WO2001089083A1

WO2001089083A1 - Complementary transfer function design of crossover filters in loudspeaker systems

Info

Publication number: WO2001089083A1
Application number: PCT/US2000/013667
Authority: WO
Inventors: Daniel P. Anagnos
Original assignee: Sony Electronics Inc.
Priority date: 2000-05-18
Filing date: 2000-05-18
Publication date: 2001-11-22
Also published as: AU2000251414A1

Abstract

The invention provides a system and method of designing complementary transfer functions (CTF) [109] for crossover filters that adjust for various crossover considerations. According to the invention, a computer-implemented system and procedure, in a specific audio environment, processes crossover factors [110], designs a CTF filter in a computer simulation [111], correlates measurements of crossover parameters with the computer simulations [112], and adjusts the CTF design [14] until there is a high or predetermined degree of correlation between the computer simulation and the crossover measurements. In particular, the inventive system and method achieves proper blending between chosen sets of subwoofers [105] and satellite speakers [101] in loudspeaker systems.

Description

COMPLEMENTARY TRANSFER FUNCTION DESIGN OF CROSSOVER FILTERS IN LOUDSPEAKER SYSTEMS

FIELD OF INVENTION

The present invention is in the field of crossover filter design in

loudspeaker systems.

BACKGROUND OF INVENTION

The frequency range of human hearing spans from about 20 Hertz (Hz) to

20 kilo-Hertz (kHz). In order to reproduce music in this frequency range, it is

desirable for the loudspeaker system to reproduce the full audio spectrum.

However, most loudspeakers are incapable of accurately reproducing the entire

audio spectrum in the range of 20 Hz to 20 kHz. One approach to this problem is

the use of a combination of one or more low-frequency (LF) speakers, i.e.,

subwoofers, which are designed specifically for clear sound reproduction in the LF

range, and one or more satellite speakers, which are designed to maintain a wide

dispersion pattern in the mid- and Mgh-frequency (HF) ranges. Such a

combination of speakers in different audio ranges can reproduce the full audio spectrum for human hearing. However, configuring a loudspeaker system with

separate speakers designed for different audio ranges has its drawbacks. In

general, the different speakers have different audio sensitivities and dispersion characteristics. Good LF reproducing speakers tend to perform poorly in the HF

range and, similarly, good HF reproducing speakers generally perform poorly in

the LF range. Since HF speakers tend to be more sensitive than LF speakers, the

satellite speakers would have greater loudness amplitude than the subwoofers. The

power-handling capabilities of HF speakers is also quite limited, because signals at

the LF range can drive the HF speakers through very large displacements and can

damage the speaker unit at significant power levels. Furthermore, in the range

where the frequency response of the LF and HF speakers overlap, there usually

exists either a large rise or a large dip in sound pressure amplitude (measured in

decibels, or dB) and significantly large phase anomalies and discontinuities. This

is usually perceived as a "lack of blending or coherence" between the subwoofers

and the satellite speakers.

In addition to power handling and loudness amplitude problems, crossover

design can solve problems resulting from the frequency response overlap by

filtering out the low frequencies from signals being applied to the HF speakers, and

by filtering out the high frequencies from signals being applied to the LF speakers.

There are two distinct ways to implement a loudspeaker crossover: passive and

active design. Passive crossover design employs only passive electrical

components (such as resistors, capacitors and inductors) and acts directly on the

speaker level signal from the power amplifier. The main advantage of passive

crossover design is that it allows one power amplifier to drive a complete full- range loudspeaker system. However, passive designs are usually forced to cross

over at relatively high frequencies (above 150 Hz) because of physical constraints

of component size and power handling considerations. As a result, audio

reproduction performance of loudspeaker systems with passive crossover design at

very low frequencies (below 150 Hz) is generally inferior to active crossover

designs because of component non-linearities and signal distortion.

The alternative to a passive crossover design is active design. Active

crossover design uses electrical or electronic filters to modify the frequency

response of signals being applied to the speakers. In a two-way system, for

example, a full-spectrum signal passes through a high-pass filter (HPF) on the way

to the HF speakers while the signal to the LF speakers passes through a

complementary low-pass filter (LPF). The HPF passes signal information above a

selected frequency, referred to as the crossover frequency, and attenuates the

components of the signal that fall below the crossover frequency. Similarly, the

LPF passes information below the crossover frequency and attenuates frequency

components above the crossover frequency. Proper selection of the crossover

frequency is essential for delivering a smooth, flat response in the audio spectrum.

