Anda di halaman 1dari 61

Sound Field Speaker Coverage Modeling Study

Prepared for: Phonic Ear Inc. 3880 Cypress Drive Petaluma, CA File: 5325 June 11, 2004 McSquared System Design Group, Inc.
#102 145 West 15th Street, North Vancouver, BC V7M 1R9 Ph 604.986.8181 Fax 604.988.9751 website: http://www.mcsquared.com e-mail: info@mcsquared.com

Phonic Ear Sound Field Speaker Coverage Study

Page 2

Purpose of the Study..................................................................................................... 3 The Sample Classroom .................................................................................................. 4 The room model of Upper Lynn Valley Elementary room #211 ......................................... 5 Speaker Configurations Modeled ..................................................................................... 6 Direct Sound versus Total SPL..................................................................................... 6 RaSTI Predictions ...................................................................................................... 7 C50 (Clarity Ratio) .................................................................................................... 8 Conclusions and Recommendations.............................................................................. 8 Sound Field Configuration: ToGo corner mounted at 7 to the center of speaker............. 11 Sound Field Configuration: ToGo on Desk or Table top .................................................. 27 Sound Field Configuration: single small speaker (flat panel, etc.) mounted in corner ......... 44 ToGo 3D polar response prediction produced by EASE 4.0 ............................................. 60

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 3

Purpose of the Study


The purpose of the study was to look at predictions of sound field speaker performance using EASE 4.0 computer modeling software, one of the most widely used sound system design packages in the industry. The study would include the most common speaker system configurations and common speaker models employed in the sound field application. The study would be based on acoustical measurements taken in a real elementary school classroom currently using a sound field speaker system, and then the existing acoustical conditions would be extended in the model to include two higher levels of background noise and two longer values for reverberation time to examine which speaker system solution worked best in varying acoustical conditions that are the primary factors in reducing speech intelligibility. The results of interest include: 1. Direct sound coverage of the seated ear plane using phase information to calculate expected interference patterns to show the uniformity of direct sound, which affects both intelligibility and gain before feedback. 2. Total SPL (direct + reverberant sound energy) at three different reverberation times to assist in showing the relationship between direct sound coverage and the reverberant sound level, and how that can be misleading in predicting improved intelligibility. 3. RaSTI (Rapid Speech Transmission Index) predictions for three different background noise levels and three different reverberation times to assist in showing which speaker system configuration is the best for rooms where excess noise or excess reverberation time are the dominant problems. The study would help to identify common errors in speaker placement and application; speaker system configurations that reduce the effectiveness of a speech reinforcement system; speaker configurations that will reduce gain-before-feedback when used with an open microphone; and the benefits of one type of speaker over another in providing effective coverage. These issues have been studied extensively over the past 20 years since the advent of accessible and affordable time domain measurement capability appeared in the form of the Techron TEF analyzer. This measurement system allowed sophisticated Time Delay Spectrometry measurement methods to be applied to sound systems. It would not be an overstatement to say that more has been learned about the behavior of loudspeakers in an acoustic environment in the past 20 years than had been understood for the previous 60 years, since Rice and Kellog developed the moving coil loudspeaker in 1925. The TEF analyzer plainly disclosed the cause of problems with sound systems that could not be corrected with an equalizer because the response variations were created by a non-minimum phase system, in other words, a speaker system that had signal delays present in excess of a wavelength. That single breakthrough in observation of speaker system performance changed the understanding of what happens when sound sources interact with each other in the acoustic environment. When sound is cancelled out by destructive interference of the sound waves, it cant be repaired using an equalizer. When the sound is cancelled out, it also begins to have a response that looks very similar to someone with a significant 4kHz hearing notch. It is possible to install speaker systems with a frequency response that replicates hearing loss, and presents a different response at every seat in the house. And thats not a good thing.
McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 4

The Sample Classroom


The classroom is a 24x30 room with a 9 T-bar ceiling, and an entry vestibule area in the front corner. There are windows on two outside walls, with operable windows. The neighborhood is very quiet with no through traffic nearby so exterior noise is minimal. The most significant noise source was an air pump on the small fish tank on the south exterior wall.

