JP2002091496A - Voice transmitting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice transmitting device which can perform procedures for freely selecting packaged plural standard CODECs during an operation and for selecting and determining plural CODECs in a short time. SOLUTION: The device has a function to download a CODEC program from a CPU 105 to DSPs 104 and to download the program to an arbitrary DSP with an arbitrary timing. Furthermore, the device has a queue to control the CODECs for every kind and has a function to select a DSP having a CODEC to be used and to conduct a switching. Audio channels 102 and the DSPs 104 are connected by a time division multiplexing switch 103. Thus, the connection between the channels 102 and the DSPs 104 is instantaneously switched by the setting change.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio transmission apparatus, and more particularly, to a DSP for encoding / decoding audio signals.
The present invention relates to an audio transmission device including

[0002]

2. Description of the Related Art An example of a conventional audio transmission apparatus is disclosed in Japanese Patent Application Laid-Open No. 9-81514. This audio processing device exchanges data between a plurality of DSPs using data buffer means connected to a common bus, and controls the order of data exchange between DSPs with an interrupt signal. By doing so, there is no restriction on the exchange of data between the DSPs, and it is easy to change the control program of the DSPs, and CODECs of different standards with the same configuration are realized.

As another conventional example, Japanese Patent Application Laid-Open No. 9-153
No. 847 discloses this. This audio processing device has an encoder function processing for encoding audio data in a predetermined audio CODEC system, and a decoder function for decoding encoded audio data in order to realize audio signal processing with the same DSP. The processing and the echo canceller function processing for realizing the echo canceller function are executed by the same DSP device, and the coefficient update frequency in the echo canceller function processing is changed according to the type of the audio CODEC system applied at that time. .

[0004]

Since the conventional apparatus does not have a function of changing a DSP function by changing a program during operation, it is difficult to implement a CODEC of a plurality of standards, and the apparatus is applied to a flexible network configuration. There was a problem that it was difficult.

[0005] Also, the functions of the DSP are limited by the amount of internal memory and are limited, and there is no function of rewriting in software during operation. However, there is a problem that the hardware configuration becomes complicated and the cost increases.

Therefore, an object of the present invention is to freely select a plurality of standardized CODECs during operation, and to select a plurality of CDECs.
An object of the present invention is to provide an audio transmission device capable of performing a procedure for selecting and determining an ODEC in a short time.

Another object of the present invention is to provide a multi-standard CO
The purpose is to realize an apparatus having a DEC function with an inexpensive hardware configuration.

[0008]

SUMMARY OF THE INVENTION The present invention relates to an audio transmission apparatus having a plurality of DSPs for encoding and decoding an input audio signal by a built-in CODEC program. Multiple CODE
It is characterized by comprising storage means for storing a C program and program transfer means for transferring the CODEC program to the DSP.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of a voice transmission apparatus according to the present invention will be described in detail with reference to the drawings. FIG.
FIG. 2 is a block diagram illustrating a configuration of a voice transmission device according to the present embodiment. The audio card 101 accommodates a plurality of audio channels 102 and can accommodate a PBX, an analog telephone, or the like.

[0010] The time division multiplex switch 103 is connected between the audio channel 102 and the DSP 104 and switches the connection. Voice channel 102 and time division multiplex switch 103
The time-division multiplex switch 103 and the DSP 104 are connected by a time-division multiplex bus, and an arbitrary audio channel 102 and an arbitrary DSP 104 are connected by designating each audio channel or a time slot used by the DSP. be able to. This connection destination setting is performed by the CPU 105
Thus, it can be performed instantaneously by rewriting the memory of the time division multiplex switch 103.

The DSP 104 has a function of encoding / decoding audio data by CODEC, and is mounted in the same number as the audio channels 102. The CPU 105 controls call connection procedures, voice data transmission / reception, and the like, and manages the use status of the DSP to which the CODEC program has been downloaded.

The RAM 106 includes the DSP 104 and the CPU 1
A buffer for transmitting and receiving control commands and voice data to and from the remote controller 05, and the transmitting side and the receiving side exist independently,
Each has a plurality of regions of a fixed size. Each area is connected in a ring shape and used sequentially. DSP104
Is transferred to the CPU 105 via a transmission / reception buffer on the RAM 106,
Is transmitted to the connected device connected via the external interface 109. On the other hand, the audio data received from the external interface 109 is passed by the CPU 105 to the DSP 104 via the RAM 106, and is output to the audio channel 102 after being combined.

The external storage device 108 stores a plurality of CODEC programs of different standards, and is loaded onto the RAM 106 when the device is started. A plurality of audio signal processing methods based on CODEC are standardized by ITU-T recommendations. The DMA controller 107 downloads the CODEC program stored in the external storage device 108 to the DSP 104 at an arbitrary timing.

