JP5161848B2 - Multi-channel sound reproduction apparatus and multi-channel sound reproduction program - Google Patents

Description

  The present invention relates to acoustic signal processing and speaker driving methods, and in particular, multi-channel acoustic reproduction that realizes multi-channel acoustic reproduction using a smaller number of amplifiers than the number of speakers and signal processing based on principal component analysis. The present invention relates to a device and a multi-channel sound reproduction program.

  In recent years, with the spread of digital broadcasting and the increase in capacity of DVDs and the like, opportunities to use multi-channel audio devices in the home are increasing. Multi-channel sound is a system that reproduces a high acoustic presence, represented by 5.1ch surround, and is a sound reproduction method that uses a plurality of sound signals and speakers.

  In multi-channel sound, one speaker is required for one channel, but generally one amplifier (amplifier / drive device) is required for one speaker, so multi-channel sound reproduction. To achieve this, the same number of amplifiers as speakers are required.

  FIG. 9 is a block diagram showing a configuration of a conventional multi-channel sound reproducing apparatus. In FIG. 9, one amplifier is connected to one speaker, and the same number of acoustic signals (channels) as the number of speakers are output to the speakers via the amplifier.

However, in multi-channel sound, it is known that not all channels are used evenly, and there is a certain degree of bias regardless of the type of content (Non-Patent Document 1).
In Non-Patent Document 1, principal component analysis is performed on several sound sources (contents) using a plurality of channels as variables at predetermined time intervals. Then, when principal components having a high average contribution rate are selected and the weight coefficient pattern is examined for them, it is shown that the predetermined channel is used as a particularly important channel compared with other channels.

  Further, Patent Document 1 proposes multi-channel driving by one amplifier based on this point. This is a method of analyzing the signal component of each channel and allocating the driving time weighted by the signal level to each speaker.

JP 2003-304125 A

Takashi Nakayama, Kimio Amagasaki, Akio Ando, "22.2 Analysis of inter-channel statistics of multi-channel audio content", Proceedings of the Acoustical Society of Japan, 2-1-6, pp.447-448, September 2006

  As shown in Non-Patent Document 1, in multi-channel sound, there is a bias in channel use regardless of the type of content. For this reason, there is a problem that a method of using amplifiers for all channels (speakers) and making them stand by regardless of the operation frequency is inefficient in terms of both power and room utilization.

  In addition, Patent Document 1 proposes multi-channel driving by one amplifier based on this point. However, for example, the Japan Broadcasting Corporation, which is the present applicant, has a driving method using one amplifier. Speakers such as 22.2ch sound and live sound field reproduction methods (Amagasaki et al., “Overview of high-quality live sound field reproduction method”, PROSOUND, April 2009, pp.49-54) When the number increases significantly, there is a problem that the silent time when no sound is output from a certain speaker increases.

  The present invention has been made to solve such problems, and efficiently and effectively distributes signals and power by using a smaller number of amplifiers than the number of driving speakers and signal processing by principal component analysis. An object of the present invention is to provide a multi-channel sound reproducing apparatus and a multi-channel sound reproducing program that realize simple and economical multi-channel sound.

  The present invention was devised to solve the above-mentioned problems. First, a multi-channel sound reproduction apparatus according to claim 1 is composed of speakers having the same number as the number of channels and amplifiers having a number smaller than the number of channels. A multi-channel sound reproducing apparatus that reproduces multi-channel sound, comprising a time division unit, a principal component analysis unit, a switching unit, and a weighting unit.

  In such a configuration, the multi-channel sound reproducing apparatus samples the input multi-channel sound signal by a time division unit a predetermined number of times at predetermined time intervals. Here, the predetermined time is a time that is sufficiently short for hearing, and is, for example, 2048 samples when the sampling frequency is 48 kHz.

  Then, the multi-channel sound reproduction apparatus performs principal component analysis on the sample data sampled for each channel by the time division means by the principal component analysis means, and the contribution rate obtained from the eigenvalues of the variance / covariance matrix and The weighting factor that is a component of the eigenvector corresponding to the eigenvalue and the principal component corresponding thereto are obtained, and a predetermined number of principal components having a high contribution rate are output as principal component signals at predetermined time intervals. At the same time, the contribution ratio and weighting factor of the output principal component signal are output as a weighting signal at each predetermined time interval. Here, the predetermined number refers to, for example, the number of principal components with a cumulative contribution rate of 80% or over 95%.

  Then, the multi-channel sound reproducing apparatus determines the distribution time of the principal component signal input from the principal component analysis means according to the contribution rate included in the weighted signal input from the principal component analysis means by the switching means, Switching is performed at each distribution time, and the principal component signal is output to an amplifier that is assigned to the speaker in advance.

  Further, the multi-channel sound reproduction apparatus amplifies by the amplifier using the weighting factor of the channel corresponding to one connected speaker among the weighting factors included in the weighting signal input from the principal component analysis unit by the weighting unit. The weighted principal component signal is weighted and output to the speaker.

The multi-channel sound reproduction apparatus according to claim 2 is the multi-channel sound reproduction apparatus according to claim 1, wherein the speaker is driven by a DC bias, and the weighting means is derived from the principal component analysis means. Among the weighting factors included in the input weighting signal, the alternating current amplified according to the weighting factor of the channel corresponding to one connected speaker is superimposed on the direct current bias voltage applied by the amplifier, and It is a bias circuit that outputs to a speaker.
With this configuration, it is possible to reproduce multi-channel sound with a smaller number of amplifiers than the number of speakers, even for speakers driven by DC bias.

  The multi-channel sound reproducing apparatus according to claim 3 is a multi-channel sound reproducing apparatus that reproduces multi-channel sound by using the same number of speakers as the number of channels and a number of amplifiers smaller than the number of channels. Means, principal component analysis means, switching means, and second switching means.