There is a low bookshelf beneath the windows on the east side and a collection of shelves and tables below the windows on the south side. Measured Background Noise The background noise measurements were made with an unoccupied room, the windows open and the aquarium pump running. Measurements were made in a number of locations and the highest levels were used near the aquarium to capture the worst-case noise condition. 125Hz 48 dB 250Hz 36 dB 500Hz 35 dB 1000Hz 33 dB 2000Hz 26 dB 4000Hz 21 dB 8000Hz 19 dB

Measured Reverberation Time We measured the reverberation time with the room unoccupied. There was a small area carpet on the floor near an upholstered couch, and a substantial amount of books and materials in the shelves around the perimeter on two sides, so the reverberation time was quite short in the voice bands. 125Hz 0.50 Sec 250Hz 0.30 Sec 500Hz 0.32 Sec 1000Hz 0.30 Sec 2000Hz 0.5 Sec 4000Hz 0.45 Sec 8000Hz 0.3 Sec

The room model is shown on the next page. It features the major architectural shape features, and the ear plane is set at 36 above the floor to reflect the shorter listeners. The image of the room model shows the ear plane where the coverage and STI plot was calculated. The STI plots have the room model boundaries turned off for clarity of display, but they were part of the calculations.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study The room model of Upper Lynn Valley Elementary room #211

Page 5

Upper Lynn Valley classroom #211 ear plane off

Upper Lynn Valley classroom #211 ear plane on

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 6

Speaker Configurations Modeled


We modeled several configurations and types of loudspeakers, including: 1. Four speakers mounted on side walls (including variations in mounting and product) 2. Four ceiling loudspeakers 3. Six ceiling loudspeakers 4. Soundsphere ceiling mounted speaker 5. TOA F121C ceiling speaker 6. Ceiling speaker cluster with 4x 8 drivers in a 24 tile replacement 7. FrontRow ToGo mounted in room corner at 7 to speaker center 8. FrontRow ToGo sitting on desk or table top 9. Phonic Ear Logia mounted in corner 10. One 5 2-way speaker in room corner Most of the device files were available in EASE 4.0, the exceptions were the ToGo and Logia devices. We created these devices as clusters of existing measured single full range drivers using the driver size and spacing in the drawings provided to us by Phonic Ear and EASE created the polar balloons as shown in the section at the end of the document. In the Logia model we shaded the line array by rolling off the response of the upper and lower drivers above 1000z to reduce the lobing and interference at higher frequencies, this produced a tight vertical pattern. There may be some slight variation between the predicted and actual coverage of the ToGo and Logia, but most of the line array coverage behavior is a result of driver spacing and interference, not individual driver characteristics, so the variations should not be significant. Direct Sound versus Total SPL Weve plotted both the direct sound and the total SPL for each speaker configuration. The direct sound is the direct output from the loudspeaker without reflections or room reverberation considered. The direct sound is the most important aspect of a speakers behavior, as the human brain attempts to process and isolate the direct sound for localization and for timbre or sound quality. All the other reflections that arrive up to 50 milliseconds after the direct sound can augment the direct sound but those reflections dont improve the audible sound quality, even though they may add to the perceived sound level and intelligibility (when the reflections contain useful bandwidth). The reflections also affect sound system function such as gain-before-feedback, as the reflections and lobes can create hot spots that lower the feedback threshold. A 6dB bump in the frequency response can lower the feedback threshold at that spot by 6dB. The smoother and more uniform the direct sound coverage, the better behaved the sound system will be around an active microphone. Where the teacher will be walking through various lobes in the direct coverage, there is a chance that the system will be sent into feedback by the sudden increase in level at a few specific frequencies. The direct sound coverage is affected by the interference generated by having multiple sound sources separated in physical space and time. The distance between speakers creates nonminimum phase response variations that are different depending on the frequency involved. Where the direct coverage is subject to wide variations over a wide range of frequencies, weve
McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 7

