首页 / 专利库 / 电脑编程 / 临时别名 / Method for implementing a fractional sample rate converter (F-SRC) and corresponding converter architecture

Method for implementing a fractional sample rate converter (F-SRC) and corresponding converter architecture

阅读:965发布:2020-06-18

专利汇可以提供Method for implementing a fractional sample rate converter (F-SRC) and corresponding converter architecture专利检索,专利查询,专利分析的服务。并且An up-sampled data stream, upsampled times a factor P, is generated by the input block providing signal samples having a frequency rate, while an intermediate data stream is generated by a Rate Adapting Stage providing signal samples adapted to an intermediate frequency rate. An output data stream is delivered by a final low pass filter, and includes M signal samples having a desired output sample frequency rate (fOUT=M/Ts). The method provides generation of a provisional stream in the Rate Adapting Stage, wherein this provisional stream is affected by aliases falling within the output Nyquist band [−fOUT/2, fOUT/2], although being adapted to the output frequency rate by a direct insertion of zero samples into the processed stream when L M. Then these aliases are suppressed in the Rate Adapting Stage by weighting the provisional stream via a set of weights.,下面是Method for implementing a fractional sample rate converter (F-SRC) and corresponding converter architecture专利的具体信息内容。

That which is claimed is:1. A method for implementing an L:M Fractional Sample Rate Converter (F-SRC), for controlled and direct insertion and cancellation of samples in a processed data stream, the method comprising:receiving an input data stream (DATA_IN) by an input block, processing L signal samples (x(m)) having an input data rate (fIN), in a predetermined time interval (Ts);generating an interpolated data stream (DATA_IN1), upsampled times a factor P (P<M), with the input block providing signal samples (x1(h)) having a data rate PfIN;generating an intermediate data stream (DATA_IMD1) with a rate adapting stage receiving the signal samples from the input block, and providing signal samples (y1(k)) adapted to an intermediate data rate (PfOUT));delivering an output data stream (DATA_OUT) with a low pass and P:1 decimation filter receiving the signal samples from the rate adapting stage, including M signal samples, z(k), having a desired output sample rate (fOUT=M/Ts);adapting the input data rate in the rate adapting stage to the output data rate by a direct insertion of repeated samples into the interpolated stream (DATA_IN1) when L<M, and by a direct cancellation of samples when L>M, thus generating a provisional intermediate stream DATA_DTY of samples w2(k) that are already adapted to an intermediate output rate PfOUT, although being affected by aliases falling within an output Nyquist band [−fOUT/2, fOUT/2]; andgenerating the intermediate stream (DATA_IMD1) in the rate adapting stage by weighting the signal samples (w2(k)) of the provisional intermediate data stream (DATA_DTY) to suppress aliases from the output Nyquist band [−fOUT/2, fOUT/2].2. The method according to claim 1, wherein the rate adapting stage includes:an up-sampling mode, when L<M and fIN<fOUT, generating an intermediate stream (DATA_IMD1), with signal samples y1(k) at the intermediate data rate (PfOUT), by directly inserting M−L repeated samples each L input samples of the interpolated stream (DATA_IN1); anda down-sampling mode, when L>M and fIN>fOUT, generating the intermediate stream (DATA_IMD1), with signal samples y1(k) at the intermediate data rate (PfOUT), by cancelling L−M samples each L input samples of the interpolated stream (DATA_IN1).3. The method according to claim 2, wherein in the up-sampling mode (L<M), repeated samples are inserted in locations of the interpolated stream (DATA_IN1), with signal samples x1(h)at a data rate PfIN, according to the following formula: w2(k)=x1(div(kL,M) where div(a,b) denotes the integer part of a/b; thus generating samples w2(k) of the provisional intermediate data stream (DATA_DTY), adapted to the intermediate rate PFOUT.4. The method according to claim 2, wherein is the down sampling mode, when L>M, data are cancelled from the interpolated stream (DATA_IN1), with signal samples w2(k) at a data rate PfIN, according to the following formula: w2(k)=x1(div(kL,M); thus generating samples of the provisional intermediate data stream (DATA_DTY), adapted to the intermediate rate PfOUT.5. The method according to claim 1, wherein weighting the samples w2(k) of the provisional intermediate stream (DATA_DTY) in said Rate Adapting Stage suppresses aliases falling within the output Nyquist bad [−fOUT/2, fOUT/2]of the provisional intermediate stream (DATA_DTY), generated by the samples insertions and cancellations; a set of weights Wg(k) is applied to the signal samples in an order, different than a natural order, thus generating the samples y1(k) of the intermediate stream DATA_IMD1 according to the following equation:
