[rockadanny]

I went to see a live jazz group (singer, stand up bass, piano, drums) at our little "Arts Center" which used to be a small church - alter area converted to stage, seating converted to theatre seats, small balcony. I've been there for some live performances but this is the first live music I've seen there. Godawmighty this venue is horrid for music! Piano rang like a bell, bass was soft and mushy, drums were over-powering, and voice was intolerably electronic sounding. Blech!! The mics and possibly speakers sucked as well. It was sheer pain.

Anyway, I was wondering if a room (building) can be treated for live music like you would an audio room dedicated for recorded music? Such as corner bass traps (wall/wall; wall/ceiling; wall/floor), reflection points, front and back walls?

p.s. Is this the dumbest question asked on this forum to date?

[bpape]

Not at all a dumb question.

Yes - a room like that can and should be treated.  It's done a bit differently than a home listening room.  Large space acoutics are a bit different overall.  One of the biggest things is that you don't really normally hit the bass all that hard.  Lots of people make great bass absorbtion.   You also tend to use more diffusion, there aren't specific 'reflection points' per se, etc.

Bryan

[jimdgoulding]

I can remember some of my best and worst.  There used to be a club in Houston that was oh so hip.  It was a converted bank with a dome ceiling.  The acoustics in there was so bad the one time I went I got my money back.  Right across the street was a club that had been converted from a hardware store or something and was more irregular inside without all those hard surfaces.  Worked just fine.  Saw Son Seals at Tramps in downtown NY with a horn section.  It was one of those deep and narrow rooms.  I remember thinking somebody done something right here because the sound was notably great.  Round venues suck hopelessly and I think some roadies dial in way too much reverb.  Saw Segovia in a theater in SF from the cheap seats in the balcony and was amazed at how well you could hear him.  Him, his guitar, and a chair were all that occupied the stage.  You could hear a pin drop.  If I were a club owner I would probably have a consultant help out with the room and I would master the console myself.  But not many are as demanding as you and me.

[nathanm]

For every crappy-sounding show I've been to I've left fantasizing about owning my own club designed from the ground up with textbook acoustic design.  Although I doubt the rest of the punters share my snobby high-fidelity views.  Sometimes I think I must have genetically inferior, wimpy ears, because they don't seem to want to put up with the abuse others seem to take for granted.    I need to upgrade to heavy duty ear bones with kevlar coating.

Am I nuts or wouldn't the average movie theater be like 10 times better for hosting live bands?  Usually you can hear people whisper from just about any seat in the place.  Seems like you get some pretty awesome acoustics for your $6-10 movie ticket whereas the $50 big live concert is loaded with noise, feedback and nasty echoes.

[youngho]

I could be wrong, but based on my own personal experience, performance venues are very different from typical listening rooms. I think the first and most important difference is in the dimensions. The much larger venues support much lower frequency modes, with less gain below the lowest fundamental and less noticeable problems with nodes and antinodes because they're spaced so far apart. The second is in the difference between reverberant fields. Generally, with listening rooms, I believe that you'll want much smaller reverberation times so that your room doesn't overwhelm the recorded venue's acoustics. Obviously, with live performance in a particular venue, it's a totally different story. The approach to acoustics will be quite different, too, since listening rooms are generally optimized for a relatively small listening position.

A few examples: The BSO in Symphony Hall. Shoebox shape, but the dimensions are so vast that it's totally different from any listening room. Several groups in the Village Vanguard. Rather triangular shape, like a horn, but providing wonderful acoustic amplification for many different listening positions. A few jazz groups at the Blue Note and Regattabar, an electrified trio in a few different venues. The direct signal overwhelms the reflected signal.

Young-Ho

[bpape]

Absolutely correct.  In large live venues, you WANT a nice reverberant field.  You also want the shape of the room to assist in minimizing direct 'echo' but keep the sound lively.  The reflections are SO late and hopefully somewhat dampened by the crowd and sprayed around by the shapes and diffusion, that when it comes back, it's dwarfed by the direct sound and at a much lower magnituede.

In smaller rooms, in the same amount of time, the sound could have bounced around the room 5, 10, maybe 20 times in the time it will bounce and come back at you only once in a live venue.