plotted direct coverage as low as 250Hz (approximately middle C on a musical scale, which sounds like high frequencies to most listeners) to show vulnerabilities to early feedback when a microphone is used. The Total SPL (direct plus reverberant levels) is indicative of what would be measured if a sound level meter were used to evaluate the uniformity of the sound field coverage in a room. This is not representative of the speech intelligibility that would result from the sound system. The 2000Hz STI plots show that the STI values are closely related to the direct sound and not the Total SPL. This is the equivalent comparison of measuring a sound system using a TEF analyzer and a sound level meter, by isolating the direct sound we get a much better idea of what we can expect for intelligibility. Any manufacturer using a sound level meter to promote the uniformity of their sound field system is not presenting an accurate indication of how the output from the system will be perceived. The reader should notice that as the Total SPL becomes more uniform with increasing reverberation time and the difference between maximum and minimum SPL values is reduced, the RaSTI values generally increase in spread between maximum and minimum values, and the maximum STI values go down. RaSTI Predictions The 2000Hz RaSTI predictions show the predicted intelligibility, factoring in the direct sound coverage, the room reverberation and the room background noise. Both background noise and excess reverberation will degrade speech intelligibility. Weve plotted the RaSTI prediction for the existing room conditions as measured, plus higher levels of background noise (NC-45 and NC-55) and longer reverberation times (0.75 seconds and 1.0 seconds). For each speaker configuration this shows the trend in speech intelligibility performance loss. Some speaker configurations will produce better results in rooms where the acoustical problem is excess reverberation, and some speaker solutions may be better suited to a room where the problem is excess noise. In general, where the acoustical problem is excess reverberation a speaker configuration that places the speakers close to the listener usually gives the best results. Where the acoustical problem is excess background noise, any speaker system that can improve the listeners signal to noise experience will improve intelligibility. Then the key design issue becomes whether or not good uniformity in sound level can be achieved so that when the sound system is loud enough for the furthest listener, it is not too loud for listeners closer to the source. The existing test room acoustical conditions are such that almost any solution works adequately well, even an unaided talker can function in the room and provide adequate level at the furthest listener. As noise and reverberation are increased, the spread in STI values increases between worst and best seats. Thats the key thing to look at in the RaSTI plots for each scenario, look at the minimum STI values and the spread between the maximum and minimum STI values. The best speaker configuration is the one that provides the least difference between best and worst seats, and the highest overall STI rating. Those speaker configurations will be the ones that have the most uniform coverage and the highest direct-to-reverberant ratio. A good STI design target for a classroom would be 0.60 (fair) or higher.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study C50 (Clarity Ratio)

Page 8

The C50 or clarity ratio is the ratio in dB of the sound arriving in the first 50 milliseconds compared to all the sound arriving later. This is roughly equivalent to the U50 rating that John Bradley at the National Research Council in Canada promotes as a measure of Useful Energy in a classroom. The most beneficial energy arrives within the primary Haas zone of 35 milliseconds, but energy up to 50 milliseconds can provide additional level to speech as long as these late arrivals have enough spectral content to assist the speech information. Conclusions and Recommendations As with any speech reinforcement system, the goal of these classroom speech reinforcement systems is to provide superior speech intelligibility performance to unaided speech. Those goals would include: uniform speech levels for all seats; uniform frequency response of reinforced sound in all seats; improved signal to noise ratio; improved direct to reverberant ratio; and improved articulation of consonants or modulation transfer function. For a sound system designer, that list of goals has to be balanced with practical architectural and budget limitations, so there is always some degree of compromise between the perfect solution and the available solution. The key to good sound system design is to understand what the actual effects of a design compromise will be, and to avoid making compromises that affect the ability to deliver the minimum required performance. The limits of those design compromises are set by functional requirements of the users and the environment the system operates in. In a swimming pool, where the minimum requirement might be to understand a simple command such as Everyone clear the pool a speech reinforcement could have an STI rating as low as 0.45 and it could still provide effective communication. In a classroom where the minimum requirement may be to successfully communicate the properties and uses of polytetrafluoroethylene, a minimum STI rating may need to be as high as 0.65. If the speech reinforcement system fails to meet the design goals, it fails as a useful item, no matter how inexpensive or easy to install it might have been. This would be equivalent to using strings of Christmas lights for general room lighting instead of fluorescent lights, because the Christmas lights are cheaper and easier to install and do not require an electrician or other specialized technician. The installation may have been inexpensive, but they arent very effective in providing a suitable work light level. Speaker System Issues Of the speaker types and configurations reviewed in this study, there are a few consistent behaviors that were observed. Two Way Surface Mount Speakers While there are some minor variations between the various brands and models of small two way speakers (5 woofer and a separate dome tweeter), but they all share a common problem of a collapse of the vertical coverage pattern at the crossover point where both the woofer and tweeter are operating at the same level. This is typically an octave wide band at 3-4kHz, right in one of the more critical bands for articulation of consonants. The only small speakers that dont behave in this fashion are devices with the tweeter coaxially located in the centre of the woofer. The effect of that vertical coverage collapse is often an octave wide frequency band that can have a coverage pattern as narrow as 20 degrees vertical. If these boxes are turned sideways and mounted horizontally, as is commonly done when yoke mounting them, that becomes a very narrow horizontal coverage pattern in that critical octave. This makes placement and aiming of these small two way speakers very critical if full speech band performance is going to be delivered
McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 9

