Early Reflections in Home Theaters: A Different Perspective
No, I am
not sitting in my home theater at 6:00 a.m., ruminating about how the pattern
on the wall from the lighting sconce looks different from the back row. This article is about sound reflections. Specifically, early sound reflections in home theaters. I have been wondering whether there are better
objective measures to help with addressing early reflections; specifically to
help determine (a) whether they are problems and (b) what to do about them if
they are. To begin to address these
issues, it might be beneficial to examine the state-of-the-science. Floyd Toole pointed out in a recent JAES
article that the acoustical design of small spaces has its origins “in investigations
of performance spaces—concert halls and auditoriums” .
Toole goes on to question this logic and, in my opinion, his skepticism is well-founded. How many home theaters sound like concert halls? How many concert halls sound like home theaters? Is either one capable of sounding like the other? Is that even a desirable design basis? The answers to these (hopefully) rhetorical questions would seem to indicate that any metric used for a large space would, at the least, need to be modified for use as a small room metric. I have argued for a number of years—along with others—that the large room metrics we’re accustomed to should not be used in small room design.
Borrowing Richard Rives Bird’s love of home theater/kitchen comparisons , consider the following: When baking a cake, it not be prudent to measure out the amount of salt in fractions of a gallon. Nor would it make sense to calculate out fuel economy in miles per teaspoon. These may seem like exaggerations, but consider the volumetric ratio of a concert hall and a home theater can typically be on the order of 700:1. Know how many teaspoons are in a gallon? Answer: 768. Shall we choose to equip our kitchens with measuring spoons or 768-teaspoon jugs? Similarly, shall we design our home theaters using RT, or shall we consider some other metrics that more accurately ascertain the relevant time domain characteristics of a significantly smaller space?
The answers to these (hopefully) rhetorical questions would seem to indicate that any metric used for a large space would, at the least, need to be modified for use as a small room metric.
What about Reverberation Time?
Reverberation time (RT) has a long and glorious history in room acoustics. Starting with Sabine  and working through to more recent works like those of Beranek , we have volumes (excuse the pun) of empirical data showing what constitutes “good” and “bad” for concert hall and auditorium acoustical design, with many presentations of RT data. Armed with this sort of evidence and a good scientific brain, one could conceivably design a fantastic concert hall using nothing more than RT as a measuring stick. Considering organ music? Target the RT in the 3.0 to 5.0 second range for the midband frequencies and the room is probably well on its way to aural ambrosia.
But home theaters don’t benefit from this approach. Using only a target midband RT of 0.30 to 0.50 seconds for a home theater does not guarantee a good sounding room.
Editorial Note on Volume / Time Relationship
Some may point out an apparent discrepancy between the 700:1 volumetric ratio and the 10:1 target RT ratio. The relationship between volume and time in acoustics is equivalent to the relationship between volume and distance. Knowing this, and that the cube-root of 700 is roughly 9, the discrepancy should be resolved.
Once we start looking at the decay of sound to 1/100ths of a second, our design parameters begin to border on the ridiculous. If the measured RT is within 10% of the target RT in a home theater design—a typical RT design target for concert halls—the result carries far less weight than the RT resulting from a similar design target in a concert hall. In other words, if the target RT is 3.0 seconds in a concert hall and the result is 2.7 seconds, the room should sound more or less as intended. If the result is 4.0 seconds (a 33% overshoot), then the room probably does not sound as intended. Conversely, if the target RT for a home theater is 0.30 seconds and the result is 0.27 seconds, who cares? If the result is 0.40 seconds (same 33% overshoot), again, who cares? Regardless of the result, the room could sound good…or bad!
To examine some alternatives to RT, besides the metrics and methods covered quite thoroughly in Toole’s journal article , I delved into the work of Yoichi Ando. One of the parameters he has established as a metric for concert hall design evaluation is called the “effective duration of the envelope of the normalized auto-correlation function” . No, I am not making this up. For the purposes of this discussion, I will abbreviate this to τe, which is the same abbreviation Ando uses. Despite the verbose description, I believe τe to be a useful metric. It is basically a measure of the repetitiveness or the reverberation contained in a source signal. Its relationship to preferable room acoustics qualities and quantities is derived from subjective tests. The τe of the source material can be used in designing spaces that are intended to reproduce that source material. According to Ando, speech signals generally carry a τe value of about 10 milliseconds (ms). What this ultimately means in terms of the science is beyond the purview of this article. For comparison, slow classical music pieces tend to have a longer τe, generally in the range of 100-200 ms.