y1(k)=w2(k)*Wg(mod(kL,M)), k=0, 1, 2, . . . ,
where mod(a,b) denotes the remainder of the division a/b.
6. The method according to claim 5, wherein the set of weights Wg(k) is computed according to the following equation:
Wg(k)=Ko|q(k)*(1−q(k))|Q; q(k)=k/(M−1), h=0, 1, . . . , (M−1);
Ko being a normalization factor and Q a parameter selected to obtain the desired alias suppression.
7. The method according to claim 1, wherein the input block, includes an up-sampler (1:P) for interpolating the signal samples (x(m)) of the input signal times a P up-sampling factor, thus increasing the input data rate from fIN to P*fIN, where P≧Pmin=M/|M−L|, P<M.8. The method according to claim 1, wherein the rate adapting stage, receives a data stream with a data rate of P*fIN, and outputs a data stream with a data rate of (M/L)*P*fIN=P*fOUT.9. An L:M fractional sample rate converter (F-SRC) architecture for controlled and direct insertion and cancellation of samples in a processed data stream, comprising:an up-sampler (1:P) input block receiving an input data stream (DATA_IN) and processing L signal samples (x(m)) having an input data rate (fIN) in a predetermined time interval (Ts), and providing an interpolated data stream (DATA_IN1), upsampled times a factor P, P<M, therefore having a PfIN data rate;a rate adapting stage generating an intermediate data stream (DATA_IMD1) with signal samples (y1(k)) adapted to an intermediate data rate (PfOUT); anda low pass and decimation filter delivering an output data stream (DATA_OUT) including M signal samples wherein (z(n)) having a desired output sample rate (fOUT=M/Ts);the rate adapting stage weighting the signal samples (x1(h)) of the interpolated data stream; andthe rate adapting stage adapting the input data rate to the output data rate by a direct insertion of repeated samples into the processed stream when L<M, and by a direct cancellation of samples when L>M.10. The L:M Fractional Sample Rate Converter (F-SRC) architecture according to claim 9, wherein said low-pass and decimation filter is structured to decimate the intermediate data stream times a factor P thus reducing the output rate down to (M/L)*fIN=fOUT.11. The L:M Fractional Sample Rate Converter (F-SRC) architecture according to claim 9, wherein said low-pass and decimation filter is structured to suppress possible aliases falling out of a Nyquist output band [−fIN/2, fIN/2] in the intermediate data stream before a P:1 decimation is executed.12. The L:M Fractional Sample Rate Converter (F-SRC) architecture according to claim 9, wherein said up-sampler (1:P) input block provides interpolation by an up-sampling factor P<M, P>Pmin=M/|M−L|.13. The L:M Fractional Sample Rate Converter (F-SRC) architecture according to claim 10, wherein the decimation factor P of the low-pass and decimation filter is equal to the up-sampling factor P of the up-sampler input block.14. The L:M Fractional Sample Rate Converter (F-SRC) architecture according to claim 10, wherein when the decimation factor of the low-pass and decimation filter is different from the up-sampling factor P of the up-sampler input block, a conversion rate is given by (LP)/(MPOUT), where POUT denotes the decimation factor of the low-pass and decimation filter in the case POUT≠P.15. An L:M fractional sample rate converter (F-SRC) comprising:an up-sampler receiving an input data stream (DATA_IN) and processing L signal samples (x(m)) having an input data rate (fIN) in a predetermined time interval (Ts), and providing an interpolated data stream (DATA_IN1), upsampled by a factor P, P<M, and having a PfIN data rate;a rate adapting stage weighting the signal samples (x1(h)) of the interpolated data stream and generating an intermediate data stream, the rate adapting stage adapting the input data rate to the output data rate by insertion of repeated samples into the data stream when L<M, and by a cancellation of samples when L>M; anda decimation filter receiving the intermediate data stream and delivering an output data stream (DATA_OUT) including M signal samples (z(n)) having a desired output sample rate (fOUT=M/Ts).16. The L:M Fractional Sample Rate Converter (F-SRC) according to claim 15, wherein said decimation filter decimates the intermediate data stream by a factor P thus reducing the output rate down to (M/L)*fIN=FOUT.17. The L:M Fractional Sample Rate Converter (F-SRC) according to claim 15, wherein said decimation filter suppresses aliases falling out of a Nyquist output band [−fIN/2, fIN/2] in the intermediate data stream before a P:1 decimation is executed.18. The L:M Fractional Sample Rate Converter (F-SRC) according to claim 15, wherein said up-sampler provides interpolation by an up-sampling factor P<M, P≧Pmin=M/|M−L|.19. The L:M Fractional Sample Rate Converter (F-SRC) according to claim 16, wherein the decimation factor P of the decimation filter is equal to the up-sampling factor P of the up-sampler.