In an ideal system, the crossover filter directs all signal information above

the crossover frequency to the HF speakers and all signal information below the

crossover frequency to the LF speakers. The crossover filter ideally would serve to

attenuate the HF response of the subwoofers ~ by minimizing the levels of signals input to the subwoofer at frequencies above the crossover frequency. However, in

passive crossover designs, the LPF, has a significant series resistance which affects

the LF response of the subwoofer. This series resistance can act to reduce the

damping factor of the amplifier and therefore raise the total "Q" factor of the

subwoofer system. More significantly, however, is the effect upon the subwoofer

transducer itself, skewing the Thiele/Small design parameters such that the system

becomes de-tuned. The results of these two effects can produce a peaked or

shelved LF response, which is highly undesirable.

Another consideration is the "slope" of the crossover filter. Crossover

filters have the characteristic that for each frequency octave beyond the crossover point, the response of the filter is reduced by a certain amount. This characteristic

is referred to as the slope of the filter. Typical crossover filters have slopes of 6,

12, 18 or 24 dB per octave. In other words, the crossover filter reduces the

response by 6, 12, 18 or 24 dB with each octave change in frequency beyond the

crossover point. As an example, in a system with a pair of filters with a crossover

frequency at 1 kHz and a filter slope of 18 dB per octave, the response of each

filter is down 3 dB at the crossover point. Above 1 kHz the crossover slope

reduces the response of the LF speaker at a rate of 18 dB per octave. This means

that at 2 kHz the response of the woofer is reduced by 18 dB and at 4 kHz the

response is reduced by 36 dB, compared to the response at the crossover frequency at -3 dB. Likewise, the response of the HF speaker is reduced by 18 dB at 500 Hz and by 36 dB at 250 Hz. In comparison, a 6-dB-per octave crossover slope reduces the response of the LF speaker only by 12 dB at 4 kHz and the response of the HF

speaker is down only by 12 dB at 250 Hz. Because of power handling

considerations in the HF speaker, the sharper crossover slope of 18 dB per octave

allows the use of a lower crossover frequency than either 6- or 12-dB-per-octave

filter slopes. In other words, because of filter slope, for a given crossover

frequency, the smaller the filter slope utilized in the crossover, the greater the

power handling capability required for an HF speaker.

Existing loudspeaker systems with active crossover design are capable of

establishing a crossover point at very low frequencies (as low as 30 Hz) and offer varying degree of adjustability in filter crossover frequency, level, phase, slope and

Q. This adjustability feature is essential for ensuring a flat response, or audio

blending of sounds in different frequency ranges in multi-way loudspeaker

systems. The problem with such existing systems is that, even with significant

adjustment freedom and flexibility, the transfer functions of the crossover filters

are not correct for proper blending. General purpose transfer functions are utilized

because of the complexity and difficulty in designing and implementing

specialized transfer functions. Exemplary general purpose transfer functions

include Bessel, Butterworth and Linkwitz-Riley. See pp. 12-5 through 12-29,

Electronics Engineers' Handbook, 3d Ed., Donald G. Fink and Donald

Christiansen, eds., which is incorporated by reference herein. However, general purpose transfer functions are exceedingly difficult to adjust, e.g., by the end user,

in a subwoofer/satellite loudspeaker system in a specific audio environment, to

obtain satisfactory blending and overall performance. The adjustment process

often involves a great deal of empirical data gathering and trial-and-error tests in order to configure a crossover filter best suited for a specific audio environment.

Conducting empirical analysis and tests becomes increasingly cumbersome and

time-consuming in more complex systems. Because the final system adjustability

cannot account for the transfer functions of the individual loudspeakers (which will

combine with the transfer functions of the filters), it remains unlikely that satisfactory performance will be attained using generalized filters.

Specialized transfer functions designed for specific audio environments or

fixed sets of LF and HF speakers can partially eliminate some of the

aforementioned disadvantages of general purpose transfer functions. However,

configuring specialized transfer functions, as they are designed with a fixed set of

parameters, also limits the choice of satellite speakers. Pre-configured specialized

transfer function filters for crossover purposes put constraints on system design

when the existing system provides only predetermined choices of subwoofers and

satellite speakers.

There is therefore a need in the art of loudspeaker system design for a

computer-implemented system and method that can create complementary transfer

functions (CTF) and associated filters with adjustability in respect to factors affecting CTF and filter considerations. There is a need for a system and method

of design procedure that allows ideal blending between chosen sets of subwoofers

and satellite loudspeakers. In short, a speaker design system and method is needed

that self-corrects CTF design parameters when configuring crossover filters,

without the aforementioned disadvantages of cumbersome empirical analysis and

tests.

SUMMARY OF THE INVENTION

The present invention is related to crossover filter design in loudspeaker

systems. In particular, the invention is a method of designing complementary

transfer functions in crossover filters for subwoofer and satellite loudspeakers

which in combination deliver a flat amplitude response and smooth, integrated

phase response across the audio spectrum.