Measurements – Subjectivists and Objectivists Unite
Armed with this information, I went about measuring the τe of some movie dialogue just to get a sense of whether Ando’s estimate for the τe of speech was accurate for my intended application. Using some software packages available from Yoshimasa Electronic Inc. called “Realtime Analyzer RAD” and “Sound Analyzer RAD” , I was able to process an excerpt of center-channel dialogue from the Sam Gamgee character (played by Sean Astin) in the second LOTR movie . I measured a τe between approximately 4.7 and 12.2 ms.
The next step was to dig deeper into what Ando had to say about room design based on τe. Ando’s specialty is large spaces. However, much of Ando’s research is based around correlations between objective metrics and (subjective) listener preferences. This is in stark contrast to many other architectural acoustics teachings that revolve around room function. Ando’s subjective research has led to methods that are used to match listener profiles with concert hall seats based on some simple screening tests. Since his studies are heavily weighted towards the subjective, I wondered whether his metrics would be more applicable to smaller spaces than the traditional objective metrics like RT. The first thing I investigated was something that actually involved RT. (Drat.) One of the basic “rules-of-thumb” Ando developed was that the RT should be roughly 23 times greater than τe. From my measurement of the LOTR dialog, this means Sam Gamgee would likely sound the best in a room with an RT between 0.11 and 0.28 seconds. If I were to go against my own philosophy and actually use RT to design a home theater room, this range of RT values for the midband frequencies would probably be on the low side, but not unreasonable.
To Treat or Not to Treat – First Reflections
However, I would prefer not to rely on a large room metric. Exploring Ando a bit more, he spends a lot of time on the “preferred delay of a single reflection,” or, more practically, the delay associated with the first reflection arriving at a listener’s ears. This would be synonymous with the initial-time-delay gap, or ITDG, from Beranek . To abbreviate, I will again borrow from Ando and use Δt1 for the delay of the first reflection. When applied to speech signals, Ando found empirically that there is beautifully simple relationship between the τe of the source material, the preferred Δt1, and the relative pressure amplitude of the first reflection. Once again using the results of my dialogue analysis above, the result is that the preferred Δt1 for the LOTR excerpt averages out to be roughly 10 ms, but depends on how much treatment there is on the surface providing the reflection. The less treatment, the lower the preferred Δt1 will be. The more treatment, the higher it will be. For small rooms, this translates to how far the “best” seats are from the nearest surfaces, most notably the side walls. A good designer should be able to apply some basic math and gauge easily whether the listener will prefer treated or untreated side walls.
To illustrate this using my LOTR dialogue analysis, the ideal Δt1 for untreated side walls (say, gypsum wallboard) could be approximately 4.8-12.5 ms [eqn. (4.1) from ref. 5]. Similarly, if absorbers are to be considered for the side walls, the ideal Δt1 could be in the 6.1-15.9 ms range. Recalculating again for diffusive walls could provide an optimum Δt1 of 5.3-13.7 ms. These ranges can be compared to the geometry of a room to figure out what the most appropriate early reflection treatments for the side walls might be. For example, if the average listening distance (DL, from the center channel) is 10 feet and the width of the room (W) is 15 feet, Δt1 works out to be roughly 7.1 ms. Based on our calculations above, any solution—reflection, absorption, or diffusion—could result in a subjectively pleasing result at the listening positions. For a smaller room where DL is 5 feet and W is 10 feet, Δt1 works out to be approximately 5.5 ms. In this case, reflection or diffusion might be the preferred treatment options, but absorption may not be desirable. Finally, changing DL again to 13 feet and W to 25 feet will give us a Δt1 of about 13.4 ms. Hence, absorption and/or diffusion, but not reflection, might be the best treatment to consider. The above is summarized in the Figure. A simplified delay calculation is presented in the upper portion of the Figure. Once the Δt1 is calculated, it can be compared to the line graph on the lower part of the Figure. Possible treatment methods can then be evaluated. A spreadsheet is also available for download that can further simplify the process of treatment selection.