to the seats. The mounting positions that would provide the best overall level uniformity are not necessarily the mounting positions that work best for 3-4kHz. Single Speaker Solutions with Spherical Spreading Wavefronts All single speaker solutions that have a coverage pattern that approximates a spherical wavefront, and therefore inverse square law attenuation of the direct sound, have the same basic limitation in covering a large area uniformly. This would include the single TOA F121C speaker, the Soundsphere, a small speaker in one corner, the four speaker pyramid ceiling speaker, an NXT flat panel, etc. This is the same basic behavior as an unaided talker. The only way to be loud enough in the back is to talk louder at the front. For any single speaker installation there will be a significant drop in direct sound level with increasing distance, and this will always result in having the sound levels be much higher for listener close to the speaker than the most distant listener would experience. If the distance is large enough it may be too loud for the closest listener when it is adequately loud at the most distant listener. These systems will always have an issue where the feedback threshold with an open microphone will limit the maximum available sound level. By the time the system gain is increased to the point where it would be louder than the unaided voice, it will likely be prone to feedback when the teacher was close to the speaker and will operate safely when the teacher is furthest away from the speaker. If this is the only way the system is used, where the single loudspeaker is on the opposite side of the room from the teacher, then there can be some small benefit to the increased sound level at the greatest distance from the teacher, but that also limits how close the teacher can get to the speaker and the students on that side of the room. Typically any room that can be covered by a single small speaker could just as effectively be covered by a talker with louder voice, and the primary acoustical problem would be elevated background noise levels and not excess reverberation. Single Speakers with Shaped Coverage The ToGo and Logia line arrays show some real promise for good sound level uniformity over a moderate size of classroom. The key aspect in using any line array is to make sure the audience seating is kept far enough away from the loudspeaker so that they are beyond the nearfield anomalies in a line array, and instead they are all seated in the area where there is minimal change in level or frequency response. The other major issue to consider when placing line arrays is that they do have a nominal cylindrical fan shaped coverage pattern. That speaker coverage should not be pointed at the rear and side walls, the fan shaped coverage should be tilted down so that the centre of the coverage pattern aims at the most distant seated listener. In that scenario, most of the sound is directed into the seating and not at the rear and side walls. All the seats are typically within the vertical coverage if the most distant seat is the selected target. The speaker should be mounted high enough that the entire coverage pattern does not have to make its way through a forest of heads to get to the rear row. For elementary school seating, the bottom of the line array should be at or just above seated student head height and at least 6-8 away from the closest listener. If the line array is tilted down from that position and aimed at the most distant seat, the direct sound coverage should be fairly uniform. Even line arrays may not be the best choice when excess reverberation is the primary acoustical problem as they are still limited by the available direct-to-reverberant ratio that a single speaker can provide in a reverberant room. They might have 3-5dB better D/R than a conventional single speaker, but that may not be adequate in all reverberant classroom settings.
McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Multiple Speaker Locations and Positioning

