JPH0879879A - Audio signal processing unit - Google Patents

Abstract

PURPOSE: To obtain an audio signal processing unit which is mounted on a speaker enclosure in which a size and a group delay of an inverse filter to correct a response characteristic of a transmission system including a speaker to avoid production of a distorted sound are reduced to attain miniaturization and low cost. CONSTITUTION: Inverse filters 2L, 2R correcting a response characteristic of a transmission system including a speaker are formed by 1st filters 3L, 3R with IIR filters having a bell characteristic at a specific frequency band and 2nd filters 4L, 4R with FIR filters acting like convolvers in cascade connection the 1st filters 3L, 3R is selected so that the filter coefficient length of the 2nd filters 4L, 4R is optimized to be a prescribed length.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio signal processing apparatus for correcting a response characteristic of a transmission system including a speaker by an inverse filter so that a distorted sound is not generated.

[0002]

2. Description of the Related Art In recent years, in multimedia, it is required to improve the quality of audio sound accompanying an image. This is not a mere sound as audio with a wide dynamic range by CD-ROM has spread.
Because Fi audio became indispensable. In this case, the speaker is required to be small and have high performance. For such a purpose, on the one hand, a faithful reproduction of a three-dimensional sound field including a sound field and a sound image such as a sound field (surround) and a sound image (3D virtual reality) for HiFi audio is sought, and on the other hand, a speaker. Is being pursued, that is, audio reproduction in which the distortion of the transmission path is not heard by correcting the transmission path of the speaker by the inverse filter is pursued, and various attempts have been made in the latter.

Conventionally, for example, Japanese Patent Laid-Open No. 3-143195 discloses a linear phase FIR type filter (FIR: Finite Impulse Respo) for each of a plurality of frequency band channels.
(nse) is provided, and a so-called multi-way speaker system is disclosed in which the characteristics as a speaker system and the sound quality are improved by performing digital signal processing suitable for a speaker unit corresponding to each of a plurality of frequency bands. ing.

As digital filters, FIR type filters and IIR type filters (IIR: Infinite Impul)
se response) is used at the same time.
-79725 and JP-A-63-221708. Among these, Japanese Patent Laid-Open No. 57-79725 discloses that the FIR type filter is used to remove the resonance characteristic in which the amplitude changes near the roll-off frequency outside the pass band of the IIR type filter. Further, Japanese Patent Laid-Open No. 63-221708 discloses an FIR type filter that variably adjusts the amplitude and an all-pass type IIR type filter that variably adjusts the group delay amount.

[0005]

However, in the above-mentioned publication, if the speaker characteristics are corrected only by the FIR type filter, the filter scale (filter coefficient length) must be made extremely large, and conversely, Smaller sizes (smaller ones) have to sacrifice performance, and therefore very large, or poorly performant, filter size (coefficient length) and group delay amounts. There is no disclosure of making it short, compact, and built into the speaker enclosure.

The present invention has been made in order to solve the above-mentioned problems associated with the conventional example. The filter scale of an inverse filter for correcting response characteristics of a transmission system including a speaker and preventing distorted sound is An object of the present invention is to provide an audio signal processing device which has a small group delay amount, is small and inexpensive, and can be incorporated in a speaker enclosure.

[0007]

In order to achieve the above object, an audio signal processing device according to the present invention includes an inverse filter for correcting the response characteristic of a transmission system including a speaker, such as a bell type filter in a specific frequency band. And a second filter constituted by an FIR type filter that operates as a convolver are connected in cascade, and the characteristic of the first filter is The filter coefficient length of the second filter is set to be optimized to a predetermined length.

Further, the present invention is characterized in that such an inverse filter is arranged in each system of a stereophonic sound field control section comprising a pair of convolvers.

[0009]

【Example】

First Embodiment FIG. 1 is a system configuration diagram showing an audio signal processing device according to the first embodiment. In FIG. 1, two-channel stereo sound sources 1L and 1R generate desired audio signals. The inverse filter 2L has first and second filters 3L and 4L connected in cascade for correcting speaker output characteristics to realize speaker reproduction with flat sound pressure characteristics and phase characteristics. ) Channel system. The first filter 3L is composed of an IIR type filter having a bell-shaped characteristic that is enhanced / attenuated with respect to a 0 dB reference value by setting a center frequency and Q (sharpness) in a specific frequency band. On the other hand, the second filter 4L
Is constituted by a convolver which is a convolution operation processing circuit composed of an FIR type filter for correcting the residual amplitude and phase characteristics of the first filter 3L excluding the bell type characteristic. Further, the characteristic of the first filter 3L is set so as to optimize the filter coefficient length of the second filter 4L to a predetermined length. Reference numeral 2R is an inverse filter of the right (R) channel system, which includes first and second filters 3R and 4R similar to the left channel system.

