JP5757093B2 - Signal processing device - Google Patents

Description

  The present invention relates to a signal processing apparatus that performs virtual sound localization processing.

  Conventionally, for example, Patent Document 1 and Patent Document 2 are known as processes that give a sense of sound spread with a monaural speaker.

  In Patent Literature 1, only a low frequency component is extracted from a difference component between an L channel input signal and an R channel input signal, added to an L + R signal with a predetermined gain, and a component that does not contribute to localization is emphasized to give a sense of spread. Things are listed.

  Patent Document 2 describes that a predetermined delay (a delay that can be perceived by humans) is given to a difference component between an L channel input signal and an R channel input signal, and a sound field is given by adding the difference component to an L + R signal. Has been.

Japanese Patent No. 4526757 JP-A-3-266598

  In the method of Patent Document 1, since the low frequency range of the non-in-phase component is emphasized, the L channel and the R channel are not separated.

  In the method of Patent Document 2, the sound image remains monaural and the L channel and the R channel are not clearly localized.

  Accordingly, the present invention provides a signal processing device that can make a sense of spread by giving individual clear localization to a plurality of signal input channels even when a plurality of channels of audio signals are synthesized and output from a speaker. The purpose is to provide.

  The signal processing apparatus according to the present invention has a difference characteristic of an input unit that inputs a plurality of channels of sound signals and a plurality of head related transfer functions that localize a sound image in a direction having an elevation angle difference of a predetermined angle or more. A characteristic adding unit for adding to the signal, and an output unit for synthesizing and outputting audio signals of a plurality of channels to which the characteristic adding unit has added the differential characteristic, and the characteristic adding unit has different differential characteristics for each channel. Give.

  For example, the difference characteristic between the head-related transfer function that localizes the sound image in the direction of the front and elevation angles and the head-related transfer function that determines the sound image of the front and elevation angles of +30 degrees is given to any one channel (for example, L channel). Then, a difference characteristic between the head-related transfer function that localizes the sound image in the direction of the front and elevation angles of 0 degrees and the head-related transfer function that determines the sound image of the front and elevation angles of -30 degrees is given to other channels (for example, the R channel). . By giving the difference characteristic, only the component contributing to the elevation angle is extracted from the sound source localization, and the localization can be given more than when the head-related transfer function of the elevation angle of the front (for example, +30 degrees) is given. Therefore, even if it is a monaural speaker, individual clear localization can be given to a plurality of signal input channels. Note that if the elevation angle difference is less than 15 degrees, it is difficult to recognize that each channel is clearly separated.

  The difference characteristic is based on the head-related transfer function, and thus has a peak or dip at a predetermined frequency. Since these peaks or dips have a large deviation from the flat characteristics and have a great influence on the sound quality as they are, it is desirable to relax them to such an extent that they do not affect the sound source localization (for example, a gain of about ± 6 dB).

The signal processing apparatus of the present invention adds the audio signals of a plurality of channels by the input unit inputted, comprising a separation unit for separating a predetermined band following the audio signal from the audio signal the sum, and the output unit, the It is desirable that the characteristic adding unit synthesizes and outputs the audio signals of a plurality of channels to which the differential characteristic is added and the separated audio signals. The predetermined band is preferably a frequency band that does not affect localization.

  As described above, the influence on the sound quality can be reduced by separating the frequency bands that do not affect the localization and synthesizing them without adding the difference characteristics. In particular, the gain of the separated audio signal can be controlled separately from the audio signal to which the differential characteristic is added, so that the adjustment according to the preference of the source or the listener can be performed. For example, if the source contains human speech, the gain of the separated audio signal is increased to increase the monaural component (L channel + R channel), thereby relatively enhancing the in-phase (center localization) component. can do. Conversely, if the source is BGM, the gain of the separated audio signal can be made relatively low, and the localization feeling of the L channel and the R channel can be strengthened.

  Note that the frequency band that does not affect localization is approximately 700 to 1.5 kHz or less, and actually depends on the characteristics of the speaker. For example, in the case of a speaker with sharp directivity, the speaker itself is likely to be localized as a sound source. Therefore, the frequency to be separated is lowered, and the localization of the sound image depends on the differential characteristics. On the contrary, if the speaker is omnidirectional, the feeling of localization of the speaker itself is weakened, so that the feeling of localization of the sound image of each channel can be maintained even if the frequency to be separated is increased.