The invention is a unique system and method of designing complementary

transfer functions (CTF) for crossover filters that adjust for various crossover

considerations. According to the invention, a computer-implemented system and

procedure, in a specific audio environment, processes crossover factors, designs a

CTF filter in a computer model, correlates measurements of crossover parameters

with the computer model, and adjusts the CTF design until there is a high or

predetermined degree of correlation between the computer model and the crossover

measurements. The invention eliminates the need for lengthy, cumbersome

empirical data gathering and trial-and-error analysis in the final end-user adjustment process. In particular, the inventive system and method achieves

proper blending between chosen sets of subwoofers and satellite speakers in

loudspeaker systems. The invention greatly reduces development time, while

improving the quality and completeness of the final design result.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 illustrates the method and procedure for complementary transfer

function (CTF) design in crossover filters with low-pass filters only.

Figure 2 illustrates the method and procedure for CTF design in crossover

filters with low-pass filters and high-pass filters.

Figures 3 A and 3B are schematics of different embodiments of the invention in an analog and a digital loudspeaker system, respectively.

DETAILED DESCRIPTION

Preferred embodiments of the invention are hereinafter described. It is

noted that, while the system and method of designing complementary transfer

functions (CTF) for crossover filters is implemented at the electrical circuit level

for the preferred embodiments, that should be viewed as a particular way of

implementing the invention which does not to limit the scope of the invention. It

will readily be seen by those skilled in the art that this CTF design system and

method could be implemented with analog loudspeaker systems as well as with digital signal processing (DSP) in digital systems. In addition, the preferred embodiments are presented in the context of a two-way loudspeaker system in a

fixed audio environment such as a typical half-space room. It is intended that the

principles of the invention be applied for multi-way loudspeaker systems in any

audio environment or channelization arrangement.

The preferred embodiments disclosed hereinafter operate in a certain audio

environment with pre-determined parameters. Specific parameters such as the

damping factor, level are pre-determined before implementing the design system

and method according to the invention. Such pre-determined parameters also

include operation in a typical home environment such as a half-space room in which the wavelengths of sound near the crossover frequency region are large

relative to the room dimensions. For the preferred embodiment, all acoustical

measurements are made at a fixed reference point in an anechoic environment for

frequencies above 200 Hz, such as an exemplary reference point at 3 meters, on-

axis, and at a listening height of one meter. Moreover, half-space acoustical

loading of speaker output is assumed for both subwoofers and satellite speakers at

all frequencies below 200 Hz (including the system crossover frequency). The

crossover frequency at which the preferred embodiments are implemented is set at

or below 80 Hz. The difference in phase response between corner loading and

center room loading is minimal as a result of the aforementioned parameters. Note that satellite speakers in all preferred embodiments hereinafter disclosed can either

be passive or active. A passive satellite speaker only uses passive components (e.g., resistors, capacitors, inductors) to divide the energy (supplied from an

amplifier) between the individual transducers in the loudspeaker system. In

contrast, an active satellite speaker uses low-level, electronic components

(including transistors, integrated circuits or vacuum tubes) to divide the energy

between individual amplifiers and transducers in a loudspeaker system.

The system and method of the invention are hereinafter particularly

described with respect to several illustrative embodiments. The first such

illustrative embodiment, a system and method of crossover design in CTF filters,

utilizing low-pass filters only, is depicted schematically in Figure 1. In this

embodiment, the active filtering is provided only for the subwoofer, utilizing a

low-pass filter (LPF).

As will be seen in the figure, the methodology for this embodiment begins

with a measurement of the acoustical response of an LF and an HF speaker to be

operated in combination and comparison of such responses with predicted

responses based on a computer model for that speaker using a trial transfer

function. Considering the specific steps of the methodology, in Step 101,

measurements are obtained for the acoustical response of the satellite (HF) speaker

of a given two-way loudspeaker system in a typical half-space room.

Measurements are taken according to the close-miking technique that approximates

half-space loading in-room conditions. The invention uses an audio analyzer that

performs an MLS-type (i.e., maximum length sequence) acoustical measurement

implemented with a pseudo-random noise source and correlated fast Fourier transform (FFT) calculations. In an exemplary measurement according to the

close-miking technique, microphones are placed at 1.5 mm from the center of each

low-frequency (LF) transducer in the speaker, with an input voltage of 2.83 volts.

The acoustical parameters measured are amplitude and phase versus frequency

from 10 Hz to 500 Hz. The data are then exported from the audio analyzer to the

correlation stage in Step 103 for comparison with the response generated by the

computer simulated model in Step 102.

