design went into the circular file (with the building).

Jim (continued): In that church, the balcony was quite isolated from the rest of the church by a follow-spot room and a catwalk, so I needed to do something drastic to make them feel a part of the worship. So for that, I put in three LCR groups (a total of nine loudspeakers) covering the three sections across the balcony.

JT: Which LCR capable consoles do you like?

Jim: Good question. I liked the old DDA Q-II, which was based in large part on an AES paper I did in the late 80's. That board is now discontinued. I saw a very nice console in the Crest booth at NSCA that offers a very nice LCR implementation. It wasn't clear to me if it offers LCR monitoring of the VCA groups the way I need to hear it - Chuck wasn't there at the time - but it seems to be a very nice board. Don't recall model numbers, but it's the VCA board in the $30K price range. I don't like any LCR console that doesn't offer the operator the opportunity to listen to VCA groups or physical groups in place.

Kevin Potts: Have you had a chance to take a look at the Innova-Son consoles and do you see them as a good digital console for the church market?

Jim: Hi, Kevin. No, I haven't seen them, nor do I know anything about them.

stephen: You mentioned positioning a sub down from the ceiling to create a bass array with a null point below it. That's a pretty nifty idea! Tell us more.

Jim: That's a technique that I learned from Craig Janssen, who shared it with us at a Syn-Aud-Con workshop about seven years ago. I've used some variation of it in nearly every system I've done since.

Jim (continued): You probably know about the concept of a pair of spaced woofers creating a null directly beneath them at a frequency where the spacing is one half wavelength. At the same time, those loudspeakers will sum coherently on axis (i.e., directly in front of them), so even though the null going down is rather narrow, the 6 dB gain is pretty broad.

Jim (continued): Now, consider a single woofer spaced from a hard wall. The hard wall will reflect the sound from the woofer, and the result will be a mathematically created image of the real loudspeaker on the other side of the wall. The real and image loudspeaker will have the same directivity as two real ones at the same spacing (taking into account that the image is just as far behind the wall as the real one is in front of it).

Jim (continued): So with this sub spaced from the ceiling, there will be some nulling going down, and some gain going forward. How much? Not a lot, but at least 6 dB. And 6 dB of directivity at low frequencies is awfully tough to come by. So the result should tighten up the bass a bit. I've never done this at the peak of a roof before, but I have done it with a relatively flat ceiling and it works quite well. If you get the spacing right…

Kevin Potts: Jim, I see designs of sound systems in churches with the subs on the floor and with subs flown. Could you just give a quick study of the plusses and minuses of the different designs?

Jim: Good question, Kevin. I generally prefer subs on the ceiling if we can find suitable places to put them. Several reasons. I've seen some pretty nasty cancellation happen with subs on the floor crammed in where they will fit instead of where they belong.

Jim (continued): Second, when you put them on the ceiling (and against a good boundary), you get all the benefits of the boundary that you would have gotten on the floor, but you also get much more even distribution across the audience. There is also, of course, the issue of isolating them from the stage.

stephen: Getting back to the previous point: Over what kind of frequency range will it work?

Jim: You mean the bass array idea? Over roughly an octave and a half for any given pair of loudspeakers. That doesn't mean that you can't run them over a wider frequency range, but above that range they will get “lobey”, and below that range they will get wide (just like any other woofer). But in the concept that Craig developed, he used a closely spaced pair for an octave and a half, then a more widely spaced pair for the frequency range below the first pair.

Jim (continued): And here’s another thought. If you space a pair of woofer pretty wide in comparison to their size, you will see a very pronounced null at the half wave frequency. To observe that, hang a spaced pair above a stage, and sweep them with an oscillator. You'll hear the null quite clearly.

Jim (continued): Now, do the same with a closely spaced pair (i.e., nearly touching). With this pair, you probably won't ever find the null. Why? Because the woofers are large in comparison to their spacing, so they aren't acting enough like point sources. That doesn't mean the array isn't helping you. It is.

Jim (continued): How? Well, you still have the coherent summation coming forward (i.e., +6 dB), and the null coming down will be much broader but not very deep (perhaps -3 dB) so what you get is 9 dB between forward and down! Now, do that twice (i.e., in both the horizontal and vertical planes) and you have 12 dB coming forward! Ralph Heinz has built several sets of special loudspeakers for me that way, and they work great!

