Achieving Flat Response in a Home Theater, Part 2

 

Producing audio in a home theater listening room is all about illusion. We would like to create the audible illusion of actually being present on the sound stage or at the location where a movie or other video material was originally produced. To come as close as possible to this sensory nirvana, our sound reproduction system should re-create every sound at our listening position at the same sound level as it was mixed by the recording engineer, no matter its frequency. The uninitiated might wonder what the problem is. As long as we purchase a high quality processor and amplifier, and speakers with tight specs, won't we reproduce great sound at the listening position(s), with flat frequency response? The answer is almost definitely "No," unless we go to a little more effort to equalize the frequency response of the entire sound reproduction system at the listening position(s).

The key is that the acoustics of the listening room are a major component of the sound reproduction system. The room acoustics are likely to negatively affect the audio frequency response, especially at bass frequencies. Whereas the electronic signal processing/ amplification equipment may have a ±1-3 dB frequency response and the speakers may have a ±1-6 dB response, the acoustic room effects may easily cause ±12 dB, or greater, variations in the bass frequency response.

 

Fig. 1: The way in which sound is affected by the room (acoustics) has more effect on frequency response at a room listening position, as shown by these RTA plots, than does the electronic reproduction system.

 

When the system response at one or more frequencies in the bass range is much more than about 3 dB above the level of other bass frequencies, your ear tends to hear any program material at those frequencies with an emphasis that overwhelms the other bass sounds. The effect is bass that sounds very one-note and boomy, and tends to hang for longer than normal periods, rather than being tight and punchy. The boomy effect is overwhelming and tiring and tends to cause us to turn down the bass level. With flat low frequency response we get bass that is natural, rich, powerful, and punchy when played at levels that match the rest of the audio spectrum, without hanging bass notes or overwhelming boominess.

Uneven frequency response is only one acoustic room problem that should be addressed in a home theater; some others are:

 

·         background noise

·         speaker boundary interference

·         poor diffusion

·         early reflections

·         comb filtering

·         echoes

Uneven frequency response, though, is one of the major room effects that must be remedied if the sound system is to accurately re-create the original recording. We want to hear the music sing; we don’t want to hear the room resonances singing.

Room Resonance Modes

The enclosed space of any room naturally resonates, much the same way that a glass soda bottle or organ pipe resonates when excited with moving air. Each of the facing boundary surfaces of a room (sidewalls, endwalls, ceiling/floor) resonate at multiple frequencies. The lowest frequency sound that resonates between facing wall surfaces has a wavelength equal to twice the wall spacing. This frequency is the fundamental mode for that surface spacing. Since sound wavelengths are long at low frequencies (approximately 28 feet at 40 Hz), the fundamental room resonance mode for facing surfaces in small home theater listening rooms tends to be in the 20 Hz to 80 Hz bass range.

A sound at this lowest resonant frequency travels exactly one wavelength as it reflects off both facing surfaces and returns to its originating point. The relationship between a sound reflected at right angles off the first surface, and the sound reflected back off the facing surface, is constant at every point along the reflection path. This causes the two reflected sounds, which are meeting each other head-on, to have a constant reinforcing effect at some points along the path and a constant canceling effect at other points along the path. At the fundamental mode frequency, (Fr) reinforcement takes place at every point along the reflecting surfaces and cancellation takes place at every point that is exactly halfway between the reflecting surfaces. A listener positioned halfway between the two walls would hardly hear a single tone played into the room at the fundamental mode frequency (assuming the other room surfaces weren’t present to modify the situation).

 

Fig. 2: At the fundamental resonance frequency of facing walls, sound level peaks (maximum pressure) are produced at the walls and a cancellation null is produced halfway between the walls.

 

Any facing room surfaces also produce resonant modes at multiples of the fundamental mode frequency. Room sidewalls spaced 14 feet apart have a fundamental resonance mode at 40 Hz, plus second, third, and fourth harmonic modes at 80 Hz, 120 Hz, and 160 Hz. These harmonic modes produce two, three, and four cancellation zones, respectively, between the facing sidewalls, with reinforcement zones at the walls and halfway between adjacent cancellation zones. With a single tone of 160 Hz played into the room, for example, a listener walking from one sidewall to the other would walk through four sound null zones.

 

Fig. 3: At the fourth resonance mode frequency of 40 Hz (160 Hz), four cancellation zones are created between the facing walls.

