JPH0775438B2 - Signal processing method for converting stereophonic signal from monophonic signal - Google Patents

Description

The present invention relates to a signal processing method for converting a monophonic signal into a stereophonic signal.

(Prior Art) It is not possible to record and reproduce an acoustic signal radiated from a large number of sound sources distributed in space as a monophonic signal (hereinafter also referred to as a monaural signal) using a cylindrical record or a disc record. It has been put to practical use for a long time, and even music played in the past by vocalists and performers can be heard at the present point in time by reproducing the record in which it is stored.

By the way, we, who live our daily lives in a three-dimensional sound field formed by acoustic signals radiated from a large number of sound sources distributed in space, usually use the acoustic signals radiated from the aforesaid large number of sound sources. Although it is heard as having a stereoscopic effect due to the level difference and phase difference of the sound in the ear,
Originally, music reproduced from a record in which acoustic signals radiated from a plurality of sound sources forming a three-dimensional sound field were recorded as monaural signals was unnatural to the sense of hearing.

Therefore, a stereophonic signal (hereinafter, referred to as a stereophonic signal that can form a stereoscopic reproduction sound field that gives a listener a stereophonic effect)
It is well known that disc records (stereo records) that record (also referred to as stereo signals) have been put into practical use by various stereophonic recording / reproducing systems. In addition to stereo records, recorded and reproduced stereo signals using recorded magnetic tape, other optical high-density recording medium discs, high-density recording medium discs with change detection type of capacitance value, etc. Transmission and reception (transmission) has been put into practical use in the past.

The reproduction of stereophonic sound in the stereophonic sound recording / reproducing (or transmitting / receiving, transmitting) method that has been most commonly performed in the past is performed by two reproduction sound sources arranged in front of the listener. Recording and reproduction of a left channel signal L and a right channel signal R (used to generate the above left and right acoustic signals) by radiating a signal to form a stereoscopic reproduction sound field in space ( Or send / receive, transmit)
Conventionally, the following two aspects are known as typical aspects.

One of them is to record the left channel signal L on the left wall surface of the record sound groove and the right channel signal R on the right wall surface of the record sound groove, for example, as implemented in a 45-45 stereo record. That is, the recording / reproducing (or transmitting / receiving / transmitting) is performed without performing any calculation on the left and right channel signals L and R, and the other one is, for example, FM. As carried out in broadcasting, from the left and right channel signals L and R, as a sum signal W shown as (left channel signal L + right channel signal R) and as (left channel signal L-right channel signal R) The difference signal S shown is formed to record (transmit) both of the above signals, and the left channel signal L = (sum signal W + difference signal S) with respect to the above sum signal W and difference signal S. / 2 right channel signal = (Sum signal W- difference signal S) / 2
The left and right channel signals are calculated by the calculation according to the above formulas.
Recording and reproduction (or transmission / reception, transmission) was performed based on the principle of restoring L and R so that left and right acoustic signals are reproduced.

(Problems to be Solved by the Invention) Now, when it becomes common to listen to reproduced music in a stereophonic sound reproduction sound field provided by a stereophonic record or other stereophonic sound information program source, the stereophonic sound information program source becomes Recorded by a monaural signal that was widely used before it became popular, and the reproduced sound from other program sources became unsatisfactory.For that reason, a recorded recording medium (for example, an old mono. It is also well known that a monaural signal reproduced from a (record) is converted into a stereo signal and, for example, a stereo record is made by using the stereo signal.

As described above, as a signal processing method that has been conventionally applied when converting a monaural signal into a stereo signal, (1) various filters are applied to the monaural signal to separate the sounds of the respective musical instruments and the respective musical instruments. Localize the sound image of.
(2) A phase shift circuit (all-pass filter) is applied to a monaural signal to give a sense of expansion to the reproduced sound field. (3) Reverberation and reverberation are added to the monaural signal so that a stereophonic expansive feeling can be felt. (4) The monaural signal is applied in combination with the steps (1) to (3) described above. Are known as typical ones, and the stereo signal created as described above is called "pseudo stereo signal", and a recorded recording medium on which the pseudo stereo signal is recorded is also commercially available. However, in the pseudo-stereo signal obtained by converting the monaural signal into a stereo signal by the conventional means (1) to (4), with respect to the above (1), the localization of the musical instrument is unnatural and unstable. There is a problem of poor localization such that there is no localization at all, and with respect to (2) above, there is a problem that the phase of the left and right reproduced sounds is greatly different and the listener feels tired. Regarding the above (3), when the sum signal in the pseudo stereo signal is heard as a monaural signal, it is reproduced as a sound completely different from the original monaural signal. There are, furthermore, the above-mentioned (4) is divided various disadvantages in the above (1) to (3) a table.

(Means for Solving Problems) The present invention is a signal processing method for converting a monophonic signal into a stereophonic signal by applying a digital signal processing technique, and a monophonic signal that is a target of stereophonic signal conversion, A means for obtaining a time-varying transfer function by dividing a difference signal between a sum signal and a difference signal formed by a right channel signal and a left channel signal of a stereophonic signal of two channels of the same music by a sum signal; The sum signal of the stereophonic signal and the monophonic signal targeted for stereophonic signal conversion of the monophonic signal targeted for phonic signal synchronization are synchronized with the time-varying transfer function described above. By comparing the envelope, the stereophonic signal Of the stereophonic signal by deriving a non-linear function indicating a time-axis synchronization relationship between the signal and the stereophonic signal, and changing the time axis of the time-varying transfer function based on the non-linear function. Means for obtaining a new time-varying transfer function in a synchronous relationship with the monophonic signal to be signaled, and the above-mentioned new time-varying variable with respect to the monophonic signal to be stereophonic signalized. Of a monophonic signal consisting of a means for obtaining a pseudo difference signal by applying a transfer function, a monophonic signal targeted for stereophonic signal conversion, and a means for performing an operation to restore the stereophonic signal from the pseudo difference signal A signal processing method for stereophonic signal conversion is provided.

(Examples) Hereinafter, specific contents of a signal processing method for converting a monophonic signal into a stereophonic signal of the present invention will be described in detail with reference to the accompanying drawings. First, a stereophonic signal of a monophonic signal of the present invention will be described. The configuration principle of the signal processing method for signalization will be described.

As described above, in the stereophonic sound recording (transmission) system such as the FM stereo broadcast, the sum signal W in which the left channel signal L and the right channel signal R are represented as (left channel signal L + right channel signal R), That is, W = L + R ...
(1) and a difference signal S represented as (left channel signal L-right channel signal R), that is, S = LR ... (2) are recorded and transmitted as two types of signals, and are also described above. The sum signal W = (L + R) and the difference signal S = (LR) are subjected to calculations as in the following equations (3) and (4) to restore the left channel signal L and the right channel signal R. It is made possible.