说明书全文

FIELD OF THE INVENTION

The present invention relates generally to a method for the Implementation of a Fractional Sample Rate Converter (F-SRC), based on a controlled, direct insertion/cancellation of samples in a processed data stream. In particular, the method according to the present invention is applicable both to VLSI and real-time software implementations of F-SRCs. The present invention relates also to a Fractional Sample Rate Converter architecture (F-SRC) adapted for applying the above-mentioned method.

BACKGROUND OF THE INVENTION

Sample Rate Converters (SRC) are important VLSI (Very Large Scale of Integration) devices used in many segments of the consumer market, such as Car Audio, Home Video/Home Theatre, Hi-Fi, and need to adapt the sample rates used by some standards to the rates required by emerging standards. A well-known example is the 44.1 KHz:48 KHz SRC, converting the earlier audio data rate of the CD format to the more qualitative data rate of DAT (Digital Audio Tape) and DVD (Digital Video Broadcasting).

A L:M Fractional SRC (F-SRC) receives L samples in a predefined interval Ts, and outputs M samples in the same interval Ts, being L, M integers: therefore, if L<M, it somehow ‘creates’ and ‘inserts’ M−L samples each L input samples into the output stream, or deletes L−M each L if L>M. These operations require special care and cannot be executed by directly inserting/canceling the |M−L| samples, because of the generation of images and alias signals into the output stream.

Therefore classical and known implementations of F-SRCs require an interpolator, anti-imaging/anti-aliasing FIR (Finite Impulsive Response) filter, working at the oversampling rate of L*M/Ts samples/sec. Finally a L:1 decimation stage lowers the output rate down to M/Ts samples/sec. In the following it is assumed that L, M, are coprime integers.

For example, in FIG. 1 is illustrated a typical and known implementation of a sampling rate conversion based on a polyphase filter structure. This is a bank of M FIR filters with identical number of taps, the same gain, but different phase delays. Each of the FIR banks is called a phase of the polyphase filter. The data stream is filtered with the M banks in parallel, therefore M stream are generated in the same time interval Ts, and the data rate is increased by a factor M. In a Rate Adapting Stage, the output of the polyphase filter is decimated times a factor L with a selection logic that selects 1 data each L in a round-robin, thus the final output rate turns out to be multiplied times a factor M/L. If the bandwidth and the phase delays of the polyphase filter are suitably chosen this processing performs the sample rate conversion from L/Ts samples/sec to M/Ts samples/sec preserving the spectral quality of the input signal.

This known method, although being simple, has however some drawbacks. First of all it is really expensive if L, M, are large and high quality of signal is required. As an example the case of a 44.1 KHz to 48 KHz SRC can be considered, where L=147 and M=160. If a 64 tap FIR filter is selected for each phase (and in some cases this may not be enough to meet high quality requirements), a total of 64*160=10240 filtering coefficient are required. This will greatly increase both the hardware size in terms of gates and the power consumption of the device and it is not feasible in practical applications.

SUMMARY OF THE INVENTION

An object of the present invention is to provide a method for the Implementation of a Fractional Sample Rate Converter (F-SRC) having a overall very low complexity, and having such respective features as to overcome the drawbacks mentioned with reference to the prior art.

The method according to the present invention will use process to adapt the input rate to the output rate based on direct insertion of repeated samples into the input stream when L<M, or direct cancellation of samples when L>M. The processing is refined with a final low-pass filter to obtain the desired spectral characteristics of the output stream.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features and advantages of the method and the converter architecture according to the present invention will be more apparent from the description of an embodiment thereof, given hereafter with reference to the attached drawings given by way of an indicative, non limiting example.

FIG. 1 is a schematic diagram illustrating a prior art polyphase Sample Rate Converter;

FIG. 2 is a schematic diagram illustrating a reference architecture of a Fractional Sample Rate Converter related to the present invention;

FIG. 3 is an example of a timing diagram relating to an “intermediate stream generation via the direct insertions of repeated samples (up sampling)” in the particular case L=5, M=7;

FIG. 4 is an example of a timing diagram relating to an “intermediate stream generation via the direct sample cancellations (down sampling)” in the particular case L=7, M=5;

FIG. 5 is a schematic diagram illustrating a theoretical scheme for the direct insertion of repeated samples and direct samples cancellation processing;

FIG. 6 is a operations diagram illustrating operations for the block diagram of FIG. 5, when in up-sampling mode (L=5, M=7);

FIG. 7 is a operations diagram illustrating operations for the block diagram of FIG. 5, when in down-sampling mode (L=5, M=7);

FIG. 8 is a schematic diagram illustrating an example of generation of image signals and related aliases for down-sampling mode (L=7, M=5);

FIG. 9 is a graph illustrating spectral characteristics of an accumulate-decimate sync filter for the case of a 44.1 Khz to 48.0 Khz F-SRC (normalized to input frequency fIN);