In Step 102, the computer LF model of the selected satellite loudspeaker in

such a typical half-space room is formulated. Such a computer model can be

formulated with conventional computer-simulated loudspeaker analysis tools such

as LEAP (Loudspeaker Enclosure Analysis Program), sold by Linear X Systems,

Inc. of 9500 Tualatin-Sherwood Road, Tualatin, Oregon 97062 USA. In Step 103,

the response generated by the computer simulated model is compared and

correlated with the acoustical response of the satellite speaker. Based on such

correlation, the computer LF model formulated in Step 102 is adjusted at Step 10

until a high or predetermined degree of correlation is reached. Adjustments to the

computer simulated model which are made to obtain an acceptable correlation with

the acoustical measurements are performed manually. At this point, parameters of

the computer model in the simulation program (such as LEAP) are adjusted by the

system designer. However, adjustments to the complementary transfer function

parameters are made automatically by the computer simulation program. Acceptable correlation is reached when the measurements agree within ±1

dB in amplitude with the simulation of the computer model. Moreover, phase

measurements must agree within ±15 degrees with the simulation of the computer

model. Note that the correlation bandwidth is 10 Hz to 500 Hz. In Step 104,

because there is a high or predetermined degree of correlation, parameters in the

computer model become the basis for configuring the LF transfer function of the

satellite loudspeaker in the two-way system. The overall acoustical transfer

function of the satellite speaker is determined and accounted for in the design

embodied in the present invention.

Similarly, considering the specific steps of the methodology in Step 105,

measurements are obtained for the acoustical response of the subwoofer (LF)

speaker of the two-way loudspeaker system in a typical half-space room.

Acoustical response measurements of the subwoofer speaker are taken in a similar

fashion as those of the subwoofer satellite speaker. Measurements are taken

according to the close-miking technique that approximates half-space loading in-

room conditions. The invention uses an audio analyzer that performs an MLS-type

(i.e., maximum length sequence) acoustical measurement implemented with a

pseudo-random noise source and correlated fast Fourier transform (FFT)

calculations. In an exemplary measurement according to the close-miking

technique, microphones are placed at 1.5 mm from the center of each low-

frequency (LF) transducer in the speaker, with an input voltage of 2.83 volts. The

acoustical parameters measured are amplitude and phase versus frequency data from 10 Hz to 500 Hz. The data are then exported from the audio analyzer to the

correlation stage in Step 107 for comparison with the response generated by the

computer simulated model in Step 106.

In Step 106, the computer LF model of the selected subwoofer loudspeaker

in such a typical half-space room is formulated. In Step 107, the response

generated by the computer simulated model is compared and correlated with the

acoustical response of the subwoofer speaker. Based on such correlation, the

computer LF model formulated in Step 106 is adjusted at Step 12 until a high or

predetermined degree of correlation is reached. Adjustments to the computer

simulated model which are made to obtain an acceptable correlation with the

acoustical measurements are performed manually. At this point, parameters of the

computer model in the simulation program (such as LEAP) are adjusted by the

system designer. However, adjustments to the complementary transfer function

parameters are made automatically by the computer simulation program.

Acceptable correlation is reached when the measurements agree within ±1

dB in amplitude with the simulation of the computer model. In addition, phase

measurements must agree within ±15 degrees with the simulation of the computer

model. Note that the correlation bandwidth is 10 Hz to 500 Hz. In Step 108,

computer model become the basis for configuring the LF transfer function of the

subwoofer loudspeaker in the two-way system. The overall acoustical transfer function of the subwoofer speaker is determined and accounted for in the design embodied in the present invention.

In Step 109, based on the adjusted LF transfer functions formulated in

Steps 104 and 108, a complementary transfer function (CTF) low-pass filter (LPF)

circuit is developed. In Step 110, measurements of the acoustical response of the

satellite/subwoofer loudspeaker system implemented with the CTF LPF circuit

developed in Step 109 are taken in a typical half-space room audio environment.

In Step 111, a computer model is formulated for a satellite/subwoofer loudspeaker

system implemented with the CTF LPF circuit developed in Step 109. In Step 112,

the response generated by the computer simulated model is compared and

correlated with the acoustical response of the satellite/subwoofer loudspeaker

system. Based on such correlation, the computer model formulated in Step 111 is

adjusted at Step 14 until a high or predetermined degree of correlation is reached.

Adjustments to the computer simulated model which are made to obtain an

acceptable correlation with the acoustical measurements are performed manually.

At this point, parameters of the computer model in the simulation program (such as

LEAP) are adjusted by the system designer. However, adjustments to the

complementary transfer function parameters are made automatically by the

computer simulation program. In Step 113, a qualitative assessment of audio reproduction performance is

conducted on the satellite/subwoofer system being developed. This step examines quality factors such as harmonic and intermodulation distortion, dynamic range (e.g., maximum volume of the speaker), transient response (e.g., quickness of bass

response), clarity of bass and midrange reproduction and resolution.

