KR20050017088A - Sinusoidal audio coding - Google Patents

Abstract

Encoding (1) the audio signal (x) comprises providing each set of sampled signal values for each of the plurality of sequential segments. The sampled signal values are analyzed 130 to generate one or more sinusoidal components for each of the plurality of sequential segments. Sinusoidal components are connected across a plurality of sequential segments. Sinusoidal signs CS comprise tracks of connected sinusoidal components for each of the plurality of sequential segments. Each track includes the frequency and amplitude for the sinusoidal component in the start segment of the track, while the selected tracks include an indicator that no phase is included for the start segment.

Description

Sinusoidal audio coding

The present invention relates to encoding and decoding audio signals.

Parametric coding schemes are described in particular in sinusoidal coders in WO 00 / 79519-A1 (representative reference PHN 017502) and PCT Patent Application No. IB02 / 01297 (representative reference PHNL010252). In this encoder, the audio segment or frame is designed by a sinusoidal encoder using a number of sinusoidal functions expressed in amplitude, frequency and phase parameters. Once the sinusoids for the segment are estimated, the tracking algorithm begins. This algorithm attempts to connect the sine functions to each other in a segment-to-segment fashion. Thus sine function parameters from appropriate sine functions from sequential segments are concatenated to obtain the so-called tracks. The connection standard is based on the frequencies of the two subsequent segments, but also amplitude and / or phase information can be used. This information is combined into a cost function that determines the sine functions to be concatenated. The tracking algorithm thus becomes sinusoidal tracks starting at a particular time, and for a certain amount of time for multiple time segments, then stops.

In that way, for a sinusoidal track, an initial phase is transmitted and the phases of the other sinusoids in the track are retrieved from the frequencies of the initial phase and the other sinusoids. The amplitude and frequency of the sinusoid can also be coded differentially with respect to the previous sinusoids. Moreover, very short tracks can be eliminated. As such, due to the tracking, the bit rate of the sinusoidal encoder can be significantly lowered.

1 shows an embodiment of an audio encoder according to the invention.

2 shows an embodiment of an audio player according to the invention.

3 shows a system comprising an audio encoder and an audio player according to the invention.

According to the present invention there is provided a method for encoding an audio signal according to claim 1.

Description of the Preferred Embodiments

In a preferred embodiment of the invention, Fig. 1, the encoder is a sinusoidal encoder of the type described in WO 01 / 69593-A1 (representative reference PHNL000120). The operation of this encoder and its corresponding decoder is well described and only the description relating to the present invention is provided herein.

In both the conventional case and the preferred embodiment, the audio encoder 1 samples the input audio signal at a particular sampling frequency which is the digital representation x (t) of the audio signal. The encoder 1 then separates the sampled input signal into three components: transient signal components, sustain decision components, and sustain probability components. The audio encoder 1 includes a transient encoder 11, a sinusoidal encoder 13 and a noise encoder 14. The audio encoder optionally includes a gain compression mechanism (GC) 12.

The transient encoder 11 includes a transient detector (TD) 110, a transient analyzer (TA) 111, and a transient synthesizer (TS) 112. First, the signal x (t) enters the transient detector 110. This detector 110 estimates whether there is a transient signal component and its position. This information is supplied to the transient analyzer 111. Once the position of the transient signal component is determined, the transient analyzer 111 attempts to extract the transient signal component (the main part of). It matches the shape function to the signal segment that preferably starts at the estimated starting position, and determines the content below the shape function, for example by using (somewhat) a number of sinusoidal components. This information is included in the transient code CT and more detailed information is provided in WO 01 / 69593-A1 when generating the transient code CT.

Transient code CT is given to transient synthesizer 112. The synthesized transient signal component is subtracted from the input signal x (t) in the subtractor 16, resulting in a signal x1. Considering the case, GC 12 is omitted and x1 = x2.

The signal x2 is given to a sinusoidal encoder 13, which is where it is analyzed in a sinusoidal analyzer (SA) 130, which determines the (deterministic) sinusoidal components. Thus, although the presence of a transient analyzer is required, it will be appreciated that the analyzer is not necessary and that the present invention can be implemented without such an analyzer. In either case, the final result of the sinusoidal coding is sinusoidal sign CS and a more detailed example illustrating the conventional generation of an exemplary sinusoidal sign CS is provided in WO 00 / 79519-A1.

In short, however, such a sinusoidal encoder encodes the input signal x2 as tracks of sinusoidal components that are connected next in one frame segment. In the prior art, the tracks are initially represented by a start frequency, start amplitude and start phase for a sinusoid that starts at a given segment-generation.

In a preferred embodiment of the invention, the starting phase is selectively encoded for the track as a function of the track's length. More specifically, the start-phase is only used for long tracks. This is because long duration tracks are probably assumed to encode tonal information, and in such cases, it is important to preserve as many of the tonal characteristics of the track as possible by transmitting the starting phase of the track. Short duration tracks are assumed to encode non-tonal information and thus transmitting a starting phase with such tracks may actually add a tonal feature to the track, thus recognizing distortion when playing the coded bitstream. You can also give

It will be appreciated that there may be a significant saving in bit rate by not transmitting the start-phase for short tracks when the overhead of the start-phase data for the short track is proportionally higher than for the longer track.