FIG. 10 is a schematic diagram illustrating benefits of increasing input sampling frequency on the images and aliases generated in the output stream;

FIG. 11 is a schematic diagram illustrating a possible theoretical implementation of the inventive F-SRC architecture;

FIG. 12 is a schematic diagram illustrating a F-SRC architecture according to the present invention;

FIG. 13 is a schematic diagram illustrating circuit implementing a F-SRC architecture according to the present invention;

FIG. 14 is a graph illustrating Windowing functions for the weighting window depicted in FIG. 11;

FIG. 15 is a table of specification for 44.1 kHz to 48 KHz SRC mode for an embodiment of the invention;

FIG. 16 is a graph illustrating spectral characteristics for 44.1 kHz to 48 KHz SRC whose specification are given in FIG. 15: DATA_OUT signal when the input is the sum of 10 different sinusoidal signals, with same spectral separation; and

FIG. 17 is a graph illustrating spectral characteristics for 44.1 kHz to 48 KHz SRC whose specification are given in FIG. 15: Pulse response (input FIR+output IIR).

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

With specific reference to the embodiment of FIG. 12, a general block diagram of a Fractional Sample Rate Converter (F-SRC) architecture according to the invention is illustrated and which includes the following main blocks: 1:P upsampler stage; a Low-Complexity Rate Adapting Stage including a block executing a controlled and direct insertion/cancellation of samples, and a block executing an Aliasing Post-compensation; and a Low-pass output filter and P:1 decimator, wherein the Low-Complexity Rate Adapting Stage is the core of the present invention.

The Aliasing Precompensation Block used in the F-SRC architecture of the invention is based on a weighting technique whose features will be further disclosed. In the meantime, for a better understanding of the invention and to analyze each above block and its function, particularly the Rate Adapting Stage, it is convenient to first consider a reference architecture of a Fractional Sample Rate Converter (FIG. 2).

With reference to FIG. 2, an input stream DATA_IN is provided, whereby L signal samples in a time interval Ts are delivered x(h) corresponds to the samples of DATA_IN, whose rate is fIN, and whose bandwidth is fs. In other words fIN means the input sample frequency, i.e. the frequency of the input data clock, say CLK_IN, (fIN=L/Ts). The method provides an output stream DATA_OUT, whereby M signal samples in the same time interval Ts are delivered. Let fOUT be the output sample frequency, i.e. the frequency of the output data clock, say CLK_OUT, (fOUT=M/Ts). To satisfy the Nyquist condition, the following equation shall hold:



fs≦min(fIN, fOUT)  (1)

Further the Rate Adapting Stage is provided to generate an intermediate stream DATA_IMD, already adapted to the final rate fOUT, whose samples are called y(k). Moreover, as depicted in FIG. 2, the stream DATA_IMD may require a further filtering process to achieve an output stream DATA_OUT with the desired spectral characteristics. According to the conventional terminology, the Rate Adapting Stage provides two operation modes, which we will distinguish up-sampling mode, i.e.: L<M, and down-sampling mode, i.e.: L>M. M, L, are supposed to be co-prime.

In the reference architecture of FIG. 2, the Rate Adapting Stage, when in Up-Sampling mode (L<M, fIN<fOUT), provides a simple but raw way to generate the intermediate stream y(k), at the output rate fOUT, by directly inserting M−L repeated samples each L input samples of x(h) in suitable locations, as described in the following: since fIN<fOUT, if we resample the input stream x(h), whose rate is fIN, with a clock at frequency fOUT, some data are sampled twice, and repeated samples are generated. The related timing is shown in FIG. 3 for the particular case L=5, M=7.

The Rate Adapting Stage, when in down-sampling mode (L>M, fIN>fOUT) should provide a direct cancellation of L−M samples. In this case a simple but raw rate conversion may be achieved by the consideration that, if we resample the input stream x(h), with a clock at frequency fOUT, L−M samples are skipped each L samples, and they will miss from the output stream. The related timing is shown in FIG. 4 for the particular case L=7, M=5.

It is seen that these simple theoretical schemes for the up-sampling mode and down sampling mode work only under limited conditions, but a similar processing scheme will work even with L, M, suitable in many practical application. To clarify this point, we consider the Rate Adapting Stage depicted in FIG. 5. In FIG. 5 the block named ‘1:M zero-padding’ inserts M−1 zeros between two consecutive samples of DATA_IN, thus achieving an up-sampled zero-padded stream, say DATA_OVS, whose samples are called w(j), at rate MfIN. The block named ‘1:M zero-order hold’ inserts M−1 repeated samples between two consecutive samples of DATA_IN, thus achieving an up-sampled stream, say DATA_OVSa, whose samples are called wa (k), at rate MfIN. This blcok is built by a 1:P zero padding block, cascaded with a M-length moving window whose length is M, as depicted in the FIG. 5, wherein the moving window, executes the sum of the last received M samples.