Page 10

When installing multiple ceiling or wall mounted speakers, the two most important considerations are having the distance between the speaker and the listener be as short as practically possible; and having the density of the speakers be high enough to keep the notches caused by destructive interference narrower than 1/3 octave in any seat. If the speaker spacing is equal to the ceiling height, that will often provide enough overlap in coverage to keep the cancellation notches below 1/3 of an octave. Some ceiling loudspeakers have a broad enough 2-4kHz coverage pattern to allow center to center spacing up to 1.5 times the ceiling height, but not all ceiling speakers are created equal. The installation of multiple ceiling speakers will often provide a higher D/R ratio in a reverberant environment, primarily because the speaker is closer to the listener, and all listeners have a similar speaker to listener distance. When the reverberation time becomes very long, increasing the number of speakers can begin to reduce to the intelligibility. This can be calculated in advance of installation using conventional spreadsheet sound system calculations, software is not required. Room Acoustics and Speaker Selection In rooms where excess background noise is the major problem and reverberation time is fairly short and controlled, almost any loudspeaker can provide an improvement in speech intelligibility by making speech louder and improving signal to noise. The primary design consideration is uniform sound level and frequency response. That might be provided using a single line array, or multiple ceiling speakers, or even the side wall mounted speakers in a narrow enough room. In rooms were excess reverberation is the primary problem and there are also problems with elevated background noise exacerbated by the excess reverberation time, the speaker selection process needs to be considered more carefully. Ceiling speakers are often a good choice in this setting. Having the speakers closer to the listeners definitely helps improve intelligibility by improving both S/N ratio and D/R ratio. Single speaker solutions may not deliver adequate performance in highly reverberant environments. Available Amplifier Power One of the most significant performance aspects of sound system design is having adequate headroom for the signals being reproduced. Many small loudspeakers have very low efficiency, often in the range of 80-85dB @1watt @1metre. To deliver reinforced speech levels higher than unaided speech with a dynamic signal like speech often requires much more power than is provided. One of the observations we have made of several sound field systems is that they are drastically underpowered and so they often exhibit severe clipping just as they are producing a sound level that would be louder than an unaided talker. It is not unreasonable to assume a 20dB dynamic range for a speech source unless it is heavily compressed. Any speech reinforcement system should be designed so that it will not clip in normal operation. The system gain limit should be the feedback threshold and not the amplifier output power.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Sound Field Configuration: ToGo corner mounted at 7 to the center of speaker

Page 11

500Hz direct sound field The ToGo is mounted at 7 to the center of the speaker and the speaker has a 10 degree downtilt so that the axis of the speaker is aimed at the ear plane in the opposite corner of the room. The ToGo achieves a large area of uniform coverage even though there are some large variations close to the speaker. The maximum level is 91.28dB and the minimum level is 76.6dB for a total variation of14.68dB. The seating area is within 6-7dB in uniformity.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 12

2000Hz direct sound field At 2000hz there is a very large uniform area of coverage. The maximum level is 88.04dB and the minimum level is 67.77dB for a total variation of 20.77dB. The bulk of the seating area is within 5-6dB of level uniformity.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 13

4000Hz direct sound field At 4000Hz the patch is a different shape but similar in overall size. The maximum level is 89.61dB and the minimum level is 68.07 for a total variation of 21.54dB. The bulk of the seating is within 4-5dB although there are some large blue areas near the seating on the side near the speaker system.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 14

2000Hz Total SPL (direct + reverberant) at current reverberation time The maximum level is 88.75dB and the minimum level is 80.76dB for a total variation of 7.99dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 15

2000Hz Total SPL (direct + reverberant) at 0.75 second reverberation time With the 0.75 second reverb time the level variation drops to 3.99dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 16

2000Hz Total SPL (direct + reverberant) at 1.0 second reverberation time At 1.0 second reverb time the level variation drops 3.19dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 17

RaSTI with current background noise and current reverb time The single ToGo has a very large area of superb STI value. The maximum is 0.77 and the minimum is 0.72, but the area of 0.74 to .75 is very large.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 18

RaSTI with NC-45 background noise and current reverb time Increasing the noise level does not affect the STI values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 19

RaSTI with NC-55 background noise and current reverb time Increasing the background noise to NC-55 reduces the minimum STI value but has no significant effect on the area that scores good values of 0.74 to 0.75.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 20

RaSTI with NC-45 background noise and 0.75 second reverb time Increasing the reverb time to 0.75 seconds reduces the maximum STI values and increases the range of STI values. The maximum value is 0.69 and the minimum value is .58. The area that achieves 0.62 to 0.65 is quite large and uniform.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 21

RaSTI with NC-55 background noise and 0.75 second reverb time Increasing the noise level to NC-55 has no significant affect on the STI values

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 22

RaSTI with NC-45 background noise and 1.0 second reverb time Increasing the reverb time to 1.0 seconds reduces STI values overall. The maximum value is 0.65 and the minimum is 0.54. The uniformity of the STI values is noticeably reduced from the 0.75 second values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 23

RaSTI with NC-55 background noise and 1.0 second reverb time Increasing the noise level to NC-55 does not significantly affect the STI values, there is only a slight reduction. Any single source speaker solution is more affected by excess reverberation than a distributed speaker solution.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 24

2000Hz C50 with the current reverb time The ToGo provides very uniform C50 values over a large area, most of the seating is between 1115dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 25

