Why It Shouldn't Be So Darn Loud

More than likely you have been to one of those loudspeaker demonstrations where the system is being played so loud that it's actually painful. You know the type. They usually start with some guy saying something like, "Let me show you how powerful this systems is!" Or, "Let's see how this explosion 'feels'."
Publish date:
Social count:

More than likely you have been to one of those loudspeaker demonstrations where the system is being played so loud that it's actually painful. You know the type. They usually start with some guy saying something like, "Let me show you how powerful this systems is!" Or, "Let's see how this explosion 'feels'."

Over the past decade the manufacturers have made immense improvements in every aspect of presenting the "home theater" experience. Comparing the systems available today to the components offered even just 3-5 years ago is almost unfair. Those improvements have dramatically reduced the kind of "cues" we and our customers are used to listening for when the sound levels get too high; for example obvious distortion, or the mechanical pain noises loudspeakers make when they are pushed too far and all the similar sonic markers. Both visually and audibly, the capabilities we can offer a client today often exceed the performance of "professional" systems, especially when it comes to areas like distortion and clarity, and can at least match them in sound level and image brightness capabilities. In fact we more often than not can provide a better experience than anyone can get "at the movies" or any other commercial venue. That's why people come to us, and that's why technologies like DVD, fixed-pixel-imaging devices, and innovative loudspeaker technologies are selling in the huge numbers that they are.

From all market and research indicators, this will only improve as more and more technology and new materials are applied to the task of re-creating the experience of film, sports and other programming.

But, there is a dark side to all this wonderment, and that is that the systems we can now install with their increased performance, are also capable of producing sound pressure levels that are unequivocally dangerous to the hearing of everyone exposed. I am sure you have seen the specifications, and often the touting of them in advertising or sales pitches, showing that the system is capable of producing 120 dB or even more. That level of sound pressure delivered to the ears of a listener:

1. Will cause hearing damage

2. Used to be the exclusive provenance of the professional touring PA guys and the huge, 20, 30, 40 and more thousand dollar audio monitors of the professional studios and dubbing stages, but not anymore! Therefore, we have now reached a complex and often invisible technological precipice. If we are not careful we will fall off that high cliff and into a literal quagmire of legal quicksand, which will definitely produce the kind of problems that keep business owners up late at night. They are called personal injury/product liability lawsuits--currently a multi-billion dollar "business" in the U.S.

Let's examine this situation more closely and look at some ways to keep things under control. I presume that most of you have been to some kind of a live popular music show where the massive stacks or flying arrays of speaker cabinets were producing a literal "wall of sound" at levels approaching a military jet in full afterburner (120dB+). You remember what that sounded like and felt like--the chest crushing bass and that ringing in your ears when you left (a warning sign to everyone that it was way too loud). What you probably didn't know is that a lot of the "pros" who were working that night were wearing specially designed "musician" earplugs that drop those massive 115dB+ levels by 9, 12, 15 even 20 dB, without having a significant effect on the spectral balance or tonality.

They weren't listening to that show at anywhere near the levels to which the audience was exposed. Theirs was more like a much safer 90-95dB max, because they know how dangerous such levels can be to their ears.

Now, it is unrealistic to expect your customers to use such protection, and even more unrealistic to expect them to realize how dangerous highsound levels can be. Exposure to high sound pressures is cumulative. For example, if your customers watch just one movie a week of 2.5 hours in duration that's 10 hours of exposure monthly, or 120 hours a year. While that is nowhere near the exposure time experienced by touring PA professionals or recording engineers, it's still quite a lot when combined with all the other high-SPL exposure they are likely to have during the course of a year.

Let's add into the equation the fact that your customer is probably over 30, and already has some considerable accumulated high-level exposure to begin with, and the concurrent damage that produces. Add in what your system is capable of and the picture gets pretty ugly, and quickly.

In fact it's quite likely that your client (male or female) has some noise-induced hearing loss, probably right in that 300 to 4,000 Hz speech intelligibility band along with some concurrent high-frequency suppression above 6-10K. You, however, have no way of knowing how bad their hearing already is when you meet them.