It may be shown that the raw direct insertion of repeated samples technique and the raw direct cancellation technique described above, are equivalent to process the input stream with this processing scheme. To this end, let's limit for the moment our attention to the case:



0.5<M/L<2;  (2)

In this case, for up-sampling mode, the samples of the stream DATA_IMD output from the Rate Adapting Stage of the block diagram of FIG. 5 are given by the following equations:



y(k)=x(div(kL,M)),  (3)



div(a,b) being the integer part of a/b.

These equation may be obtained by considering that each M by M summation executed by the moving window always contains just a single non-zero sample of x(h). Therefore the moving window output is a train of sequences, each built by M repeated samples thus providing in output M identical samples; moreover, the M samples of the h-th sequence are all equal to x(h). (see FIG. 6). Since the L:1 decimator selects 1 samples each L among the output of the moving window, and the h-th sequence is built by M samples equal to x(h), the k-th decimated samples turns out to be provided by x(div(kL,M)).

As a final result of this process, |L−M| repeated samples each L are added to the DATA_IN stream, due to the fact that fIN<fOUT. It is important to point out that the quantity div(kL, M) even represents the number of clock cycles of CLK_IN occurring within the k-th clock cycle of CLK_OUT or, from an equivalent point of view, a data running at rate L/Ts resampled at rate M/Ts. Hence it may be understood why the timing diagram of FIG. 3 even represents operations of the block diagram of FIG. 5 when in up-sampling mode.

For down-sampling operation mode, the same equations and the same considerations hold. In this case |L−M| samples each L are cancelled from the resampled stream since they are skipped, due to the fact that fIN>fOUT. Hence it may be understood why the timing diagram of FIG. 4 even represents operations of the block diagram drawn in FIG. 5 when in down-sampling mode.

FIG. 6 and FIG. 7 show examples of operation for the block diagram of FIG. 5, for the case of L=5, M=7, and L=7, M=5. It may be seen that the output streams of this figures are the same as in FIG. 3, 4. We now investigate why the Rate Adapting Stage of FIG. 5 works only in a very limited number of cases. To this end, first consider the spectrum of DATA_OVS (see FIG. 8). This spectrum is built by M replicas (‘images’) of the baseband signal whose distance is fIN each other, and whose whole bandwidth is MfIN. For a generic frequency fo, images are generated at frequencies (see equation (4)):



fi(k)=fo+kfINk=±1, ±2, ±3, . . . ±div(M,2)  (4)



After L:1 decimation such images will generate aliased frequencies, say fa(k), into the output Nyquist band [−fOUT/2, fOUT/2], resulting in spectral degradation.

Locations of such aliases into the output Nyquist band may by obtained from fi(k) by subtracting an integer multiple of FOUT, such that the result is reduced within the output Nyquist band. Therefore the relationship between the image frequencies fi(k) in the input spectrum of DATA_IN and the related aliased frequencies fa(k) in the intermediate spectrum DATA_IMD is given (see the following equation):



fa(k)=fi(k)−u(kL/M)fout=fo+kfin−u(kL/M)foutk=±1, ±2, ±3, . . . , +div(M,2)  (5)



where u(kL/M) is an integer number defined by:



kL/M−½≦u(kL/M)<kL/M+½ if k≧0



kL/M−½<u(kL/M)≦kL/M+½ if k<0  (6)



An example is shown in FIG. 8 for down-sampling mode (up-sampling is very similar).

Therefore, to correctly reconstruct the signal spectrum at the output frequency fOUT, the DATA_OVS stream should be filtered before the L:1 decimation with a filter whose stop-band should suppress all the generated images. On the contrary, the only applied filter to DATA_OVS in the Rate Adapting Stage of FIG. 5 is the moving window, having a frequency response H(f) given by (in the fOUT frequency domain):

H

(

f

)

=

sin

c

(

π

f

f

IN

)

sin

c

(

π

f

M

f

IN

)

(

7

)



with sin c(x)=sin(x)/x. This function has a poor low-pass characteristic in most of practical cases, with not enough attenuation in the stop-band to suppress the images, and a great distortion in the pass-band. Spectral characteristics of this filter are plotted in FIG. 9, for the case of a 44.1 Khz to 48.0 Khz Sample Rate Converter.