  In such a configuration, the multi-channel sound reproduction device samples the multi-channel sound signal a predetermined number of times at predetermined time intervals by the time division means. Further, the multi-channel sound reproducing device performs principal component analysis based on the sample data by the principal component analysis means, and outputs a principal component signal and a weighting signal. Further, the multi-channel sound reproducing device switches the main component signal by the distribution time by the switching means and outputs it.

  Furthermore, the multi-channel sound reproducing device uses the weighting factor for each channel corresponding to a plurality of connected speakers among the weighting factors included in the weighting signal input from the principal component analyzing unit by the second switching unit. The main component signal amplified by the amplifier is output to the connected speaker by switching the distribution time according to the weighting factor.

The multi-channel sound reproduction apparatus according to claim 4 is the multi-channel sound reproduction apparatus according to any one of claims 1 to 3, wherein the switching means switches the main component signal at a zero cross point. It is characterized by.
With such a configuration, the multi-channel sound reproducing device can suppress abnormal noise from the speakers.

  According to a fifth aspect of the present invention, there is provided a multi-channel sound reproduction program for reproducing a multi-channel sound with a speaker having the same number as the number of channels and an amplifier having a number smaller than the number of channels. It is characterized by functioning as component analysis means, switching means, weighting means.

  In such a configuration, the multi-channel sound reproduction program samples the multi-channel sound signal a predetermined number of times at predetermined time intervals by the time division means. The multi-channel sound reproduction program performs principal component analysis based on the sample data by the principal component analysis means, and outputs a principal component signal and a weighted signal. Further, the multi-channel sound reproduction program switches the main component signal according to the distribution time by the switching means and outputs it to the amplifier. The multi-channel sound reproduction program weights the input principal component signal for each channel by the weighting means and outputs the result to the speaker.

  According to a sixth aspect of the present invention, there is provided a multi-channel sound reproduction program for causing a computer to reproduce multi-channel sound with a number of speakers equal to the number of channels and a number of amplifiers smaller than the number of channels. It functions as a component analysis means, a switching means, and a second switching means.

  In such a configuration, the multi-channel sound reproduction program samples the multi-channel sound signal a predetermined number of times at predetermined time intervals by the time division means. The multi-channel sound reproduction program performs principal component analysis based on the sample data by the principal component analysis means, and outputs a principal component signal and a weighted signal. Further, the multi-channel sound reproduction program switches the main component signal according to the distribution time by the switching means and outputs it to the amplifier. In addition, the multi-channel sound reproduction program switches the main component signal at a time corresponding to the weighting coefficient by the second switching means and outputs it to each speaker.

  According to the first and fifth aspects of the present invention, the principal component analysis is performed on the multi-channel acoustic signal, and the number of the acoustic signals is reduced to be the principal component signal, so that the number of speakers smaller than the number of speakers is obtained. It can be amplified with an amplifier to reproduce multi-channel sound. For this reason, the power consumption by an amplifier can be reduced and the space by an amplifier can be reduced. On the other hand, even if the number of speakers is greatly increased, silence time during which no sound is output from the speakers can be reduced by not reducing the number of amplifiers extremely.

  According to the second aspect of the present invention, multi-channel sound can be reproduced with a smaller number of amplifiers than the number of speakers, even for speakers driven by a DC bias.

  According to the third and sixth aspects of the present invention, the principal component analysis is performed on the multi-channel acoustic signal, and the number of acoustic signals is reduced to be the principal component signal. It can be amplified with an amplifier to reproduce multi-channel sound. For this reason, the power consumption by an amplifier can be reduced and the space by an amplifier can be reduced. On the other hand, even if the number of speakers is greatly increased, silence time during which no sound is output from the speakers can be reduced by not reducing the number of amplifiers extremely.

  According to the fourth aspect of the present invention, since the switching unit switches the main component signal at the zero cross point, it is possible to suppress the generation of abnormal noise from the speaker.

1 is a block diagram illustrating a configuration of a multi-channel sound reproducing device according to a first embodiment of the present invention. It is a block diagram which shows the structure of the multi-channel sound reproduction apparatus which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the multichannel sound reproduction apparatus which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structure of the multi-channel sound reproduction apparatus which concerns on the 4th Embodiment of this invention. It is a figure which shows the switching method of the some main component signal by the switching means which concerns on the 1st thru | or 4th embodiment. It is a figure which shows the switching method to each speaker of the main component signal by the 2nd switching means which concerns on 3rd Embodiment. It is a flowchart which shows operation | movement of the multi-channel sound reproduction apparatus which concerns on the 1st Embodiment and 2nd Embodiment of this invention. It is a flowchart which shows operation | movement of the multi-channel sound reproduction apparatus which concerns on the 3rd Embodiment and 4th Embodiment of this invention. It is a block diagram which shows the structure of the conventional multichannel sound reproduction apparatus.

  Embodiments of the present invention will be described below with reference to the drawings.

[Configuration of multi-channel sound reproduction device]
[[Configuration of Multi-Channel Sound Reproducing Device (First Embodiment)]]
With reference to FIG. 1, the structure of the multi-channel sound reproduction apparatus 1 (1A) according to the first embodiment of the present invention will be described.

The multi-channel sound reproducing apparatus 1A according to the first embodiment performs principal component analysis on a multi-channel sound signal, amplifies only the main component having a high contribution rate with an amplifier, and weights and outputs it to a speaker.
Specifically, the multi-channel sound reproducing device 1A performs a principal component analysis by sampling a plurality of sound signals at time intervals that are sufficiently short for hearing. Then, the multi-channel sound reproducing device 1A switches only the main component having a high contribution rate among the main components at a time corresponding to the contribution rate, and outputs it to one amplifier. Further, the multi-channel sound reproducing device 1A weights the amplified output from the amplifier according to the weighting coefficient of each main component, and outputs the weighted signal to the speaker.