Bryan

[Duke]

I remember seeing a two-part article on room acoustics written by Pisha and Billelo for Audio Magazine back in the mid-80's.  They had a couple of graphics depicting intensity vs time for a good concert hall, and for a typical home listening room.

The concert hall graphic showed a fairly long interval between the first-arrival sound and the onset of reflections.   When the reflections arrived, they set up a very powerful, dense, and diffuse reverberant field that lasted for a long time.

In contrast, as expected in the home there was a much shorter time interval before the onset of reflections.  Then the reflections tended to be strong & distinct, as opposed to highly diffuse.  The reverberant field had much less energy than in the concert hall (like maybe 1/10 as much), and it died away much faster. 

Now there are of course different schools of thought as to what role (if any) the room's reverberant energy should play in a home stereo playback system, but I'm of the opinion that a relatively late-arriving, highly diffuse, dense, well-energized reverberant field is desirable. 

Approximating large-room low-frequeny performance in a small room is a challenge, and one that I'm still investigating.   If anybody is interested I could describe some of the issues and possible solutions. 

Duke

[youngho]

Duke, for my own learning, I'd be interested in hearing some of your experience.

For direct-radiating speakers, I think that your post reflects the general approach to acoustics: absorption at the first reflection points on all room boundaries (helps to minimize first reflections, resulting in a relatively late-arriving reverberant field), bass trapping (which helps to reduce, widen, and slightly delay modes), and judicious use of diffusion at most of the other locations (rather than absorption, hence the reverberant field will be relatively diffuse and well-energized). I believe the potential issue with overabsorption of highs (and, to a lesser extent, mids) is why products like TubeTraps or RealTraps include a semi-reflective membrane.

Young-Ho

[bpape]

I'd be very interested in your thoughts on the low frequency aspect Duke.

Bryan

[jimdgoulding]

This is interesting to me, too.  I listen in a 12x15x8 room.  I think that I am getting strong reverberation and I get the sense that it works very well in conjunction with my speaker placement (after a lot of experimentation).  I’ve given some thought to doing my own version of those Shakti diffuser things as, in my case, I’m adverse to damping (except, of course, on my speaker faces for reasons explained elsewhere).   I just don’t seem to be in a hurry cause everything seems so right regardless of what I’m listening to.  How can this be?   One thing that I always think makes me odd man out is that my speakers are electronically time aligned in their crossovers.  They are internally bi-amped Meridian M20’s although I use 20” after market stands for higher placement (those are not my speakers in my avatar).  But I am interested in this discussion even if not for myself, for others.  Maybe for me.  Thanks.

[Duke]

Okay I'm not claiming any of this is gospel, but it's my understanding.

First a bit of psychoacoustics for background purposes.  The ear integrates sound over intervals roughly 1/3 octave wide, such that if there are peaks and dips that average out over a 1/3 octave interval the ear will only hear the average.  If there is a large peak that is not balanced out by a corresponding nearby (within 1/3 octave) dip, or vice versa, then we hear it.  But we still hear the 1/3 octave average, so a very narrow peak or dip usually looks much worse on paper than it actually sounds - unless it's associated with ringing (we won't go into the psychoacoustic implications of ringing at this time). 

Now imagine you're sitting in a chair in the "sweet spot" in a home listening room.   Let's imagine the bass is produced by a single subwoofer.  There will be peaks and dips in the bass region due to how the subwoofer interacts with the room, and how that interaction occurs at the listening position.  Now the problem is not that there are too many of these peaks and dips - the problem is that there are too few!  With the room dimensions comparable to or even smaller than bass wavelengths, the room-interaction peak-and-dip pattern will have large peaks and dips more than 1/3 octave apart.  So, the ear's averaging mechanism can't cope with them - and the result is unnatural-sounding bass.  Peaks are more audible than dips, and impart a heaviness or tubbiness to the bass.  Note that we have such room-interaction peaks and dips at midrange and treble frequencies as well, but because the wavelenghts are short compared to the room's dimensions we get lots of narrow peaks and dips all bunched up so they easily average out over 1/3 octave intervals.