L = (sum signal W + difference signal S) / 2 (3) R = (sum signal W-difference signal S) / 2 (4) The sum signal W represented by the above formula (1) is monaural. It is well known that the sum signal W has substantially the same content as the signal and can be used when the stereo signal is reproduced as the monaural signal M, and the difference expressed by the equation (2) described above. The signal S is a left and right channel signal by a simple matrix calculation performed between the signal S and the sum signal W of the above-mentioned formula (1), which has a signal content such that compatibility between monaural reproduction and stereo reproduction is obtained.
The signal contents are such that L and R can be easily obtained.

Now, the above-mentioned monaural signal M and left and right channel signals L and R in the stereo signal are produced by mixing the electric signals obtained by acoustoelectrically converting the acoustic signals radiated from a plurality of sound producing bodies by a plurality of microphones. However, the left and right channel signals L and R in the stereo signal are
And the left and right channel signals L and R in the stereo signal
The sum signal W and the difference signal S are in the relationship represented by the above equations (3) and (4), and the monaural signal M is the sum signal W of the left and right channel signals L and R in the stereo signal. Since the signals are almost the same, the left and right channel signals in the stereo signal shown in the above-mentioned equation (1)
The sum signal W of L and R can be replaced with the monaural signal M as in the above equation (5), and therefore the sum signal in the above equations (3) and (4) When the monaural signal M shown in the above formula (5) is used instead of W, the left and right channel signals L and R in the stereo signal become those shown in the following formulas (6) and (7). Become.

L = (monaural signal M + difference signal S) / 2 (6) R = (monaural signal M-difference signal S) / 2 (7) However, the monaural signal M is the left and right channel signals L in the stereo signal, Assuming that it is the same as the sum signal W of R, even if the monaural signal M shown in the above equation (5) is used instead of the sum signal W in the above equations (3) and (4), the monaural signal M From the above, it is possible to obtain the difference signal S in the above equations (6) and (7), that is, the signal S which is the difference between the left and right channel signals L and R in the stereo signal forming the monaural signal M. Since this is not possible, it is clear that the left and right channel signals L and R in the stereo signal cannot be generated simply by giving the monaural signal M, due to the relationships of the expressions (6) and (7).

Therefore, in order to obtain the left and right channel signals L and R in the stereo signal based on the monaural signal M, that is, in order to convert the monaural signal into a stereo signal, the above-mentioned method (6) is used. ), It is necessary to generate a pseudo difference signal G that can be used in place of the difference signal S in equations (7) and use it as in equations (8) and (9).

Ls = (monaural signal M + pseudo difference signal G) / 2 (8) Rs = (monaural signal M-pseudo difference signal G) / 2 (9) In equations (8) and (9), Ls and Rs are respectively The left channel signal Ls and the right channel signal Rs in the pseudo stereo signal obtained by converting a monaural signal into a stereo signal are shown.

Now, the left channel signal L and the right channel signal R in the stereo signal, the sum signal W and the difference signal S of the left and right channel signals L and R in the stereo signal, the monaural signal M, the pseudo difference signal G, and the monaural signal are stereo. Left channel signal Ls in the pseudo stereo signal obtained by signalization,
When the right channel signal Rs in the pseudo stereo signal obtained by converting the monaural signal into a stereo signal is treated as a function of frequency, each of the above signals becomes as shown below.

The left channel signal L in the stereo signal is L (jω) The right channel signal R in the stereo signal is R (jω) The sum signal W of the left and right channel signals L and R in the stereo signal is W (jω) in the stereo signal The difference signal S of the left and right channel signals L and R is: S (jω) Monaural signal M, (jω) Pseudo difference signal G is the left of the pseudo stereo signal obtained by converting the G (jω) monaural signal into a stereo signal. The channel signal Ls is Ls (jω) The right channel signal Rs in the pseudo stereo signal obtained by converting the monaural signal into a stereo signal is Rs (jω). Therefore, the above equations (1) to (9) are 1ω) to (9ω).

W (jω) = L (jω) + R (jω) ... (1ω) S (jω) = L (jω) −R (jω) ... (2ω) L (jω) = {W (jω) + S (jω)} /2...(3ω) R (jω) = {W (jω) −S (jω)} / 2 ... (4ω) M (jω) = L (jω) + R (jω) ・ ・ ・ (5ω) L (jω) = {M (jω) + S (jω)} / 2 ... (6ω) R (jω) = {M (jω) −S (jω)} / 2 ... (7ω) Ls (jω) = {M (jω) + G ( jω)} / 2 ... (8ω) Rs (jω) = {M (jω) −G (jω)} / 2 ... (9ω) What has been described with reference to the equations (6) to (9) is 6ω)
The same applies to the expressions (9ω) to
In order to obtain the left and right channel signals L (jω) and R (jω) in the stereo signal based on (jω), that is, to convert the monaural signal into a stereo signal, some means is used. (6ω), (7
Generate a pseudo difference signal G (jω) that can be used instead of the difference signal S (jω) in the expression (ω), and generate it by (8
ω) and (9ω) are required to be used, but the problem is how to generate the pseudo difference signal G (jω).

The inventor has found that the sum signal W (jω) and the difference signal S of the left and right channel signals L (jω) and R (jω) in the stereo signal.
As a relationship with (jω), a function F (jω) obtained by dividing the difference signal S (jω) by the sum signal W (jω), that is, F (jω) = S (jω) / W (jω) ) Considering (10ω), the transfer function F as shown in the above equation (10ω)
If a filter having (jω) can be realized, this (10)
The difference signal S (jω) between the left and right channel signals L (jω) and R (jω) in the stereo signal is
Left and right channel signals L (j
ω), R (jω) sum signal W (jω) and the filter transfer function F (jω) described above: S (jω) = W (jω) × F (jω) (11ω) We paid attention to the fact that it can be obtained by the equation (11ω).

Then, as described above with reference to the equations (1) and (5), the left and right channel signals L (jω) and R (jω) in the stereo signal shown in (1ω) are Japanese parent W (jω), W (jω) = L (jω) + R (jω) (1ω), and monaural signals M (jω) and M (jω) = L (shown in the formula (5ω). jω) + R (jω) (5ω) are substantially the same, and therefore the monaural signal M (jω) is used instead of the new sum signal W (jω) in the equation (11ω).
By using, the necessary pseudo difference signal G (jω) can be obtained by G (jω) = M (jω) × F (jω) (12ω) by the above equation (12ω). It is the first point of the basic concept of a signal processing method for converting a monophonic signal into a stereophonic signal of the invention.