The multi-channel sound reproduction apparatus 1A includes a time division unit 10, a principal component analysis unit 12, a switching unit 14, an amplifier 16, a weighting unit 18, and a speaker 20. The weighting means 18 and the speaker 20 are respectively assigned to the weighting means 18 ₁ ,..., The weighting means 18 _N , and the speakers 20 ₁ ,. , it consists of the speaker 20 _N.

The time division means 10 samples a multi-channel acoustic signal at time intervals that are sufficiently short for hearing.
Specifically, the time division means 10 samples the acoustic signals (Ch1,..., ChN) of each channel input from the outside a predetermined number of times at predetermined time intervals. Then, the time division unit 10 outputs a predetermined number of sampled sample data to the principal component analysis unit 12 for each channel.
For example, when the sampling frequency is 48 kHz, the time division unit 10 samples 2048 sample data with 2048 samples as one predetermined time interval.

The principal component analysis means 12 performs principal component analysis based on the sample data and outputs a principal component signal.
Specifically, the principal component analysis unit 12 inputs M sample data for each of N channels from the time division unit 10. The principal component analysis unit 12 performs principal component analysis on the N × M sample data, where N is the number of variables and M is the sample size. Further, the principal component analyzing means 12 is a switching means that outputs a predetermined number of principal components as output signals (principal component signals) from the principal components having a high contribution rate at predetermined time intervals when the time division means 10 samples the sample data. 14 for output.
The principal component analysis method will be described later.

  For example, the principal component analysis unit 12 may output eight principal components, and it is known that the cumulative contribution rate in this case exceeds 95% (Non-Patent Document 1).

  The principal component analysis means 12 also outputs the contribution ratio and weighting factor (principal component) of the output principal component signal obtained by the principal component analysis for each predetermined time interval in which a predetermined number of principal component signals are output. The eigenvalues and eigenvectors obtained in the calculation of analysis) are output to the switching means 14 and the weighting means 18 as weighting signals.

The switching means 14 switches the main component signal according to the distribution time and outputs it.
Specifically, the switching unit 14 sets the distribution time of a predetermined number of principal component signals input from the principal component analysis unit 12 in proportion to the contribution rate included in the weighted signal input from the principal component analysis unit 12. The calculation is switched for each allocation time, and the corresponding principal component signal is output to the amplifier 16. Alternatively, the switching unit 14 may mechanically switch a plurality of turns sequentially at equal time intervals of, for example, about 1 MHz.
In the first embodiment, one main component signal is output.

With reference to FIG. 5, the method by which the switching means 14 switches a main component signal is demonstrated.
FIG. 5A shows the first principal component signal, and the area delimited by the left dotted line represents the distribution time allocated to the first principal component signal by the switching unit 14. Similarly, FIG. 5 (b) represents the second principal component signal, and the area divided by the dotted line in the middle represents the allocation time allocated to the second principal component signal by the switching means 14. FIG. 5 (c) represents the third principal component signal, and the area delimited by the right dotted line represents the distribution time allocated to the third principal component signal by the switching means 14.
A point where each principal component signal intersects the time axis is called a zero cross point.

The switching means 14 switches the principal component signal at the zero cross point. Therefore, first, the switching means 14, to release the first principal component signal at the zero-crossing point Z ₁ of the nearest point of the allocation time. Then, the switching means 14, in the zero-crossing point Z ₂ of the second principal component signal beyond the allocation time of the first principal component signal, switch to the second principal component signal. Similarly, switching means 14, the second principal component signal released by the zero-crossing point Z ₃ in nearest neighbor of the distribution time, the third principal component signal zero crossing point Z ₄ of exceeding the allocation time of the second principal component signal Thus, the third principal component signal is switched.
As a result, the principal component signal output from the switching unit 14 becomes a composite output on the right side of FIG. By doing so, the switching unit 14 can suppress the generation of abnormal noise from the speaker 20.

Returning to FIG. 1, the description of the configuration of the multi-channel sound reproducing device 1A will be continued.
The amplifier 16 amplifies the input main component signal.
Specifically, the amplifier 16 amplifies the principal component signal input from the switching means 14 and outputs the amplified signal to the weighting means 18 (weighting means 18 ₁ ,..., Weighting means 18 _N ). Here, the principal component signals output to each weighting means 18 _i (1 ≦ i ≦ N) are the same.
In the first embodiment, there is one amplifier 16.

The weighting means 18 weights the input principal component signal for each channel and outputs it, thereby realizing a volume ratio suitable for each speaker.
Specifically, the weighting means 18 _i (1 ≦ i ≦ N) inputs the weighting signal related to the principal component signal from the principal component analysis means 12 while the principal component signal is output from the amplifier 16, and The principal component signal is weighted by the weighting coefficient of channel i included in the weighted signal and output to the speaker 20 _i .

  The weighting means 18 is composed of a passive attenuator (for example, a variable resistor or a transformer). “Weighing the principal component signal with a weighting factor” means that the weighting means 18 has an internal ratio proportional to the weighting factor. This means that the amplitude (intensity) of the main component signal is changed by changing the resistance value or the like included in the signal.

The speaker 20 converts the input principal component signal into sound and outputs the sound.
Specifically, the speaker 20 _i (1 ≦ i ≦ N) receives the main component signal input from the weighting means 18 _i , converts it into sound, and outputs it.
This speaker 20 is provided for the number of acoustic signals.

  The speaker 20 may be a full frequency speaker or an LFE (Low Frequency Effect) speaker for effectively adding bass.