We may say that at midrange and high frequencies, the soundfields are "de-correlated".  Conversely, at low frequencies the soundfield is "correlated" - that is, coherent waves of reverberant energy arrive (and/or interact) throughout the room producing those undesirable peak-and-dip pattern.   In a very large room like a recital hall, the bass wavelengths are small in comparison with the room's dimensions so the peak-and-dip pattern is much more dense, and the bass energy is effectively de-correlated.  So the bass sounds smoother and we can discern pitch better in a good large room. 

Back to our small room with a subwoofer.  If we move either the subwoofer or the listening position, the peak-and-dip pattern changes.  In my opinion, herein lies the opportunity to get smooth bass in a small room - use multiple subs in different locations, so that their inherently somewhat dissimilar peak-and-dip patterns smooth one another out!  If we have two subwoofers spaced somewhat apart, they produce different peak-and-dip patterns at the listening position - not only smoothing the net result, but increasing the density of peaks and dips so that the ear has a better chance to average them out.   Most people who have gone from one sub to two report smoother, more natural-sounding bass, and I believe the beneficial interaction of the two subs' differing peak-and-dip patterns is largely responsible.  In fact, the improvement doesn't stop there; the more independent low frequency sources we have in a room, the smoother the in-room bass will be:

http://www.audioasylum.com/audio/speakers/messages/208728.html

I learned from Earl Geddes the technique of using scattered multiple low frequency sources to smooth the in-room bass.  Researcher Todd Welti published a non-technical paper on the subject, in which he recommends using four symmetrically-placed subwoofers:

http://www.harman.com/wp/pdf/multsubs.pdf  (while aimed more at home theater, the acoustic principles don't change)

Earl ran a simulation that indicates asymmetrical multisub placement gives smoother bass than symmetical placement.  This isn't intended to be a scholarly paper - just an illustration:

http://www.gedlee.com/downloads/sub%20study%20.pdf

What scattered multiple subs do then is more closely mimic the acoustics we'd hope for in a much larger room.  I am presently investigating mixed-polarity scattered multisub arrangements, which should improve de-correlation but I want to see if there's a significant downside.

Room gain is another issue in small rooms, and one I'm still learning about.

Duke

[youngho]

The psychoacoustics you discuss (spectral integration?) sound rather like the section from Everest's Handbook on critical bands (http://books.google.com/books?id=sgwg1Vwm9VUC&pg=PA58&lpg=PA58&dq=acoustics+everest+critical+band+rectangular&source=web&ots=JGsEOw5IOz&sig=Bj8ooaXiML6Ah2MHFw0ozJ3AkEc#PPA58,M1), although these were closer to 1/6th octave bands above 1000 Hz according to the equivalent rectangular bandwidth function. In terms of bass perception, the critical bands are significantly worse than 1/3rd octave bands.

I thought that Toole's research into the inverse relationship of perceptability of peaks/dips and their Q was quite interesting, and he briefly address a few of the psychoacoustic implications (or lack thereof) ringing in a white paper on the Harman website (http://www.harmanaudio.com/all_about_audio/audio_art_science.pdf).

In terms of multiple subwoofers, Linkwitz references the Welti paper (which is really only applicable for rectangular, symmetric rooms with certain dimension ratios, but does include real world measurements, unlike the Geddes one, which did not compare random subwoofer placement with Welti's suggested placement but rather placing the four subwoofers in the corners of the room), as well as one by Greisinger (http://world.std.com/~griesngr/asa05.pdf), which makes for very interesting reading. I believe the top of the line Lexicon approach is to use two subwoofers, one on each side of the listener, with signals 90 degrees out of phase. I would like to point out that the use of parametric filters in Greisinger's paper did significantly improve time response in a real-world room, supporting Toole's contention that modes are essentially a minimum phase phenomena, which is something that has been challenged here in this forum. However, l lack the experience or expertise on this to comment further. Anyway, Linkwitz argues strongly for the use of dipole bass, and it sounds like you're looking into that yourself, but frequencies below the lowest room mode could probably be handled by just two true subwoofers.

Young-Ho

[youngho]

I wanted to add that I'm not trying to be argumentative but rather to use the opportunity for open discussion to add to my exceedingly limited familiarity with acoustics.

