JPH0747039Y2 - Headphone listening correction device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to a device for correcting sound image localization when a stereo signal is listened to by headphones, and facilitates correction operation.

[Conventional technology]

When listening to stereo signals produced for normal speaker reproduction with headphones, the sound image is not localized properly. In particular,
A sound stage may be created inside the head, which may even be uncomfortable.

Binaural recording is performed using a dummy head, and since the effect of the human head transfer function is naturally included in the recorded signal, relatively good headphone reproduction is possible.

However, due to the characteristics of the dummy head, the subtle difference in the listener's head, the difference in the characteristics of the headphones to be listened to, etc., a problem is likely to occur in front localization, which is particularly important in music reproduction. In addition, it is not possible to support a source for normal speaker reproduction.

Therefore, when a stereo signal for normal speaker reproduction is heard through headphones, correction is performed to give a audible impression as if the stereo signal was reproduced through the speaker, that is, the stereo signal is electrically A method has been proposed in which the transfer functions from both speakers to both ears are approximated by dynamic processing. This is a transfer function for correcting a two-channel stereo signal as shown in FIG.
Pass the filters 11 to 14 of HLL, HLR, HRL, HRR to adder 16,
The left and right channel signals are added at 18 and supplied to the left and right units 20, 22 of the headphones.

[Problems to be solved by the device]

In the approximation method by electrical processing shown in FIG. 2, the transfer function is
When HLL, HLR, HRL, and HRR are set to average values, there are subtle differences in the listener's head, subtle differences in the listener's head, etc. However, good results were not always obtained. In addition, transfer functions HLL, HLR, HRL,
Even if the HRR value can be adjusted, it was difficult to adjust to the optimum characteristics for each case using a normal music signal.

This invention solves the above-mentioned drawbacks of the prior art,
It is an object of the present invention to provide a headphone listening correction apparatus that can easily adjust the optimum characteristics according to each case when using an approximation method by electrical processing.

[Means for Solving the Problems]

A headphone listening correction device for correcting the sound image localization of left and right channel stereo signals and supplying the left and right units of the headphones to the left and right units, and a first variable filter for correcting the left channel signal for supplying the left headphone unit, A second variable filter that corrects the left channel signal for supplying the right headphone unit, a third variable filter that corrects the right channel signal for supplying the left headphone unit, and a right headphone unit for the right channel signal. A fourth variable filter that performs correction for supply, a first adder that adds the output signals of the first and third variable filters and supplies to the left headphone unit, and the second and fourth Of the output signals of the variable filter of the second A plurality of combinations of parameters to be set in the first to fourth variable filters are stored as parameters defining a filter characteristic for correcting a sound image localization when a left and right two-channel stereo signal is heard by the headphones. The coefficient memory and the headphone listener's selection operation selects any one of a plurality of combinations of parameters stored in the coefficient memory to select the first parameter.
~ Parameter selection operation means for defining the characteristics of the fourth variable filter by the selected parameter, and operation of the listener of the headphones to generate a test signal in place of a normal stereo signal to generate the test signal. A test signal generator for inputting to a variable filter, wherein a headphone listener can select a desired sound image localization state by operating the parameter selection operation means while listening to the test signal with the headphones. To do.

[Action]

According to this invention, the listener generates a test signal from the test signal generator and inputs it to the first to fourth variable filters prior to normal use, and supplies the output to the headphones for listening. At this time, by selecting a combination of parameters from the coefficient memory by the parameter selecting operation means and defining the characteristics of the first to fourth variable filters, it is possible to set the characteristics most suitable for oneself.

According to this, since it is possible to select the desired parameter combination from the plurality of parameter combinations stored in the memory by the parameter selection operation means, it is possible to define the characteristics of the first to fourth variable filters. The characteristics of the first to fourth variable filters can be easily adjusted. Also,
Since the adjustment is performed using the dedicated test signal, it is possible to easily adjust to the optimum characteristic as compared with the case of using the normal music signal.

〔Example〕

An embodiment of this invention is shown in FIG. The 2-channel stereo signal is input to the first to fourth variable filters 27 to 30 via the contacts a of the switches 24 and 26, where the desired characteristics are obtained.
After adding HLL to HRR, the left and right channel signals are added by adders 32 and 34, and the left and right units 36 and 38 of the headphones are added.
Is supplied to.

