KR950003627B1 - Bit allocating method using threshold audibility of digital audio machine - Google Patents

Abstract

No content.

Description

Bit Allocation Method Using Audible Threshold of Digital Audio Equipment

1A and 1B are graphs for explaining a data compression method using a conventional audible threshold curve.

2 is a block diagram of a device associated with a conventional bit allocation scheme.

3 is a graph illustrating a bit allocation method using a reference point of an audible threshold curve according to the present invention.

4 is a block diagram of an apparatus related to a bit allocation method according to the present invention.

* Explanation of symbols for main parts of the drawings

1: Pulse Code Modulator (PCM) 2: Register

3: code converter 4: audible threshold detector

5: ROM 6: Recording medium

10: audible threshold data correction unit

The present invention relates to a bit allocation method using an audible threshold point of a digital audio device. In particular, the present invention relates to a difference between an original sound generated when compressing data using a threshold of audibility during recording and playback of a digital audio device. It relates to a bit allocation method to minimize and improve sound quality.

In the conventional data compression method using an audible threshold curve, as shown in FIG. 1A, an audible threshold curve is applied to block data (1, 2, ..., N) generated by converting N digital input signals (time axis) to a frequency axis. After θ (N) = θ (1,2,…, N), only θ (N) satisfying θ (N) ≥ audible threshold curve is extracted and the first b according to the bit allocation stored in the ROM 5 is obtained. As shown in the figure, recording is performed after quantization of θ (2), θ (i), θ (K), and θ (L) from which datamins having an audible threshold or higher are extracted.

That is, in the apparatus shown in FIG. 2, the input signal X (t) is sampled in units of 16 bits through the pulse code modulator 1 (hereinafter referred to as PCM), and then N pieces of data are stored in the register 2. N data inputted to the audible threshold detector 4 after being recorded and then converted by the code matrix 3 by the first N * N primary matrix are detected by the audible threshold detector 4 through the audible threshold detector 4. Q1 to QN are quantized according to the number of bits of the ROM 5 that are assigned in advance and then recorded on the recording medium 6.

As described above, the bit allocation stored in the ROM 5 is determined as follows.

… … … … … … (One)

Where Ri is the number of bits allocated to the i-th data of the block whose time axis is converted to the frequency axis, sigma ² is the variation of the i-th block (variance) of the block i-time with the time axis being converted to the frequency axis, and Ro is the average number of bits allocated in the block. to be.

In this way, the data recorded on the recording medium 6 is not quantized according to the number of bits of the ROM 5 stored in advance below the audible threshold, so that the data is not recorded on the recording medium 6, and the N block of the quantized data is stored in the N block. The position is recorded as Side Information (SI).

As described above, the i-th change amount σ i ² of the block in which the time axis is converted to the frequency axis in Equation (1) used to allocate bits is a conventional equation.

… … … … … … (2)

It can be obtained as

Here, K is the K-th position of the block in which the time axis is converted to the frequency axis, and i is the block that is converted in the i-th.

In other words, in the above equation (1), the bit change increases when the change amount is large, and the bit allocation decreases when the change amount is small.

However, such a conventional data compression method uses the bit allocation method that does not consider the audible threshold at all, and the existing data stored in the ROM 5 is stored in the state where the (i) th data, which has been given low bit allocation, is higher than the audible threshold. In the case of quantization by bit allocation, as the quantization error occurs frequently, the compressed and reproduced sound quality is remarkably degraded compared to the original sound quality.

The present invention has been made to solve the problems of the prior art. When the data passing through the audible threshold detector is recorded by the data correction unit, the audible threshold value is set as a negative value. The purpose of the present invention is to provide a bit allocation method for processing the data, thereby preventing the sound quality from being degraded by the data compression ratio and improving the compression ratio relatively in the case of the sound quality reduction.

The present invention for achieving the above object is a signal conversion step of converting the audio signal from the time domain to the frequency domain after PCM, detecting an audible threshold point for the signal converted into the frequency domain, during recording and playback According to the data correction step of subtracting the audible threshold value from the value below the audible threshold point during recording, and adding the audible threshold value to the value below the audible threshold point during playback, and stored in the ROM according to the value calculated in the data correction step. It consists of the steps of quantizing with the assigned bits to record or reproduce.

Hereinafter, described in detail by the accompanying drawings of the present invention.

3 shows a graph for explaining a bit allocation method using a reference point of the audible threshold curve according to the present invention, which converts the time axis to the frequency side without using the horizontal axis as a reference point (zero point). The i-th change amount σ i ² is obtained.

In other words, if the masking level value belonging to the Kth in the block is Mk, the new average value Is

… … … … … … (3)

Becomes

If (X _ki -M _k ) <0, then X _ki -M _k = 0.

That is, it means that there is no value greater than the audible threshold value at any audible threshold point based on the audible threshold curve.

That is, in the present invention, only the data whose time domain is converted into the frequency domain is larger than the masking level is detected and applied as in the following equation (4).

here, Therefore, the above-mentioned? K ² is substituted into the above equation (1) to allocate bits.

This series of processes will be described with reference to the block diagram of the apparatus shown in FIG.

As described above, the input signal X (t) is sampled in units of 16 bits through the PCM (1), and then N data is written to the register 2, and then N * N unitary matrixes are provided. N data inputted to the audible threshold detector 4 after being converted by the code converter 3 are applied to the data correction unit 10 after detecting the audible threshold through the audible threshold detector 4. .

At this time, the data correction unit 10 sets the audible threshold value as negative when recording and a positive value when playing.

The data corrected value is quantized to the allocated bits stored in the ROM 5 by the bit allocation method according to the present invention.

Here, the bit allocation method according to the present invention typically obtains the i-th change amount σ i ² of the block in which the time axis is converted into the frequency axis, assigns bits according to Equation (1) above, and then uses the masking curve as a reference point. Bit allocation is performed by obtaining the i-th change amount? I ² of the block converted to the new frequency axis.

As described above, the i-th change amount σ i ² of the block in which the new time axis is converted to the frequency axis has a value σ i ^{2 based} on the audible threshold curve as a reference point instead of the horizontal axis as a reference point.

In other words, assuming that the masking level value belonging to the K-th in the block is Mk, the new average value mK is expressed by the above equation (3).

The position in the data N block quantized by the allocated bits stored in the ROM 5 is recorded in the recording medium 6 as sub information SI.

As described above, according to the present invention, in the case of digital audio apparatus, when the data converted below the audible threshold is recorded and the data converted above the audible threshold is recorded by the number of quantization bits, the allocation of bits required for quantization is performed. By allocating bits on the basis of the audible threshold, the sound quality is prevented from being deteriorated due to the data compression ratio, and when the sound quality is at the same level, the compression ratio can be relatively increased to increase the recording amount.

Claims (2)

A signal conversion step of converting an audio signal into a frequency domain after PCM, detecting an audible threshold point for the signal converted into the frequency domain, and recording the audio threshold at or below the audible threshold during recording and playback. Subtracting from the value of, and during reproduction, a data correction step of adding the audible threshold to a value less than or equal to an audible threshold, and quantizing and recording or reproducing the allocated bit stored in the ROM 5 according to the value calculated in the data correction step. Bit allocation method using an audible threshold of a digital audio device, characterized in that made. The data correction step of claim 1, wherein the i-th change amount? I ² of the block transformed into the frequency domain on the time axis is calculated. According to Bit allocation method using an audible threshold point of a digital audio device, characterized in that the bit is calculated by calculating as follows.