I wonder whether you might feel that a reasonable if rather expensive approach in a multichannel listening room would be to set all speakers to small, split the subwoofer signal twice, and hook up four smaller and less expensive subwoofers like the Hsu MBM high-pass filtered to the lowest room mode frequency and randomly located around the room, two of which would be daisy-chained with a low pass filter set to the lowest room mode with these lower frequencies sent to two larger subwoofers located to the sides of the listening position and set so that one is 90 degrees out of phase with the other. This would be a logical extension of Hsu's suggestion that the MBM be placed close to the listening position and a subwoofer low-passed and placed in a corner, which would seem to be a cost-effective solution that would work reasonably well for rooms without a dimension larger than 22 feet or so, but with the additional input of Geddes', Greisinger's, and Linkwitz. Lastly, perhaps parametric filters could help to get the last bit of smoothness in terms of bass response.

My main concern with placing four subwoofers randomly in a room, with one nearer the ceiling, is that real-world considerations often preclude random placement of subwoofers. Also, I don't know how one would go about setting the phase on these. What have you found?

Young-Ho

[satfrat]

This is interesting to me, too.  I listen in a 12x15x8 room.  I think that I am getting strong reverberation and I get the sense that it works very well in conjunction with my speaker placement (after a lot of experimentation).  I’ve given some thought to doing my own version of those Shakti diffuser things as, in my case, I’m adverse to damping (except, of course, on my speaker faces for reasons explained elsewhere).   I just don’t seem to be in a hurry cause everything seems so right regardless of what I’m listening to.  How can this be?   One thing that I always think makes me odd man out is that my speakers are electronically time aligned in their crossovers.  They are internally bi-amped Meridian M20’s although I use 20” after market stands for higher placement (those are not my speakers in my avatar).  But I am interested in this discussion even if not for myself, for others.  Maybe for me.  Thanks.

I'm sure that church room was a whole lot larger,,, or at least had higher ceilings than your smallish room Jim. I could also be that you've just grown accustom to your sound so you don't realize "what could be". I know that's the way it was with me anyways and your rooms are simular. Everytime I added something to the walls or ceilings, the music either got more clearer or the sound stage depth got deeper, the latter being especially true when I installed at diffusion panel on the ceiling between the main speakers at the first reflection point on my 7' ceilings. If a munufacterer has a 30 day MBG Jim,it might be worth the price of return shippping to audition room treatment and hear for yourself what it can or can't do for you. Just looking at how sound waves travel, it's an exceptional room that won't benefit from some form of room treatement. 

Cheers,
Robin

[Duke]

Hi Young-Ho,

Thanks for your in-depth reply, and don't worry I'm not the type to take offense if someone disagrees with me (which I didn't read into your post anyway).  And thanks for giving references with links!

I did over-simplify my description of critical bands - couldn't remember the details, so I probably should have dug them out.   

I don't recall what Floyd Toole said about ringing so I'll have to take a look.  My understanding is that sounds which last longer (like due to ringing) are perceived as being louder, even if they don't measure higher in actual sound pressure level.

The brief Geddes study was not written in direct response to Welti; it was written in response to a discussion Earl was having on another board in which the other person claimed four subs in the corners was ideal.  My recollection, though, is that Welti makes two suggestions regarding placement of subs at the end of his article:  Four subs each in a corners; and four subs each at a wall midpoint. 

I think the concept of spread out multiple subs smoothing the bass is applicable for room geometries and sizes other than what Welti used for his paper, and from my conversations with Earl Geddes I think he'd agree with this. 

Griesinger's approach is certainly interesting.  I use sort of a hybrid Greisinger/Geddes physical arrangement.  It wasn't until today, actually, that I had the parts on hand to try Greisinger's 90-degree phasing - so I'll be giving it a try.

I have lived with dipole speakers for the past thirteen years (Quads, SoundLabs, Gradients, Maggies, and homebrews).  They have lovely pitch definition but lack impact, so I don't think they are the ultimate answer.  James Kates wrote an AES paper on the in-room low-frequency behavior of dipoles, and he found that dipoles give smoother in-room bass than monopoles - similar to what Welti found from using multiple monopoles.   I hypothesize that a single dipole can be modelled as two monopoles, with the polarity of one reversed relative to the other.   