  According to the present invention, even when a plurality of channels of audio signals are synthesized and output from a speaker, a broad sense of surround can be provided by giving individual clear localization to the plurality of signal input channels.

It is a block diagram which shows the structure of a virtual sound image localization system. It is a block diagram which shows the structure of a signal processing apparatus. It is a figure which shows an elevation angle +/- 30 degree | times and the frequency characteristic of a difference. It is a figure which shows the frequency characteristic of an elevation angle +/- 15 degree | times and a difference.

  FIG. 1 is a block diagram showing a configuration of a virtual sound image localization system including a signal processing device of the present invention. The virtual sound image localization system of this embodiment performs multi-channel reproduction using a monaural speaker. As shown in FIG. 1, the virtual sound image localization system in this embodiment includes a signal processing device 1 and a speaker 2. In this embodiment, unless otherwise specified, all audio signals are described as digital signals, and the configurations of the DA converter and the power amplifier are omitted.

  The signal processing apparatus 1 uses a head-related transfer function (hereinafter referred to as HRTF) that expresses the difference in sound volume, arrival time, and frequency characteristics from a plurality of speakers installed at a certain position to the listener's ear. By applying the signal processing based on the audio signal and supplying it to the speaker 2, it is a monaural speaker installed on the median plane (in front of the listener), but spreads by giving individual clear localization to multiple signal input channels It can give a certain surround feeling.

  As shown in FIG. 2, the signal processing apparatus 1 includes an input interface (I / F) 11, a high pass filter (HPF) 12, an FIR filter 13, an HPF 14, an FIR filter 15, a level adjuster 16, a level adjuster 17, and an addition. A unit 18, an LPF 19, a level adjuster 20, an adder 21, and an output I / F 22 are provided.

  The input I / F 11 acquires an audio signal from another device or a content reproduction unit (not shown) in the device itself. Of the acquired audio signals, the left channel (L) signal is supplied to the HPF 12 and the level adjuster 16, and the right channel (R) signal is supplied to the HPF 14 and the level adjuster 17.

  The L signal input to the HPF 12 is output to the adder 21 through the FIR filter 13. The L signal input to the HPF 14 is output to the adder 21 through the FIR filter 15.

  The FIR filter 13 and the FIR filter 15 correspond to a characteristic providing unit of the present invention, and give predetermined transfer characteristics to the L signal and the R signal, respectively. The FIR filter 13 is a filter coefficient for realizing a differential characteristic between the HRTF that localizes the sound image to the VSL installed at the front elevation angle +30 degrees shown in FIG. 1 and the HRTF that localizes the sound image to the VSC installed at the front elevation angle 0 degrees. Is set. The FIR filter 15 is a filter for realizing a differential characteristic between the HRTF for sound image localization in the VSR installed at a front elevation angle of −30 degrees shown in FIG. 1 and the HRTF for sound image localization in the VSC installed at a front elevation angle of 0 degrees. The coefficient is set. As a result, the L channel signal can be localized in the elevation angle plus direction, and the R channel signal can be localized in the elevation angle minus direction. Detailed characteristics will be described later with reference to FIG.

  On the other hand, the L signal input to the level adjuster 16 and the R signal input to the level adjuster 17 are added by the adder 18 and input to the LPF 19 as an L + R signal. The gains of the level adjuster 16 and the level adjuster 17 are set to 0.5, respectively, so that the levels before and after the addition of the in-phase components are maintained.

  The level of the L + R signal output from the LPF 19 is adjusted by the level adjuster 20 and output to the adder 21. The gain of the level adjuster 20 is an arbitrary value of 0 or more. The gain of the level adjuster 20 is set for relative level adjustment with the L signal and the R signal that are output signals of the FIR filter 13 and the FIR filter 15. If the gain of the level adjuster 20 is increased, the L + R signal is enhanced and the in-phase component is enhanced. For example, if a human voice is included in the input voice signal (source), the gain of the level adjuster 20 may be increased to emphasize the human voice, and conversely, the source is BGM. If so, the gain of the level adjuster 20 may be lowered to emphasize the L signal and the R signal (output signal of the FIR filter), thereby enhancing the sense of localization. Or you may set according to a listener's liking.