For example, pink noise (continuous noise) is generated from the test tone generator 40 as a test tone for characteristic adjustment. When a normal music signal is used as a test tone for adjusting the characteristics, the frequency components instantaneously included in the music signal are limited and change from moment to moment. Even if it feels good, it is extremely difficult to judge which is the best for the listener, such as feeling different characteristics at other moments. On the other hand, if a signal including frequency components from low frequency to high frequency such as pink noise is used, optimum characteristics can be easily selected.

The test tone generated from the test tone generator 40 is supplied to the contacts b of the switches 24 and 26 via the switches 42 and 44, respectively. When the test key 49 is turned off, the switches 24 and 26 are connected to the contact a when the reproduction mode is selected, and the 2-channel stereo signal is input to reproduce.
When the test key 49 is turned on, it is connected to the contact b and a test tone is input to adjust the characteristics.

When the test key 49 is turned on, the switches 42 and 44 are automatically switched to contact a → b → c → a → b → ... In a cycle of about 3 seconds. When connected to the contact a, the test tone is supplied only to the left channel, when connected to the contact b, the test tone is supplied only to the right channel, and when connected to the contact c, the test tone is supplied. Attenuator 4
It is attenuated by 3 dB at 5,47 and supplied to both channels.

The coefficient memory 46 includes the characteristics HLL to HRR of the variable filters 27 to 28.
There are stored a plurality of parameters defining Transfer characteristic
HLL to HRR take into consideration (a) assumed speaker characteristics, (b) spatial transfer characteristics of the room from both speakers to both ears, (c) head transfer characteristics, (d) headphone characteristics, etc. Determined. Of these, (a) and (b) may be fixed or may be set according to preference. (C) and (d) require individual adjustment when listening. In particular, (c) has a significant effect on sound image localization.

As parameters, these conditions (a) to (d) can be set variously, an optimum parameter combination can be obtained by actual measurement for each combination of the conditions, and stored in the coefficient memory 46. Each condition can be selected from the following, for example.

(A) Speaker characteristics Select from various speaker characteristics (b) Spatial transfer characteristics of the room i) Size: small, large (c) Head transfer characteristics i) Gender: male, female ii) Old and young: children, young , Middle-aged, old man iii) Hair: Shaved, short, long (right divided, left divided) iv) Head shape: round, corner v) Pinna shape: large, medium, small (d) Headphone characteristics i) Form: It is sufficient to approximate the head-related transfer characteristics of the open type and the close contact type (c) if the impulse response of several milliseconds is reproduced. In the case of a high fidelity music signal, convolution of about 512 points is performed, for example, by digital signal processing. It should be done by. When the operation of the variable filters 27 to 28 is executed by a convolution device of 16-bit signal data and 16-bit coefficient data, 512 × 4 × 2 = 4096 bytes of memory (ROM) capacity is required to reproduce one measured data. I wish I had it. It is easy to store dozens of such data in the coefficient memory 46.

Also, the characteristic correction of the headphone of (d) is short as an impulse response, is several milliseconds or less, and can be performed by the same convolution.

Further, in order to strictly set the spatial transfer characteristics of the room (b), it is necessary to convolve the impulse response of a relatively long time (about 0.2 to 1.5 seconds) corresponding to the residual sound time of the reproduction room. . For this, convolution of about 65536 points is required. However, in normal music reproduction, superposition of reproduced sound fields is not always necessary, and this effect may be ignored. Alternatively, it is practically sufficient to perform a convolution of about 16,000 points, focusing on the fact that the early reflection sound that arrives within 150 milliseconds immediately after the direct sound determines the feeling of room expansion.

The optimum parameter combination under each condition obtained by actual measurement is stored in the coefficient memory 46 and read by the selection operation of the parameter selection operator 48. Variable filter 27 ~
The characteristic of 30 is specified by the parameter read from the coefficient memory 46, and the frequency characteristic is given to the input signal by the convolution operation with the input signal.

A configuration example of the variable filters 27 to 30 is shown in FIG. This is composed of digital filters, and the parameters hLL0 to hLLn, hLR0 to hLRn, hRL0 to are read from the coefficient memory 46.
Coefficient values are set in hRLn, hRR0 to hRRn.