If the quality assessment suggests further adjustment in the CTF LPF

circuit, Step 109 is re-initiated and the design process continues through Steps 109,

110, 111, 112 and 113. Typically, the crossover frequency or center frequency of

the CTF filter is adjusted. The order or slope of the CTF filters can be adjusted as

well. When these parameters are adjusted, the computer model determines the

correct filter Q and gain. The CTF is determined to be correct when the acoustical

output from the subwoofer loudspeaker and the acoustical output of the satellite

loudspeaker delivers a sum flat response in terms of sound pressure amplitude, of

less than plus or minus 3 dB variation and exhibits a smooth combined phase

response. If the quality assessment in Step 113 results in no more adjustments,

then the design process is complete.

The second illustrative embodiment, a system and method of crossover

design in CTF filters, utilizing low-pass filters and high-pass filters, is depicted

schematically in Figure 2. Considering the specific steps of the methodology in

Step 201, measurements are obtained for the acoustical response of the satellite

speaker of a given two-way loudspeaker system in a typical half-space room.

performs an MLS-type (i.e., maximum length sequence) acoustical measurement

The acoustical parameters measured are amplitude and phase versus frequency data

from 10 Hz to 500 Hz. The data are then exported from the audio analyzer to the

correlation stage in Step 203 for comparison with the response generated by the

computer simμlated model in Step 202.

In Step 202, the computer LF (low frequency) model of the selected

satellite loudspeaker in such a typical half-space room is formulated. In Step 203,

the response generated by the computer simulated model is compared and

correlated with the acoustical response of the satellite speaker. Based on such

correlation, the computer LF model formulated in Step 202 is adjusted until a high or predetermined degree of correlation is reached. Adjustments to the computer

simulated model which are made to obtain an acceptable correlation with the

computer model in the simulation program (such as LEAP) are adjusted by the

system designer. However, adjustments to the complementary transfer function

parameters are made automatically by the computer simulation program.

Acceptable correlation is reached when the measurements agree within ±1

dB in amplitude with the simulation of the computer model. Moreover, phase measurements must agree within ±15 degrees with the simulation of the computer model. Note that the correlation bandwidth is 10 Hz to 500 Hz. In Step 204, because there is a high or predetermined degree of correlation, parameters in the

computer model become the basis for configuring the LF transfer function of the

satellite loudspeaker in the two-way system. The overall acoustical transfer

function of the satellite speaker is determined and accounted for in the design

embodied in the present invention.

Similarly, considering the specific steps of the methodology in Step 205,

measurements are obtained for the acoustical response of the subwoofer (LF)

speaker of the two-way loudspeaker system in a typical half-space room. In Step

206, the computer LF (low frequency) model of the subwoofer loudspeaker in such

a typical half-space room is formulated. In Step 207, the response generated by the

computer simulated model is compared and correlated with the acoustical response

of the subwoofer speaker. Based on such correlation, the computer LF model

formulated in Step 206 is adjusted until a high or predetermined degree of

correlation is reached. Adjustments to the computer simulated model which are

made to obtain an acceptable correlation with the acoustical measurements are

performed manually. At this point, parameters of the computer model in the

simulation program (such as LEAP) are adjusted by the system designer.

However, adjustments to the complementary transfer function parameters are made

automatically by the computer simulation program.

Acceptable correlation is reached when the measurements agree within ±1

dB in amplitude with the simulation of the computer model. Moreover, phase

measurements must agree within ±15 degrees with the simulation of the computer model. Note that the correlation bandwidth is 10 Hz to 500 Hz. In Step 208,

computer model become the basis for configuring the LF transfer function of the

subwoofer loudspeaker in the two-way system. The overall acoustical transfer

function of the subwoofer speaker is determined and accounted for in the design

embodied in the present invention.

In Step 209, based on the adjusted LF transfer functions formulated in Step

204, a complementary transfer function (CTF) high-pass filter (HPF) circuit is

developed. Similarly, in Step 211, based on the adjusted LF transfer function

formulated in Step 208, a CTF low-pass filter (LPF) circuit is developed. In Step

210, measurements of the acoustical response of the satellite/subwoofer

loudspeaker system implemented with the CTF HPF and LPF circuits developed in

Steps 209 and 211, respectively, are taken in a typical half-space room audio

environment. In Step 212, a computer model is formulated for a

satellite/subwoofer loudspeaker system implemented with the CTF HPF and LPF

circuits developed in Steps 209 and 211, respectively. In Step 213, the response

generated by the computer simulated model is compared and correlated with the

acoustical response of the satellite/subwoofer loudspeaker system. Based on such

correlation, the computer model formulated in Step 212 is adjusted at Step 18 until

a high or predetermined degree of correlation is reached. Adjustments to the

the acoustical measurements are performed manually. At this point, parameters of the computer model in the simulation program (such as LEAP) are adjusted by the

system designer. However, adjustments to the complementary transfer function

parameters are made automatically by the computer simulation program.

In Step 214, a qualitative assessment of audio reproduction performance is

conducted on the satellite/subwoofer system being developed. This step examines

quality factors such as harmonic and intermodulation distortion, dynamic range

(e.g., maximum volume of the speaker), transient response (e.g., quickness of bass

response), clarity of bass and midrange reproduction and resolution.