The output of the left channel convolver 4L is
The amplifiers are connected to a plurality of amplifiers 5L1 to 5L3 forming a network provided for each channel of each frequency band of the left-channel multiway system. Also,
These amplifiers 5L1 to 5L3 are connected to the corresponding speakers 6L1 to 6L3, respectively. 5R1 to 5R3 and 6
R1 to 6R3 are a plurality of amplifiers and speakers that form a network provided for each channel in each frequency band of the right channel multiway system, and these left channel system and right channel system speakers are each frequency band. The above-mentioned one inverse filters 2L and 2L for the left channel system and the right channel system without being divided for each channel.
The output characteristics of all the speakers are respectively corrected by R, and speaker reproduction with flat sound pressure characteristics and phase characteristics is realized.

Further, the control unit 7 includes the inverse filters 2L and 2R.
Of the coefficient memory 8 such as a ROM.
Are connected to the control unit 7. In this coefficient memory 8,
The second filter 4 (collective term for 4L and 4R) stores in advance coefficients relating to the output characteristics of the speaker when performing the convolution operation, and the control unit 7 controls the first filter 3 to store the coefficients.
Bell frequency characteristic center frequency and Q
Is set and the coefficient is read from the memory 8 to perform convolution calculation processing on the time axis based on the coefficient, and the output characteristics of the speakers 6L1 to 6L3 and 6R1 to 6R3 are corrected by the inverse characteristics thereof and the sound pressure characteristics are corrected. Also, the reproduction of a speaker having a flat phase characteristic is realized.

That is, as shown in FIG. 1, a speaker correction system (audio signal processing apparatus) according to the present invention.
Is a speaker system of a two-channel stereo system, and includes left and right channel inverse filters 2L and 2R, a control unit 7, and a memory 8. The left (L) channel speaker correction system (audio signal processing device) is
The sound source 1L, the inverse filter 2L, the amplifiers 5L1 to 5L3 and the speakers 6L1 to 6L3, and the speaker correction system for the right (R) channel system includes the sound source 1R, the inverse filter 2R, the amplifiers 5R1 to 5R3 and the speakers 6R1 to 6L.
Composed of R3.

By the way, in the above structure, it is general to use an FIR type filter as the inverse filter. However, if the inverse filters 2L and 2R are constituted by only the FIR type filter, the present inventors have conducted experiments, for example. In the example (characteristic shown in FIG. 3 described later), the sampling frequency is 32 KH.
In z, the number of taps (filter coefficient length) is about 1200.
Above (delay time = 1200 x (1/3200) x 1/2
= 0.1875 seconds) is required. In addition to this, if an FIR type filter of the same scale is used in the three-dimensional sound field control, it takes 0.1875 seconds, which results in a delay of 0.375 seconds. This is larger than the allowed 0.33 seconds per frame.

However, it is necessary to shorten the filter scale (coefficient length) in order to reduce the time lag (delay) from the image, and the advantage that it can be incorporated in the speaker if it is downsized. is there. Therefore, it is conceivable to use an IIR type filter that has no time lag auxiliaryly to significantly reduce the filter scale. In this way, the HiFi audio sound can be enjoyed even in multimedia. In this embodiment, an FIR type filter and an IIR type filter are connected in series as an inverse filter, and at this time, a bell type characteristic is adopted as the IIR type filter, and the filter characteristic is the filter coefficient length of the FIR type filter. Set to optimize for a given length. The bell-shaped characteristic does not have a sharp roll-off like the band-pass characteristic but has a gentle slope depending on Q (sharpness), so that the slope can be easily set adaptively.

Next, a method of calculating the bell-shaped characteristic of the IIR filter will be described. 2 shows an amplitude characteristic before correction (A) of the transmission system in the case where the inverse filter is not provided in FIG. 1, and FIG. 3 shows a case where only the second filter 4 is provided as the inverse filter 2 (general term for 2L and 2R). 3 shows corrected amplitude characteristics (B) and (C) of the transmission system of FIG. 3, respectively, and in FIG. 3, the amplitude characteristics (B) is the number of taps (filter coefficient length) 1
The amplitude characteristic and the amplitude characteristic (C) when realized by the inverse filter 200 of FIG. Further, FIG. 4 shows a bell-shaped characteristic, and FIG. 5 shows a target characteristic A for correction.
6 shows the target characteristic B of the new correction, and FIG. 6 shows the number of taps (filter coefficient length) 300 shown in comparison with the amplitude characteristic (B) when only the second filter 4 having the number of taps (filter coefficient length) 1200 is provided. 5 shows the amplitude characteristic C corrected by the inverse filter 2 having the second filter 4 and the first filter 3 having the bell-shaped characteristic shown in FIG.