The left and right channel inputs are input via the A / D converters 50, 52, the frequency characteristics are given by the convolution calculation by the variable filters 27 to 30, and the output is output via the A / D converters 54, 56. .

When using the correction device of FIG. 1, first, the characteristics of the variable filters 27 to 30 are set using a test tone. That is, the listener puts the headphones on his head, puts the headphone units 36, 38 on his ears, and turns on the test key 49. As a result, the switches 24 and 26 are connected to the contact b. Also, a test tone is generated from the test tone generator 40. At this time, the switches 42 and 44 are automatically switched at a cycle of about 3 seconds, and the test tones are repeatedly sent to the variable filters 27 to 30 in the order of left channel → right channel → center localization → left channel → right channel → .... Be done. The variable filters 27 to 30 have their characteristics set by the parameters selected by the parameter selection operator 48 and read from the coefficient memory 46, and give the frequency characteristics to the test tone by the convolution operation. The listener switches the parameter selection operator while listening to this test tone, and selects the parameter for which the front localization is most clear.

When the parameter selection is completed, the test key 49 is turned off. As a result, the generation of the test tone is stopped, the switches 24 and 26 are connected to the contact a, the two-channel stereo signal is input, and the convolution operation is performed in the variable filters 27 to 30 with the parameters set above. , Listeners can enjoy music with optimal sound localization.

[Modification]

In the above embodiment, pink noise was used as the test tone, but any reproducible sound can be used as the test signal. For example, various actual sounds stored in the waveform memory so that they can be repeatedly reproduced may be used. Impact sounds and rubbing sounds with a wide frequency spectrum are suitable as sounds that allow easy confirmation of localization, but in consideration of comfort, flute sounds including a rising portion and drum sounds may be used.

[Effect of device]

As described above, according to the present invention, the listener generates a test signal from the test signal generator and inputs it to the first to fourth variable filters prior to normal use, and supplies the output to the headphones. Then, the parameter selection operation means selects a combination of parameters from the coefficient memory to define the characteristics of the first to fourth variable filters, thereby setting the characteristics most suitable for oneself. it can.

According to this, since it is possible to select the desired parameter combination from the plurality of parameter combinations stored in the memory by the parameter selection operation means, it is possible to define the characteristics of the first to fourth variable filters. The characteristics of the first to fourth variable filters can be easily adjusted. Also,
Since the adjustment is performed using the dedicated test signal, it is possible to easily adjust to the optimum characteristic as compared with the case of using the normal music signal.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention. FIG. 2 is a block diagram showing a conventional device. FIG. 3 is a circuit diagram showing a concrete example of the variable filters 27 to 30 shown in FIG. 27-30 ... 1st-4th variable filters, 32, 34 ... 1st, 2nd adder, 36, 38 ... Headphone unit, 40
…… Test signal generator, 46 …… Coefficient memory, 48 …… Parameter selection operator (parameter selection operation means).

Claims (1)

[Scope of utility model registration request] 1. A headphone listening correction apparatus for correcting the sound image localization of left and right channel stereo signals and supplying the left and right units to the left and right units of the headphone, wherein the left channel signal is corrected for supplying the left headphone unit. Variable filter, a second variable filter that corrects the left channel signal for supplying the right headphone unit, a third variable filter that corrects the right channel signal for supplying the left headphone unit, and the right channel signal A fourth variable filter that performs correction for supplying the right headphone unit, a first adder that adds output signals of the first and third variable filters and supplies to the left headphone unit, The right headphone unit by adding the output signals of the second and fourth variable filters. The second adder to be supplied and the parameters to be set in the first to fourth variable filters as parameters for defining the filter characteristics for correcting the sound image localization when the left and right two-channel stereo signals are received by the headphones. A coefficient memory storing a plurality of combinations and a characteristic of each of the first to fourth variable filters selected by a headphone listener's selection operation from a combination of a plurality of parameters stored in the coefficient memory. And a parameter selection operation means for defining the selected parameter, and a test signal generation for generating a test signal in place of a normal stereo signal and inputting the test signal to the first to fourth variable filters by an operation of a listener of headphones. The headphone listener listens to the test signal with the headphones and operates the parameter selection operation. Headphone listening correction apparatus is characterized in that by operating the stage to be able to select a desired sound image localization state.