  According to the multi-channel sound reproducing apparatus 1A, the multi-channel sound signal is sampled by the time division means 10, and the sampled multi-channel sound signal is subjected to principal component analysis by the principal component analysis means 12, and a predetermined number is obtained. The main component signal is output by the switching means 14 according to the distribution time, the main component signal is amplified by the amplifier 16, the main component signal is weighted by the weighting signal 18, and the speaker 20 To output sound. For this reason, multi-channel sound can be effectively reproduced with one amplifier.

[[Configuration according to multi-channel sound reproducing apparatus (second embodiment)]]
With reference to FIG. 2, the structure of the multi-channel sound reproduction apparatus 1 (1B) according to the second embodiment of the present invention will be described.

The multi-channel sound reproduction apparatus 1B of the second embodiment performs principal component analysis on a multi-channel sound signal, amplifies only a main component having a high contribution rate by a plurality of amplifiers, and weights and outputs the amplified signal to a speaker. is there.
Specifically, the multi-channel sound reproducing device 1B performs a principal component analysis by sampling a plurality of sound signals at time intervals that are sufficiently short for hearing. Then, the multi-channel sound reproducing device 1B sequentially switches the main component having a high contribution rate among the main components at a time corresponding to the contribution rate, or mechanically, for example, at equal time intervals of about 1 MHz, sequentially several times. Output to multiple amplifiers. Further, the multi-channel sound reproducing device 1B weights the amplified output from each amplifier according to the weighting coefficient of each main component, and outputs the result to the speaker.

  The multi-channel sound reproduction apparatus 1B differs from the multi-channel sound reproduction apparatus 1A in that a plurality of amplifiers are provided, and each amplifier receives a predetermined number of principal component signals from the switching means via a predetermined number of weighting means. It is a point to output to the speaker.

The multi-channel sound reproduction apparatus 1B includes a time division unit 10, a principal component analysis unit 12, a switching unit 14B, an amplifier 16B, a weighting unit 18, and a speaker 20. The weighting means 18 and the speaker 20 are respectively assigned to the weighting means 18 ₁ ,..., The weighting means 18 _N , and the speakers 20 ₁ ,. , it consists of the speaker 20 _N. The amplifier 16B includes a plurality of amplifiers 16B ₁ ,..., And amplifiers 16B _K that are smaller than the number of speakers (number of channels) N (K <N). In FIG. 2, the number of amplifiers K is exemplified as 3.
Moreover, about the structure similar to 1 A of multichannel sound reproduction apparatuses demonstrated in FIG. 1, the same code | symbol is attached | subjected and the description is abbreviate | omitted.

The switching means 14B switches the main component signal according to the distribution time and outputs it.
In the second embodiment, the switching unit 14B is connected to a plurality of amplifiers 16B (amplifiers 16B ₁ ,..., Amplifiers 16B _K ) (K <N).

Then, the switching unit 14B obtains a contribution rate of each principal component using a weighted signal input from the principal component analysis unit 12 for a predetermined number of principal component signals input from the principal component analysis unit 12, and is proportional to the contribution rate. allocation time that, or to calculate the equivalent time interval of about mechanically example 1 MHz, switch to each its distribution time, the corresponding principal component signals plurality of amplifiers 16B _1, and outputs ..., to the amplifier 16B _K . Here, the switching means 14B outputs the same switched principal component signal to each connected amplifier 16B.

The amplifier 16B amplifies the input main component signal.
In the second embodiment, the amplifier 16B includes a plurality of amplifiers 16B ₁ ,..., And amplifiers 16B _K smaller than the number of speakers (number of channels) N (K <N). Each amplifier B _i (1 ≦ i ≦ K) is connected to a plurality of different weighting means 18 (weighting means 18 _i1 ,..., Weighting means 18 _{ij (i)} ) (Σ _i j (i) = N).

Then, the amplifier 16B _i (1 ≦ i ≦ K) amplifies the principal component signal input from the switching unit 14B, and weighting unit 18 (weighting unit 18 _i1 ,..., Weighting unit 18 _{ij (i)} ). (Σ _i j (i) = N). Here, the amplifier 16B _i outputs the same principal component signal to each connected weighting means 18.

  According to the multi-channel sound reproducing apparatus 1B, the multi-channel sound signal is sampled by the time division means 10, and the sampled multi-channel sound signal is subjected to principal component analysis by the principal component analysis means 12, and a predetermined number is obtained. The main component signal is output by the switching means 14B, the main component signal is switched according to the distribution time, the main component signal is amplified by the plurality of amplifiers 16B, the weighting signal 18 is weighted, Sound is output from the speaker 20. For this reason, multi-channel sound can be effectively reproduced by a plurality of amplifiers.

[[Configuration of Multi-Channel Sound Reproducing Device (Third Embodiment)]]
With reference to FIG. 3, the structure of the multi-channel sound reproduction apparatus 1 (1C) according to the third embodiment of the present invention will be described.

The multi-channel sound reproduction apparatus 1C according to the third embodiment performs principal component analysis on a multi-channel sound signal, amplifies only the main component having a high contribution rate with a single amplifier, and switches to a speaker by time division. It is.
Specifically, the multi-channel sound reproducing device 1C performs a principal component analysis by sampling a plurality of sound signals at time intervals that are sufficiently short for hearing. Then, the multi-channel sound reproducing device 1 </ b> C switches only the main component having a high contribution rate among the main components at a time corresponding to the contribution rate or mechanically at an equal time interval of, for example, about 1 MHz. Output to the amplifier. Furthermore, the multi-channel sound reproducing device 1C switches the amplified output from the amplifier at a time corresponding to the weighting factor for each channel of each main component, and outputs the main component to the speaker.

  The difference between the multi-channel sound reproduction apparatus 1A and the multi-channel sound reproduction apparatus 1A is that a second switching means 22 is provided in place of the weighting means 18, and the second switching means 22 is used for the main component signal from the amplifier. The point is that the output is switched in time division and output to the speaker.