When the pseudo signal G (jω) required for converting a monophonic signal into a stereophonic signal is obtained as described above, the relationship between (8ω) and (9ω) described above,
That is, Ls (jω) = {M (jω) + G (jω)} / 2 ... (8ω) Rs (jω) = {M (jω) −G (jω)} / 2 ... (9ω) Left signal Ls (jω) and right signal Rs in a pseudo stereo signal obtained by converting (jω) into a stereo signal.
(Jω) can be generated.

By the way, as described above, the pseudo signal G (jω) is converted into the monaural signal M (jω) by the transfer function of the filter, as is clear from the equation: G (jω) = M (jω) × F (jω) (12ω). Since it is obtained by weighting F (jω), the necessary pseudo difference signal G (j) is obtained from the monaural signal M (jω) that is the target of stereo signalization.
A filter with a transfer function F (jω) such that ω) is obtained is needed.

The transfer function F (jω) of the filter is (10
ω), the function obtained by dividing the difference signal S (jω) between the left and right channel signals L (jω) and R (jω) in the stereo signal by the sum signal W (jω) F (j
ω), but the difference signal S (j
[omega]) and the sum signal W (j [omega]) are so-called time-varying spectrums whose contents change with time, so that the function F (j [omega]) described above also changes with time. It becomes a time-varying transfer function.

Therefore, various signals which have been described as a function of frequency, that is, the left channel signal L (jω) and the right channel signal R (j in the above-mentioned stereo signal are described.
ω), the left and right channel signals L and R in the stereo signal
Sum signal W (jω) and difference signal S (jω), monaural signal M (jω), pseudo difference signal G (jω), and left channel signal Ls (jω) in a pseudo stereo signal obtained by converting the monaural signal into a stereo signal. , Right channel signal Rs in pseudo stereo signal obtained by converting monaural signal to stereo signal
(Jω) and the like are treated as time-varying functions, and the display according to the following display method is performed, that is, the left channel signal L (jω, t) in the stereo signal described above.
And the right channel signal R (jω, t), the sum signal W (jω, t) of the left and right channel signals L, R in the stereo signal and the difference signal S (jω, t), the monaural signal M (jω, t), The pseudo difference signal G (jω, t) and the left channel signal Ls (j in the pseudo stereo signal obtained by converting the monaural signal into a stereo signal.
ω, t), and a right channel signal Rs (jω, t) in a pseudo stereo signal obtained by converting a monaural signal into a stereo signal is displayed.

Therefore, the above equations (1ω) to (12ω) are
ωt) to (12ωt).

W (jω, t) = L (jω, t) + R (jω, t) (1ωt) S (jω, t) = L (jω, t) −R (jω, t) (2ωt) L (jω , t) = {W (jω, t) + S (jω, t)} / 2 (3ωt) R (jω, t) = {W (jω, t) −S (jω, t)} / 2 (( 4ωt) M (jω, t) = L (jω, t) + R (jω, t) (5ωt) L (jω, t) = {M (jω, t) + S (jω, t)} / 2 ( 6ωt) R (jω, t) = {M (jω, t) -S (jω, t)} / 2 (7ωt) Ls (jω, t) = {M (jω, t) + G (jω, t) } / 2 (8ωt) Rs (jω, t) = {M (jω, t) −G (jω, t)} / 2 (9ωt) F (jω, t) = S (jω, t) / W (Jω, t) ... (10ωt) S (jω, t) = W (jω, t) × F (jω, t)… (11ω
t) G (jω, t) = M (jω, t) × F (jω, t) (12ω
t) By the way, the characteristics of the filter having the time-varying transfer function shown in the above equation (10ωt) change with the progression of the music mainly depending on the organization and arrangement of the musical instrument,
Except for special cases, it is virtually impossible to apply the law to the monaural signal M (jω, t) to calculate the time-varying transfer function F (jω, t).

Therefore, in the method of the present invention, the same music (the same music as the monaural signal M (jω, t) targeted for stereo signal conversion (the same music), and the music of the same music is A time-varying transfer function F (jω, t) = S (jω) as shown in the above equation (10ωt) from a stereo signal of the music of a specific song obtained by playing based on the score of the same content). , t) / W (jω, t) ... (10ωt) is derived and the time-varying transfer function F (jω, t) is applied to the monaural signal M (jω, t) which is the target of the Steres signalization. , (12ωt) equation which has already described the pseudo difference signal G (jω, t) of the monaural signal M (jω, t) which is the object of stereo signalization, that is, G (jω, t) = M (jω, t) × F (jω, t)… (12ω
The point obtained by using the relationship of t) is the second point of the basic concept of the signal processing method for converting the monophonic signal into the stereophonic signal of the present invention.

FIG. 2 is a diagram for explaining the waveforms of various signals generated based on the acoustic signals obtained when the performer A plays the music of the specific song α, the time-varying transfer function of the filter, and the like. 2A shows the time waveform of the sum signal W (jω, t) generated based on the acoustic signal obtained when the player A plays the music of the program α. FIG. 2B is a diagram showing the envelope of the sum signal W (jω, t) generated based on the acoustic signal obtained when the performer A plays the music of the program α. is there. Further, FIG. 2C shows a time-varying transfer function F (j of the filter applied to the acoustic signal obtained when the performer A plays the music of the program α.
ω, t), that is, when the time-varying transfer function F (jω, t) of the filter, which is a two-variable function of jω and t, is displayed using a representation method similar to a sonargram (voiceprint) of a voice. It is a figure, and the amplitude of the time-varying transfer function F (jω, t) of the filter is represented by shading similarly to the sonarlam (voiceprint) of the voice (however, in the accompanying drawings attached to the specification of the patent application, Since shaded drawings are prohibited, shades are not shown in the figure).

FIG. 2D is a diagram showing a time waveform of the difference signal S (jω, t) generated based on the acoustic signal when the performer A plays the music of the program α.

Now, the difference signal S (jω, t) of the left and right channel signals in the stereo signal is expressed by the above-mentioned equation (11ω, t), S (jω, t) = W (jω, t) × F (jω, t) )… (11ω
t), the time-varying transfer function F is added to the sum signal W (jω, t) of the left and right channel signals in the stereo signal.
It can be obtained by applying (jω, t), but in order to correctly obtain the difference signal S (jω, t) according to the above-mentioned equation (11ω, t), the sum signal W (jω, t) It is necessary that the time-varying transfer function F (jω, t) is synchronized.

Next, the time-varying transfer function F (j) is added to the monaural signal M (jω, t).
ω, t) is applied to generate a pseudo difference signal G (jω, t) of the monaural signal M (jω, t) that is the target of stereo signal generation.

FIG. 3 shows the performer A mentioned in the explanation of the contents of FIG.
The music of the music piece α played by the player A is reproduced by the acoustic signal when the musician of the music piece α plays the music of the music piece α by using the same score as that used by the music player A when playing the music of the music piece α. FIG. 3 is a diagram showing a time waveform {(a) in FIG. 3} and an envelope {(b) in FIG. 3} of a monaural signal M (jω, t) (any performer, but here Is assumed to be B).