Kevin Potts: I would guess that my biggest battle is with stage volume. I have seen some unique ideas from Craig at Acoustic Dimensions using in-ear monitor etc., but what are you using to monitor large groups such as church orchestra and large church bands? Ever used stuff from Listen Tech as affordable in-ear monitoring?

Jim: Craig is generally blessed with much larger projects (and budgets) than I am, so I haven't used any of that stuff yet. I certainly do like to advise my clients to go as far down the road to in-ear monitors as their budget permits. Shure has some nice stuff that fits a good price point with a combination of wired and wireless components. I suspect the Listen Tech units would work well too.

Jim (continued): Also, I have adopted in-ear monitors from Etymotic Research for all of my own mixing and live work. They really help a lot with leakage, and I feel pretty confident monitoring with them now. Those earphones, in combination with a good in-ear system (i.e., the right mixing for it, including room mics, etc.) are undoubtedly the right way to go for monitoring.

Moderator: Can you give us your views on the importance of modeling, such as EASE?

Jim: Yes, I think it is a very important part of what we need to do for a client. It allows us to design better systems, to avoid mistakes in systems, to try various wild and wonderful ideas before we build them, to help installations or load-ins go a lot quicker, to get systems tuned more quickly.

Jim (continued): For example, I can generally do a rough model of most rooms in EASE in half a day, and the simple ones in a lot less time. Then I can try various designs and turn them into a budget that means something, i.e., I know I can build it and it will work.

Jim (continued): Then, I can work through the details, and give the contractor all the delay times for a complicated system, and they will be right. So all we have to do is get the system in, get the levels right, get stuff aimed right, and get it tuned. The only delays we have to adjust are those for cluster alignments (and some tweaky listening for precedence). This saves everyone a lot of time.

Jim (continued): I rarely do extremely detailed acoustic models - that's what the acoustics guy is paid to do. I concentrate on loudspeaker coverage, imaging, keeping sound off surfaces where it's going to bounce somewhere and cause a problem. And I don't need a detailed model to do that.

Dave D: How closely does measured data match calculated data?

Jim: Good question, Dave. I find EASE to be quite accurate at what I ask it to predict, which is direct sound coverage and arrival times. I almost never build detailed rooms, so I don't usually ask EASE to keep track of reverberation or intelligibility. So I have no experience with how well it does that. But it is right on with direct sound and delays (as long as I'm careful to feed it good loudspeaker data).

Charlie Hughes: How much to temperature changes and gradients in rooms typically affect the system alignment?

Jim: A lot more than we would like to think, I think! Dennis Bohn posted or wrote something to this effect several years ago, and I suspect he is right. I have never had the opportunity to verify it, but his math looked right. One thing I have experienced with humidity (I think it was humidity that was responsible) is that a system will sound less bright during the winter here in Chicago, as it gets dry. I've had critical listeners identify it. One was the chief engineer at WFMT, for whose church I had designed a system.

Charlie Hughes: Yes. Low humidity causes high frequencies to roll off more quickly.

Jim: Yes.

Charlie Hughes: Where can Dennis' article be found?

Jim: I don't recall where he published it. Probably call him - I just got a flyer from them talking about them celebrating their 20th year, so they may still be celebrating. Funny - in one photo, Dennis looks just like I did in those days!

Moderator: Jim, we'll be wrapping it up in a few minutes. Can you tell us what you look for first in loudspeakers?

Jim: The first thing I look for is decent polar response, i.e., it should sound the same in every seat (or at least enough seats to make it worth using). A system that only works in a sweet spot at the console isn't worth much.

Jim (continued): I use polar measurements as a weeding out process, because so many loudspeakers fail it, and because I can do it quickly. Listening takes a lot longer (days, sometimes, and a lot of work for big stuff), so I only want to spend my time on those that have some chance of being useful (i.e., if I make them sound good on axis, they will also sound good most other places).

Jim (continued): And, of course, I want loudspeakers that solve problems - some that array well, some that fit into tough places, some that have very narrow patterns, some with very wide coverage, etc. Not every loudspeaker has to go from DC to daylight - we need underbalcony fills, paging in dressing rooms, etc. And there's the usual stuff about power handling, maintainability, riggability, etc.

Jim (continued): And I won't waste my time with a product for which I don't have EASE data, because I can't design with it.