Small Room Considerations
The above is not intended to be used for the design of an actual home theater. I have simplified Ando’s methods and made some rather large assumptions, chief among them being that Ando’s methods and techniques for room design can actually be applied to small rooms. I think they can be, but that’s my opinion. This may seem counterintuitive to some—it certainly flies in the face of the conventional wisdom on treating early reflection in small rooms, which has traditionally been approached with the “absorb the heck out of everything” mindset. However, the writings and researches of folks such as Toole , Moulton , and others, would seem to indicate that the conventional wisdom may not always be wise, and is certainly becoming less conventional with the advent of better sound reproduction technology (loudspeakers, processing, etc.), center-channel dominance of movie dialog, the ubiquity of heavy carpet (and pad) and “acoustical” ceiling tiles in many home theaters, etc. The potential desire for early side wall reflections is real. Careful consideration of the objective measure of the source material being considered for reproduction in a home theater—and the fact that the faithful reproduction of soundtrack dialogue is (arguably) the #1 goal of the home theater—would seem to indicate that this “new” school of thought is not far off the mark when it challenges the use of heavy absorbers on the side walls. That is not to say everyone should tear down their absorbers. For larger listening rooms, a certain amount of absorption and/or diffusion is necessary. However, for smaller spaces (and equally smaller budgets), the conventional “necessity” for side wall absorption (or diffusion) is debatable. The absorption of early reflections in small rooms is largely a carryover of conventional home studio design (and this is not errant). The wealth of literature and online advice concerning the acoustical treatment of small sound studios will make it difficult for any paradigm shift for home theater acoustics to take hold.
Fabric-wrapped panels and absorber systems are relatively poor off-axis absorbers.
Additionally, it is not surprising that the home theaters that have been treated with side wall absorption—and there are many—are perceived as good-sounding rooms. If one considers that the overwhelming majority of absorptive panels used in home theaters are of the “fabric-stretched-over-fuzz” variety, then the widespread acceptance of this technique falls right in line with what has been discussed here. Fabric-wrapped panels and absorber systems are relatively poor off-axis absorbers. That is, when measured off-axis, there is less absorption. When a person is sitting at a listening position with reflections arriving at some angle from the side wall treatments, the panels are actually reflecting a good deal of sound. This could easily lead to a pleasant sounding environment that falls neatly in line with what I’ve discussed above. The absorption is working best on-axis, resulting in absorption of many reflective artifacts, such as flutter echoes, that would otherwise decrease sonic quality. However, when looking at one crucial element, that of the arrival of the first reflection, the “fabric-stretched-over-fuzz” absorbers are worse performers, much to the benefit of the average home theater.
I really haven’t delved into anything groundbreaking or “new” here. I’ve simply taken another approach to reaffirming the work of others. The main point is that absorption may not be the only option for treating the average home theater. There is certainly ample research indicating that leaving the early reflections alone, or treating them mildly with some diffusers of “fabric-stretched-over-fuzz”-type panels, would not yield poor results. In fact, many of the “bass traps” folks are selling could serve the purpose of controlling low-end and midrange problems, but reflect enough sound to the main listening area so that the objective results are consistent with the subjective expectations. (Indeed, this is already the case in many home theaters.)
Early reflections may or may not be major problems in home theaters. Addressing them through the use of heavy absorption should be pursued with caution. I believe the issue of early reflections and their relative merits (or lack thereof) in any home theater should never be ignored. It is my hope that continued discussion and further research on subjective preferences in small acoustic spaces will result in the development of better objective measures. New and better metrics or methods—possibly including applications of Ando’s work similar to what I’ve discussed above—would help home theater designers provide more definitive answers to the questions about early reflections.
Thanks to Erikk Lee at Auralex Acoustics, Inc. for extracting the LOTR center channel audio for use in my analyses, and to Masatsugu Sakurai of Yoshimasa Electronic Inc. for his wonderful programs and his assistance in making them do what I wanted them to do.
1. “Loudspeakers and Rooms for Sound Reproduction—A Scientific Review” by Floyd E. Toole, Journal of the Audio Engineering Society, Vol. 54, No. 6, June 2006.
2. See, for example, A New Way to Think About Room Acoustics
3. Collected Papers on Acoustics by Wallace Clement Sabine, 1992 Peninsula Publishing republication of the 1922 Harvard University Press publication. Available from http://asa.aip.org/publications.html.