In the following equations we analyze how to modify the Rate Adapting Stage of FIG. 5 and to make it work even with M, L, values suitable in industrial applications. To this end, consider the aliases generated by the images of the fo=0 frequency. They may be computed from (5), by substituting fo=0. Thus aliases fa (k) of fo=0 are given:



fa(k)=kfIN−u(kL/M)fOUT k=±1, ±2, ±3, . . . , ±div(M,2)



or, after few arithmetical manipulations:



fa(k)=v(kL,M)·(fOUT/M)k=±1, ±2, ±3, . . . , ±div(M,2)  ( 8)



where:



v(kL,M)=kL−u(kL/M)M  (9)



From the equations (6), (8), (9), it may by verified that v(kL,M) ranges from −div(M,2) to div(M,2) step 1. Therefore we can characterize the lower aliased frequency, the higher aliased frequency, and the inter-alias minimum spectral distance as follows:



min(|fa(k)|)=fOUT/M; (lower aliased frequency of fi=0)



max(|fa(k)|)=(fOUT/M)·div(M,2); (higher aliased frequency of fi=0)



Δfa=fOUT/M; (minimum inter-aliases spectral distance)



k=±1, ±2, ±3, . . . , +div(M,2)  (10)

We suppose now to up-sample the input signal DATA_IN times a factor P and to execute the described processing. This will cause two main effects: from the equation (4), it is seen that images are generated at frequency distance PfIN each other rather then fIN; from the equation (10), it is seen that some aliases are kicked out of the signal bandwidth [−fs/2, fs/2].

The situation is depicted in FIG. 10. Note that the aliases of the fo=0 falling within the new Nyquist output band [−PfOUT/2, PfOUT/2], have the new inter-alias spectral distance PΔfa rather then Δfa. On the contrary the signal spectrum remains unchanged (the signal is just up-sampled) and it will occupy just 1/P of the output bandwidth. This fact results in the reduction by a factor 1/P of the number of aliases falling inside the signal spectrum [−fs/2, fs/2].

Therefore, once a P factor has been chosen greater then 1, two groups of aliases are generated: the aliases falling out of the signal spectrum [−fs/2, fs/2]; and the aliases falling within the signal spectrum [−fs/2, fs/2]. The aliases falling out of the signal spectrum may be suppressed by filtering the intermediate signal y(k) with a suitable output filter. The aliases falling within the signal spectrum must be suppressed in advance, by filtering the images generating them before the L:1 decimation is executed. This may be done by suitably substituting the moving window with an improved weighting window when processing the signal w(j). This weighting window has improved spectral characteristics with respect to the spectral characteristics of the moving window whose spectral characteristics were expressed by equation (7) and FIG. 9.

The processing scheme of FIG. 5 may be modified now, according to the abovementioned guidelines, to make it work in many cases of practical interest. In this respect, a purely theoretical block diagram for a modified processing is shown in FIG. 11. This block diagram includes the following main blocks: 1:P Up-sampler: to implement the described method, we need the samples of the input signal up-sampled times a factor P. Since we are starting from the signal samples x(h) running at rate fIN, this first 1:P up-sampler stage executes the task to increase the input data rate from fIN to P*fIN. This 1:P Up-sampler is not architecture dependent and may be implemented via whichever up-sampling technique, as interpolation or even via a classical polyphase FIR filter, analogous to the one depicted in FIG. 1, but with lowered complexity owing to the fact that, as it will be shown further, it will result P<<M in many cases of practical interest. A Rate Adapting Stage: this is the core of the inventive processing. It receives a stream whose rate is P*fIN, and outputs a stream whose rate is (M/L)*P*fIN=P*FOUT. It performs a similar processing to the one depicted in FIG. 5, except that an improved weighting window is applied to the stream before the L:1 decimation, substituting the Moving Window of FIG. 5. This improved weighting window is needed to suppress those images in DATA_OVS1 that would generate aliases within the spectrum of the signal, [−fs/2, fs/2], after the first L:1 decimation. The difference between the Moving Window of FIG. 5 and the improved Weighting Window of FIG. 11 is that, while in the first case the samples w(j) are simply accumulated, in the second case the samples w1(k) are accumulated after having been weighted via a set of weights Wg(k) suitably chosen, in such a way to achieve the desired suppression of images before the L:1 decimation is executed. A Low-Pass output filter and P:1 decimator: this stage suppresses the aliases falling out of the spectrum of the signal, [−fs/2, fs/2], before the second decimation (P:1) is executed. Moreover, it will decimate the stream times a factor P thus reducing the output rate down to (M/L)*fIN=FOUT. It is not architecture dependent and may be implemented both in a FIR or IIR fashion.

It may be noted that the decimation factor P of the Low-Pass and decimation output filter may be equal to the up-sampling factor P of the up-sampler input block. However, when the decimation factor P of the Low-Pass and decimation output filter is different from the up-sampling factor P of the up-sampler input block the conversion rate is given by (L P)/(M POUT), where POUT denotes the output decimation factor, POUT≠P.

