JPH0619901B2 - Leveling method for in-record voice data - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to a magnetic disk, an optical disk, etc.
The audio data storage medium, a table for designating and managing, in record units (quantitative units when accessing the storage medium) a data read position or a data write position on the medium, and a control unit, the control unit including the table The present invention relates to a voice storage device capable of recording or reproducing voice data in real time by storing or reading voice data in a record unit in the storage medium.

More specifically, as a result of the editing process of the stored data in the storage medium, when the address is rewritten in record units in the table, the voice data in the record for leveling the number of voice data in each record It is related to the leveling method of.

[Conventional technology]

FIG. 5 is a block diagram showing a general hardware configuration of an audio storage device using a disk as a storage medium.

In the figure, 1 is an A / D converter, 3 is a D / A converter,
3 is a buffer memory (hereinafter simply referred to as a buffer), 4
Is a disk controller, 5 is a disk, and 6 is a RAM
(Random access memory) 7 is a table in the RAM 6 called a directory.

A voice signal converted into an electric signal by a microphone (not shown) is converted into a digital signal by the A / D converter 1 and stored in the buffer 3 as voice data. The disk controller 4 refers to the directory 7 and writes the audio data from the buffer 3 to the disk 5 on a record-by-record basis in accordance with the address specified in the directory 7 on a record-by-record basis. Alternatively, conversely, the disk controller 4 refers to the directory 7, reads audio data from the disk 5 in record units according to the address specified in record units therein, and stores the audio data in the buffer 3
After being stored in, the D / A converter 2 converts it into an analog signal,
Output to a speaker not shown.

In this way, it is possible to record and reproduce the voice in real time. When the audio data written in the disk 5 is edited, as a result, the address is rewritten on a record-by-record basis in the directory 7, so that the recorded and reproduced audio as edited can be reproduced. Is possible.

As described above, the access to the audio data on the disk 5 is generally performed in record units. In order to efficiently store a plurality of voice messages on the disk 5, each record constituting the message is arranged randomly, and the directory 7 on the RAM 6 defines and manages the order relation of each record.

FIG. 6 is an explanatory diagram showing such a voice data management system. In the figure, the voice data for one message, for example, stored in the disk 5 is divided in record units and stored at random, and the order relation of each record is such that each cell C _{1 to} C ₆ of the directory 7 ... It is managed by writing the address of each record in.

By the way, in the voice storage device as shown in FIG. 5, when performing real-time voice recording and reproduction, it is required to access the disc at high speed so as not to interrupt the voice. In the case of reproduction, it is necessary to read the data of the next record from the disk before the reproduction of the data of the previous record is completed so that the sound is not interrupted.

The time required to read the data of one record from the disk does not depend on the length of the valid data in the record such as seek and search, and the length of the valid data in the record such as transfer time. In a general magnetic disk, for example, the former has a dominant length over the latter in the time required for reading. Therefore, as the length of the valid data in the record becomes shorter, the time required to finish reproducing the data of the previous record becomes shorter, but the time required to read the next record remains almost unchanged. Therefore, there is a drawback in that the reading of the next record cannot catch up with the reproduction of the previous data, and the reproduced sound is interrupted.

On the other hand, it has been attempted to add an edit processing function of deleting a part of the stored voice message or inserting a phrase in the voice storage device. Also in this case,
It can be said that the method of managing the audio data as shown in FIG. 6 is an advantageous method because a high-speed editing process can be performed only by rewriting the directory 7. However, in this case, the precondition is that the editing unit matches the record length, but if this is the case, it will not be possible to realize fine editing processing in units of less than that length.
It had the disadvantage of being inconvenient.

[Problems to be solved by the invention]

Therefore, the present invention solves the problem that in the voice storage device, there is no interruption of the reproduced voice, and when the voice message editing process is performed, a fine editing process in units of a short length equal to or less than the record length is possible. It should be a problem. Therefore, it is an object of the present invention to provide a leveling system for voice data in a record in a voice storage device that enables the above.