4. Concert and Opera Halls: How They Sound by Leo Beranek, Acoustical Society of America, 1996.
5. Architectural Acoustics: Blending Sound Sources, Sound Fields, and Listeners by Yoichi Ando, Springer, 1998.
6. See http://www.ymec.com/.
7. DVD: The Lord of the Rings: The Two Towers, Chapter 50, “The Tales that Really Mattered…”, ©MMIII New Line Home Entertainment, Inc.
8. “The Significance of Early High-Frequency Reflections from Loudspeakers in Listening Rooms” by David Moulton, 99th AES Convention (1995), Preprint No. 4094.
Jeff D. Szymanski is an acoustical engineer with over 12 years of experience in the music, room acoustics, and noise control industries. He is a member of the Audio Engineering Society, the Acoustical Society of America, the Institute for Noise Control Engineering, and Synergetic Audio Concepts. He co-holds two U.S. patents for acoustical treatment devices, is an accomplished musician, and is a resident of the beautiful state of Kansas.
Early reflections may or may not be major problems in home theaters. Addressing them through the use of heavy absorption should be pursued with caution. I believe the issue of early reflections and their relative merits (or lack thereof) in any home theater should never be ignored. This article opens a dialog on how to best deal with them and how they differ from large listening spaces.
Discuss "Early Reflections in Home Theaters: A Different Perspective" here. Read the article [audioholics.com].
... I am sure, the early reflections the core of each acoustic. Its correct spatial distribution is one of the most important cues for the spatial impression. The main difference regarding the genuine event is the reduction of its spatial distribution upon a few single channels in the conventionally surround procedures. Moreover the surround formats including no information regarding the elevation level of the mirror sources, which causing the early reflections in the recording room from all directions, mostly outside the horizontal level.
A new approach enables to synthesize its correct distribution by the reproduction side. Thats very expensive, but the reproduction by far more accurate as the conventionally methods. I am work on the topic since some years; the principle is described by easily understandable animations on my website. Unfortunately the procedure practically realized until today only by horizontal loudspeaker rows around the listener, wherefore the reproduction remains two dimensional still.
Hence, the website www.syntheticwave.de [syntheticwave.de] is only a description; the system isnt purchasable until today. But the Problem description seems my important for all home cinema users. It becomes possible to enlarging the playback room virtually by the loudspeaker wall.
Kind regards Helmut
I would note however that Jeff's point is narrow in that he addresses the useful of early reflection control for speech intelligibility, and as such, properly focuses on calculations regarding the center channel. But there are other purposes beyond speech intelligibility and the center channel that can be beneficially served when you install early reflection control absorptive treatments... such as enhancing that clarity/depth of the stereo image... diminishing the likelihood the listener will be exposed to flutter echo... making the "sweet spot" larger... etcetera.
Still I think Jeff has an excellent point that absorptive treatment can easily be over done - and trying to apply irrational / inapplicable concepts like RT60 to small rooms is a practice that should stop.
I think if you dig into the Audioholics University section, you will find some relevant articles.
I think that you may be missing the gist of the article though. What we have been exposed to, as far as room treatments, was largely based on large room calculation scenarios. Since home theaters are not large when compared to public concert halls these original formulas were based on, we are now starting to question their validity of these calculation in the modern home theater application. This means we must rethink everything including acoustic treatments when addressing our own acoustic issues.
I hate to say it, but the only real reliable way to diagnose room acoustics is in room measurements with equipment capable of very high sampling rates. Not cheap or readily availble to the average sound designer, much less the audio enthusiast.
The only other real solution is to experiment with acoustic panels and place them according to what sounds best to you. You can try the tradional locations like corners and edges or primary reflection points, but as the article pointed out, we may be doing the right things for all the wrong reasons.
This article should be a real eye opener and I hope everyone who reads it walks away with a new found appreciation for the complexities of audio science.
And as you mentioned Dr. Toole, it seems that Harman International and its subsidiaries may actually be designing their speakers with early reflections in mind and trying to offset their influences from the get go.
Other speaker designs like Klipsch have used controlled dispersion technology to also avoid early reflections and side panel placement in a tradional manner may be mute at shorter seating distances.
All very good perspectives that should give some people something to think about. In the end, it is what provides the flattest frequency response in your room and\or sounds good to you.
Keeping an open mind is the real key, IMO.