2000Hz C50 with 0.75 second reverb time Increasing the reverb time to a 0.75 second drops to a range of 2-7dB, with the bulk of the seating area ranging from 3-5dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo corner mounted

Page 26

2000Hz C50 with a 1.0 second reverb time As with any single source speaker system configuration, the C50 values drop off with increasing reverb time. At 1.0 seconds the range of C50 values drops from 5-0dB. The bulk of the seating has a value between 0-3dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Sound Field Configuration: ToGo on Desk or Table top

Page 27

500Hz direct sound field With the ToGo speaker closer to the ear plane and closer to the listeners, the near field anomalies are more likely to interact with the teachers microphone. The maximum level is 95.77dB and the minimum level is 77.21dB for a total variation 18.56dB. The bulk of the level difference happens within the first 4 radius of the loudspeaker. The bulk of the seating area is covered with a uniformity of 6-7dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 28

1000Hz direct sound field The 1000Hz coverage is more constrained at this mounting height. The maximum level is 88.66dB and the minimum level is 68.8dB for a total variation of 19.86dB. The bulk of the level variation occurs in the nearfield of the speaker. There is about 10dB of variation throughout the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 29

2000Hz direct sound field The 2000Hz coverage pattern appears as a largely square footprint at this mounting height. The maximum level is 93dB and the minimum level is 58.91dB for a total variation of 34.04dB, with the bulk of the variation in the nearfield of the speaker location. The bulk of the seating area is covered to a uniformity of about 6dB. The nearfield exhibits some extreme behavior that could generate feedback when the microphone is in close proximity to the loudspeaker.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 30

4000Hz direct sound field The 4000Hz coverage almost grazes the listening plane with the speaker sitting on a desktop and no tilt, there are some significant level variations in this band. The maximum level is 92.98dB and the minimum level is 66.18dB for a total level variation 26.8dB. A large area of the seating is within 3dB but there are also several zones within the seating that are about 7-8dB lower in level.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 31

2000Hz Total SPL (direct + reverberant) at current reverb time The level variation in the Total SPL is exaggerated by the very high level in close proximity to the speaker painted on the listening plane. The maximum level is 93.24dB and the minimum level is 80.57dB for a total variation of 12.67dB. Most of the seating area is within 4dB in level/

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 32

2000Hz Total SPL (direct + reverberant) at 0.75 second reverb time At 0.75 second reverb time the far field coverage exhibits less variation. The maximum level is 93.82dB and the minimum level is 86.17dB for a total variation of 7.65dB. Most of the seating area has a level variation of about 2dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 33

2000Hz Total SPL (direct + reverberant) at 1.0 second reverb time Increasing the reverb time to 1.0 seconds reduces the total variation further. The maximum level is 94.1dB and the minimum level is 87.6dB for a total variation of 6.5dB with less 2dB of variation over the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 34

RaSTI with current background noise and current reverb time The ToGo provides similar STI performance to the corner mounting, the maximum level is artificially high as the highs core appears on the desk top the speaker is sitting on. The maximum STI value is 0.78 and the minimum value is 0.71. A large area of the seating is still 0.74 to 0.75 in STI which is very good.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 35

RaSTI with NC-45 background noise and current reverb time Increasing the noise level does not have a significant effect on STI values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 36

RaSTI with NC-55 background noise and current reverb time Increasing the noise level to NC-55 reduces the minimum STI values, but its important to note that those low scores are behind the loudspeaker.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 37

RaSTI with NC-45 background noise and 0.75 second reverb time Increasing the reverb time to 0.75 seconds drops the overall STI levels, the maximum being 0.72 and the minimum being 0.58. Significantly, the STI values in the primary coverage area have dropped from 0.74 to 0.66-0.62. This is typical of the effects of reverberation on any single source speaker configuration.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 38

RaSTI with NC-55 background noise and 0.75 second reverb time Increasing the noise level to NC-55 produces no further reduction in STI showing that the performance is most affected by excess reverberation.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 39

RaSTI with NC-45 background noise and 1.0 second reverb time Increasing the reverb time to 1.0 seconds produces another major drop in STI values overall. The maximum value is 0.69 and the minimum value is 0.53. Significantly, the area inside the direct coverage has dropped from 0.66 to 0.62 down to 0.62 to 0.57. The maximum and minimum values are all very close to the loudspeaker. Excess reverberation is the major limiting acoustical parameter for maintaining good speech intelligibility with the ToGo.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 40