As described above, in the signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention, a monaural signal M (jω,
(10) from the stereo signal in the same song as the above (10Ω
t) A time-varying transfer function F (jω, t) = S (jω, t) / W (jω, t) ... (10ωt) is derived, and the time-varying transfer function F (jω, t) is derived. Is applied to the monaural signal M (jω, t) targeted for stereo signalization, and the pseudo difference signal G (jω, t) of the monaural signal M (jω, t) targeted for stereo signalization is applied. (12ωt) equation described above, that is, Gg (jω, t) = M (jω, t) × F (jω, t) (12ωt)
However, even when the pseudo difference signal G (jω, t) is generated by applying the equation (12ωt), the monaural signal M (jω, t) and the time-varying transfer function F (j
ω, t) is required to be synchronized.

If the monaural signal M (jω, t) and the time-varying transfer function F (jω, t) are in good synchronization, the time-varying transfer function F (jω) is added to the monaural signal M (jω, t). , t) is applied to the pseudo difference signal G (jω, t) to obtain the equations (8ωt) and (9ωt), that is, Ls (jω, t) = {M (jω, t) + G (jω, t)} / 2 ... 8 (ωt) Rs (jω, t) = {M (jω, t) −G (jω, t)} / 2 (9ωt), so that the left and right channels of the pseudo stereo signal are obtained. Signal Ls (j
ω, t), Rs (jω, t) can be created.

As described above, when the performer B plays the music of the specific song α to convert the monaural signal M (jω, t) into a pseudo stereo signal, the performer B plays the music of the song α. The sum signal W (jω, t) in the stereo signal obtained by the performer A playing the music of the same music as the above-mentioned specific music α using the same content and music as the music used when )
And the time-varying transfer function F obtained based on the difference signal S (jω, t)
The pseudo difference signal G is obtained by applying (jω, t) to the monaural signal M (jω, t) which is the target of the stereo signalization.
(Jω, t) and the left and right channel signals Ls (jω, t) and Rs (jω, t) of the pseudo stereo signal are generated.
The effect of the stereo reproduction by t) and Rs (jω, t) is that when the performer A plays the same music as the specific song α, the performer B plays the music of the specific song α. It will have a stereo effect.

However, as described above, the reason why the monaural signal can be converted into a good stereo signal is that the monaural signal M (jω, t) is already described.
And the time-varying transfer function F (jω, t) is in a good synchronization state, and the monaural signal M (jω, t) and the time-varying transfer function F (jω, t) can be synchronized. If it is not, stereo signalization in a good state cannot be realized.

As a practical matter, even if the same performer plays the same piece of music a plurality of times, the same result may not always be obtained, but the monaural signal M (jω) described above is not always obtained. , t) and the time-varying transfer function F (jω, t) are not always in good synchronization.

Therefore, in the signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention, it is substantially the same as the music that generated the monaural signal M (jω, t) that is the target of the stereophonic signal conversion. Music (track, score,
A time-varying transfer function F (jω, t) = S (as shown in the above equation (10ωt) derived from a stereo signal generated by music when various elements such as musical instruments are played in the same manner) jω, t) / W (jω, t) ... (10ωt) is a monaural signal M that is the subject of stereo signalization
(Jω, t), so that the time-varying transfer function F (j
ω, t) is expanded / contracted on the time axis, and this is the third point of the basic concept of the signal processing method of the present invention for converting a monophonic signal into a stereophonic signal.

Next, referring to FIG. 4, music that is substantially the same as the music that has generated the monaural signal M (jω, t) that is the target of stereo signal conversion (various items such as music, musical score, and type of musical instrument). The time-varying transfer function F (jω, t) as shown in the above equation (10ωt) derived from the stereo signal generated by the music (when the elements are played the same) is targeted for stereo signalization. A case will be described in which the time-varying transfer function F (jω, t) is expanded / contracted on the time axis so as to be synchronized with the monaural signal M (jω, t).

FIG. 4 (a) is a sum signal W (j) generated based on the acoustic signal obtained when the performer A plays the music of the piece α.
Fig. 4 (b) showing the envelope of ω, t).
FIG. 4 is a diagram showing an envelope of a monaural signal M (jω, t) generated based on an acoustic signal obtained when the performer B plays the music of the program α, and further, FIG. c) is applied to the acoustic signal obtained when the performer A plays the music of the program α, using the same score as that used by the performer B when playing the music of the program α. Time-varying transfer function F (jω, t) of the filter, that is, j
Time-varying transfer function F of the filter, which is a two-variable function of ω and t
(Jω, t) is a diagram in the case where it is displayed by adopting an expression method similar to a sound sonargram (speech), and the amplitude of the time-varying transfer function F (jω, t) of the filter is the sound sonargram ( Voice print)
Similarly, it is represented by shading (however, in the accompanying drawings attached to the specification of the patent application, it is forbidden to describe drawings that are shaded to fill, so the shading is not shown in the drawings. ).

In FIG. 4, ta indicates a time axis when the performer A performs, and tb indicates a time axis when the performer B performs. Moreover, between (a) to (c) of FIG. 4 and (a) of FIG.
The dotted line between (d) to (d) is for indicating the same time point.

As described above, the moral signal can be converted into a good stereo signal only when the monaural signal M (jω, t) and the time-varying transfer function F (jω, t) are in a good synchronization state. However, the monaural signal M (jω, t) shown in FIG. 4 (b) as an example of the monaural signal to be stereophonically converted, that is, the program α
Of the monaural signal M (jω, t) generated based on the acoustic signal obtained when the performer B plays the music of
When the above-mentioned signal amplitude variation mode and the above-mentioned monaural signal M (jω, t) to be stereo-signaled are converted into stereo signals, the monaural signal M (jω, t)
The sum signal W (jω, t) of the stereo signals used to create the time-varying transfer function F (jω, t) to be applied to, that is, the curve α whose envelope is illustrated in FIG. 4 (a). Is generated based on the acoustic signal obtained when the performer A plays the music of the song α, using the score of the same content as that used by the performer B when playing the music of the song α. The manner in which the signal amplitude of the added sum signal W (jω, t) changes on the time axis ta differs from each other as indicated by the solid arrow between (a) and (b) in FIG. In the case of the sum signal W (j) shown in FIG.
ω, t) is obtained, that is, the music of the program α is used by the performer A to reproduce the music of the program α using the same score as that used by the performer B when playing the music of the program α. A time-varying transfer function F (jω, t) created from a stereo signal obtained based on the acoustic signal obtained when playing music (similar to a sonargram of voice in (c) of FIG. 4) It is displayed using the expression method} (b) of FIG. 4 described above.
When applied to the monaural signal M (jω, t) shown in FIG. 1, a good stereo signal cannot be generated.