A free DSP queue is constructed on the RAM 106 by the CPU 105, and the unused DSP 10
4 are classified and connected for each downloaded CODEC type. FIG. 2 is a state diagram showing the configuration of an empty DSP queue. The empty DSP queue is a buffer for storing the number of the empty DSP 104, and each buffer has an area for storing a pointer indicating the connection to the next buffer. In addition, there is an area for storing a pointer indicating the position of the first buffer, and a plurality of buffers are connected with the first pointer at the top. In the tail buffer, a tail code is stored in the area of the next pointer to indicate the end of the buffer connection. The CPU 105 stores such a queue in C
It is created and managed for each ODEC type.

FIG. 2 shows an example in which four types of CODECs are used. CODEC-A has four DSPs, and CODEC-B has D type.
SP3 and CODEC-C indicate that 5 DSPs are vacant. In CODEC-D, an empty DSP
In this case, the end code is stored in the area for storing the head buffer pointer. CPU10
5 selects a CODEC to be used, and selects an unused DSP to be used by referring to a free DSP queue corresponding to the CODEC type.

Next, the operation of the embodiment of the present invention will be described in detail with reference to the drawings. In FIG. 1, the connection between the audio channel 102 and the DSP 104 is set by the CPU 105, but the initial state immediately after the device is started is ch1 and the DSP.
1, such as ch2 and DSP2, so that the numbers are the same. Further, the CPU 105 downloads a CODEC program to each DSP at the time of starting the apparatus. At this time, which CODEC is to be downloaded depends on the configuration given in advance from the outside.

Therefore, the distribution of the number of CODECs to be downloaded by the configuration can be freely determined according to the frequency of use or the like. For example, if the number of audio channels is 30 and the number of supported CODECs is 4, the number of DSPs to be downloaded is set to 1 according to the frequency of use.
Configurations can be created so as to be distributed as 6, 8, 4, and 2.

The CPU 105 sends R
After copying the CODEC program on AM106,
Using the DMA controller 107, the CODEC program is downloaded to the DSP 104 in a distribution according to the configuration. After that, the CPU 105
A free DSP queue for each CODEC type is constructed on the 106.

FIG. 3 is a sequence diagram showing a voice calling / incoming call procedure in this embodiment. The voice transmission / reception procedure shown in FIG. 3 shows a call connection procedure when a call is made from the calling apparatus to the called apparatus in the voice transmission apparatus to which the PBX is connected in the embodiment shown in FIG.

First, when there is a call from the calling side PBX, S
The DSP 104 is notified when the S signal is turned on.
The DSP 104 turns on / off the SR signal and urges the PBX to transmit a selection signal. When the selection signal is notified from the PBX, the DSP 104 transmits a call instruction to the CPU 105. The call instruction is passed to the CPU 105 via a transmission / reception buffer on the RAM 106.

When receiving a call instruction from the DSP 104, the CPU 105 notifies the CPU 105 of the connected device via the external interface 109 of the call using a predetermined procedure. CPU1 of the called device that receives the call
05 is a DSP using the transmission / reception buffer on the RAM 106
An incoming call notification is transmitted to 104. DSP receiving the incoming call notification
Reference numeral 104 indicates an incoming call by turning on the SR signal to the PBX, and confirms the incoming call by detecting ON of the SS signal from the PBX.

The DSP 104 confirming the incoming call is stored in the RAM 1
An incoming call confirmation is sent to the CPU 105 using the transmission / reception buffer on 06, and the CPU 105 receives this and sends a standby notification (ALERT message) to the calling CPU 105 using a predetermined procedure. Calling CPU 105
Receives this, notifies the DSP 105 via the transmission / reception buffer on the RAM 106 of the call confirmation.

Next, the CPUs 105 of both devices exchange information on the CODECs owned by each other using a predetermined procedure, and select the CODEC determined to be optimal. After that, CPU1
05 is a DSP in which this CODEC is downloaded because audio data is processed by the selected CODEC.
Is determined.

FIG. 4 is a flowchart showing the procedure of the CODEC determination process of the present embodiment. First, the CPU 10
5 is a codec selected by referring to an empty DSP queue.
It is determined whether or not there is a free DSP for which has been downloaded (S401). If there is, the DSP registered at the head of the queue is determined as the DSP for processing the voice call currently being sent and received (YES in S402), and this DSP is determined.
Then, a process of connecting the voice channel and the voice channel is performed (S403). That is, the DSP that has performed the call connection procedure up to now and the DSP that performs call connection and voice transmission thereafter are switched.

FIG. 5 is a state diagram showing a DSP switching procedure, and is an example of a procedure for switching from the state 1 in which the DSP 4 is connected to the audio channel 2 to the state 2 in which the DSP 3 is connected. In this case, since the switching destination DSP 3 is connected to the audio channel 3 in the state 1, the released DSP 4 is connected to the audio channel 3.
The switching is completed by connecting to.

This switching is performed by changing the setting of the time division multiplex switch 103 from the CPU 105.
Switching is completed instantaneously because only minimal changes are required.