  The adder 21 combines the L signal output from the FIR filter 13, the R signal output from the FIR filter 15, and the L + R signal output from the level adjuster 20, and outputs the resultant signal to the output I / F 22. The output I / F 22 outputs the combined signal to the speaker 2 (not shown, but the combined signal is D / A converted, amplified by a power amplifier, and then output). Thereby, from the speaker 2, the L signal to which the difference characteristic between the front elevation angle 0 degree and the front elevation angle +30 degrees is given, the R signal to which the difference characteristic between the front elevation angle 0 degree and the front elevation angle -30 degrees is given, and the gain. The adjusted L + R signal is output.

  Note that a level adjuster may be provided at the subsequent stage of the FIR filter 13 and the FIR filter 15 to individually adjust the gain.

  The HPF 12 and the HPF 14 have the same characteristics, and the same cutoff frequency is set. The cut-off frequency of the LPF 19 is set corresponding to the HPF 12 and the HPF 14. In this embodiment, the cutoff frequencies of the HPF 12, HPF 14, and LPF 19 are all set to 1.5 kHz, and the FIR filter 13 and the FIR filter 15 provide differential characteristics only for frequency bands that affect localization, and the localization is not affected. The effect on the sound quality is mitigated by separating the frequency band and adjusting the gain without synthesizing the difference characteristic. In particular, as described above, since the separated L + R signal can control the gain separately from the L signal and the R signal that give the differential characteristics, the gain is adjusted according to the preference of the source or the listener. be able to. However, the cut-off frequencies of the HPF 12 and the HPF 14 may be lowered, the cut-off frequency of the LPF 19 may be increased, and the bands may be overlapped. Conversely, the cut-off frequencies of the HPF 12 and the HPF 14 are raised, and the LPF 19 The band may be separated by lowering the cut-off frequency.

  The frequency band that does not affect the localization due to the frequency level characteristic difference is approximately 700 to 1.5 kHz or less, and actually depends on the characteristics of the speaker. For example, in the case of a speaker with sharp directivity, since the speaker itself is likely to be localized as a sound source, the frequency to be separated is lowered, and the localization of the sound image is influenced by the difference characteristics. On the contrary, if the speaker is omnidirectional, the localization of the speaker itself is weak, so that the localization of the sound image of each channel can be maintained even if the frequency to be separated is increased.

  Next, frequency characteristics realized by the FIR filter 13 and the FIR filter 15 will be described with reference to FIG. FIG. 3A is a diagram showing a HRTF (frequency characteristic) with a front elevation angle of 0 degrees, a HRTF with a front elevation angle of +30 degrees, and a difference characteristic. FIG. 3B shows a HRTF with a front elevation angle of 0 degrees, a front elevation angle− It is a figure which shows HRTF of 30 degree | times, and a difference characteristic.

  As shown in these figures, the HRTF with a front elevation angle of 0 degrees has a large dip of about -20 to -25 dB in the vicinity of 10 kHz, and also has an increase or decrease in gain at other frequencies. An HRTF with a front elevation angle of +30 degrees has a large dip of about -5 to -10 dB around 7 kHz and about -10 dB at a few dozen kHz, and also has an increase or decrease in gain at other frequencies. Similarly, an HRTF with a front elevation angle of −30 degrees has a large dip of about −15 to −20 dB in the vicinity of 8 to 9 kHz, and a gain of about −15 to −20 dB at a little less than 20 kHz, and gain increases at other frequencies. Have a slight depression.

  These complicated characteristics include, for example, those affected by the influence of the dummy head used when measuring the HRTF. That is, the HRTF that localizes the sound image at various elevation angles has various characteristics other than the peak and dip that affect the elevation angle localization. In the signal processing device 1 of the present embodiment, the difference characteristics between the HRTF with a front elevation angle of 0 degrees and the HRTF with a front elevation angle of +30 degrees, and the difference characteristics between the HRTF with a front elevation angle of 0 degrees and the HRTF with a front elevation angle of -30 degrees, respectively. By adding only the components that affect the elevation angle localization by adding them to the L signal and the R signal, it is possible to give a sense of localization more effectively than simply applying the HRTF having the same elevation angle. Therefore, even if it is a monaural speaker, a clear localization can be given to the L channel and the R channel, and a broad surround feeling can be given.

  The differential characteristic imparted to the L signal has a peak of about 15 to 20 dB around 10 kHz, and the differential characteristic imparted to the R signal has a peak of about 15 to 20 dB around 10 kHz and at 8 to 9 kHz (18 kHz Also has a dip of about -15 dB. Since these peaks or dips have a great influence on the sound quality as they are, it is desirable to relax them so as not to affect the sound source localization (for example, gain of about ± 6 dB).