If the quality assessment necessitates further adjustment in the CTF HPF or

LPF circuit, Steps 209 and 211 are re-initiated and the design process continues

through Steps 209, 210, 211, 212, 213 and 214. Typically, the crossover frequency

or center frequency of the CTF filter is adjusted. The order or slope of the CTF

filters can be adjusted as well. When these parameters are adjusted, the computer

model determines the correct filter Q and gain. The CTF is determined to be

correct when the acoustical output from the subwoofer loudspeaker and the

acoustical output of the satellite loudspeaker delivers a sum flat response in terms

of sound pressure amplitude, of less than plus or minus 3 dB variation, and exhibits

a smooth combined phase response. If the quality assessment in Step 214 results in

no more adjustments, then the design process is complete.

The satellite loudspeaker in the embodiment in Figure 2 is high-pass

filtered or cut off below the crossover frequency. Active filtering is provided for

both the subwoofer and the satellite loudspeakers. Acoustical overlap between the subwoofer and satellite speakers is minimal in this embodiment.. Moreover, power

handling performance of the satellite speaker is greatly improved in this

implementation, while signal distortion is significantly reduced.

Figure 3A is an embodiment of the invention in an analog loudspeaker

system. In this embodiment, the complementary transfer function (CTF) filters are

designed as active analog electronic circuits which are located within the powered

subwoofer. The CTF filters are implemented in the subwoofer itself or within an

external processor. After the CTF designs are determined for each of the filters

according to the invention, they are implemented in electronic circuits that replace

conventional general purpose filters for the subwoofer and satellite speakers in the

analog system.

Referring to Figure 3A, the Analog Source 301 inputs signal information

into the loudspeaker system. The signal information passes through the Active

Analog Circuit for processing in 302, which is a CTF low-pass filter (LPF). The

CTF is used in place of the conventional general purpose filter in the subwoofer,

before the subwoofer amplifier 303. After amplification at 303, the signal

information is output by the subwoofer loudspeaker 304.

Similarly, in an analog loudspeaker system with LPF and high-pass filter

(HPF) functionalities, the Analog Source 301 inputs signal information into the

system and sends it to the Active Analog Circuit at 305. The CTF HPF in 305 is designed according to the method of the invention as embodied in Figure 2 and is

implemented in analog electronic circuitry in the Active Analog Circuit at 305. After filtering at 305, the signal information is passed on to Amplifier 306 for

amplification. The Satellite Loudspeaker at 307 then outputs the signal

information.

A variation of the analog loudspeaker system (as in Figure 3A) that can

utilize all elements of the present invention is the digital loudspeaker system in

Figure 3B. In this digital embodiment, all of the CTF filters are implemented

using digital signal processing (DSP), either within a powered subwoofer or within

an outboard processor. A DSP implementation of the CTF filters are developed

using integrated circuits and software development platforms in accordance to the

method of the invention as disclosed hereinabove.

Referring to Figure 3B, the Digital Source at 311 inputs signal information

into the system for processing at 312, which is a CTF LPF. After processing and

filtering at 312, the signal information is transmitted to the Digital to Analog

Converter (DAC) at 313 for conversion into analog information. The analog signal

information is amplified at Amplifier 314. The amplified signal information is

then output by the Subwoofer Loudspeaker at 315.

Similarly, in a digital loudspeaker system with LPF and HPF

functionalities, the Digital Source 311 inputs signal information into the system

and sends it to the CTF HPF at 316. The CTF HPF in 316 is designed according to

the method of the invention as embodied in Figure 2 and is implemented in digital

form. After processing and filtering at 316, the signal information is converted to analog form by the DAC at 317. The Amplifier at 318 amplifies the signal

information for output by the Satellite Loudspeaker at 319.

Through DSP, the digital embodiment of the invention can execute the CTF

more precisely and has less errors in matching the required transfer function

characteristics than its counterpart in analog form. In addition, the digital

embodiment avoids many of the disadvantages of analog circuitry such as penalty

in noise, signal distortion, and circuit complexity. Moreover, in the digital format,

a limitless number of speakers can be accommodated by the design since the

required CTF parameters are completely software determined and easily stored and

downloaded to the DSP. The CTF can be continually updated or modified and

distributed in software form, e.g., over the Internet or by floppy diskette.

CONCLUSION

The present invention provides a system and method of designing

complementary transfer functions for crossover filters in a specific audio

environment by processing crossover factors, simulating CTF filters in a computer

model, correlating measurements of crossover parameters with the computer

model, and adjusting the CTF design until there is a high or predetermined degree

of correlation between the computer model and the crossover measurements. The

invention eliminates the need for lengthy, cumbersome empirical data gathering

and trial-and-error analysis in the final end-user adjustment process. In particular,

the invention provides proper blending between chosen sets of subwoofers and satellite speakers in loudspeaker systems. The invention greatly reduces

development time, while improving the quality and completeness of the final

design result.