Therefore, in the signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention, the envelope of the monaural signal M (jω, t), which is the target of stereo signalization, and the target of stereophonic signal conversion, The stereo signal generated by the music that generated the monaural signal M (jω, t) that is generated and the music that is substantially the same (the music when the musical composition, the musical score, the type of the musical instrument, and other elements are the same). By comparing with the envelope of, the monaural signal M (jω, t), which is the target of stereo signalization,
Time axis tb and the music that has generated the monaural signal M (jω, t) targeted for stereo signal conversion are substantially the same as each other (various elements such as music, musical score, and type of musical instrument are the same). The relationship between "advance" and "delay" with respect to the time axis ta of a stereo signal generated by the To the time-varying transfer function F (jω, t)
The time axis of (jω, t) is expanded / contracted to obtain the time-varying transfer function F ′ (jω, t) as illustrated in (d) of FIG. The time-varying transfer function F '(jω, t) in the state in which the time axis correction is performed is illustrated in the envelope of FIG. 4 (b) as a monaural signal to be stereo-converted. Signal M (jω,
t) In other words, as described above, the music signal of the same content of the same music α (the music of the music α played using the score of the same music) is added to the acoustic signal obtained when the performer B plays. The pseudo difference signal in a state of being synchronized with the monaural signal M (jω, t) that is the target of stereo signal conversion by applying the monaural signal M (jω, t) generated based on the following equation. G (jω, t) = M (jω, t) × F ′ (jω, t) ... (12 ′
ωt) is created.

FIG. 5 shows the sum signal W (jω, t) of the stereo signals shown in FIG. 4 (a) and the stereo signal shown in FIG. 4 (b). The non-linear function for giving the matching of the time axis with the monaural signal M (jω, t) which is present is shown as an example. This is, for example, that of the stereo signal shown in (a) of FIG. Sum signal W (j
The peaks and valleys in the envelope of ω, t) and (b) in Fig. 4.
It can be obtained by comparing with the peaks and valleys in the envelope of the monaural signal M (jω, t) targeted for the stereo signal ratio shown in FIG. In FIG. 4, the horizontal axis ta shows the time passage of the sum signal W (jω, t) of the stereo signals, and the vertical axis shows the time passage of the monaural signal M (jω, t) which is the target of the steres signal. Is shown.

If the sum signal W (jω, t) of the stereo signals and the monaural signal M (jω, t) to be converted into the stereo signal are in a synchronous relationship, it is shown in (a) of FIG. , Etc. are all located on the straight line shown in FIG. 5, but in reality, the sum signal W (jω, t) of the stereo signal and the object of stereo signalization Since it does not necessarily have a synchronous relationship with the monaural signal M (jω, t) which is set to, the sum signal W (jω) of the stereo signals located at time t1 in FIG. , t) corresponding to the index point, the point corresponding to the index of the monaural signal M (jω, t) which is the object of stereo signalization is as shown in FIG. 4 (b). It is located at time t1 '. Then, from each point in FIG.
The solid arrow drawn toward each point in (b) of the figure is for clarifying the correspondence relationship between the two points corresponding to each other in the above-mentioned figures.

Therefore, in FIG. 5, the coordinate point corresponding to the above-mentioned index is located outside the straight line, and in the same manner as the case of the above-mentioned index, it corresponds to many other indexes such as ... The coordinate points to be plotted are shown in FIG. 5 and the respective coordinate points are connected to obtain a non-linear function as shown by the dotted curve in FIG.

Therefore, the time-varying transfer function F derived from the stereo signal
For (jω, t), for example, the operation of moving the time-varying transfer function F (jω, t) corresponding to the index to the time t1 ′ is performed by the non-linear operation shown by the dotted line in FIG. By carrying out based on the function, the monaural signal which is the target of stereo signalization corresponding to the index point in the sum signal W (jω, t) of the stereo signals located at time t1 in FIG. 4 (a). A new time-varying transfer function F ′ (jω, t) having a synchronous relationship with the signal M (jω, t) is obtained. The above operation is shown by the relationship between FIG. 4 (c) and FIG. 4 (d).

The envelope of the sum signal W (jω, t) of the actual stereo signal and the monaural signal M (jω, t) to be converted into the stereo signal is necessarily (a) in FIG. (B)
Since it does not necessarily have a clear correspondence relationship as exemplified in the above, by applying a technique such as linear matching or DP matching which is practically used in the speech recognition technology,
The above object can be achieved. The above-mentioned linear matching or DP matching method is a well-known technical means that is widely used in the field of speech recognition, and a detailed description thereof will be omitted. In short, the time-varying transfer function F (jω, t). ) Is time-aligned (corrected) as shown in FIG.
The time axis of (jω, t) is expanded / contracted to form a new time-varying transfer function F ′ (jω, t) as illustrated in (d) of FIG. 4 with the time axis corrected. Let them do it.

So far, the sum signal W (jω, t) of the steres signal and the monaural signal M (jω,
Paying attention to the envelope with t), the time-varying transfer function F (j
Although the case of operating the time axis with respect to ω, t) has been described, the time axis of the time-varying transfer function F (jω, t) is manipulated to create a new time-varying transfer function F ′ (jω, t). t) is formed by, for example, the sum signal W (jω, t) of stereo signals and the monaural signal M (j
ω, t) is closely observed, a non-linear function is derived from the contrast between the two, and the time axis of the time-varying transfer function F (jω, t) is manipulated to create a new time-varying function. It is also possible to form the transfer function F '(jω, t).

The configuration principle of the signal processing method for converting the monophonic signal of the present invention into a stereophonic signal is as described above. Next, referring to FIG. 1 and other drawings, the stereophonic signal of the monophonic signal of the present invention will be described. A specific description will be given of a signal processing method for conversion.

FIG. 1 is a block diagram of a signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention.
In FIG. 1, 1 is a signal source of a monaural signal obtained by the performer B playing the music of the program α, and 2 is a performer A.
Is a signal source of a stereo signal obtained by the performer A playing the music of the program α using the music of the program α as the music used by the performer B when playing the music of the program α. Is.

The waveform of the monaural signal M (jω, t), which is obtained by the performer B playing the music of the program α from the signal source 1 of the monaural signal, is converted into a digital signal by the analog-digital converter 3. It is stored in the memory 5 as data.

In addition, the stereo signal from the signal source 2 of the stereo signal, that is, the music of the program α by the performer A is changed to the performer B.
, The left and right channel signals L (jω, t), R of the stereo signal obtained by the performer A playing the music of the music α using the music having the same contents as the music used when the music of the music α was played.
(Jω, t) is stored in the memory 6 as waveform data of the left and right channel signals L (jω, t) and R (jω, t) of the stereo signal converted into the digital signal by the analog / digital converter 4. .

