Acoustical Measurements - what are they?
By Richard Bird, Rives Audio
In this second article of our introductory series on room acoustics [click here for the first installment], we want to continue building some fundamentals that will be very important in understanding what could be wrong with the listening room; and ultimately: How can I fix the listening room? I have heard some people claim "I just listen and walk around the room and clap my hands and I know what to do." I would say this is another myth. True, the final tuning of a room often does require extensive and subjective listening. But to even get into the ball park, you will need basic measurements.
In this article we will discuss frequency response; reverberation time; energy time curves; waterfall plots; and the much debated psycho-acoustical response curves.
Our original definition was pretty simple: Frequency response encompasses the range of human hearing, namely 20Hz to 20kHz. So what should the goal be? A flat frequency response from 20Hz to 20kHz, right? Maybe. It depends on both how it's being measured and subjective preferences. First off, what is flat frequency response? It admittedly does sound like a simple question but it isn't. On a linear frequency scale, you probably would not like the sound of a flat frequency response. In fact, it would be rather bright because our hearing is logarithmic, not linear.
We generally graph frequency response on a logarithmic scale. In that context, a linear or near-linear response is the desired goal. But beware the difference: Linear on a linear scale is not what we want! While today's article is not about taking measurements (that's for next month), let's quickly look at "White Noise vs Pink Noise" (remember, all the technical terms were defined in article #1). White noise, in a perfectly responding system, would give us flat frequency response when we graph said response on a linear scale. Pink noise is more often used as its goal is a flat response on a logarithmic scale. Since the latter is what we strive for, almost no one uses white noise generators anymore.
Below is a graph of a linear frequency response. You will note how jagged it is but don't be alarmed - this is normal comb filtering which occurs in almost all listening rooms. What is more important to note is that overall, the response is gently sloping down in the treble - decaying. Thus there's a little more energy in the lower octaves relative to the treble.
I mentioned subjective preferences. Do we always want flat? The human ear is not flat. By that I mean that we do not perceive every frequency at the same volume while listening to a flat frequency sweep. Simply put, the human ear and the brain to which it connects for data processing is designed with one primary goal in mind: To reproduce and interpret speech. This means specialization. It means a narrow band of frequencies that we hear more clearly than others. The generally accepted range of these 'speech' frequencies is from 500Hz to 5Khz. Because of this phenomena, we do not hear the registers above and below this range within the total 20-20,000 bandwidth as easily. Because of this, most people actually don't want a measurable flat frequency response. Also, because lower frequencies are harder to hear at lower levels, most people want a little added boost in the bottom end and, in many cases, gently tapering off as one gets into the upper octaves. If you've heard a system that you subjectively really liked, it would be interesting to graph its response and see how flat it really measured. Conversely, we've had clients that knew how their hearing in the higher octaves was not as good as it used to be. We actually designed rooms that allow for more of the energy -- and recommended speakers that excel -- in the higher octaves. Some people would potentially frown on this approach but music is ultimately for enjoyment, not measurement statistics. Measurements should be used as informational tools to attain that level of enjoyment. They should not dictate what constitutes enjoyment.
This is the time it takes for an initial sound to decay in volume to a certain level. The technical definition of RT-60 or reverberation time is the time it takes a sound to decay 60dB or shrink to 1,000,000th of its initial impact or sound pressure level. RT-30 and RT-90 measurements are also used, albeit less frequently. It is important to preface this discussion by stating that reverberation time measurements were developed for large venues such as concert halls and churches. These venues have large and diffuse sound fields in the audience sitting areas. It is these diffuse sound fields that make up the context for critical RT-60 measurements.
Our listening rooms are naturally much smaller and as such, do not contain diffuse sound fields. Thus many claim that the notion of reverberation times in consumer listening spaces is invalid. However, our experience has shown us that not only are reverberation times in a small field valid, they are probably the second most useful measurement for us. It's important to appreciate that the ideal reverberation times for listening rooms, theaters (both large and small), churches, recording studios and control/mixing rooms vary greatly. Each of these types of spaces serve different purposes and as such should have reverberation times optimized for their intended usage. Unlike the general goal for frequency response, reverberation times can and will vary across the frequency range depending on the type of room and its primary use. Given this, you can already predict that the overall or average reverberation times will be different for different venues. Let's consider some examples.
Have you ever been in a church and listened to a chorus? Hopefully the church was large and had many hard surfaces. The reverberation times (RT-60) in a church like this could reach up to a second or even more! When you listen to the chorus, individual notes seem to carry on forever. More importantly, in an acoustically well-designed church, they decay so evenly across the frequency spectrum without favoring certain ranges. Hence the midrange of the human voice is well balanced, the frequency response is relatively flat but the decay time for each note is rather long. For music-making and listening enjoyment, this can be a nice effect. It creates an enveloping feel of music all around us. When we record the music in such a venue, this natural reverberation time is preserved on the recording. With top systems during playback, we can obtain a good sense for the space in which the recording took place.
Recording studios can be quite different. Some may have long reverberation times, others such as voice-over post production studios might require very short reverberation times, in part to protect overall voice intelligibility, in part to allow the recorded vocals/voices to later be electronically manipulated without interference from room colorations. It really depends on what is being recorded and what the artist/producer's intentions are.
Control/mixing rooms generally have very short reverberation times. Here the recording engineer wants to hear the music being produced without any room colorations. He isn't interested in pretty or enveloping but the 'truth' (as Dynaudio would say). He is listening 'near field' which has more to do with the percentage of direct sound vs indirect sound than with how far removed his ears are from the monitors. Distance to the monitors does effect this, but more important is that he be listening to direct sound without the adverse acoustic involvement of the room, making listener distance from the monitor speakers a function of room size, room treatment and placement within the room. Given these specialized conditions, the reverberation times in a control room may be as short as 0.2 seconds. (In a future article, we will discuss studio, control room and listening room differences and their individual goals in more detail.)
Let's return to our listening rooms. Relative to our church example, listening rooms have short reverberation times. Relative to our control room example, they have long reverberation times. Reverberation times for listening rooms vary depending on listener preferences, listening levels and whether or not it is designed for multi-channel or 2-channel use. This is an area where we can scientifically measure the reverberation times but it is somewhat of an art to dial in the desired reverberation times for a particular listener just so. Not unlike the differences with studios and their primary applications, we must ask what the goals are. What do we want to achieve? Some audiophiles believe that the listening room should not interact with the speakers at all. They believe it should perform much like a control room. Having listened in rooms designed like this, I can tell you unequivocally that I completely disagree - this design completely destroys the kinetic energy of the music. Having no interaction with the room at all produces a very unnatural 'dead' sound. In general, we would like to achieve an RT-60 of 0.34 to 0.39 seconds from ca. 200Hz on up. RT-60 measurements in small room acoustics below 200Hz have little significance as they are flawed by the energy build-up of room modes. To properly evaluate bass response below 200Hz, we need to use other methods. Below is the reverberation time for an ideal 2- channel listening room.