Changes and modifications in the specifically described embodiments can

be carried out without departing from the scope of the invention. In particular, the

present invention may be used in designing complementary transfer functions (CTF) in crossover filters in a loudspeaker system with chosen sets of subwoofers

and satellite speakers in a fixed audio environment. Although preferred

embodiments are disclosed herein, they do not limit the scope of the present

invention.

Claims

I CLAIM:

1. A method of complementary transfer function design of crossover filters for

a loudspeaker system including at least one speaker arranged to reproduce audio

frequencies above a crossover point ("HF speaker") and at least one speaker

arranged to reproduce audio frequencies below the crossover point ("LF speaker"),

the method comprising the steps of:

measuring acoustical response of the at least one HF speaker [101];

providing a computer simulation of acoustical response of the at least one

HF speaker [102];

correlating the measured acoustical response and the computer simulation

of the at least one HF speaker [103];

measuring acoustical response of the at least one LF speaker [105];

providing a computer simulation of acoustical response of the at least one

LF speaker [106];

correlating the measured acoustical response and the computer simulation

of the at least one LF speaker [107]; and developing a complementary transfer function filter for the loudspeaker

system based on resultants of correlating steps for the at least one HF speaker and

for the at least one LF speaker [104] [108] [109].

2. The method in claim 1, wherein the correlating step for the at least one HF

speaker further comprises the substeps of:

determining whether a degree of correlation at or below a predetermined

threshold level has been reached between the acoustical response measurements

and the computer simulation of the at least one HF speaker [103]; and,

adjusting the computer simulation of the at least one HF speaker [10] if the

threshold correlation has not been reached in the determining step.

3. The method in claim 1, wherein the correlating step for the at least one LF

speaker further comprises the substeps of:

determining whether a degree of correlation at or below a predetermined

threshold level has been reached between the acoustical response measurements

and the computer simulation of the at least one LF speaker [107]; and,

adjusting the computer simulation of the at least one LF speaker [12] if the

threshold correlation has not been reached in the determining step for the at least

one LF speaker.

4. The method in claim 1 , further comprising the steps of:

measuring acoustical response of the loudspeaker system using the

developed complementary transfer function filter [110];

providing a computer simulation for the loudspeaker system with the

developed complementary transfer function filter [111];

determining whether a degree of correlation at or below a predetermined

threshold level has been reached between the measured acoustical response and the

computer simulation of the loudspeaker system [112]; and

adjusting the computer simulation of the loudspeaker system [14] if a

degree of correlation at or below a predetermined threshold level has not been

reached in the determining step for the loudspeaker system.

5. The method in claim 4, further comprising the steps of:

performing an assessment of the loudspeaker system performance relative

to predetermined performance criteria [113]; and

adjusting the complementary transfer function filter based on the

performance assessment if adjustment is warranted [109] .

6. The method in claim 5, wherein the measuring steps for the at least one HF

speaker, the at least one LF speaker, and the loudspeaker system, respectively, are

performed in a low-frequency region in a typical half-space room [101] [105].

7. The method in claim 1, wherein the complementary transfer function filter

in the developing step is a low pass filter [109].

8. The method in claim 1, wherein the developing step develops a

complementary transfer function high pass filter [209] and a complementary transfer function low pass filter [211].

9. The method in claim 8, further comprising the steps of:

measuring acoustical response of the loudspeaker system with the

developed complementary transfer function high pass filter and the developed

complementary transfer function low pass filter [210];

providing a computer simulation for the loudspeaker system with the

developed complementary transfer function high pass filter and the developed

complementary transfer function low pass filter [212]; determining whether a degree of correlation at or below a predetermined threshold level has been reached between the measured acoustical response and the

computer simulation of the loudspeaker system [213]; and

adjusting the computer simulation of the loudspeaker system [18] if the

threshold correlation has not been reached in the determining step for the loudspeaker system.

10. The method in claim 9, further comprising the steps of:

performing an assessment of the loudspeaker system performance relative

to predetermined performance criteria[214];

adjusting the developed complementary transfer function high pass filter

based on the performance assessment if adjustment is warranted [209]; and,

adjusting the developed complementary transfer function low pass filter

based on the performance assessment if adjustment is warranted [211].

11. The method in claim 10, wherein the measuring steps for the at least one

HF speaker, the at least one LF speaker, and the loudspeaker system, respectively,

are performed in a low-frequency region in a typical half-space room [201] [205]

[301].

12. The method in claim 1, wherein the loudspeaker system is an analog system

in which the complementary transfer function filter is implemented with active

analog electronic circuitry [302].

13. The method in claim 1, wherein the complementary transfer function filter

is implemented using digital signal processing [312].