When a signal source capable of transmitting a monaural signal M (jω, t) in the form of a digital signal and left and right channel signals L (jω, t), R (jω, t) is used, the above-mentioned analog is used. Needless to say, the digital converters 3 and 4 are unnecessary.

The waveform data of the monaural signal M (jω, t) in the form of a digital signal stored in the memory 5 is read out from the memory 5 and the matrix calculation circuit 20, the digital filter processing circuit 19, and the envelope calculation circuit 7 are read. Is supplied to.

In the envelope calculation circuit 7, the monaural signal M (jω, t) is read from the waveform data of the digital signal form monaural signal M (jω, t) read from the memory 5.
t) envelope data is created and stored in the memory 8 used as a storage device for the envelope data of the monaural signal M (jω, t).

Also, the left and right channel signals L (read out from the memory 6 which stores the waveform data of the left and right channel signals L (jω, t), R (jω, t) of the stereo signal converted into the digital signal described above. The waveform data of jω, t) and R (jω, t) are supplied to the calculation circuit 14 for the time-varying transfer function F (jω, t) and the calculation circuit 10 for the sum signal W (jω, t) of stereo signals. To be done.

Calculation circuit for sum signal W (jω, t) of stereo signals described above
Sum signal W (jω, t) of stereo signals output from 10
Are stored in the memory 11 of the waveform data of the sum signal W (jω, t).

The waveform data of the sum signal W (jω, t) of the digital signal type stereo signal stored in the memory 11 is read from the memory 11 and supplied to the envelope calculation circuit 12.

In the envelope calculation circuit 12, the envelope data of the sum signal W (jω, t) of the stereo signal is obtained from the waveform data of the sum signal W (jω, t) of the stereo signal in the digital signal form read from the memory 11. Are stored in the memory 13 used as a storage device for the envelope data of the sum signal W (jω, t) of the stereo signals.

The monaural signal M (j
The envelope data of ω, t) and the envelope data of the sum signal W (jω, t) of the stereo signals read from the memory 13 are supplied to the time-axis matching calculation circuit 9, where, for example, DP matching or the like is performed. Thus, the time axis matching calculation is performed, and the output signal from the time axis adjusting calculation circuit 9 is stored in the memory 16 as the data of the time axis matching (non-linear) function.

Further, in the calculation circuit 14 of the time-varying transfer function F (jω, t) to which the waveform data of the left and right channel signals L (jω, t) and R (jω, t) read from the memory 6 is supplied, The left and right channel signals L (jω,
t), R (jω, t) waveform data, the time-varying transfer function F (j
ω, t) is calculated and the time-varying transfer function F output from it is calculated.
Make (jω, t).

That is, in the calculation circuit 14 for the time-varying transfer function F (jω, t), the left and right channel signals L (j
From the waveform data of ω, t) and R (jω, t), (1ω
t) and (2ωt), a sum signal W (jω, t) and a difference signal S (jω, t) are generated, and then the sum signal W (j
By using (ω, t) and the difference signal S (jω, t), the time variable transfer function F (jω, t) is created by the equation (10ω, t).

W (jω, t) = L (jω, t) + R (jω, t) (1ω
t) S (jω, 5) = L (jω, t) −R (jω, t) (2ω
t) F (jω, t) = S (jω, t) / W (jω, t) ... (10ωt) Time-varying transfer function F (jω, t) generated by the calculation circuit 14 of the time-varying transfer function F (jω, t). jω, t) is stored in the memory 15, and the time-varying transfer function F (jω, t) output from the memory 15 is supplied to the time-axis matching processing circuit 17.

The time axis matching processing circuit 17 described above includes the memory described above.
The data of the time-axis matching (non-linear) function read from 16 is also supplied.
The time axis of the time-varying transfer function F (jω, t) read from
Time axis alignment (non-linear) read from the memory 16 described above
A new time-varying transfer function F ′ (jω, t) in a state where the time axis is corrected is output by expanding and contracting with the function data, and a new time-varying transfer function in the state where the time axis is corrected is output. F '
(Jω, t) is stored in the memory 18.

FIG. 6 illustrates the contents of the time-axis matching processing performed in the time-axis matching processing circuit 17 described above. FIG. 6 (a) shows the time-varying transfer function F (jω, t). It is a three-dimensional display (only the amplitude characteristic is shown), and FIG. 6B shows the time-varying transfer function F (jω, t) shown in FIG. The new time-varying transfer function F ′ (jω, t) obtained by applying the above, that is, the time-varying transfer function F (jω, t) shown in FIG. 6 shows a state in which the time axis is expanded and contracted by the time axis matching function described with reference to FIGS. 4 and 5.

That is, the time-varying transfer function F (jω, t1) at time t1 shown in (a) of FIG. 6 is F (jω) obtained by performing time expansion / contraction in FIG. 6 (b). , t1 ′)
, The time is changed from t1 to time t1 ′.

FIG. 6 (c) is obtained by performing interpolation processing on the new time-varying transfer function F ′ (jω, t) shown in FIG. The new time-varying transfer function F ′ (jω, t) is shown, and the new time-varying transfer function F ′ (jω, t) thus expanded / contracted in the time axis is shown.
Interpolation processing is performed on t) because the signal processing is performed by a discrete digital signal, so that the new time-varying transfer function F ′ (jω, t) is represented by time t1, t2, t3 ... This is because it is required to be converted into a time-varying transfer function that is synchronized with the time of the waveform data. For example, a new time at the time t1 shown in ′ of FIG. The variable transfer function F ′ (jω, t1) is obtained.

The new time-varying transfer function F ′ (jω, t) read out from the memory 18 in the time-axis corrected state is supplied to the digital filter processing circuit 19.
As described above, the filter processing circuit 19 is also supplied with the waveform data of the monaural signal M (jω, t) output from the memory 5, and the digital filter processing circuit 19 performs time-axis correction. When the player B plays the monaural signal M (jω, t), that is, the music of the song α, which is the new time-varying transfer function F ′ (jω, t) in the above state. Is applied to the monaural signal M (jω, t) generated on the basis of the acoustic signal obtained as described below, and is synchronized with the monaural signal M (jω, t) that is the target of stereo signalization. Pseudo difference signal G (jω, t) = M (jω, t) × F ′ (jω, t) ... (12 ′
ωt) is generated, and the waveform data of the pseudo difference signal G (jω, t) is stored in the memory 21.

In the digital filter processing circuit 19, a monaural signal M (jω, t) generated based on the acoustic signal obtained when the performer B plays the music of the music piece α to be stereophonicized is next generated. The pseudo difference signal G (jω, t) = M (jω, t) × F in a state of being synchronized with the monaural signal M (jω, t), which is the object of stereo signalization, by applying the equation ′ (Jω, t)… (12 ′
FIG. 7 illustrates the procedure for producing ωt).