Since the 1:P Up-Sampler may be implemented via a polyphase FIR filter, and the Rate Adapting Stage depicted in FIG. 11 is something similar to a FIR filter, it could be difficult to immediately understand why this scheme achieves a simplification of the classical polyphase Sample Rate Converter depicted FIG. 1. The main advantages are in many practical cases the 1:P up-sampler may require an up-sampling factor P<<M (tenths, while M may be hundreds). This will be better explained further.

The Rate Adapting Stage may be implemented in a simple way. In fact the scheme of FIG. 11 is only a purely theoretical processing scheme, while in practical implementation the Rate Adapting Stage may directly apply eq. (3), reducing to the block of FIG. 12, such that: the block named ‘Direct insertion/cancellation of samples’ executes the following processing: a) directly. inserts repeated samples in suitable locations of the processed stream (DATA_IN1) when up-sampling, by direct application of equation (3); and b) directly cancels some samples from the processed stream (DATA_IN1) in suitable locations when down-sampling, by direct application of equation (3).

This operation generates the stream DATA_DTY, whose samples are called w2(h), that we call provisional since it is affected by aliases falling even within the signal spectrum [−fs/2, fs/2], although this provisional stream is adapted to the intermediate rate PfOUT. The Aliasing Precompensation block just weights each sample of the provisional stream (DATA_DTY) at the rate P*fIN, instead of executing the weighting window on the stream (DATA_OVS1) at the rate P*M*fs, provided that the weights are applied in an order that is suitably changed from their natural order, as it will be shown later (see equation (11) below). This way the stream DATA_IMD1 is generated.

In other words, it is seen that the block diagram of FIG. 12 yields a direct implementation of equations (3), and moreover executes the Aliasing Post-compensation to suppress the aliases generated by the direct insertion/cancellation of samples: this is sufficient to provide the Sampling Rate Conversion via the direct M−L direct insertion of repeated samples and the direct L−M cancellation technique, with L, M, suitable in many practical cases, under condition (2), suppressing the aliases generated in the provisional stream DATA_DTY within the signal band [−fs/2, fs/2].

The Rate Adapting Stage provided in the digital signal processing chain of FIG. 11 is analyzed as well as the reasons underlying the modifications leading to the simplified Low-Complexity Rate Adapting Stage depicted in FIG. 12. To understand this point, it can be noticed that the weighting window in the Rate Adapting Stage in block diagram of FIG. 11 shall be applied to the data of the stream DATA_OVS1. Since the most of these samples are zeros (it is a zero-padded stream), this allows the further great simplification of this block leading to the simplified block diagram of FIG. 12.

In fact the Aliasing Post-Compensation block just applies the set of weights of the improved weighting window only onto the samples (w2(k)) of the stream DATA_DTY at the rate PfIN, instead of applying them on the zero padded stream DATA_OVS1 at the rate PMfIN. This is possible if the weights application is executed according to a predetermined order that is different than their natural order. This fact requires a specific procedure in defining the permutation of the weighting coefficients that we are going to analyze in the following: in fact, it can be considered that the input to the weighting window block (DATA_OVS1) is a zero-padded signal based on the sequence of samples x1(h), such that each non-zero padded sample is followed by M−1 zeros. Therefore the output of the weighting window block is a series of M-length pulses, each one having a duration of M samples, the time shaping of the window itself, and multiplied times x1(h).

This stream is to be decimated L:1: thus the indexes of the decimated stream DATA_IMD1 is given by the remainders of the divisions of kL by M, i.e.: mod(kL,M): this defines the order in which the set of weights are to be applied to the stream DATA_DTY and the operations in the Aliasing Post-compensation block, in particular it means that the k-th sample of the stream DATA_DTY shall be multiplied times the weight Wg′(h)=Wg (mod(hM, L)), yielding the samples y1(k) of the stream DATA_IMD1 (see FIG. 12).



y1(k)=Wg(mod(kL, M))*w2(k);  (11)



The post-processing expressed by Eq. (11) allows suppression of the alias generated after the samples insertion/cancellation is executed. The above equation obviously holds whichever weighting window is selected.

The criteria for selecting the weights to be used will now be considered. As a matter of fact any weighting window could be used in the Aliasing Precompensation block of the inventive architecture depicted in FIG. 11 and FIG. 12, provided that it guarantees sufficient aliasing suppression. The window to be used should be shaped as a generic, finite energy pulse. Although a wide variety of weighting windows is available in the technical literature, the following family of windows has been chosen:



Wg(j)=(22Q)[x(j)*(1−x(j))]Q, x(j)=j/(M−1), j=0, 1, . . . , (M−1);  (12)



which provides the best performances.