14. The method in claim 1 implemented with predetermined parameters

including damping factor and level.

15. The method in claim 5, wherein the performing step [113] examines quality

factors including harmonic and intermodulation distortion, dynamic range,

transient response, clarity of bass and midrange reproduction and resolution.

16. A system of complementary transfer function crossover filters for a

loudspeaker system including at least one speaker arranged to reproduce audio

frequencies above a crossover point ("HF speaker") and at least one speaker

comprising:

means for measuring acoustical response of the at least one HF speaker

[101]; means for providing a computer simulation of acoustical response of the at

least one HF speaker [102];

means for correlating the measured acoustical response and the computer

simulation of the at least one HF speaker[103];

means for measuring acoustical response of the at least one LF speaker

[105];

means for providing a computer simulation of acoustical response of the at

least one LF speaker [106];

means for correlating the measured acoustical response and the computer

simulation of the at least one LF speaker [107]; and

means for developing a complementary transfer function filter for the

loudspeaker system based on resultants of the correlation of the at least one HF

speaker and the at least one LF speaker [104] [108] [109].

17. The system in claim 16, wherein the correlating means of the at least one

HF speaker further comprises:

means for determining whether a degree of correlation at or below a

predetermined threshold level has been reached between the acoustical response

measurements and the computer simulation of the at least one HF speaker [103]; means for adjusting the computer simulation of the at least one HF speaker

[10] if the threshold correlation has not been reached; and

means for configuring a transfer function for the at least one HF speaker if

the threshold correlation has been reached [104].

18. The system in claim 16, wherein the correlating means of the at least one

LF speaker further comprises:

means for determining whether the threshold correlation has been reached

between the acoustical response measurements and the computer simulation of the

at least one LF speaker [107];

means for adjusting the computer simulation of the at least one LF speaker

[12] if the threshold correlation has not been reached for the at least one LF

speaker; and

means for configuring a transfer function for the at least one LF speaker if

the correlation threshold has been reached for the at least one LF speaker [108].

19. The system in claim 16, further comprising:

means for measuring acoustical response of the loudspeaker system using

the developed complementary transfer function filter [110]; means for providing a computer simulation for the loudspeaker system with

the complementary transfer function filter [111];

means for determining whether a degree of correlation at or below a

predetermined threshold level has been reached between the measured acoustical

response and the computer simulation of the loudspeaker system [112] ; and

means for adjusting the computer simulation of the loudspeaker system if a

degree of correlation at or below a predetermined threshold level has not been

reached for the loudspeaker system [14].

20. The system in claim 19, further comprising:

means for performing an assessment of the loudspeaker system

performance relative to predetermined performance criteria [113]; and,

means for adjusting the complementary transfer function filter based on the

performance assessment if adjustment is warranted [ 109] .

21. The system in claim 20, wherein the acoustical response measurements for

the at least one HF speaker, the at least one LF speaker, and the loudspeaker

system, respectively, are performed in a low-frequency region in a typical half-

space room [101] [105].

22. The system in claim 16, wherein the complementary transfer function filter

developed by the developing means is a low pass filter [109].

23. The system in claim 16, wherein the developing means develops a

complementary transfer function high pass filter [209] and a complementary

transfer function low pass filter [211].

24. The system in claim 23, further comprising:

means for measuring acoustical response of the loudspeaker system with

the developed complementary transfer function high pass filter and the developed

complementary transfer function low pass filter [210];

means for providing a computer simulation for the loudspeaker system with

complementary transfer function low pass filter [212];

means for determining whether a degree of correlation at or below a

predetermined threshold level has been reached between the measured acoustical

response and the computer simulation of the loudspeaker system [213]; and

means for adjusting the computer simulation of the loudspeaker system if

the threshold correlation has not been reached for the loudspeaker system [18].

25. The system in claim 24, further comprising:

means for performing an assessment of the loudspeaker system

performance relative to predetermined performance criteria [214]; and,

means for adjusting the developed complementary transfer function high

pass filter based on the performance assessment if adjustment is warranted [209].

26. The system in claim 25, wherein the acoustical response measurements for

the at least one HF speaker, the at least one LF speaker, and the loudspeaker system, respectively, are performed in a low-frequency region in a typical half-

space room [201] [205].

27. The system in claim 16, wherein the loudspeaker system is an analog

system in which the complementary transfer function filter is implemented with

active analog electronic circuitry [301] [305].

28. The system in claim 16, wherein the complementary transfer function filter

is implemented using digital signal processing [312] [316].

29. The system in claim 16 implemented with predetermined parameters including damping factor and level.

30. The system in claim 20, wherein the means for performing a qualitative

assessment [214] examines quality factors including harmonic and intermodulation distortion, dynamic range, transient response, clarity of bass and midrange

reproduction and resolution.