The "random" placement advocated by Geddes and myself isn't truly random - it's mainly an attempt to avoid symmetry in the horizontal plane; the subs are still positioned along the walls.  In practice, it's actually easier to do than a truly symmetrical placement.  Earl achieves vertical scattering by using one or more bandpass subs whose output exits through a tall vertical port that's above the centerline of the room; I leave all my subs on the floor and use a single band of parametric EQ to address the floor-to-ceiling standing wave mode (whose intensity depends on what the floor and ceiling are made of). 

I wasn't aware of the HSU MBM approach - very interesting.   Indeed, an alternative way to minimize the room's acoustic influence on a sub's output is to listen in the sub's nearfield so that the direct sound is significantly louder than the reverberant sound - but psychoacoustically the difference may be insignificant in the deep bass.  I read somewhere that the ear does not even register deep bass before more than one wavelength has passed, so in a home listening room the concept of "nearfield" probably doesn't apply in the deep bass region (unfortunatley I don't remember where I read this, or in what frequency region the transition occurs).  Dr. Hsu must have read a similar paper, because he resorts to more conventional subwoofer placement for reproducing the bass below 50 Hz.  Still, I like his approach - it is innovative and takes psychoacoustics into account.  My reservation about his nearfield MBM placement is that it only works for a small listening area.  Your idea of using four of them spread around the room is of course different from what Dr. Hsu had in mind, but might make more sense depending on one's priorities.

Note that all of these unorthodox subwoofer placement schemes demand a steep-slope low-pass filter, like fourth order; a second-order low-pass filter won't do.  If upper-bass/lower-midrange energy leaks through, the ear will be able to hear the location of the subwoofer(s). 

As for the phasing of our hypotherical quartet of semi-randomly placed subs, my understanding is that Earl leaves them in phase but Griesinger and others advocate different phasing configurations.  Once I have reached some conclusions from my experiments I'll post an update. 

Duke

edit:  I'm getting dangerously close to hijacking the thread here, so when I post about my experiments I'll probably start a thread in Industry News.

[youngho]

Toole's research suggests that low Q resonances are much more audible than high Q ones, despite the ringing associated with high Q resonances. Your understanding reflects the conventional wisdom, but Toole questions this in the linked paper.

Welti seems to suggest the corner placement of four subs as inferior to midwall placement of either two or four subs, and the modelled metrics support this. However, I disagree with you about the room geometries, and let me explain why. The Welti placement of the subwoofers relies on two phenomena: mode cancellation and the fact the subwoofer will not excite a mode when it is located in a node (hence, midwall placement will avoid exciting odd-order modes). This depends on a rectangular room, along with some degree of symmetry in terms of fenestrations and construction, or else the subwoofers would have to be shifted to compensate. In a non-rectangular and/or significantly nonsymmetrical room, I would strongly argue that empirical measurements would be preferred over any rule-of-thumb recommendation. Importantly, the vast majority of the Welti presupposes the use of equalization. I would draw your attention to the investigation #8, where no equalization was used, and you'll see that corner placement of a single unequalized subwoofer was actually smoother than eight configurations involving two or more unequalized subwoofers. In fact, four single unequalized subwoofers in the corners was remarkably similar to one unequalized subwoofer in a corner in terms of the metrics. Geddes' study did not use equalization, so I believe that you cannot relate the results of one study with the other.

Dipole bass will require significantly more power input or amplification because of cancellation, which is the primary disadvantage.

I absolutely agree about the need for higher order filters, like fourth-order, which I actually think of as a virtual prerequisite in bass management.

Because the randomly placed subwoofers will vary in location from the listener, it seems to me that any phase setting other than that which would result in all signals arriving at the subwoofer at the same would be desirable, but I get confused very quickly when phase is involved.

Sorry, have to go to dinner,

Young-Ho

[jimdgoulding]

Sat-  Thanks for the info.  While I'm aware that there may be benefits from what you say, I don't know if I want my stage to be deeper.  The breadth and depth from where I sit (literally) is very impressive.  Instruments are sharp, clearly separated and very nearly sculpted.  It IS a live room, but good live to me as in live event.  You do make it sound tempting.  Maybe one of these days.