FIG. 7 shows what is called digital filtering in the frequency domain. The monaural signal M (jω, t) is cut out at a predetermined timing for a predetermined time width. And undergo Hanning window processing.

Next, the partial waveform of the monaural signal M (jω, t) is Fourier-transformed, and the time-varying transfer function F'corresponding to the timing of the Fourier transform (time t2 in the illustrated example).
After being weighted by (jω, t), the inverse Fourier transform is performed, and at this stage, the pseudo difference signal G (j
The partial waveform of ω, t) is obtained.

The pseudo difference signal G (jω, t) obtained by sequentially performing the same signal processing as described above at successive times
The pseudo difference signal G (jω, t) is formed by connecting the subsequent partial waveforms of, and is stored in the memory 21.

The waveform data of the pseudo difference signal G (jω, t) read from the memory 21 and the waveform data of the monaural signal M (jω, t) read from the memory 5 are combined in the matrix calculation circuit 20.
And Ls (jω, t) = {M (jω, t) + G (jω, t)} / 2 (8ωt) Rs (jω, t) = {M (jω, t) in the matrix calculation circuit 20. -G (jω, t)} / 2 (9ωt) is performed to generate left and right channel signals Ls (jω, t) and Rs (jω, t) of the pseudo stereo signal, and the pseudo steres signal Left and right channel signals Ls (jω, t), Rs
The waveform data of (jω, t) is stored in the memory 22. The waveform data of the left and right channel signals Ls (jω, t) and Rs (jω, t) of the pseudo stereo signal read from the memory 22 is converted into an analog signal by the digital / analog converter 23. Is the monaural signal M (jω, t) of the music of the program α played by performer B,
Using a musical score having the same contents as the musical score used by the performer B when playing the music of the song α, a pseudo stereo signal added with a stereo sound effect in the stereo signal of the music of the song α played by the performer A It is done.

Since the pseudo-stereo signal stored in the memory 21 is a digital signal, it goes without saying that the digital / analog converter 23 is not necessary when the succeeding recording device is a digital tape recorder. .

(Effects of the Invention) As is clear from the above description, in order to perform a signal processing method for converting a monophonic signal into a stereophonic signal by applying a digital signal processing technique, it is targeted for stereophonic signal conversion. Time-varying transfer function by dividing the difference signal between the sum signal and the difference signal formed by the right channel signal and the left channel signal of the stereophonic signal by the second channel of the same music To obtain the sum signal of the stereophonic signals and the monophonic signals targeted for stereophonic signal synchronization so as to synchronize with the time-varying transfer function described above for the monophonic signals targeted for stereophonic signal conversion. By comparing each envelope with the signal , Deriving a non-linear function indicating a time axis synchronous relationship between the sum signal of the stereophonic signals and the monophonic signal that is the target of stereophonic signal conversion, and based on the non-linear function, the time axis of the time-varying transfer function described above. To obtain a new time-varying transfer function in synchronization with the monophonic signal that is the target of stereophonic signal conversion, and for the monophonic signal that is the target of stereophonic signal conversion, Means for applying the new time-varying transfer function described above to obtain a pseudo difference signal;
A signal processing method for converting a monophonic signal that is a target of stereophonic signal conversion to a monophonic signal that is a target of stereophonic signal conversion and a means for performing an operation of restoring the stereophonic signal from the pseudo difference signal is a stereo signal. In the relationship between the sum signal W (jω, t) of the left and right channel signals L (jω, t), R (jω, t) and the difference signal S (jω, t) in FIG.
A function F (jω, t) obtained by dividing (jω, t) by the sum signal W (jω, t), that is, F (jω, t) = S (jω, t) / W (jω, t) Considering (A), the transfer function F as shown in the above equation (A)
If a filter having (jω, t) can be realized, the difference signal S between the left and right channel signals L (jω, t) and R (jω, t) in the stereo signal is obtained by using the relationship of the equation (A).
(Jω, t) is the sum signal W (jω, t) of the left and right channel signals L (jω, t), R (jω, t) in the stereo signal.
And the transfer function F (jω, t) of the filter described above, S (jω, t) = W (jω, t) × F (jω, t) (B) As in the formula (B), Focusing on the fact that it can be obtained, the left and right channel signals L in the stereo signal
(Jω, t), R (jω, t) sum signal W (jω, t), W (jω, t) = L (jω, t) + R (jω, t) (C), and monaural signal M (Jω, t), M (jω, t) = L (jω, t) + R (jω, t) (D) are substantially the same, and therefore the sum signal W in the above equation (B) is Instead of (jω, t), monaural signal M (jω, t)
To obtain the required pseudo difference signal G (jω, t) = G (jω, t) = M (jω, t) × F (jω, t) (E) by the above equation (E) When the pseudo signal G (jω, t) required for converting the monophonic signal to the stereophonic signal is obtained as described above, Ls (jω, t) = {M (Jω, t) + G (jω, t)} / 2 (F) Rs (jω, t) = {M (jω, t) −G (jω, t)} / 2 (G) The left signal Ls (j in the pseudo stereo signal obtained by converting the monaural signal M (jω, t) into a stereo signal by
ω, t) and the right signal Rs (jω, t) can be generated, but the pseudo signal G (jω, t) is a monaural signal, as is apparent from the equation (E). M (j
.omega., t) is obtained by weighting the transfer function F (j.omega., t) of the filter, so that the pseudo difference signal required from the monaural signal M (j.omega., t) targeted for stereo signalization is obtained. Transfer function F such that G (jω, t) is obtained
A filter with (jω, t) is needed. And
The transfer function F (jω, t) described above is, as shown in equation (A), the difference signal S (of the left and right channel signals L (jω, t), R (jω, t) in the stereo signal. jω, t) is a transfer function F (jω, t) obtained by dividing the sum signal W (jω, t) by the sum signal W (jω, t).
Since the sum signal, the sum signal W (jω, t), and the like are so-called time-varying spectra whose contents change with time, the function F (jω, t) described above also changes with time. The characteristics of the filter having the time-varying transfer function shown in the above equation (A) change with the progression of the music mainly depending on the organization and arrangement of the musical instrument. However, except for a special case, it is practically impossible to calculate the time-varying transfer function F (jω, t) by applying some law to the monaural signal M (jω, t). It is one.

