JPS58178395A - Time axis extension for voice signal - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION When the present invention reads audio signal data digitally stored in a storage device and outputs the audio signal, the audio signal is expanded on the time axis in order to lengthen the sounding time of the audio signal. Regarding the method.

When an audio signal within a limited time is converted into a digital signal, stored in a storage device, and the digital signal is arbitrarily read out to reproduce the stored audio signal, the stored original audio signal In some cases, you may want to extend the length of the sound.

For example, if you read the stored data at a shorter period than the sampling period of the audio signal and increase the reproduction frequency of the audio signal, the sounding time of the audio signal will be shortened, so you want to compensate for this, or you want to extend the original audio signal to make it sound longer. etc.

Conventionally, when there is an audio signal in which a waveform such as that shown in FIG. 1 is stored, the audio signal is expanded in the time axis by repeatedly reproducing a portion of the waveform within an appropriate period T.

However, in the conventional method as described above, if the period T is set relatively long, the repetition period is long and the seams may be audible, and if the original audio signal changes over time, the timbre and tone of the original audio signal may change. Because the amplitude changes repeatedly, the expanded audio signal may become unnatural compared to the original audio signal, and if the period T is set short, the expanded audio signal may become monotonous without any changes in timbre. This has the disadvantage that the pitch may be slightly distorted.

By the way, the above-mentioned case where the pitch is slightly out of order is due to the fact that the cycle for repeatedly reading out the stored digital signal is an integral multiple of the sampling cycle, and there is no specific relationship between the cycle of the original audio signal and the sampling cycle, resulting in an error. This is the case when this occurs.

For example, as shown in Figure 2, the signal waveform 3 of 400Hz is
．． The sampling value pulling value is stored in a storage device as a digital signal at an interval of 4 m5ec, and the sampling point T1 closest to the first zero-crossing point and the second When the sampling value between the sampling point T13 closest to the zero crossing point of is read out in the same period as the sampling period, and the reading operation is repeated, the sampling point T1 to TI3
The repetition period for reading is 0.4m5ec x 13
= 5.2m5ec, and the pitch of the reproduced audio signal is the second harmonic component of the pitch of the repetition period, 1/(5,2X10'')X2 = 384.6Hz.
, and an error occurs between the pitch of the reproduced audio signal and the original audio signal.

Therefore, the present invention has been made in order to eliminate the above-mentioned drawbacks, leave the timbre characteristics of the original audio signal intact, and extend the time axis of the audio signal.

Hereinafter, the present invention will be explained in detail with reference to the drawings.

FIG. 3 is a block diagram of an audio signal generation device that implements the audio signal time axis expansion method according to the present invention.

1 is a storage device that converts audio signals into digital signals and stores them; 2 is a digital-to-analog conversion device (hereinafter referred to as DAC) that converts the digital signals output from the storage device 1 into analog signals; 4 is a storage device that converts audio signals into digital signals and stores them; An output terminal 3 outputs the analog signal, a control device 3 controls the storage device 1 and the DAC 2, and a control device 5 controls the start and end of sounding the audio signal, and the control device 3 detects the operating state.

Now, assuming that the storage device 1 stores a digital signal representing the audio signal waveform as shown in FIG. When you operate the control device 3,
The digital signal between the sampling points closest to the zero cross point representing the first period A of the audio signal waveform (4-1) stored in the storage device 1 is sampled a desired number of times (in this specification, twice). ) reading, then the second
The digital signal between the sampling points closest to the zero-crossing point representing the period B part of is also read out twice, and the period C of bundle 3 is similarly read out and input to the DAC 2 (4-
As shown in 2), an audio signal in which the time axis of the audio signal waveform (4-1) is expanded is output.

The expanded audio signal has an error because the zero crossing point and the sampling point do not match as described above in each period part, but the error occurs around the period of the original audio signal and is always stored. By repeatedly reading out the waveform the same number of times, errors are absorbed as a whole, and when listened to as an audio signal, the pitch of the original audio signal can be reproduced.

In the above explanation, the pitch of the original audio signal and the reproduced audio signal are the same, but by changing the readout period, the period of the reproduced audio signal can also be changed.1. It goes without saying that the pitch of the reproduced audio signal can be changed freely.

As another example, a plurality of periods are set as a unit period as shown in FIG. A similar effect can be obtained by repeating.

As described above, the present invention is a method for reading an audio signal stored in a storage device while expanding it on the time axis, and the audio signal expanded by the present invention is a natural sound signal that does not lose the tonal characteristics of the original audio signal. Electronic music Figures 1 and 2 are waveform diagrams for explaining the conventional time axis expansion method, and Figure 3 is an audio signal generation diagram that implements the time axis expansion method according to the present invention. Block diagram of the device, No. 4
5 are waveform diagrams for explaining the time axis extension method of the present invention.

1...Storage device, 2...Digital-
Analog converter (DAC), 3...control device,
4... Output terminal, 5... Operating device.

Patent applicant: Roland Co., Ltd. Representative: Kadashi Ikutabe 11 Figure

Claims (1)

[Claims] An audio signal generation device comprising a storage means, a digital-to-analog conversion means, and a control means for reading out stored data of an audio signal stored in the storage means and inputting it to the digital-to-analog conversion means. By repeatedly reading each of the stored data corresponding to one to several cycles of the audio signal waveform the same number of times and inputting it to the digital-to-analog conversion means, the audio signal stored in the storage means is converted into a time signal. A time axis expansion method for an audio signal, wherein an audio signal expanded along the axis is obtained from the output side of the digital-to-analog conversion means.