[Means and Actions for Solving Problems]

In order to achieve the above object, the present invention provides a storage medium for storing audio data, and a table for designating and managing a data read position or a write position in the storage medium by a record which is a quantitative unit when accessing the medium. And a control unit, wherein the control unit stores audio data in the storage medium or reads the audio data by referring to the table, in a process of editing stored data in the storage medium. As a result, when the address is rewritten in record units in the table, the control unit determines whether the number of audio data contained in each record in the table after rewriting is less than or equal to a certain threshold m. investigated whether, in the case where less, the l _i to the number of audio data contained in the record, the Adds the number l _{i + 1} of the audio data contained in the adjacent record adjacent on Buru, checks whether the sum is exceed the number of data that can be accommodated in the unit record, when around above, the neighbor Of the audio data included in the record,
At least the first (m-l _i ) pieces of data are extracted from the adjacent record and added next to the audio data included in the record, and if not exceeded, the audio data included in the record is added next to the adjacent audio data. All the voice data contained in the record is taken out and added, and the adjacent record is deleted from the table, and the number of voice data contained in each record is equalized after the editing process in this way. It is characterized by that.

The point that the present invention is different from the prior art is that the record length and the editing unit length do not necessarily have to be the same, which enables fine editing processing, and the number of audio data included in each record. It can be said that this is to prevent the interruption of the reproduced voice by equalizing the above.

〔Example〕

 Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is an explanatory diagram of a voice data management system and a voice data deletion processing system according to the present invention.

That is, as shown in FIG. 1, according to the present invention, the audio data management method is such that the number of audio data contained in each record or the start address and end address of the data in each record is set to each cell in the directory 7. It is a method of storing in. Specifically, in the directory 7 on the RAM, not only the record address of each voice record, but also the actual number of data l contained in each record 1
_i (however, i = 1, 2, 3, ...) Is stored.

For example, when partial deletion of voice data is performed in a message composed of a set of records, and the deletion start position and end position do not match record boundaries, both ends of the deletion point include odd data numbers. It allows and allows records to occur. However, the process of packing all the data in each record to the beginning is performed.

As described above, since the position where the editing process is performed does not have to coincide with the record boundary in the voice message, the fine editing process can be performed in units of short length.

The data management method described above is a general method for a text editor of a computer, but the voice storage device has not been realized so far because of the following problems.

In other words, if there is a record with an extremely small amount of data among the "half-length" records generated by the editing process, as explained earlier in that part, the real-time property of audio reproduction becomes unsatisfactory, that is, there is a break. Will occur.

In order to solve this problem, according to the present invention, a record having an extremely small number of data is previously searched for before the reproduction is started after the series of editing processes is completed. Next, by moving all or part of the data contained in the record adjacent to that record to the record concerned, the process of leveling the number of data contained in each record to at least a certain fixed value or more I do. Note that the records that are completely empty are deleted (repealed).

FIG. 2 is an explanatory diagram showing the situation. In the figure (a), i
When the number of data l _i included in the th record is smaller than a preset threshold m, the number of data l _{i + 1} included in the (i + 1) th record is combined into one record, and (i + 1 )
The second record is empty, so I will remove it from the directory.

In this case, (l _i + l _{i + 1} ) ≦ N must hold. However, N is the maximum number of data that can be included in one record.

FIG. 2 (b) shows the number of data l included in the i-th record.
_i shows the case where the sum of the number of data l _{i + 1} contained in the (i + 1) th record is (l _i + l _{i + 1} )> N. In this case, from the (i + 1) th record, (m-
l _i) by binding to the data l _i contained in the i-th record transferred pieces of data, the number of data of the i-th record l _{i 'is} m, (i + 1) th record data number l of _{i + 1} 'is - becomes _{[l i + 1 (m-} l i)].

The algorithm of the voice data leveling processing operation within a record described above is shown in the flowchart of FIG.

In FIG. 3, steps (a) to (g) need not be explained again.

Then return to FIG. 2, in the case of FIG. (A), the number of data l _i contained in the i-th record, combined added data number l _{i + 1} included in the (i + 1) th record Result (l _i + l _{i + 1} )
May be (l _i + l _{i + 1} ) <m. In this case, the same processing is further performed with the next adjacent record ((i + 2) th record before the directory is rewritten).

Further, in the case of FIG. 2 (b), the i-th record has the number of data l _i ′ contained therein l _i ′ = m, but in the (i + 1) -th record, it is included therein. There is a possibility that the number of data l _{i + 1} ′ decreases and l _{i + 1} ′ <m. in this case,
Under the worst condition, the processing moves to the adjacent record in a chain reaction. But l _i + _{1 '=} - if _{[l i (m-l i} )] = (l i + 1 + l i -m)> (N-m) and is because mN / 2, does not occur in this chain reaction I understand.