[Duke]

Young-Ho,

Thanks for taking the time once again.

I wasn't trying to characterize low-Q vs high-Q resonances, which apparently Toole addresses in his paper.  I was just commenting on the relationship between the duration of a signal and its audibility.  In fact, pitch recognition at low frequencies relies upon the signal having sufficient duration for the ear/brain system to identify the pitch.

Granted Welti's symmetrical configurations are intended to address room modes (which calls for a symmetrical room), but the overall smoothing effect of multiple subs does not depend on room symmetry.  Note that all of the four-sub configurations Welti examined are symmetrical about one or more axes, so he didn't try asymmetrical placement.

If corner placement is ideal, then it should have outperformed the "random" placement in Earl's study.  It clearly did not. 

I have nothing against empirical measurements but for most people that's not a practical method for positioning subwoofers, and the improvement to be gained from investing in precise placement is marginal.  From a post by Earl over on DIYaudio.com: 

"I recommend the use of several subs - as does just about everybody these days since it works so well. Place them around the room - random almost works as well as strategic placement. Putting one off of the floor helps too."

And a related quote:

"What you are likely hearing [when a subwoofer is moved a few inches]... is a different modal excitation pattern, which points out the extreme sensitivity to location when only a single sub is used. Use multiple subs and you can move any one of them quite a bit and there will be no noticable effect on the sound."

Griesinger's recommendation to use two laterally-placed subs and shift the phase between them by 90 degrees is (as I understand it) based primarily on psychoacoustic considerations.  The ear interprets an interaural phase difference in the deep bass as "spaciousness" or "envelopment" - the sense of being in a large acoustic space, and this geometry maximizes the perception of any phase difference between the left and right channel.  I don't recall why he chose a 90 degree phase shift; I'll try to look it up.  I speculate that a phase difference between the subs in the room might further de-correlate the bass energy, but the math to analyze that is way beyond me.

Duke

[youngho]

Hi Duke,

I brought up Toole's research because you had mentioned narrow peaks, audibility, and ringing. I just thought it was interesting that shallower but broader peaks are more audible at lower amplitudes than narrower peaks, despite the concerns about ringing.

I wasn't trying to argue that corner placement is ideal. I wanted to point out that corner placement of four unequalized subs measured just as badly in terms of metrics as one unequalized sub in the corner. The use of equalization significantly changed the results in terms of metrics. Therefore, Geddes' study cannot be related to Welti's study. Also, I agree with you that Welti's study cannot be related to Geddes' study. However, Welti's study does demonstrate that a second subwoofer can often make bass performance worse. For example, I've seen placement #5 recommended before, where two unequalized subwoofers are placed along the front wall, each about 25% of the distance from the nearest sidewall. This is worse than a single unequalized subwoofer in the corner in terms of the metrics. For a person with a single unequalized subwoofer, corner placement may perform surprisingly well, but this will depend on the room dimensions.

Maybe it's just me, but I think a little bit of measurement is reasonable to suggest while recommending they buy a second or third or fourth subwoofer, particularly for the sorts of folks who are inclined to buy a second, third, or fourth sub. I'm not talking about Joe Sixpack. Often, real-world considerations will relatively limit the number of positions where a subwoofer can be placed. Many people, for example, like to position a subwoofer relatively close to one of the main channel speakers. Also, Welti's study demonstrates that a second subwoofer may make bass performance worse.

If I understand it correctly, the phase difference contributes both to psychoacoustic considerations and relatively less excitement of room modes. It's probably the compromise between mode cancellation (0 degree difference) and mode reinforcement (180 degree difference).

Young-Ho

[bpape]

I wanted to point out that corner placement of four unequalized subs measured just as badly in terms of metrics as one unequalized sub in the corner.

Young-Ho

Interesting since other research points to the 4, 1 in each corner as a good layout in terms of frequency response - at multiple seating locations.  I've not gone into all the research discussed - but are they all looking at single seat responses or multliple seat response?

Bryan