Now, as shown in FIG. 3, the number of taps (filter coefficient length) is 12 as the corrected amplitude characteristic of the transmission system when only the second filter 4 is provided as the inverse filter 2.
When the number of taps is small as in the amplitude characteristic (C) in the case of the number of taps (filter coefficient length) 300 compared to the amplitude characteristic (B) in the case of 00, the number of taps (filter It can be seen that the error between the amplitude characteristic (B) in the case of (coefficient length) 1200 and the target characteristic A shown in FIG. That is, if the number of taps (filter coefficient length) is reduced and the filter scale is reduced (downsized), the performance is adversely affected.

Therefore, from the amplitude characteristic (C) in the case of the number of taps (filter coefficient length) 300 shown in FIG. 3, a bell-shaped characteristic in which the center frequency is lowered by 5 dB at 150 Hz as shown in FIG. 4 is taken out, which is contrary to this. Perform the first correction with the first filter 3 set as the characteristic IIR filter,
Next, when a new target characteristic B for correction, which is a characteristic obtained by removing the bell characteristic from the target characteristic A for correction shown in FIG. 5, is obtained and corrected by the second filter having 300 taps,
The amplitude characteristic C shown in FIG. 6 is obtained. In particular, it can be seen that the amplitude characteristic C has a smaller error from the target characteristic than the amplitude characteristic (C) shown in FIG. That is, by setting the characteristics of the first filter 3 so as to optimize the filter coefficient length of the second filter 4 to a predetermined length, the filter scale can be reduced and the multimedia HiFi audio can be embedded in the speaker. Can allow you to enjoy comfortably.

Although not described in detail, as another characteristic calculation method for the bell-shaped characteristic, the bell target BT may be directly obtained from the amplitude characteristic (A) as shown in FIG. 2, for example. Also, here, the number of bell targets BT is not limited to one but may be two or more depending on the amplitude characteristics. In order to specify the center frequency and Q (sharpness) of such a bell type, a bell analysis routine (by sweeping each parameter within a predetermined band) is operated on the workstation to synthesize the IIR type filter and the FIR type filter. The characteristics are repeatedly obtained until finally within a predetermined deviation.

Therefore, according to the first embodiment, the inverse filter 2 for correcting the response characteristic of the transmission system including the speaker is
A first filter 3 composed of an IIR type filter having a bell-shaped characteristic in a specific frequency band and a second filter 4 composed of an FIR type filter operating as a convolver are connected in cascade, and The characteristics of the filter 3 are set so that the filter coefficient length of the second filter 4 is optimized to a predetermined length, so that the inverse filter 2 corrects the response characteristics of the transmission system including the speaker. When audio reproduction is performed so that the distorted sound of the road is not heard, the filter scale can be made small and, for example, multimedia HiFi audio can be comfortably enjoyed by being built in the speaker.

Second Embodiment Next, FIG. 7 illustrates a speaker device having an audio signal processing device according to a second embodiment. (A) is an external structural view and (b) is internal audio signal processing. It is a circuit block diagram of an apparatus. In the second embodiment, as shown in the external structure of FIG. 7A, a multi-way speaker system is adopted as in the structure of the first embodiment shown in FIG.
For example, the left channel speakers 6L1 to 6L3 are not driven by individual amplifiers, but are driven by a single amplifier 5L, and the amplifier 5L and an inverse filter are integrated by the processing circuit unit 9. It is built in the speaker device. Further, it is provided with an infrared light receiving unit 10 for receiving a sound source and a remote control (volume / on / off) command by infrared rays from a main body 20 such as a player or a display. In addition, 11 is an infrared light receiving unit 10.
A conversion unit that converts the infrared light received by the received signal into a received signal,
Reference numeral 12 denotes the converted reception signal and the control unit 7 and the inverse filter 2L
The reference numeral 13 denotes a distributor for distributing the drive signal via the amplifier 5L to the speakers 6L1 to 6L3. Further, the same structure as the left channel system is also provided in the right channel system.

Therefore, according to the second embodiment, each speaker of the multi-way speaker system is not driven by an individual amplifier, but a single amplifier 5L.
Since the amplifier 5L and the inverse filter are integrated by the processing circuit unit 9 and incorporated into the speaker device, the processing circuit unit 9 can be constructed at low cost and the speaker device can be improved in performance. Miniaturization can be achieved. In addition, since the command is given by infrared rays, the speaker device can be placed at an optimum distance for easy stereo localization with good mobility, and it is provided separately from the main body of the player, display, etc. There is no leakage, and the computer of the main body can be provided with beautiful features such as no adverse effects such as vibration. It is especially suitable for multimedia audio systems.

