JP2006146110A - Speech converting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inexpensive speech converting device capable of playing a speech at high speed without changing an interval with the use of one-chip microcomputer while suppressing deterioration of the speech at the high speed playing. <P>SOLUTION: The device is composed of a microphone for inputting a speech, a jack for outputting a speech, an ADC for converting the inputted speech to digital speech data, a DAC for reconverting the digital speech data to a speech, a memory card for recording the digital speech data, a one-chip microcomputer for implementing a program, a program memory for storing an algorithm, a display unit for displaying an operation state, and an operation part by which a user selects an operation. The digital speech data are divided into blocks, and by laying the digital speech data by block and changing the sampling frequency at the time of playing, high speed and smooth playing of the speech can be achieved while suppressing deterioration of the speech. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an audio conversion device that performs high-speed audio reproduction without changing the pitch.

  Conventionally, in order to perform high-speed playback of sound without changing the pitch, there is also a method in which digital voice data is thinned out within a range where the pitch does not change (for example, see Patent Document 1). Some are realized by performing Fourier transform using a high-performance DSP.

The conventional method of thinning out digital audio data in the range where the pitch does not change has a problem that the interval of the thinned data becomes longer as the reproduction speed increases, and the deterioration of sound during reproduction becomes severe. It was.
JP 2000-89800 A

  In addition, the conventional method of performing Fourier transform using a DSP requires a high-performance DSP and a high-speed memory, so that the apparatus becomes expensive.

  The present invention is intended to solve the problems of such a conventional method, and while suppressing deterioration of sound during high-speed playback, the sound can be reproduced without changing the pitch using a one-chip microcomputer. An object of the present invention is to realize an inexpensive audio conversion device that can perform high-speed reproduction.

  An audio conversion apparatus according to the present invention includes a microphone that inputs audio, a jack that outputs audio, an ADC that converts input audio into digital audio data, a DAC that reconverts digital audio data into audio, and a digital It consists of a memory card that records audio data, a one-chip microcomputer that executes a program, a program memory that stores an algorithm, a display unit that displays an operation state, and an operation unit that allows a user to select an operation.

  Hereinafter, embodiments of the present invention will be described with reference to FIGS.

  FIG. 1 is a block diagram of a speech conversion apparatus according to the present invention. The audio input from the microphone is converted from an analog audio signal to digital audio data by the ADC, sent to a one-chip microcomputer, and recorded on a memory card.

  The digital audio data recorded on the memory card is taken out from the memory card by a one-chip microcomputer when the user of the audio conversion device gives an instruction from the operation unit, and the speed conversion process is performed according to the algorithm written in the program memory. After being sent to the DAC, the DAC converts the digital audio data into an analog audio signal and outputs it from the jack. The display unit displays changes in these operating states.

  Next, with reference to FIGS. 2 and 3, taking the reproduction at double speed as an example, the algorithm for high-speed reproduction of sound according to the present invention will be described.

  FIG. 2-1 is a diagram in which continuous digital audio data is blocked in a cycle in which the pitch does not change even if it is thinned out. One block includes 256 words of digital audio data as shown in FIG. It shall be assumed.

  The 256-word digital audio data included in one block is rearranged by overlapping one word twice as shown in FIG. At this time, the overflowed 129th to 256th digital audio data are discarded.

  When the digital audio data that has been rearranged by overlapping one word in Fig. 2-3 is played back at the same sampling frequency as when recording, the playback speed remains at 1x speed, and the audio is lowered by one octave. . Here, when the sampling frequency at the time of reproduction is doubled, the pitch returns to the original pitch, the reproduction speed is doubled, and reproduction at double speed can be realized without changing the pitch.

  FIG. 3 is a diagram comparing a conventional method of thinning out digital audio data within a range where the pitch does not change and the algorithm of the present invention.

  In the conventional method of thinning out digital audio data within a range in which the pitch does not change, even-numbered blocks of digital audio data are thinned out at hatched portions in the period T in FIG. 3-1.

  FIG. 3B is a data block in which the 1-word digital audio data shown in FIG. 2C is overlapped and rearranged twice. In each data block, the 129th to 256th digital audio data are truncated. The period of the truncated audio data is T / 2.

  FIG. 3C is a diagram when the block of the digital audio data in FIG. 3-2 is reproduced at double speed with the sampling frequency doubled, and the digital audio data cut out because the reproduction speed is doubled. The period is T / 4, and digital audio data can be thinned out with a period of 1/4 compared with the conventional method of thinning out digital audio data within the range where the pitch does not change, and smoothness is achieved by suppressing sound deterioration during playback. Sound can be played at high speed.

  As described above, the reproduction at the double speed has been described as an example. However, it is possible to perform high-speed audio reproduction at various speeds by combining digital audio data superimposition and the sampling frequency at the time of reproduction.

The invention's effect

  As described above, the audio conversion apparatus of the present invention can thin out digital audio data at a quarter cycle compared to the conventional method of thinning out digital audio data within a range where the pitch does not change. Smooth speech can be played at high speed while suppressing deterioration of the sound. In addition, since the algorithm is simple, it is possible to realize an audio conversion device using an inexpensive one-chip microcomputer without using a high-performance DSP or a high-speed memory.

1 is a block diagram of an audio conversion device showing an embodiment of the present invention. The figure for demonstrating the superimposition of the digital audio | voice data of the algorithm in embodiment of this invention. The figure for demonstrating reproduction | regeneration in the double speed of the algorithm in embodiment of this invention.

Claims (2)

  Digital audio data sampled based on the constant velocity is divided into blocks, and the digital audio data is overlaid on a block basis, and playback is performed with varying sampling rates, thereby suppressing deterioration of the audio during playback and reducing the original sound A method for high-speed sound reproduction, characterized in that sound is reproduced at high speed without changing the frequency.   2. A voice conversion apparatus according to claim 1, wherein said voice is reproduced at high speed using an inexpensive one-chip microcomputer.

Images