  In this embodiment, the L signal is localized in the plus direction of the elevation angle and the R signal is localized in the minus direction of the elevation angle to give a sense of separation. Of course, the R signal is in the plus direction of the elevation angle, and the L signal is in the minus direction of the elevation angle. It is good also as an aspect which makes it localize and gives a feeling of separation.

  If the elevation angle difference is less than 15 degrees, it is likely to be visually influenced by the speaker, and it is difficult to recognize that each channel is clearly separated. FIG. 4A is a diagram showing a HRTF (frequency characteristic) with a front elevation angle of 0 degrees, a HRTF with a front elevation angle of 15 degrees, and a difference characteristic. FIG. 4B shows a HRTF with a front elevation angle of 0 degrees, a front elevation angle− It is a figure which shows HRTF of 15 degree | times, and a difference characteristic.

  As shown in these figures, the HRTF with a front elevation angle of 15 degrees has a large dip of about −15 dB in the vicinity of several tens of kHz, and also has an increase or decrease in gain at other frequencies. Similarly, an HRTF with a front elevation angle of -15 degrees has a large dip of more than -30 dB in the vicinity of 9 kHz, and also has gain rises and falls at other frequencies.

  In this case as well, as in the example of FIG. 3, the difference characteristics between the HRTF with a front elevation angle of 0 degrees and the HRTF with a front elevation angle of 15 degrees, and the difference characteristics between the HRTF with a front elevation angle of 0 degrees and the HRTF with a front elevation angle of −15 degrees, Is applied to the L signal and the R signal, respectively, so that only the component that affects the elevation angle localization is extracted, and the localization can be given more effectively than just adding the HRTF having the same elevation angle. Therefore, even when a difference characteristic having an elevation angle difference of about 15 degrees is given, a clear localization can be given to the L channel and the R channel, and a broad surround feeling can be given.

  Also in the example of FIG. 4, the differential characteristic imparted to the L signal has a peak of about 10 to 15 dB around 10 kHz, and the differential characteristic imparted to the R signal has a peak of about 15 dB around 10 kHz. It has a dip greater than −20 dB at 9 kHz. Since these peaks or dips have a great influence on the sound quality as they are, it is desirable to relax them to such an extent that they do not affect the sound source localization (for example, a gain of about ± 6 dB).

  In this embodiment, an example in which two-channel reproduction of the L channel and the R channel is performed using a monaural speaker is shown, but it is also possible to perform reproduction of more channels. For example, when the rear L channel and rear R channel audio signals are output from a single speaker installed on the back of the listener, the difference characteristics shown in the present embodiment are added to the rear L channel and rear R channel. Can be given a clear localization, and can provide a broad sense of surround.

DESCRIPTION OF SYMBOLS 1 ... Signal processing apparatus 2 ... Speaker 11 ... Input I / F
12 ... HPF
13 ... FIR filter 14 ... HPF
15 ... FIR filter 16 ... Level adjuster 17 ... Level adjuster 18 ... Adder 19 ... LPF
20 ... Level adjuster 21 ... Adder 22 ... Output I / F

Claims (5)

An input unit for inputting multi-channel audio signals;
A characteristic imparting unit that imparts to the audio signals of the plurality of channels the differential characteristics of a plurality of head related transfer functions that each localize a sound image in a direction having an elevation angle difference equal to or greater than a predetermined angle;
An output unit that synthesizes and outputs audio signals of a plurality of channels to which the characteristic assigning unit has added differential characteristics;
With
The said characteristic provision part is a signal processing apparatus which provides a different differential characteristic for every channel.   The signal processing apparatus according to claim 1, wherein the predetermined angle is 15 degrees. The input section adds the audio signals of a plurality of channels inputted, comprising a separation unit for separating a predetermined band following the audio signal from the audio signal the sum,
The signal processing apparatus according to claim 1, wherein the output unit synthesizes and outputs a plurality of channels of audio signals to which the characteristic adding unit has added a differential characteristic and the separated audio signals.   The signal processing apparatus according to claim 3, wherein the predetermined band is a frequency band that does not affect localization.   The signal processing apparatus according to claim 1, wherein the characteristic imparting unit relaxes a peak or a dip in the difference characteristic.

Images