Time and spectral characteristics of this family of windows are shown in FIG. 14. It may be seen that an excellent frequency behavior in terms of trade-off between selectivity of the pass-band and attenuation of the stop-band is shown. Usage of this particular window is to be intended as a preferred choice, however other weighting windows might be used to implement the method. Moreover, this family is particularly suitable to be runtime computed, due to its very simple form.

The number of used weights is M, therefore only M/2 memory locations are required to implement this windowing function (the window is symmetric). As far as the sample insertions/deletions are concerned, they occur in |M−L| cycles each L input clock cycles, thus achieving the target output rate of M/Ts sample per second. From an hardware point of view, the block named ‘Direct insertion/cancellation of |L−M| samples each L’ may be simply achieved by resampling the stream DATA_IN1, whose frequency is PfIN, via a register clocked at frequency PfOUT, as shown in FIG. 13. A SW implementation requires computation of the sequence div(kL,M), that is a periodical sequence with period M.

It's pointed out that the main drawback of the described method is that there do exist some pairs L, M, such that the reduction of the polyphase FIR filter cannot be applied, resulting P=M: this fact occurs when |M−L|=1. However the method turns out to be applicable in all other cases. To better clarify this point, consider that P may be chosen as the minimum up-sampling rate that allows pushing outside of the band [−fs/2, fs/2] only the aliases generated by those critical images too hard to be filtered by the weighting window (i.e.: the nearest images to fo=0). Hence an estimation of the P value can be achieved from Equations (8), (9), (10): in fact from these equations it can be obtained that the image fi (−1) (i.e.:the hardest to be filtered) generates the |M−L|th alias, i.e.:fa(M−L). Since this alias is located at a distance |M−L|Δfa=|M−L|*(fOUT/M) from fo=0, we can relocate it at a distance greater then fOUT from fo=0 by adopting the suitable up-sampling rate:to this end, we just need a P factor:



P≧Pmin=M/|M−L|  (13)



and this may result much lower than the M up-sampling factor required by the polyphase Sample Rate Converter architecture.

Referring now to a possible implementation when L/M>2 or L/M<0.5, the following considerations hold: When the ratio L/M>2 (up-sampling) or L/M<0.5 (down-sampling) the processing scheme needs to be slightly modified. In fact we consider two coprime integer numbers, s, p, such that:

L

M

=

pP

sP

sL

pM

and

:

0.5

<

sL

pM

<

2



These equations show that the presented method may still be applied by substituting M′=pM and L′=sL. The only difference in this case is that we must use an output decimation factor POUT different from the input up-sampling factor P.

In the following step we will consider numerical results for an example of implementation of a 44.1 KHz to 48 KHz Sample Rate Converter. The presented method may be applied to design a 16 bits, high resolution, 44.1 KHz to 48 KHz Sample Rate Converter. This implementation is useful to convert a CD data stream to DAT (Digital Audio Tape). The example of implementation requires an input polyphase FIR filter and a three-stage IIR output filter. Specifications and achieved results are listed in the FIG. 15, while spectral characteristics are shown in FIG. 16, FIG. 17. Note that FIG. 16, represents the output spectrum when the input signal is built by 10 different harmonics located within the [0–20 KHz] band, with same spectral distance each other.

As far as the memory occupation is concerned, both ROM memory and RAM (Random Access Memory) memory are required: ROM memory is used for coefficients storage, while RAM memory is used as input data buffer (tap lines). It is seen that the example implementation requires a whole of 586 ROM locations to store the FIR coefficients and the set of weights Wg(k). Note that the output IIR filter just require 18 ROM words as coefficients storage, while the remaining ROM is used for the input FIR coefficients (488 words) and the set of weights for the Aliasing Post-compensation block (80 words).

The RAM memory is used basically to implement the FIR tap lines, and it turns out to be 82 words. This is due to the fact that the input polyphase filter needs a number of memory locations equal to the taps of each single phase (64 taps) and the IIR filter, built by 3 IIR stages of 6th order, just requires 18 RAM words as data buffer.

高效检索全球专利

专利汇是专利免费检索,专利查询,专利分析-国家发明专利查询检索分析平台,是提供专利分析,专利查询,专利检索等数据服务功能的知识产权数据服务商。

我们的产品包含105个国家的1.26亿组数据,免费查、免费专利分析。

申请试用

分析报告

专利汇分析报告产品可以对行业情报数据进行梳理分析,涉及维度包括行业专利基本状况分析、地域分析、技术分析、发明人分析、申请人分析、专利权人分析、失效分析、核心专利分析、法律分析、研发重点分析、企业专利处境分析、技术处境分析、专利寿命分析、企业定位分析、引证分析等超过60个分析角度,系统通过AI智能系统对图表进行解读,只需1分钟,一键生成行业专利分析报告。

申请试用

QQ群二维码
意见反馈