The multi-channel sound reproduction apparatus 1C includes a time division unit 10, a principal component analysis unit 12, a switching unit 14, an amplifier 16C, a second switching unit 22, and a speaker 20. The speakers 20 are composed of a number of speakers 20 ₁ ,..., Speakers 20 _N corresponding to the number N of channels input to the time division means 10.
In addition, about the structure similar to 1 A of multichannel sound reproduction apparatuses demonstrated in FIG. 1, the same code | symbol is attached | subjected and the description is abbreviate | omitted.

The amplifier 16C amplifies the input main component signal.
Specifically, the amplifier 16 </ b> C amplifies the main component signal input from the switching unit 14 and outputs the amplified signal to the second switching unit 22.
In the third embodiment, there is one amplifier 16C.

The second switching means 22 switches the main component signal at a time corresponding to the weighting coefficient and outputs it to each speaker.
Specifically, the second switching means 22 uses the weighted signal corresponding to the principal component signal input from the principal component analysis means 12 for the amplified principal component signal input from the amplifier 16C, and For each (speaker), an allocation time proportional to the weighting factor or a mechanically equivalent time interval of about 1 MHz is calculated. Then, the second switching means 22 switches the principal component signal with the distribution time calculated for each channel, and outputs it to the speaker corresponding to the channel.

  The weighting signal input to the second switching means 22 is the weighting signal output from the principal component analysis means 12 to the switching means 14 and the weighting means 18 in the multi-channel sound reproducing device 1A (first embodiment), Are the same.

With reference to FIG. 6, a method of switching the main component signal to each speaker by the second switching means 22 will be described.
FIG. 6A schematically shows the first principal component signal input to the second switching means 22. FIG. 6B schematically shows the second principal component signal input to the second switching means 22 next.

"SP1" in FIG. 6 (a), second switching means 22, represented in the first principal component signal, the partial signal of the speaker 20 allocation time allocated to the channel corresponding to _1. Similarly, "SPN" is in FIG. 6 (a), the second switching means 22, represented in the first principal component signal, the partial signal of the speaker 20 allocation time allocated to channels corresponding to _N. "SP1" in FIG. 6 (b), the second switching means 22, represented in the second principal component signal, the partial signal of the speaker 20 allocation time allocated to the channel corresponding to _1.

Then, when the first principal component signal amplified from the amplifier 16C is input, the second switching means 22 calculates the distribution time for each channel corresponding to each speaker, and then as shown in FIG. , during distribution time corresponding to the "SP1", the partial signal, and outputs to the speaker 20 _1, repeating this, during the allocation period corresponding to "SPN", the partial signal, output to the speaker 20 _N To do.
Next, when the second principal component signal amplified from the amplifier 16C is input, the second switching means 22 calculates the distribution time for each channel corresponding to each speaker, and then is shown in FIG. as such, during the allocation period corresponding to "SP1", the partial signal, and outputs to the speaker 20 _1, repeating this.
Thus, the 2nd switching means 22 switches a main component signal by the calculated distribution time, and outputs it to each speaker.

Returning to FIG. 3, the description will be continued.
According to the multi-channel sound reproducing apparatus 1C, the multi-channel sound signal is sampled by the time division unit 10, and the sampled multi-channel sound signal is subjected to principal component analysis by the principal component analysis unit 12, and a predetermined number is obtained. The main component signal is output by the switching unit 14 according to the distribution time, the main component signal is amplified by the amplifier 16C, and the second switching unit 22 is time-divided into the main component signal, Sound is output from the speaker 20. For this reason, multi-channel sound can be effectively reproduced with one amplifier.

[[Configuration of Multi-Channel Sound Reproducing Device (Fourth Embodiment)]]
With reference to FIG. 4, the structure of the multi-channel sound reproduction apparatus 1 (1D) according to the fourth embodiment of the present invention will be described.

The multi-channel sound reproduction apparatus 1D according to the fourth embodiment performs principal component analysis on a multi-channel sound signal, amplifies only the main component having a high contribution rate with a plurality of amplifiers, and then switches to a speaker by time division. It is.
Specifically, the multi-channel sound reproducing device 1D performs a principal component analysis by sampling a plurality of sound signals at time intervals that are sufficiently short for hearing. Then, the multi-channel sound reproduction device 1D switches only a main component having a high contribution rate among the main components at a time corresponding to the contribution rate or mechanically at equal time intervals of, for example, about 1 MHz, Output to the amplifier. Furthermore, the multi-channel sound reproducing device 1D switches the amplified output from each amplifier at a time corresponding to the weighting factor for each main component channel and outputs the main component to the speaker.

  A difference between the multi-channel sound reproduction device 1D and the multi-channel sound reproduction device 1B is that, instead of the weighting means 18, a plurality of second switching means 22B are provided corresponding to the plurality of amplifiers 16C. The 2 switching means 22B is a point which outputs the signal to a speaker by switching the output of the principal component signal from the amplifier 16C in a time division manner.

The multi-channel sound reproducing device 1D includes a time division unit 10, a principal component analysis unit 12, a switching unit 14B, an amplifier 16C, a second switching unit 22B, and a speaker 20. The speaker 20 includes speakers 20 ₁ ,..., A speaker 20 _N according to the number N of channels input to the time division means 10. The amplifier 16C includes a plurality of amplifiers 16C ₁ ,..., And an amplifier 16C _K that are smaller than the number of speakers (number of channels) N (K <N). The second switching means 22B is composed of the same number of second switching means 22B ₁ ,..., Second switching means 22B _{K as} the number K of amplifiers 16C. In FIG. 4, the number of amplifiers 16 </ b> C and the number of second switching means 22 </ b> B are each exemplified as 3.
In addition, the same components as those of the multi-channel sound reproducing apparatuses 1A, 1B, and 1C described in FIG. 1, FIG. 2, and FIG.

