TW200410126A - Digital audio sampling scheme - Google Patents

Abstract

A digital audio sampling scheme, which includes a computer implementing a software program for computation of impulse responses for an SRC filter by the weighted least square algorithm. The weighted least square algorithm can alternately be carried into execution by a DSP or a specific IC. As such, the entire invention, can efficiently minimize the computational power in software implementation.

Description

200410126 V. Description of the invention (1) The technology of the invention belongs to a digital audio sampling method, which uses a weighted minimum root mean square algorithm to expand the impulse response of an SRC. In the application example, as shown in FIG. 1, a digital mixer 15 can now be played from, for example, a sound card 丨 丨 a source machine 丨 2, an optical disc player, and a cassette player 14 Many of the audios are mixed at different sampling rates, and through this a back-end circuit 16 and a speaker on the left and right are used to enhance the colorful digital audio. Converting these digital audio data to the required sample rate is the main function of the mixer. Figure 2 is a waveform diagram of a 32 Hz sampling rate. Fig. 3 is a frequency response diagram of the i-th waveform generated by a discontinuous Fourier transform (DFT). Fig. 4 is a frequency response diagram of Fig. 1 generated when the sampling frequency is 512 Hz. Figure 5 is a block diagram of a sample rate converter (SRC) for changing the sampling rate (for example, from 32 Hz in Figure 2 to 512 Hz in Figure 3). In Figure 5, in practice, in order to simplify the calculation, the sampling rate conversion is performed in the time domain. As shown in FIG. 5, at time τ, the input data χκ passes through an interpolation device 4 using 5 to insert zeros between each pair of samples, a low-pass filter for performing convolution calculation (DFT), and a The scaler 43 for generating the time scale of the above-mentioned convolution calculation, thereby generating an output value γκ. In order to install such an SRC function in a 1-input mixing system, a typical method is to use a finite impulse response low-pass filter (hereinafter, referred to as an F IR crossover). Use the above interpolation calculations to obtain the signal &

0702-7568TW (nl); 90P144; Sue.ptd Page 5

200410126 V. Description of the invention (2) == Generates the impulse response used in the oversampling method for the master in the field of FIR filter design. ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Some-not used No complexity and no easy installation in the invention. In view of this, one of the purposes of the present invention is a method that uses the weighted minimum mean === the coefficient required when taking the impulse response of s R C. … ^ Expand to configure one. Includes installation-software. The overall invention can effectively minimize the software response required for software configuration. This method should be used. Wang Wenzhong, similar component functions are represented by the same component number. Figure 6 shows a block diagram of the digital mixer of the present invention. In Figure 6, all input digital audio data D1_Dn are added by the corresponding sampling rate converter (SRC) finite impulse response (FIR) low-pass filter η (hereafter referred to as FIR filter), and then added by The mixer 62 mixes to generate an output waveform. As shown in Fig. 6, the above-mentioned FIR filter 61 and the adder "form a mixer 65. The configuration of the mixer 65 is the same as that of the conventional structure, but in the configuration method of the key FIR filter It is a weighted minimum root mean square algorithm. The frequency response of the F IR filter of the present invention produced by using a weighted minimum root mean square algorithm can be expressed by the following formula:

200410126 V. Description of the invention (3) Σ p ^ 'n where (1) where P (z) is a Fourier transform function (hereinafter referred to as the z-function), and the coefficient Pn is the impulse response of the FIR filter, where N Is the length (order) of the FIR filter. Let // (z) be the desired frequency response of the FI r filter and the frequency response error function E is: £ 1ζ)-Σ P «z * n-^ Cz) ^ (2) Equation (2) can be evaluated On a linear density grid from ω = 0 to Jiang. For an F IR filter with a length N, a 4 N frequency grid point is sufficient to represent the efficiency of the filter. If the frequency band edge does not fall on the set frequency grid point, it is necessary to add extra grid points to correspond to the missing frequency band edge. At this time, it can be expressed by the following vector equation: £ = Ua-(3) where E-1 [£ (z 丨), £ (z2), ... Γ (4) U-· · · · • · * · * ( 5) a = [p〇. · * · .P «JT (6) A [A (Z 丨), ram (Z2), ... Γ (7) where zi + 1 > Zj.

0702-7568TWF (nl); 90P144; Sue.ptd p. 7 200410126

In the design of the weighted minimum root mean square method of the present invention, 〒ir wide)} 'skin minimization', where ^ is the minimum root mean square weighting value (/ aSweighting value). The optimized solution can be expressed using the following equation: /, (8) Only functions with diagonals

a = (iTRUriTR ^ where R can be expressed as a diagonal matrix as follows: Values q, r2, ...:

(9)

An example of a linear phase low-pass filter uses an exponential function to substitute the linear phase term of the above equations (1) to (9), and compares it with the conventional one. The comparison result is shown in FIG. The above reference is based on the condition that the band edge of the sampling frequency is 0.15, the filter length is 51, and all diagonal values are i. It can be seen from FIG. 7 that the filter and the wave filter configured by using the minimum root mean square method of the present invention have a ripple magnei tude which is much smaller than that of the conventional method. In addition, the efficiency of the filter configured using the minimum root-mean-square method of the present invention near the edge of the frequency band, when using a larger rn value near the edge of the frequency band, can reduce more efficiency than other places The losses in turn significantly increase the efficiency of the execution here. This is the focus of the present invention using the minimum root mean square technique. As described above, the method of digital audio sampling of the present invention includes a sound program in a computer for performing weighted minimum root mean square calculation by built-in ^

0702-7568TW (nl); 90P144; Sue.ptd page 8 " " " " ------- 200410126

The method of calculating the minimum root mean square of the pulse weight of a sampling rate conversion (SRC) filter can also be configured to respond in one digit. The implementation of the add-on integrated circuit is not limited to a device or a special = the present invention can effectively minimize the software configuration: =: = Although the present invention has been limited to the present invention with a better implementation, anyone who is familiar with the spirit of this technology Within the scope, when the scope of the patent application attached as the modification and encirclement is disclosed as above, but it is not used by those who do not depart from the embellishment of the present invention, the scope of protection of the present invention shall prevail.