To compensate for this loss, without even realizing it in most cases they simply increase the playback level, and often turn up the "treble" control on their audio systems to try and give them back what they have lost, while actually increasing the overall damage level they started with, and on the circle of destruction goes. When you enter the situation, they are going to want you to do the same, and unfortunately for you the hardware you sell them is easily capable of providing what they want. So if you don't set it up to their liking they will simply turn it up when you leave and all that careful calibration you performed goes right out the window. How many times have you been back to a client's theater and found the volume controls raised up more than just a little bit?

Now let's add in the SOAF (Significant Other Acceptance Factor) to this complex situation. If that SOAF is a women they will, according to numerous scholarly studies, often have significantly less age- and noise-induced hearing loss and thus will find high levels and boosted mid-upper frequency responses more objectionable, than their male counterpart. This is in part also due to the differences in they way men and women processing acoustical information and the slight, but often, noticeable differences in specific frequency sensitivity. In fact it has been noted that one of the most-often-used arguments against a purchase of this type of system comes from the fact that the demonstration was too loud and uncomfortable for the females present, and they have no wish to have this experience repeated in their home.

So what do we do about all of this, and how do we deliver the system performance we want to provide without creating a situation that allows hearing damage? The first thing to realize is that you must carefully and precisely calibrate every system to deliver the flattest possible power response. Note that I did not say "flat frequency response," but flat power response. In the hard copy edition of August's Residential Systems Magazine, look carefully at the graphs shown in Figure 1. These are the classic Fletcher/Munson nominalized (blending of both sexes responses) human hearing sensitivity plots. Conveniently for you the structure of the human hearing system your client has is no different than the one used when these were first produced in the 1930s, based on the work of Harvey Fletcher and WA Munson at the legendary Bell Laboratories.

These curves with sound intensity (level) along the left vertical axis and frequency along the horizontal axis show the ears sensitivity to sound at specific frequencies. The lower the point on the curve at any specific frequency, the higher the ears' sensitivity is at that point.

If you look at the curves between 80 and 100 dB, you will see that they progressively flatten out, showing that the ear's sensitivity by frequency is also flattening out--that is it is becoming more (but never totally) equal across the entire perceivable frequency response. At lower sound intensitylevels (small numbers on the dB scale) the ear becomes more sensitive to the 3-500 through the 4-5 kHz area, where human speech is concentrated, and also by the way where most hearing damage initially occurs, partially because of this increased sensitivity.

If you more closely match the power response of your system to the ears ability to detect sound you will go a long way to insuring that your customer can hear everything the system is capable of without the need to generate excessive and damaging levels.

Using the 90 dB curve as a reference (because that's really the maximum continuous level you should allow for), a flat power response in a system that approximates this curve will sound clean and natural to the majority of listeners. This is not something you should be slavishly trying to achieve, because precisely matching this curve will not be needed. Just be aware of how the ear actually functions and use it to your advantage. Also, you should recognize that because of the physics of rooms and loudspeakers in rooms, most systems you install will have some type of low-frequency peaks caused by room resonances. Note the double peak and then a nice smooth response curve.

This type of system response is doing two things. First the energy peaks are accentuating the already inherent masking effect that low-frequency sounds have on higher frequency sounds, especially those above 500 Hz. The more masking that occurs the harder it is for the ear to distinguish the "detail" in a particular presentation.

Second, these peaks are distorting the apparent system response. The bumps are at a higher level than the surrounding frequencies allowing them to more easily "heard" by the ear and thus creating the impression of a shifted frequency spectrum around those areas.

This leads such a system to be perceived as "boomy" and bass heavy.

By removing these anomalies and presenting a more even response to the listener the system's overall level will perceptually increase without raising the volume. Why? Largely because the low mid- to high-frequency masking will be removed allowing the ear to present a more natural sounding, balanced spectrum to the brain, with more detail and clarity. This means things will be easier to hear and the system will apparently get clearer and louder.