The second switching means 22B switches the principal component signal to a time corresponding to the weighting factor or mechanically at an equal time interval of about 1 MHz, for example, and outputs it to each speaker.
In the fourth embodiment, the second switching means 22B includes the same number of second switching means 22B ₁ ,..., Second switching means 22B _{K as} the number K of amplifiers 16C. Each second switching means 22B _i (1 ≦ i ≦ K) is connected to a plurality of different speakers 20 (speakers 20 _i1 ,..., Speakers 20 _{ij (i)} ) (Σ _i j (i) = N).

The second switching means 22B _i (1 ≦ i ≦ K) weights the amplified principal component signal input from the amplifier 16C _i corresponding to the principal component signal input from the principal component analysis means 12. Using the signal, an allocation time proportional to the weighting factor is calculated for each channel. Then, the second switching unit 22B _i switches the principal component signal at the distribution time calculated for each channel and outputs the signal to the speaker 20 _ih corresponding to the channel (1 ≦ h ≦ j (i)).

  The weighting signal input to the second switching means 22B is the weighting signal output from the principal component analysis means 12 to the switching means 14 and the weighting means 18 in the multi-channel sound reproduction apparatus 1A (first embodiment), Are the same.

  According to this multi-channel sound reproduction apparatus 1D, the multi-channel sound signal is sampled by the time division means 10, and the sampled multi-channel sound signal is subjected to principal component analysis by the principal component analysis means 12, and a predetermined number The main component signal is output according to the distribution time by the switching unit 14B, the main component signal is amplified by the plurality of amplifiers 16C, and is converted into the main component signal by the plurality of second switching units 22B. Time-division is performed and sound is output by the speaker 20. For this reason, multi-channel sound can be effectively reproduced by a plurality of amplifiers.

The configuration of the multi-channel sound reproducing device 1 has been described above, but the present invention is not limited to this, and can be changed without departing from the spirit of the present embodiment.
For example, the multi-channel sound reproduction apparatus 1 can also implement each unit as a function program in a computer, and can also operate the multi-channel sound reproduction program by combining the function programs.

[Operation of multi-channel sound reproduction device]
Next, the operation of the multi-channel sound reproducing device 1 according to the embodiment of the present invention will be described.

Note that the multi-channel sound reproduction apparatus 1A according to the first embodiment and the multi-channel sound reproduction apparatus 1B according to the second embodiment are different in whether one or more amplifiers 16 (16B) are provided. However, since the operations of the individual amplifiers are the same, they will be described together with FIG.
Similarly, the multi-channel sound reproduction device 1C according to the third embodiment and the multi-channel sound reproduction device 1D according to the fourth embodiment include one amplifier 16C and the second switching unit 22 (22B), Since the operations of the individual amplifiers and the second switching means are the same only in the difference in whether they are provided, they will be described together with FIG.

[[Operation of Multi-Channel Sound Reproducing Device (First Embodiment and Second Embodiment)]]
First, referring to FIG. 7 (for the configuration, refer to FIGS. 1 and 2 as appropriate), the multi-channel sound reproducing apparatus 1A (1B) in the first embodiment (second embodiment) of the present invention is described. The operation will be described.

First, the multi-channel sound reproducing device 1A (1B) samples the multi-channel sound signal a predetermined number of times at predetermined time intervals by the time division means 10 (S11).
Then, the multi-channel sound reproducing device 1A (1B) performs principal component analysis based on the sample data by the principal component analysis means 12, and outputs a predetermined number of principal components having a high contribution rate for each predetermined time interval. It outputs as a component signal (S12).

  Then, the multi-channel sound reproducing device 1A (1B) uses the switching unit 14 (14B) to obtain an allocation time according to the contribution ratio of each main component, or mechanically, for example, an equal time interval of about 1 MHz (S13). . Thereafter, the multi-channel sound reproducing device 1A (1B) outputs the i-th principal component signal during the distribution time of the i-th principal component by the switching unit 14 (S14).

Then, the multi-channel sound reproducing device 1A (1B) amplifies the main component signal by the amplifier 16 (16B) (S15).
Then, the multi-channel sound reproducing device 1A (1B) weights the main component signal for each channel by the weighting means 18 (S16).

Then, the multi-channel sound reproducing device 1A (1B) converts the main component signal into sound by the speaker 20 and outputs it (S17).
Then, the multi-channel sound reproducing device 1A (1B) determines whether or not a predetermined number of principal component signals have been output by the switching means 14 (14B) (S18), and if not (No in S18). Return to S14, and if output (Yes in S18), proceed to S19.

  Then, the multi-channel sound reproducing device 1A (1B) determines whether or not the multi-channel sound signal is continued by the time division unit 10 (S19), and if it is continued (Yes in S19), the process proceeds to S11. If the operation is not continued (No in S19), the operation is terminated.

[[Operation of Multi-Channel Sound Reproducing Device (Third Embodiment and Fourth Embodiment)]]
Next, referring to FIG. 8 (for the configuration, refer to FIGS. 3 and 4 as appropriate), the multi-channel sound reproducing device 1C (1D) in the third embodiment (fourth embodiment) of the present invention. Will be described.

First, the multi-channel sound reproducing device 1C (1D) samples the multi-channel sound signal a predetermined number of times at predetermined time intervals by the time division means 10 (S21).
Then, the multi-channel sound reproducing device 1C (1D) performs principal component analysis based on the sample data by the principal component analyzing means 12, and outputs a predetermined number of principal components having a high contribution rate for each predetermined time interval. It outputs as a component signal (S22).