Third Embodiment Next, FIG. 9 shows a block diagram of a speaker device having an audio signal processing device according to a third embodiment in combination with a three-dimensional sound field control. The audio signal processing apparatus according to the third embodiment shown in FIG. 9 includes inverse filters 2L and 2R according to the first embodiment in each system of a stereoscopic sound field control unit including sound image localization filters 14L and 14R including a pair of convolvers. By adjusting the pair of sound image localization filters 14L and 14R by the control unit 7, the two speakers 6L and 6R can reproduce the virtual speaker 6I at an arbitrary position. . In addition, 15 is a CD-RO
M, 16a to 16c are sound source data of the CD-ROM 15, localization data and video data, and 17 is a display for displaying the video data.

With this configuration, the total time delay is reduced, and even when combined with the stereoscopic sound field control, the time delay can be suppressed to less than one frame of the image, which hinders multimedia reproduction. Never give. Although not described in detail here, in each of the above-described embodiments, in order to supplement the bass that is generally deficient in small speakers, for example, NFB (negative feedback) technology is adopted at a low frequency of 60 to 200 Hz, and It goes without saying that the sound quality may be improved.

[0024]

As described above, according to the present invention, the inverse filter for correcting the response characteristic of the transmission system including the speaker is constituted by the IIR type filter having the bell type characteristic in the specific frequency band. A second filter including a filter and an FIR type filter that operates as a convolver.
Since the filters are connected in cascade, and the characteristics of the first filter are set so that the filter coefficient length of the second filter is optimized to a predetermined length, the transmission including the speaker is performed by the inverse filter. An audio signal processing device including a small and inexpensive speaker device by shortening the filter scale and group delay amount when performing audio reproduction in which distortion of the transmission path is inaudible by correcting the response characteristics of the system Obtainable.

Further, by providing an inverse filter in each system of the stereophonic sound field control section consisting of a pair of convolvers, even if combined with the stereophonic sound field control, the time delay can be suppressed to less than one frame of the image. This has the effect of not hindering multimedia playback.

[Brief description of drawings]

FIG. 1 is an overall block diagram showing a first embodiment of an audio signal processing device of the present invention.

FIG. 2 is a characteristic diagram showing an amplitude characteristic (A) before correction of the transmission system in the case where an inverse filter is not provided in FIG.

FIG. 3 is a characteristic diagram showing corrected amplitude characteristics (B) and (C) of a transmission system in the case where only a second filter 4 is provided as an inverse filter 2 (general term for 2L and 2R) in FIG. And the amplitude characteristic (B) is the number of taps (filter coefficient length)
The characteristics when realized by the inverse filter of 1200 and the amplitude characteristics (C) show the characteristics when realized by the same 300.

FIG. 4 is a characteristic diagram of a bell-shaped characteristic.

FIG. 5 is a characteristic diagram showing a correction target characteristic A and a new correction target characteristic B.

FIG. 6 is a diagram showing a second filter 4 having a tap number (filter coefficient length) of 300, which is shown in comparison with an amplitude characteristic (B) in the case where only the second filter 4 having a tap number (filter coefficient length) 1200 is provided. 6 is a characteristic diagram showing an amplitude characteristic C corrected by an inverse filter 2 having a bell-shaped characteristic first filter 3 shown in FIG.

FIG. 7 is an external view of a speaker device having an audio signal processing device according to a second embodiment of the present invention and a block diagram of an internal processing circuit.

FIG. 8 is an external view of a speaker device having an audio signal processing device according to a second embodiment of the invention and a block diagram of an internal processing circuit.

FIG. 9 is an overall block diagram showing a third embodiment of the audio signal processing device of the invention.

[Explanation of symbols]

1L, 1R Sound source 2L, 2R Inverse filter 3L, 3R First filter (IIR type filter) 4L, 4R Second filter (FIR type filter, convolver) 5L1 to 5L3, 5R1 to 5R3 Amplifier 6L1 to 6L3, 6R1 to 6R3 Amplifier 7 Control unit 8 Memory 14L, 14R Convolver (sound localization filter)

Claims (2)

[Claims] 1. An audio signal processing apparatus including an inverse filter for correcting a response characteristic of a transmission system including a speaker, wherein the inverse filter is a IIR type filter having a desired characteristic in a specific frequency band. A second filter comprising a filter and an FIR type filter that is connected in cascade to the first filter and operates as a convolver.
And the characteristics of the first filter are set so that the filter coefficient length of the second filter is optimized to a predetermined length. 2. The inverse filter is arranged in each system of a three-dimensional sound field control unit configured by a pair of convolvers for controlling a three-dimensional sound field including a sound field and a sound image. Audio signal processing device.