Achieving this is simple and straightforward. By using precisely tuned parametric EQ filters, these peaks can be removed and the system's power response corrected. This is shown Figures 3 and 4 (hard copy edition of Residential Systems Magazine). In Figure 3 an AudioControl Diva unit has been set to create two steep and precise notch filters matching the two peaks show in Figure 2. This produces the much more even response shown in Figure 4, again taken with an Audio Control Iasys analyzer. If you look at the Iasys response in Figure 4 and 90 dB sensitivity plot shown in Figure 1, you will see that they are almost the inverse of each other. In other words, the system being measured is supplying an acoustic energy distribution that closely matches the ears' ability to detect it, thus providing a sound intensity presentation that sounds even and smooth. The last piece in this puzzle is the use of an appropriate smoothing EQ as shown in Figure 5.

In this example the graphic EQ has been set to provide a gentle rolloff above 4 kHz to compensate for the ears' sensitivity as shown on the Fletcher Munson graph in Figure 1 from about 4-6kHz. A very shallow "house curve" like this will help reduce some of the often substantially annoying HF sizzle and excessive energy in the system as perceived by the ear due to steep rise in sensitivity. Above about 6kHz the ears sensitivity decreases significantly, and fairly rapidly as well, with as much as a 20dB differential between 6kHz and 10kHz. However the equalization applied to a lot of program material in production attempts to compensate (or over compensate in many cases) for this natural process, and for the spaces in which the program material will be reproduced (large room volume commercial theaters for example) producing a strident and often harsh sounding presentation in the much different residential spaces. In the consumer THX processing structure, they incorporated a RE-EQ function to help correct this. While that process is still appropriate the addition of a shallow smoothing EQ will often provide that last tweak needed to make the system sound smooth to the listener. This is a suggested curve only, each system and each room will be slightly different and you must both measure and listen to achieve the best result. In some cases this may not be needed.

Too Much Dynamic Range is Not a Good Thing
Current playback media such as DVD provide a dynamic range (that is softest sound to loudest sound) that is almost double that of the motion picture systems of 25 years ago. It is usually better than any commercial theater can achieve simply because of the background noise levels founding commercial spaces vs. the background noise levels found in most residential environments. However in many we now have too much space between loud and soft to be effectively delivered in all but the very best residential spaces.

If the system is set to provide adequate level for the soft passages, it is most likely to be too loud when the higher level stuff arrives. While many processing/decoding systems offer such options as night modes and similar schemes to reduce the dynamic range, many customers do not like the overall effect produced.

There is an alternate method that not only can work better for some. It also helps you with keeping overall levels down, because it again takes advantage of the ear's sensitivity curve. It's called frequency selective compression.

There are now devices on the market like the AudioControl Diva unit that allow you to apply dynamic range compression selectively to specific channels and across specific frequency ranges. The most appropriate and sensible use of this option is to apply carefully designed compression to the Sub/LFE channel output(s). By controlling the overall dynamic range of this channel to achieve a smaller "window" between soft and loud, you can help the ears' reduced/limited sensitivity in this region to more efficiently detect the low frequency sounds. This allows you to keep the overall system level down, because you won't need raw level to make the low-frequency material, the foundation stuff audible. This will give the system a satisfying "bottom" without masking the rest of spectrum under to much LF energy.

A free benefit of this approach is that you also protect the subwoofers from excessive excursion, being overdriven, and other nasty things which cause them to bottom out, distort, or just plain fail. In most cases the hard work has been done for you, and you simply need to select from a few options to implement this process. It requires some listening to determine what scheme is best with which product in a particular room. One size does not fit all here.

Another option is to also apply some gentle compression to the center channel, which as you should know contains a very high percentage of film audio. By gently restricting the dynamic range just a bit you can often smooth out your LCR sound stage and avoid that dialog line or effect jumping out of the center speaker, especially when the center speaker is not identical to the L/R systems. Additionally this process may help improve the splice between the sub and LCR area adding additional smoothness to overall system's presentation.

Note: Readers are encouraged to check with their manufacturers to see what specific options may be available within the products they sell and how best they might be applied. This article is a recommendation only, not a set of hard and fast rules. You must explore each system and be careful to measure the systems' installed performance accurately, before attempting any correction.

Frederick J. Ampel (FrederickAmpel@compuserve.com) is president of Technology Visions in Overland Park, Kansas.