There are a number of alternative criteria for determining whether a track is long enough to require a start phase or short enough to match so as not to require a start-phase.

The simplest criterion is to take an absolute track length-it has been found experimentally that tracks less than 40ms do not require a starting phase, while longer tracks are advantageously transferred to the start-phase. In encoders with a recent interval of 8 ms this means that tracks less than five segments in length do not contain a start-phase but rather include an indicator that the start-phase is not used as a track. (It is assumed that by comparison with the start-phase value, it is more efficient to encode such an indicator.) Alternatively, the encoder assumes that the encoded signal generated by the encoder will be decoded by a compatible decoder. Need not include an indication that no start-phase is used and cannot leave it to the decoder to determine how to advance the tracks without the start-phase.

An alternative criterion is based on determining whether the time interval in which the track is located is voiced or unvoiced. In places where it is determined that the time interval is voiced, it is assumed that the time interval is originally non-tonal and so that the tracks should not contain a start-phase, and vice versa for non-voice time intervals. In the IEEE Transactions on Acoustics, Speech, and Signal Processing, published in October 1976, L. R. Rabiner, M. J. Cheng, A. E. Rosenberg, CAMcGonegal, "A Comparative Performance Study of Several Pitch Detection Algorithms," Volume ASSP-24, pages 399-417, describes how to make such a decision. By revealing and including components that implement such a method in the tracking algorithm, the tracking algorithm will include start-phase information for tracks that exist within the tonal time interval, while tracks that exist within the non-tonal time interval. For, no start phase is included in the coded bitstream. This standard assumes that in tonal time-segment, tracks will tend to be longer than in non-tonal time-segment, so that the track's final length does not need to be known prior to determination that the track should include the start-phase Or not.

An alternative method of determining whether a time interval represents a tonal or non-tonal audio signal is to consider the energy level of the noise component of the signal, discussed below. Knowing that the ratio of noise energy to sinusoidal component energy exceeds a given threshold for a given time interval, in this way the audio signal is non-tonal and the start-phase information does not need to be included in the tracks and the noise energy. It can be assumed that the ratio of the sinusoidal component energy to the inverse is less than a given threshold.

In both the preferred embodiment and the prior art, the track is represented in subsequent segments by possible phase differences (continuouss) during frequency differences, amplitude differences and long tracks up to the segment where the track ends. In practice, it may be determined that there is some gain in coding phase differences even for long tracks. Thus, phase information never needs to be coded for successives and phase information for long tracks may be regenerated using successive phase reconstruction.

As in the prior art, from a sinusoidal sign CS generated with the improved sinusoidal coder of the present invention, a sinusoidal signal component is reconstructed by a sinusoidal synthesizer (SS) 131. This signal is subtracted from subtractor 17 from input x2 to sinusoidal encoder 13, resulting in a residual signal x3 lacking (large) transient signal components and (main) crystal sinusoidal components.

The residual signal x3 is assumed to mainly contain noise and the noise analyzer 14 of the preferred embodiment is a noise representative of the noise, as described, for example, in WO 01 / 89086-A1 (representative reference: PHNL000287). Generates the sign CN. Again, it will be appreciated that the use of such an analyzer is not essential to the implementation of the invention, but nevertheless is complementary to such use.

Finally, in the multiplexer 15, an audio stream AS comprising the signs CT, CS and CN is constructed. The audio stream AS is provided for example in a data bus, an antenna system, a storage medium and the like.

2 shows an audio player 3 according to the invention. For example, the audio stream AS 'generated by the encoder according to FIG. 1 is obtained from a data bus, an antenna system, a storage medium and the like. The audio stream AS is de-multiplexed in the demultiplexer 30 to obtain the signs CT, CS and CN. These codes are provided to the transient synthesizer 31, the sinusoidal synthesizer 32 and the noise synthesizer 33, respectively. From the transient sign CT, the transient signal components are calculated in the transient synthesizer 31. If the transient sign indicates a shape function, the shape is calculated based on the received parameters. In addition, the shape content is calculated based on the frequencies and amplitudes of the sinusoidal components. If the transient sign CT indicates a step, no transient is calculated. The total transient signal yT is the sum of all transients.

Sinusoidal code CS is used to generate signal yS, described as the sum of sinusoids on a given segment. In the decoder, the phase of the sinusoid in the sinusoidal track is determined in one of two ways. As in the prior art, at a location where the track contains a start-phase, the phase is calculated from the phase of the frequencies of the outgoing sinusoidal and intermediate sinusoidal curves. In a preferred embodiment, where the track contains an indication that no starting phase is provided, the decoder generates a random starting phase for all sinusoids in the track and then synthesizes the track as before. (The decoder may optionally calculate a random start-phase for only the outgoing sine curve or calculate residual phases as in the prior art.) Where no such indication or start-phase is provided, the decoder Infer that it is required to generate a random start-phase for the sine curves of the track.