 

Consider that there are many room resonance modes between a room’s sidewalls, endwalls, and floor/ceiling, and you see that the situation exists for the sound pressure level to be either peaked or nulled at many different frequencies, at any given listening point in the room. This is true, even if the amplifiers and speakers produce perfectly flat response just in front of the speaker. To achieve optimum audio performance in any listening room, you must realize that the acoustics of the room is one of the more important components of the sound reproduction system.

Frequency Response Measurement

To evaluate the frequency response of a sound reproduction system, we view a real-time display of the sound pressure level (SPL) versus frequency, measured at the listening position(s). We do this with an instrument called a Real-Time Audio Spectrum Analyzer, or simply RTA. The RTA, rather than showing the SPL for every single frequency in the audible range, instead separates the audible spectrum into octave or sub-octave bands and displays the average SPL for each of these bands, usually with a bar-graph display. When octave band resolution is used, the RTA shows the SPL for each of the ten octave bands that make up the audible spectrum. When one-third octave resolution is used, thirty bands cover the spectrum.

Fig. 4: An RTA with 1/3rd octave resolution divides the audible spectrum into 30 bands and plots the sound pressure level for each band.

We usually use the RTA to evaluate the result of playing a standard test signal into a listening room. The standard test signal that complements the RTA is pink noise, covering the audible range of 20 Hz to 20 kHz. By definition, pink noise contains equal energy per octave. So, if pink noise were played through a theoretical sound reproduction system with perfectly flat frequency response (including the room), an RTA would measure the same SPL for every octave or sub-octave band.

 

One-third octave RTA resolution is particularly significant, because that corresponds very closely to the critical band resolution of the human ear. We measure a sound system’s frequency response, using a one-third octave RTA display, to most accurately evaluate how the system sounds to our own ear/brain combination. There are many times, however, when more detailed RTA resolution is necessary.

When we see that our sound reproduction system produces much louder or much quieter sound in a particular one-third octave band, at the listening position(s), we usually need to see the problem in more detail before we try to identify or correct it. We want to see what precise range of frequencies is being peaked or nulled within the problem one-third octave band. To show the desired detail, the RTA needs to have at least one-twelfth octave resolution. Within the 63 Hz one-third octave band, for example (which is 14.7 Hz wide), one-twelfth octave resolution allows us to see the one-third octave band split into four segments, each approximately 3.6 Hz wide. This allows us to identify or precisely treat (possibly with a narrow-band parametric equalizer) the exact band of audio frequencies causing the anomalous response, without affecting the entire one-third octave band of frequencies.

Fig. 5: An RTA with 1/12th octave resolution allows us to see each 1/3rd octave band split into four segments.

 

When we are trying to evaluate the frequency response of more than one listening position, we may at times wish to measure the average response across these positions. This type of averaged response across some room space is called a spatially averaged response. The easy way to obtain a spatially averaged frequency response of a seating area is to use a microphone multiplexer to feed multiple microphone inputs to a Real Time Spectrum Analyzer, one at a time. When the RTA is set to average the switched microphone inputs over time, an averaged frequency response of the entire sampled area is obtained. This averages out the severe peaks and nulls of any single seating position and allows us to view an averaged measurement of the entire listening area and apply room treatment which favorably affects all listening positions, without unduly favoring or harming any one position.

Fig. 6: A microphone multiplexer allows us to view an averaged RTA room response without undue emphasis on a single position’s modal response.

 

Summary

We have seen that uneven frequency response in a listening room is much more a problem at bass frequencies as compared to midrange and treble frequencies. This is due primarily to the effects of room resonance modes that cause peaks and dips in the system response, and is more pronounced at some positions in the room than at others.

 

We have discussed that an RTA with 1/3rd octave resolution is the primary tool that we will use to measure the frequency response at different locations, in a way that corresponds to the human ear perception. To correct low frequency response problems that we identify with 1/3rd octave resolution, we will switch to 1/12th octave resolution to precisely localize the problem frequency. For those cases in which we are measuring multiple listening positions, trying to flatten the response at all positions, a microphone multiplexer is a valuable aid in producing a spatially averaged response over the entire listening area.

 

Next month, in part 3, we will discuss the different methods you can use to flatten the frequency response at one or more listening positions in a home theater room.

 

For additional information or if you have questions pertaining to this article, please contact: mailto:sales@sencore.com