Kevin Potts: Do you look at the physical construction of the loudspeaker box? I have seen some interesting variations from the major manufacturers.

Jim: That part of things isn't my strong point, so I count on the observations of others in the industry who know a lot more about it. But I do concentrate on manufacturers who I'm sure know how build boxes that are well built and safe to rig. And those who have paid their engineering dues about understanding how the cabinet resonances and shaping affects both directivity and performance. JBL has done some very effective presentations on this as part of their engineering workshops for their contractors.

Alan H: What about the 'political' end of things as far as the church understanding that what they need might not be a simple as their initial budget? Do you run into trouble meeting their budget expectations or can you generally convince them what is in their best interests, long term?

Jim: Great question, Alan. Bringing budget in line with expectations is one of the toughest parts of our job, whether we're a consultant or a contractor. One of the most important things we have to do is find out what is REALLY important to any given church, then get them to spend enough to accomplish that well enough that their most important needs are met. And so that they can come back later with more money to expand the initial system rather than having to start from scratch.

Jim (continued): Another VERY important part of that is getting them to put enough money into the building (acoustics, mostly, but also power, conduit, grounding) so that they don't have to rebuild it to have a good building. On more than one occasion I've told a client to leave out the sound system, but get the acoustics right.

Jim (continued): They can buy a new sound system later, but it's awfully hard to fix lousy acoustics. And they can put a couple of cheap boxes up on poles and have church (although it may not be grand) if the acoustics are cool. But if the acoustics are lousy, neither the best sound system in the world or their praise band is going to sound good.

Moderator: ProSoundWeb will be featuring a new article, regarding a recent project at Remembrance Church in Grand Rapids, MI, on just this very topic. Look for it tomorrow on PSW Install Sound.

Jim: Great! Thanks to all for staying with us!

Kevin Potts: Thanks Jim, I look forward to reading the entire text as we West Coast guys were still at work when the Chat started.

Charlie Hughes: Thanks for your time this evening, Jim. It was very informative and enjoyable.

chat.boy: On behalf of the entire ProSoundWeb team, I'd like to thank Jim for sharing his time and knowledge.

Jim: All of your questions made it a lot of fun, and brought out some interesting thoughts.

Jim [The next morning, looking at the transcript]: In the heat of the moment, a few important points managed to escape me that I'd like to add as a postscript.

First, in deciding whether or not to do stereo in any venue, one must do what is important for ANY loudspeaker system -- make sure that the required loudspeaker locations won't set up undesirable reflections or late arrivals somewhere else. In other words, where the sound goes AFTER it reaches your listener is just as important as getting it to your listener in good shape. If it hits a rear or side wall and comes strongly late somewhere in the audience, you have a problem. So the system must fit the acoustics (or the acoustics must adapt to the reinforcement).

Second, there are two popular misconceptions about stereo reinforcement systems. One is that they cost a lot more than mono. That isn't always true -- in fact, I'm often able to do stereo for only a small cost premium. Three reasons. One is that moving a loudspeaker (or cluster) off center often means that it takes fewer boxes to get the coverage angle.

Another is that part of the cost of any system is having enough horsepower up there to meet SPL requirements. When we do stereo, that horsepower is still there, but simply split into two or three more useful pieces. We may not need as much horsepower because our system is able to provide us better clarity than a mono system of comparable size. And, of course, DSP makes all the electronics a lot cheaper.

The other misconception is that L/R systems have to have poor intelligibility. While it's certainly possible to get in trouble with intelligibility without a center channel, it isn't a given, and whether it will be a problem depends on the venue. Again, EASE will tell us this, both with a study of arrival times, echoes, and direct to reverberant ratios. If it's a good room and not too wide, L/R is likely to do just fine for speech. I've done several systems where I designed LCR but had to fall back to L/R for budget reasons. They've all worked fine.

Moderator: Be sure to join us next week for Church Talk, same time, same channel, featuring Vance Breshears. Tuesday, April 3, 9:30 pm Eastern time. Transcript of this chat session will be posted on PSW tomorrow.

chat.boy: Also, join us Monday April 2, 9:30 Eastern for Charlie Hughes; Monday, April 9 for Robert Scovill, (same time); andJames Stoffo, Wednesday, April 11th, (same time).

chat.boy: Good night all. Join us in the Live Sound Chat room for "after chat".