200410126 Brief description of the drawings In order to make the above and other objects, features, and advantages of the present invention more obvious, a preferred embodiment is given below, and in conjunction with the accompanying drawings, the detailed description is as follows: Figure 1 shows a typical Block diagram of a digital audio system; Figure 2 is a waveform diagram with a sampling rate of 32 Hz; Figure 3 is a frequency response diagram of the waveform of the i-th waveform generated by discontinuous Fourier transform (DFT); Figure 4 Fig. 1 is a frequency response diagram of Fig. 1 generated when the sampling frequency is 512 Hz; Fig. 5 is a block diagram of a sampling rate converter ⑺ ^ ㈡ for changing the sampling rate; Fig. 6 shows a digital hybrid of the present invention Fig. 7 is a block diagram of a microphone; and Fig. 7 shows a comparison diagram of a filter frequency response generated by using the conventional method and the weighted minimum root mean square method of the present invention for optimization. [Symbol description] 11 sound card 12 sound source machine 13 disc player 14 cassette player 15 digital mixer 16 rear circuit 17 left speaker 18 right speaker 41 plug-in device

0702-7568TW (nl); 90P144; Sue.ptd Page 10 200410126 Brief description of the diagram 42 Low-pass crossing wave 43 Scaler 61 Finite impulse response low-pass filter (F I R filter) 62 Adder 6 5 Mixer

0702-7568TWF (η1); 90P144; Sue.p td Page 11

Claims (1)

200410126 VI. Patented digital audio methods, including the installation of a software program using a pulsed sampling rate conversion (SRC) filter, which should be installed in a Lei m ^. 8 computer, of which, The impulse response has etches at multiples of the phase 'sampling frequency, which can be used by the chirp filter to further generate the desired frequency response. A method for digital audio such as item 1 of the scope of patent application, in which a 3 · Λ root mean square algorithm is configured in a digital signal processor. The weighted minimum root-mean-square algorithm, the method of "information" of which 4 · As in the scope of the patent application, the first | μ 1 立 t 1 ^. The method of digital audio of item 1, where the impulse response With the equation 仏 (_1 务, the common is the desired frequency response, U is 滹 turbid? 彳 p 0, the horrible tA of the city. From the response wave function 'R is the one whose diagonal is desired The most surprising square root-weighted surprise angle matrix, and T is the reverse operation. W Different 5 · As in the patent application for the 4th day of the bismuth a "* ^ Ang 4 digital audio Method, where U, R, and / or // can be expressed by the following equations: ⑺ [务 （Z 丨） ， έ (ζ2), Ί r2 0 (5) 200410126 The scope of application for patents WWW 1, 2, Z22, ..., 2 are allowed, where 'ri, I " 2, ... is the minimum root mean square Λ leaf conversion function, μ is used to convert foreign rights in the field of time Function; ζ is the frequency response required by Fourier in the frequency domain, and when the two-input two-signal signal △: Limited impulse response (FIR) low-pass filter, the length of the wave represents a digital audio sampling method A mixer comprising a means for calculating a mean using a weighted minimum root-mean-square algorithm, the mixer comprising: a plurality of parallel SRC filters, each filter and # _ & ^ $ Impulse response, and each filter is connected to a sample of different audio sources or later and the one or more impulse responses are calculated to produce the desired output coefficient to form I want to use a long-time private t-connector, which is connected to the negative parallel SRC filters, and generates an audio output with each output coefficient of a. As shown in item 6 of the scope of patent application, there is a digital audio sampling method '妁: said state, wherein the means is a digital signal processor equipped with the weighted minimum root mean square algorithm. 0. ^ 8e, as in the sixth item of the patent application, which has a digital audio sampling method: said state, wherein the means is a dedicated integrated circuit equipped with the weighted minimum root mean square algorithm. 0702.7568W (nl); 90p144; Sue.ptd page 13 200410126 VI. Application for patent scope 9. If the scope of the patent application for item 7 is ancient ^ ^ ^, it happens that there is a mix of digital audio sampling methods The sound device, the device (means), is a computer that is based on the algorithm of the alpha dagger. The weighted minimum mean square is 10. For example, the mixer with a digital audio sampling method in item 7 of the patent application scope, wherein the sample is expressed by the equation a =, where 其中 is the desired frequency response, u is a filter function, r is a diagonal matrix whose diagonal is the desired minimum root-mean-square weighted value, and T is a reverse operation. 11. The digital audio method according to item 10 of the scope of patent application, wherein the above U, R and ugly can be expressed by the following equations: (7) ram = [έ (ζ ·), έ (ζ2), ... ] r \ 0… 0 0: · 0, · 0… 0 ··· 1 -j 9, ··· ， 2 ^ where ri 'r2 is the minimum mean square (5) weighting function; z is Fourier 1 L is cleverly taken / j, called a very weighted function; z is 穑 4: The function 'is used to convert the external input in the time domain to the desired frequency response in the frequency domain, and when i = 1 to NiN When the length of the limited impulse response (FIR) low-pass filter is first-in or first-out, the input requirement is either first-in or first-out. 1 + 11 advanced first Surface 07027568TWF (η1); 90Ρ144; Sue.ptd Page 14