It will be appreciated that one feature of the present invention is to preserve non-sound quality in non-tonal audio fragments. Thus, unlike the prior art, where very short tracks are not included anywhere in the bitstream, a decoder for reproducing these short tracks with random start phases and an encoder to preserve very short tracks for non-tonal audio fragments. It may be desirable when employing the present invention.

At the same time, noise code CN is supplied to noise synthesizer NS 33, which is primarily a filter with a frequency response that approximates the spectrum of the noise. NS 33 generates noise yN that is reconstructed by filtering the white noise signal with noise code CN.

The total signal y (t) includes the sum of the product of the transient signal yT and any amplitude decompression (g) and the sum of the sinusoidal signal yS and the noise signal yN. The audio player includes two adders 36 and 37 to sum the respective signals. The combined signal is provided to an output unit 35, which output is for example a speaker.

FIG. 3 shows an audio system according to the invention comprising an audio encoder 1 as shown in FIG. 1 and an audio player 3 as shown in FIG. 2. Such a system provides playback and recording features. The audio stream AS is fed from the audio encoder to the audio player on the communication channel 2, which may be a wireless connection, a data 20 bus or a storage medium. If the communication channel 2 is a storage medium, the storage medium may be fixed to the system or may be a removable disk, a memory stick or the like. The communication channel 2 may be part of the audio system, but will often be outside the audio system.

The present invention can be used with any sinusoidal audio encoder. As such, the present invention is applicable wherever such encoders are used.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and those skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word 'comprising' does not exclude the presence of elements or steps other than those listed in a claim. The invention can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the apparatus claim enumerating several means, some of these means may be implemented by one and the same item of hardware. The simple fact that some measurements are described in mutually dependent claims does not indicate that a combination of these measurements cannot be used to advantage.

Claims (15)

In the method of encoding the audio signal (x), Providing each set of sampled signal values for each of the plurality of sequential segments; Analyzing the sampled signal values to generate one or more sinusoidal components for each of the plurality of sequential segments; Linking sinusoidal components across a plurality of sequential segments, Generating sinusoidal signs comprising tracks of linked sinusoidal components for each of the plurality of sequential segments, each track including a frequency and an amplitude for the sinusoidal component in a starting segment of the track, Generating the sinusoidal signs, the tracks not comprising a phase for the starting segment, and Generating an encoded audio stream comprising the sinusoidal codes. The method of claim 1, And the selected tracks comprise an indicator that contains no phase for the start segment. The method of claim 1, And wherein said selected tracks are less than five segments in length. The method of claim 1, And said selected tracks are less than 40 ms in length. The method of claim 1, Wherein the selected tracks represent non-tonal components of an audio signal. The method of claim 1, Wherein the selected tracks represent components of a voiced time interval in the audio signal. The method of claim 1, Wherein the selected tracks represent a component of a noisy interval in the audio signal. The method of claim 1, Wherein each track comprises a frequency and amplitude difference for each sinusoidal component in subsequent successive segments of the track. In the method of decoding an audio stream, Reading an encoded audio stream comprising sinusoidal signs comprising tracks of linked sinusoidal components for each of the plurality of sequential segments, each track having a frequency and amplitude for the sinusoidal component at the beginning segment of the track. And wherein the selected tracks do not include a phase for the start segment; Generating a random starting phase for the selected tracks, and Using the sinusoidal signs to synthesize the audio signal, comprising reconstructing sinusoidal components over a plurality of sequential segments. The method of claim 9, Wherein the generating comprises generating a random phase for each sinusoidal component of the selected tracks. In an audio encoder arranged to process each set of sampled signal values for each of a plurality of sequential segments of an audio signal (x), An analyzer arranged to analyze the sampled signal values to generate one or more sinusoidal components for each of the plurality of sequential segments; A linker arranged to link sinusoidal components across a plurality of sequential segments, A component arranged to generate sinusoidal signs comprising tracks of linked sinusoidal components for each of the plurality of sequential segments, each track having a frequency and amplitude for the sinusoidal component at the beginning segment of the track. Wherein the selected tracks do not include a phase for the start segment, and And a bit stream generator for generating an encoded audio stream comprising the sinusoidal codes. In the audio player, Means for reading an encoded audio stream comprising sinusoidal signs comprising tracks of linked sinusoidal components for each of the plurality of sequential segments, each track having a frequency for the sinusoidal component at the beginning segment of the track and Said reading means comprising an amplitude, wherein selected tracks do not comprise a phase for said starting segment; A phase generator arranged to generate a random starting phase for the selected tracks, and And a synthesizer that uses the sinusoidal signs to synthesize the audio signal, comprising reconstructing sinusoidal components across a plurality of sequential segments. An audio system comprising an audio encoder as claimed in claim 11 and an audio player as claimed in claim 12. An audio stream comprising at least sinusoidal signs representing components of an audio signal, The symbols include tracks of sinusoidal components that are linked across the plurality of sequential segments, each track including a frequency and an amplitude for a sinusoidal component in a starting segment of the track, and selected tracks in the starting segment. An audio stream that does not contain a phase for. A storage medium having stored thereon an audio stream as claimed in claim 14.