Therefore, in the signal processing method for converting a monophonic signal into a stereophonic signal of the present invention, the stereo signal S (jω, t) of the same music as the monaural signal M (jω, t) to be stereophonicized is used. From above (A)
The time-varying transfer function F (jω, t) = S (jω, t) / W (jω, t) (A) is derived as shown in the equation, and the time-varying transfer function F (jω, t) is stereophonic. When applied to the monaural signal M (jω, t) targeted for conversion, the monaural signal M (j targeted for stereo conversion is used.
ω, t) pseudo-difference signal G (jω, t) already described in equation (E),
That is, G (jω, t) = M (jω, t) × F (jω, t) (E) The following relation is used to create the time-varying monaural signal M (jω, t). A monaural signal M to be stereo-signaled by using the transfer function F (jω, t)
When creating the pseudo difference signal G (jω, t) of (jω, t), the monaural signal M that is the target of stereo signalization
(Jω, t) and the time-varying transfer function F to be applied to it
As for the point that (jω, t) is required to be synchronized, it is derived from the stereo signal of the same music as the monaural signal M (jω, t) which is the target of stereo signalization. Time-varying transfer function F as shown in equation (A)
(Jω, t) = S (jω, t) / W (jω, t) (A) is a monaural signal M (j) that is the subject of stereo signalization.
ω, t), the time-varying transfer function F (jω, t)
So that it can be expanded and contracted on the time axis, the time axis of the monaural signal M (jω, t) targeted for stereo signalization and the monaural signal M (jω, t targeted for stereo signalization The relationship between “advance” and “delay” between t) and the time axis of the stereo signal generated by playing the same music of the same song is found, and the “advance” and “delay” of both time axes are found. Is applied to the time-varying transfer function F (jω, t) to expand / contract the time axis of the time-varying transfer function F (jω, t), and a new time-varying transfer function F ′ in the state where the time axis correction is performed is performed. (J
ω, t), and the new time-varying transfer function F ′ (jω, t) in the state where the time axis correction is performed is used as a monaural signal to be stereo-signalized, and the music of the track α is The monaural signal M (jω, t) generated based on the acoustic signal obtained when the performer B plays is applied as in the following equation to obtain a monaural signal M to be converted into a stereo signal.
Pseudo difference signal in the state of being synchronized with (jω, t) G (jω, t) = M (jω, t) × F ′ (jω, t) ...
(E ') is produced. In the signal processing method for converting the monophonic signal into the stereophonic signal of the present invention, the performer B plays the music of the specific song α as described above. Obtained monaural signal M (jω, t)
Is converted into a pseudo-stereo signal by using a musical score having the same contents as the musical score used by the performer B at the time of playing the music of the specific song α, and stereo by the music of the song α played by the performer A. Time-varying transfer function F (j) obtained based on the sum signal W (jω, t) and the difference signal S (jω, t) in the signal.
ω, t) is applied to the monaural signal M (jω, t) which is the target of the stereo signalization, and the pseudo difference signal G (j
ω, t) to generate the left and right channel signals of the pseudo stereo signal
Left and right channel signals Ls (jω, t), R of the pseudo stereo signal when Ls (jω, t), Rs (jω, t) are generated
The effect of the stereo reproduction by s (jω, t) is the same as the score used by the performer B when playing the music of the specific song α, with respect to the performance of the music of the specific song α by the performer B. The music of No. α is used to give a stereo effect when the performer A plays, and the signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention is used. When the generated pseudo stereo signal is monaurally reproduced, the pseudo difference signal is canceled and is not reproduced, and the monaural signal M (jω, t)
Since only the audio signal is reproduced, the compatibility between the reproduction of the stereo signal and the reproduction of the monaural signal is perfect. According to the signal processing method for converting the monophonic signal into the stereophonic signal of the present invention, the conventional (1) monaural signal The sound of each musical instrument is localized by applying various filters to the sound of each musical instrument. (2) A transfer circuit (all-pass filter) is applied to a monaural signal to give a sense of expansion to the reproduced sound field. (3) Reverberation and reverberation are added to the monaural signal so that a stereophonic expansive feeling can be felt. (4) The above (1) to the monaural signal
It is possible to satisfactorily solve all the problems as described above in the method of converting a monophonic signal into a stereophonic signal, which is typified by applying the means of (3) in combination.

[Brief description of drawings]

FIG. 1 is a block diagram of a signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention, and FIGS. 2 to 7 are configurations of a signal processing method for converting a monophonic signal into a stereophonic signal according to the present invention. It is a figure for demonstrating a principle and an operation principle. 1 ... Signal source of monaural signal obtained by performer B playing music of song α, 2 ... Source of stereo signal obtained by performer A playing music of song α, 3, 4 ... ... analog / digital converter, 5 ... memory of waveform data of monaural signal M (jω, t) converted to digital signal, 6 ... left and right channel signals L (jω, of stereo signal converted to digital signal t), R (jω, t) waveform data, memory, 7 ... Envelope calculation circuit, 8 ... Memory used as monaural signal M (jω, t) envelope data storage device, 9 ... Time axis Matching calculation circuit, 10 ...
… Calculation circuit for sum signal (jω, t) of stereo signals, 11 ……
Memory for waveform data of sum signal W (jω, t), 12 ... Envelope calculation circuit, 13 ... Sum signal W of stereo signal
Memory used as a storage device for envelope data of (jω, t), 14 ... Calculation circuit for time-varying transfer function F (jω, t), 15 ... Memory for time-varying transfer function F (jω, t) ,
16 ... Memory of time axis matching (non-linear) function data, 17
...... Processing circuit for time-axis matching, 18 ... Memory of time-varying transfer function F '(jω, t) with time-axis correction performed, 19 ...
Digital filter processing circuit, 20 ... Matrix calculation circuit, 21 ... Memory for waveform data of pseudo difference signal G (jω, t), 22 ... Memory for pseudo stereo signal, 23 ... Digital / analog converter,

Claims (1)

[Claims] 1. A signal processing method for converting a monophonic signal into a stereophonic signal by applying a digital signal processing technique, which comprises a monophonic signal targeted for stereophonic signal conversion and two channels of the same music. A means for obtaining a time-varying transfer function by dividing the difference signal in the sum signal and the difference signal formed by the right channel signal and the left channel signal of the stereophonic signal by the sum signal, and being a subject of stereophonic signal conversion. By synchronizing the respective signals of the sum signal of the stereophonic signals and the monophonic signal which is the target of the stereophonic signal conversion so as to synchronize with the above-mentioned time-varying transfer function with respect to the monophonic signal, Stereophonic signal's Japanese parent and stereophonic signal Deriving a non-linear function representing the synchronization relationship between the time axis of the monophonic signal is subject to reduction,
A means for changing the time axis of the time-varying transfer function based on the non-linear function to obtain a new time-varying transfer function in a synchronous relationship with a monophonic signal that is a target of stereophonic signal conversion, For a monophonic signal that is the subject of stereophonic signalization, means for obtaining a pseudo difference signal by applying the new time-varying transfer function described above, and a monophonic signal that is the subject of stereophonic signalization, A signal processing method for converting a monophonic signal into a stereophonic signal, which comprises means for performing an operation of restoring the stereophonic signal from the pseudo difference signal.