Then, the multi-channel sound reproducing device 1C (1D) uses the switching unit 14 (14B) to obtain an allocation time corresponding to the contribution ratio of each main component, or mechanically, for example, an equal time interval of about 1 MHz (S23). . Thereafter, the multi-channel sound reproducing device 1C (1D) outputs the i-th principal component signal during the distribution time of the i-th principal component by the switching means 14 (14B) (S24).
Then, the multi-channel sound reproducing device 1C (1D) amplifies the main component signal by the amplifier 16C (S25).

Then, the multi-channel sound reproducing device 1C (1D) obtains an allocation time corresponding to the weighting factor for each channel or a mechanically equal time interval of, for example, about 1 MHz by the second switching unit 22 (22B) (S26). ). Thereafter, the multi-channel sound reproducing device 1C (1D) outputs the main component signal to the speaker corresponding to the channel j during the distribution time of the channel j by the second switching means 22 (22B) (S27).
Then, the multi-channel sound reproducing device 1C (1D) converts the main component signal into sound by the speaker 20 and outputs it (S28).

  Then, the multi-channel sound reproducing device 1C (1D) determines whether or not the main component signal has been output to all the connected channels by the second switching means 22 (22B) (S29) and outputs it. If not (No in S29), the process returns to S27, and if it is output (Yes in S29), the process proceeds to S30.

Then, the multi-channel sound reproducing device 1C (1D) determines whether or not a predetermined number of principal component signals have been output by the switching unit 14 (14B) (S30), and if not (No in S30). Return to S24, and if output (Yes in S30), proceed to S31.
Then, the multi-channel sound reproduction device 1C (1D) determines whether or not the multi-channel sound signal is continued by the time division unit 10 (S31), and if it is continued (Yes in S31), S21. If the operation is not continued (No in S31), the operation is terminated.

Note that the operation of the multi-channel sound reproducing device 1 is not limited to this, and can be changed without departing from the spirit of the present embodiment.
For example, when the time division means 10 determines that the multi-channel sound signal is not continued, the multi-channel sound reproduction device 1 ends the operation (S19 and S31). The operation may be terminated by providing a switch 1 and turning the switch off.

  In addition, for convenience of the flowchart, the operation in each unit is described sequentially, but a unit that can operate in parallel may operate in parallel. For example, when the switching means 14 (14B) outputs the i-th principal component signal during the i-th principal component distribution time (S14, S24), the time division means 10 samples the multi-channel acoustic signal. (S11, S21).

[About principal component analysis]
Below, the principal component analysis method in the multivariate analysis used in the principal component analysis means 12 of this embodiment is demonstrated.
Principal component analysis is a technique for obtaining a comprehensive index by synthesizing several variables that are correlated. The principal component is a composite variable of the original variable, and the respective principal components are calculated so as to be uncorrelated with each other.

Assume that there are M sample data for each of N variables (x ₁ , x ₂ ,..., X _N ). That is, it is assumed that there is sample data x _i1 , x _i2 ,..., X _iM for the variable x _i (1 ≦ i ≦ N). From such data, a composite variable indicating a comprehensive index is obtained.
Each coefficient (x ₁ , x ₂ ,..., X _N ) is assigned a coefficient (α ₁ , α ₂ ,..., Α _N ) to create a composite variable (see equation (1)). ).

  In order to find the overall index in the direction of the largest dispersion of each sample, the variance of the composite variable Y may be maximized. Therefore, the variance of the composite variable Y is obtained (see equation (2)).

Here, the upper line represents the average value of the variables. α = (α ₁ ,..., α _N ), and α ^T is a transposed matrix of α. S is a variance / covariance matrix of N variables of this data. For this coefficient α _i , the constraint condition (3) is provided.

This problem maximizes equation (2) under equation (3). This results in an eigenvalue problem of the variance / covariance matrix S when the Lagrange's undetermined multiplier method is used. The matrix S has N eigenvalues (λ ₁ ,..., Λ _N ). Among them, the composite variable Y ₁ that maximizes the variance is expressed as an equation (4) with the element of the eigenvector for the maximum eigenvalue λ ₁ as a coefficient. This synthetic variable _{Y 1} of the first principal component.

The dispersion of Y ₁ is λ ₁ and the dispersion is the largest. Therefore, it is a composite variable that most reflects the N variable data.
In the following, the composite variable Y _i (2 ≦ i ≦ N) can be expressed as the equation (5) with the element of the eigenvector for the i-th largest eigenvalue λ _i as a coefficient. This synthetic variable Y _i is called the i-th principal component.

An index of how much each principal component reflects the original data is called a contribution rate. The variance of each principal component is represented by its eigenvalue λ _i , and the sum Σ _i λ _i is equal to the sum of variances of the original variables. Therefore, the contribution ratio of the i-th principal component is expressed by equation (6).

  In addition, the total contribution rate from the first principal component to the i-th principal component is referred to as a cumulative contribution rate up to the i-th principal component.

In relation to the present embodiment, the sample data for each channel output from the time division means 10 to the principal component analysis means 12 is sample data x _i1 , x _i2 ,... For the variable x _i (1 ≦ i ≦ N). X, _iM . Using this, the principal component analysis means 12 performs the principal component analysis described above.

The principal component analysis means 12 obtains an eigenvector corresponding to the eigenvalue of the variance / covariance matrix S, and uses a predetermined number of eigenvectors corresponding to a predetermined number of eigenvalues (that is, the contribution ratio of the equation (6) is large). The main components (4) and (5) are configured. Then, the principal component analyzing means 12 calculates M sample data for each predetermined number of principal components using the sample data, and outputs the data to the switching means 14.
At this time, the principal component analysis unit 12 outputs the contribution ratio (equation (6)) of the principal component being output and the weighting factor that is an element of the corresponding eigenvector as a weighting signal.