Further, there is no effect on the audio channel and the DSP connection which are not changed, and there is no adverse effect such as interruption of the audio during the call.

In FIG. 4, the released DSP is instructed to stop the processing so far, and the DSP is returned to the initial state (S4).
04), link to the end of the free DSP queue corresponding to the CODEC type owned by the DSP (S405).

On the other hand, if there is no free space in the DSP to which the selected CODEC has been downloaded, the CPU
This is dealt with by downloading the CODEC program used for P (S406). The CODEC program is downloaded to the DSP using the DMA controller 107.

The CPU 105 constantly manages the number of CODECs downloaded to the DSP 104 for each CODEC type, and adjusts the distribution so as to maintain the distribution specified by the configuration. CODE used in the call connection procedure
If the DSP from which C was downloaded has no free space and a new download occurs, the CPU 105 keeps the distribution for each CODEC type specified in the configuration by downloading the CODEC again to the DSP 104 released after the call disconnection. .

As described above, in this embodiment, in order to use the CODEC selected during the call connection, a means for switching the connection between the voice channel and the DSP by changing the setting of the time division multiplex switch is provided. Since the download is performed in the idle state, the call connection procedure can be performed at a high speed and a plurality of CODECs can be handled.

Further, since the DSP has a function of downloading a CODEC program, an inexpensive DSP having a small amount of internal memory can be used, and a plurality of CODEC processes can be performed by one DSP. This contributes to cost reduction even for devices with many components.

Next, another embodiment of the present invention will be described in detail with reference to the drawings. In the embodiment of FIG. 1, FIG.
As shown in the figure, a queue for managing the CODEC downloaded to the DSP for each type is provided, and the DSP having the target CODEC is provided by referring to this queue.
It has the function of selecting This queue is CODE
It can be used for the purpose of managing not only C but also other programs, and can also be used together with CODEC management.

FIG. 6 is a state diagram showing the configuration of the queue when the CODEC and the FAX relay program are managed at the same time, and shows the state when two types of CODEC and two types of FAX relay are managed. One type of CODEC and one type of FAX relay are downloaded to the DSP 104 and used for voice processing. There are four combinations of these,
Which combination is to be used and in which distribution may be given in advance in the configuration.

The CPU 105 controls the CO used in the call connection procedure.
When a combination of DEC and FAX relay is selected, the respective management queues are referenced and the target CODEC, FA
Select a DSP with an X relay. In the example of FIG.
When the DEC-A and the FAX relay A are selected, the DSP 11 (thick frame in FIG. 6) has these at the same time.
The PU 105 selects and uses this.

As described above, by changing the configuration of the management queue, it is possible to selectively manage programs other than CODEC in the same manner as in CODEC. It is possible to do.

[0037]

The first effect is that a plurality of CODECs can be simultaneously handled and a flexible network can be constructed. The reason is that the CODE is sent to the DSP at any time.
This is because the C program can be downloaded.

The second effect is that a CODEC used for each call is used.
, And a high-speed call connection can be realized. The reason is that voice channels and DSP
Is connected by a time division multiplex switch, so that the connection between the voice channel and the DSP can be instantaneously switched. In addition, the CPU has a function of managing the usage status of the downloaded DSP for each CODEC type and assigning a DSP having the CODEC selected during the call connection.

A third effect is that an audio transmission apparatus that can handle a plurality of CODECs can be configured at low cost. The reason is that the DSP has a function of downloading the CODEC program at any time, so that an inexpensive DSP with a small amount of internal memory can be used.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a state diagram showing a configuration of an empty DSP queue.

FIG. 3 is a sequence diagram showing a voice calling / incoming call procedure in the embodiment.

FIG. 4 is a flowchart illustrating a procedure of a CODEC determination process according to the embodiment;

FIG. 5 is a state diagram showing a DSP switching procedure in the embodiment.

FIG. 6 is a state diagram showing a configuration of a queue when a CODEC and a FAX relay program are managed simultaneously.

[Explanation of symbols]

Reference Signs List 101 voice card 102 voice channel 103 time division multiplex switch 104 DSP 105 CPU 106 RAM (random access memory) 107 DMA controller 108 external storage device 109 external interface (serial, ether)
Etc.) 110 CPU bus 111 Time division multiplex bus

Claims (3)

[Claims] 1. An audio transmission apparatus comprising a plurality of DSPs for encoding and decoding an input audio signal by a built-in CODEC program.
Storage means for storing a DEC program;
A voice transmission device comprising: a program transfer unit that transfers a C program to the DSP. 2. The apparatus according to claim 1, further comprising a plurality of audio channels for inputting an audio signal, and a time division multiplex switch for switching connection between the plurality of audio channels and the plurality of DSPs. Voice transmission device. 3. The audio transmission apparatus according to claim 1, further comprising a management unit that manages the use status of the plurality of DSPs for each type of a built-in CODEC program.