RaSTI with NC-55 background noise and 1.0 second reverb time There is no significant effect on STI value by increasing the noise level to NC-55.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 41

2000Hz C50 with current reverb time With the ToGo at desktop or table height, the C50 values are artificially high very close to the speaker because these results are painted on an ear plane at 36 above the floor. The maximum value of 23dB happens directly in front of the speaker. The C50 values over the seating area range from 11-16dB, very similar to most other systems in the current acoustical environment.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 42

2000Hz C50 with 0.75 second reverb time Increasing the reverb time reduces the C50 values. The seating area ranges from 2-5dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo on Desk or Table top

Page 43

2000Hz C50 with 1.0 second reverb time The C50 values drop to 3-0dB over the seating area with the reverb time at 1.0 seconds.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 44

Sound Field Configuration: single small speaker (flat panel, etc.) mounted in corner

500Hz direct sound field A single speaker mounted in a corner is entirely dominated by inverse square law for direct sound level drop off and will be seriously affected by increasing reverb time. The maximum level is 92.13dB and the minimum level is 77.18dB for a total level variation of 14.95dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 45

1000Hz direct sound field The maximum level is 90.19dB and the minimum level is 74.32dB for a total level variation of 15.87dB. There is more than 10dB of level drop across the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 46

2000Hz direct sound field The maximum level is 90.03dB and the minimum level is 74.28dB for a total level variation of 15.75dB. There is almost 12dB of level variation across the seating area. Any effort to get the speaker loud enough in the far corner will result in the speaker being 12dB louder near the source. If the teacher has to walk near the speaker it is likely to be a feedback problem.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 47

2000Hz Total SPL (direct + reverberant) at current reverb time In the current reverb time there is 6.93dB of level variation in the room.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 48

2000Hz Total SPL (direct + reverberant) at 0.75 second reverb time As the reverb time increases the level variation drops to 3.36dB, an indicator of large variations in STI value with increasing reverb time.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 49

2000Hz Total SPL (direct + reverberant) at 1.0 second reverb time With the 1.0 second reverb time the level variation drops to 2.66dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 50

RaSTI with current background noise and current reverb time With the current acoustical conditions a single speaker can provide good STI values, but may still be a feedback problem.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 51

RaSTI with NC-45 background noise and current reverb time Increasing the noise level to NC-45 does not impact the STI values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 52

RaSTI with NC-55 background noise and current reverb time Increasing the noise level to NC-55 does not have a significant impact on STI values. What that means is that the speaker has enough sound level to provide the needed S/N ratio at this reverb time. S/N ratio adjustment can be achieved by level adjustment whereas performance limits caused by the reverb time in the room are a factor of direct/reverb ratio and cant be corrected by turning it up.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 53

RaSTI with NC-45 background noise and 0.75 second reverb time As expected increasing the reverb time drops the overall STI values significantly and really increases the range of STI values in the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 54

RaSTI with NC-55 background noise and 0.75 second reverb time Increasing the noise level to NC-55 doesnt have any significant impact on STI values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 55

RaSTI with NC-45 background noise and 1.0 second reverb time Increasing the reverb time to 1.0 seconds causes a further drop in STI values in the seating area and increases the spread in STI values as well.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 56

RaSTI with NC-55 background noise and 1.0 second reverb time Increasing the noise level to NC-55 has no significant effect on the STI values.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 57

2000Hz C50 with current reverb time The single speaker works reasonably well when the reverb time is short. The range of C50 through the seating area is 15-10dB.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 58

2000Hz C50 with 0.75 second reverb time Increasing the reverb time to 0.75 seconds drops the range of C50 to 5-2dB through the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study Single small speaker mounted in corner

Page 59

2000Hz C50 with 1.0 second reverb time At a 1.0 second reverb time the C50 values drop to 3-0dB through the seating area.

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study ToGo 3D polar response prediction produced by EASE 4.0

Page 60

500Hz

1000Hz

1600Hz

2000Hz

2500Hz

3150Hz

McSquared System Design Group Inc.

Phonic Ear Sound Field Speaker Coverage Study

Page 61

4000Hz

5000Hz

6300Hz

8000Hz

McSquared System Design Group Inc.

Anda mungkin juga menyukai