In FIG. 4, the record length (that is, the maximum number of data that can be stored in one record) N = 8, and the minimum required number of data in one record for real-time access (that is, no interruption) m = When a record i that does not satisfy this condition (that is, a record in which the number of data contained is 1 or 2 or 3) occurs, the level cutting process performed with the adjacent record (i + 1) is performed. It is explanatory drawing which showed the type.

In FIG. 4, the upper horizontal column indicates that the number of data l _i included in the i-th record is 1, 2, 3 and the left vertical column indicates the (i + 1) -th column. It is shown that the number of data l _{i + 1} contained in the record is 1, 2, ...

At this time, in FIG. 4, the hatched area (b) and
In the remaining area (a) except for (c), the leveling is performed by combining the data of the i-th record and the (i + 1) -th record and storing them in one record. Is shown.

In (b) the region subjected to the hatch, the result of performing the leveling process, the number of data of the i-th record l _{i 'and} (i + 1) th number of data records l _{i + 1'} and, but each column It indicates that the distribution is performed as shown in (1).

In the hatched (c) area, even if the data of the (i + 1) th record was combined with the data of the i-th record, the result did not reach the required minimum data number m = 4.
Again, it shows that the same processing is required with the next adjacent record.

Therefore, if the condition "real-time recording / playback is possible when the number of data in each record is 1/2 or more of the record length" is satisfied, the minimum unit of editing is 1/8 of the record length.
It is possible to choose. And, this assumption is an assumption that can be sufficiently satisfied even with the current commercially available magnetic disks and the like.

Further, as can be seen from FIG. 4, according to the present invention, two records may be combined into one record, which has the advantage of improving the data storage efficiency.

In the present embodiment, the case where each record and the voice data contained therein are managed by the directory on the RAM is shown, but the present invention is also applicable to the voice data having a list structure with a pointer in each record. The invention can be similarly applied.

〔The invention's effect〕

As described above, according to the present invention, the audio data in the record can be processed and managed in a unit smaller than the record length while enabling real-time recording / playback. There is an advantage that a voice storing / editing device having a function can be configured. Further, even data other than voice (for example, a video signal) is effective especially when real time input / output is required. Also, since records with a small number of data are combined into one record,
There is also an advantage that the storage medium can be effectively used.

[Brief description of drawings]

FIG. 1 is an explanatory view of a voice data management system and a voice data deletion processing system in the case of the present invention, FIG. 2 is an explanatory view showing a leveling system of voice data in a record according to the present invention, and FIG. 3 is a view of the present invention. FIG. 4 is a flow chart showing an algorithm of a voice data leveling operation within a record, FIG. 4 is an explanatory view showing the types of data movements performed between two adjacent records, and FIG. 5 is a disk as a storage medium. FIG. 6 is a block diagram showing a general hardware configuration of a voice storage device, and FIG. 6 is an explanatory diagram showing a voice data management system. Explanation of symbols 1 ... A / D converter, 2 ... D / A converter, 3
... buffer, 4 ... disk controller, 5 ... disk, 6 ... RAM, 7 ... directory

Claims (1)

[Claims] 1. A storage medium for storing audio data, a table for designating and managing a data read position or a write position in the storage medium by a record, which is a quantitative unit when accessing the medium, and a control unit. Prepare,
In a voice storage device in which the control unit stores or reads voice data in the storage medium by referring to the table, in the record unit in the table as a result of the editing process of the stored data in the storage medium. When the address is rewritten by the control unit, the control unit checks, for each rewritten record in the table, whether or not the number of audio data contained therein is less than or equal to a certain threshold value m, and in this case, the number l _i of the audio data included in the record, and adding the number l _{i + 1} of the audio data contained in the adjacent record adjacent on the table, which can accommodate the sum is in the unit record data It is checked whether or not the number exceeds the number, and if the number exceeds the number, at least the first (m-1) of the audio data included in the adjacent record is detected. _i ) The data is taken out from the adjacent record and is added next to the voice data contained in the record. If not exceeded, the voice data contained in the record is added next to the voice data contained in the record. In the record, the whole record is extracted and added, the adjacent record is deleted from the table, and the number of audio data included in each record is equalized after the editing process in this way. Leveling method for voice data.