1 (1A, 1B, 1C, 1D) Multi-channel sound reproduction apparatus 10 Time division means 12 Principal component analysis means 14, 14B Switching means 16, 16B, 16C Amplifier 18 Weighting means 20 Speakers 22, 22B Second switching means

Claims (6)

A multi-channel sound reproducing apparatus that reproduces multi-channel sound with the same number of speakers as the number of channels and a smaller number of amplifiers than the number of channels,
Time division means for sampling an input multi-channel acoustic signal a predetermined number of times at predetermined time intervals;
A principal component analysis is performed on the sample data sampled for each channel by the time division means, a contribution factor obtained from the eigenvalues of the variance / covariance matrix and a weighting factor that is a component of the eigenvector corresponding to the eigenvalues; The principal components corresponding to these are obtained, and for each predetermined time interval, a predetermined number of principal components from the high-contribution main components are output as principal component signals, and for each predetermined time interval, A principal component analysis means for outputting the contribution ratio and the weighting factor of the principal component signal being output as a weighting signal;
In accordance with a contribution rate included in the weighted signal input from the principal component analysis means, a distribution time of the principal component signal input from the principal component analysis means is determined, and the principal component signal is switched for each distribution time. Switching means for outputting a signal to the amplifier that has been assigned to the speaker in advance;
Among the weighting factors included in the weighting signal input from the principal component analyzing means, the principal component signal amplified by the amplifier is weighted by the weighting factor of the channel corresponding to one connected speaker. Weighting means for outputting to the speaker,
A multi-channel sound reproducing apparatus comprising: The speaker is a speaker driven by a DC bias,
The weighting means is
Of the weighting factors included in the weighting signal input from the principal component analysis means, an alternating current amplified according to the weighting factor of the channel corresponding to one connected speaker is applied by the amplifier. 2. The multi-channel sound reproducing apparatus according to claim 1, wherein the multi-channel sound reproducing apparatus is a bias circuit that superimposes on the DC bias voltage and outputs the superimposed signal to the speaker. A multi-channel sound reproducing apparatus that reproduces multi-channel sound with the same number of speakers as the number of channels and a smaller number of amplifiers than the number of channels,
Time division means for sampling an input multi-channel acoustic signal a predetermined number of times at predetermined time intervals;
A principal component analysis is performed on the sample data sampled for each channel by the time division means, a contribution factor obtained from the eigenvalues of the variance / covariance matrix and a weighting factor that is a component of the eigenvector corresponding to the eigenvalues; The principal components corresponding to these are obtained, and for each predetermined time interval, a predetermined number of principal components from the high-contribution main components are output as principal component signals, and for each predetermined time interval, A principal component analysis means for outputting the contribution ratio and the weighting factor of the principal component signal being output as a weighting signal;
In accordance with a contribution rate included in the weighted signal input from the principal component analysis means, a distribution time of the principal component signal input from the principal component analysis means is determined, and the principal component signal is switched for each distribution time. Switching means for outputting a signal to the amplifier that has been assigned to the speaker in advance;
Among the weighting factors included in the weighting signal input from the principal component analysis means, the weighting factor for each channel corresponding to a plurality of connected speakers is switched at a distribution time according to the weighting factor, Second switching means for outputting the main component signal amplified by the amplifier to the connected speaker;
A multi-channel sound reproducing apparatus comprising: The switching means is
The multi-channel sound reproducing apparatus according to any one of claims 1 to 3, wherein the principal component signal is switched at a zero cross point. In order to reproduce multi-channel sound with as many speakers as the number of channels and a smaller number of amplifiers than the number of channels,
Time division means for sampling an input multi-channel acoustic signal a predetermined number of times at predetermined time intervals;
A principal component analysis is performed on the sample data sampled for each channel by the time division means, a contribution factor obtained from the eigenvalues of the variance / covariance matrix and a weighting factor that is a component of the eigenvector corresponding to the eigenvalues; The principal components corresponding to these are obtained, and for each predetermined time interval, a predetermined number of principal components from the high-contribution main components are output as principal component signals, and for each predetermined time interval, A principal component analysis means for outputting the contribution ratio and the weighting factor of the output principal component signal as a weighting signal;
In accordance with a contribution rate included in the weighted signal input from the principal component analysis means, a distribution time of the principal component signal input from the principal component analysis means is determined, and the principal component signal is switched for each distribution time. Switching means for outputting a signal to the amplifier previously assigned to the speaker;
Among the weighting factors included in the weighting signal input from the principal component analyzing means, the principal component signal amplified by the amplifier is weighted by the weighting factor of the channel corresponding to one connected speaker. Weighting means for outputting to the speaker,
A multi-channel sound reproduction program characterized by functioning as In order to reproduce multi-channel sound with as many speakers as the number of channels and a smaller number of amplifiers than the number of channels,
Time division means for sampling an input multi-channel acoustic signal a predetermined number of times at predetermined time intervals;
A principal component analysis is performed on the sample data sampled for each channel by the time division means, a contribution factor obtained from the eigenvalues of the variance / covariance matrix and a weighting factor that is a component of the eigenvector corresponding to the eigenvalues; The principal components corresponding to these are obtained, and for each predetermined time interval, a predetermined number of principal components from the high-contribution main components are output as principal component signals, and for each predetermined time interval, A principal component analysis means for outputting the contribution ratio and the weighting factor of the output principal component signal as a weighting signal;
In accordance with a contribution rate included in the weighted signal input from the principal component analysis means, a distribution time of the principal component signal input from the principal component analysis means is determined, and the principal component signal is switched for each distribution time. Switching means for outputting a signal to the amplifier previously assigned to the speaker;
Among the weighting factors included in the weighting signal input from the principal component analysis means, the weighting factor for each channel corresponding to a plurality of connected speakers is switched at a distribution time according to the weighting factor, Second switching means for outputting the main component signal amplified by the amplifier to the connected speaker;
A multi-channel sound reproduction program characterized by functioning as

Images