Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
2000-10-18
2004-02-10
Dorvil, Richemond (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S214000, C704S208000, C704S226000
Reexamination Certificate
active
06691085
ABSTRACT:
FIELD OF THE INVENTION
The present invention generally relates to the field of coding and decoding synthesized speech and, more particularly, to such coding and decoding of wideband speech.
BACKGROUND OF THE INVENTION
Many methods of coding speech today are based upon linear predictive (LP) coding, which extracts perceptually significant features of a speech signal directly from a time waveform rather than from a frequency spectra of the speech signal (as does what is called a channel vocoder or what is called a formant vocoder). In LP coding, a speech waveform is first analyzed (LP analysis) to determine a time-varying model of the vocal tract excitation that caused the speech signal, and also a transfer function. A decoder (in a receiving terminal in case the coded speech signal is telecommunicated) then recreates the original speech using a synthesizer (for performing LP synthesis) that passes the excitation through a parameterized system that models the vocal tract. The parameters of the vocal tract model and the excitation of the model are both periodically updated to adapt to corresponding changes that occurred in the speaker as the speaker produced the speech signal. Between updates, i.e. during any specification interval, however, the excitation and parameters of the system are held constant, and so the process executed by the model is a linear-time-invariant process. The overall coding and decoding (distributed) system is called a codec.
In a codec using LP coding to generate speech, the decoder needs the coder to provide three inputs: a pitch period if the excitation is voiced, a gain factor and predictor coefficients. (In some codecs, the nature of the excitation, i.e. whether it is voiced or unvoiced, is also provided, but is not normally needed in case of an Algebraic Code Excited Linear Predictive (ACELP) codec, for example.) LP coding is predictive in that it uses prediction parameters based on the actual input segments of the speech waveform (during a specification interval) to which the parameters are applied, in a process of forward estimation.
Basic LP coding and decoding can be used to digitally communicate speech with a relatively low data rate, but it produces synthetic sounding speech because of its using a very simple system of excitation. A so-called Code Excited Linear Predictive (CELP) codec is an enhanced excitation codec. It is based on “residual” encoding. The modeling of the vocal tract is in terms of digital filters whose parameters are encoded in the compressed speech. These filters are driven, i.e. “excited,” by a signal that represents the vibration of the original speaker's vocal cords. A residual of an audio speech signal is the (original) audio speech signal less the digitally filtered audio speech signal. A CELP codec encodes the residual and uses it as a basis for excitation, in what is known as “residual pulse excitation.” However, instead of encoding the residual waveforms on a sample-by-sample basis, CELP uses a waveform template selected from a predetermined set of waveform templates in order to represent a block of residual samples. A codeword is determined by the coder and provided to the decoder, which then uses the codeword to select a residual sequence to represent the original residual samples.
FIG. 1
shows elements of a transmitter/encoder system and elements of a receiver/decoder system. The overall system serves as an LP codec, and could be a CELP-type codec. The transmitter accepts a sampled speech signal s(n) and provides it to an analyzer that determines LP parameters (inverse filter and synthesis filter) for a codec. s
q
(n) is the inverse filtered signal used to determine the residual x(n). The excitation search module encodes for transmission both the residual x(n), as a quantified or quantized error x
q
(n), and the synthesizer parameters and applies them to a communication channel leading to the receiver. On the receiver (decoder system) side, a decoder module extracts the synthesizer parameters from the transmitted signal and provides them to a synthesizer. The decoder module also determines the quantified error x
q
(n) from the transmitted signal. The output from the synthesizer is combined with the quantified error x
q
(n) to produce a quantified value s
q
(n) representing the original speech signal s(n).
A transmitter and receiver using a CELP-type codec functions in a similar way, except that the error x
q
(n) is transmitted as an index into a codebook representing various waveforms suitable for approximating the errors (residuals) x(n).
According to the Nyquist theorem, a speech signal with a sampling rate F
s
can represent a frequency band from 0 to 0.5F
s
. Nowadays, most speech codecs (coders-decoders) use a sampling rate of 8 kHz. If the sampling rate is increased from 8 kHz, naturalness of speech improves because higher frequencies can be represented. Today, the sampling rate of the speech signal is usually 8 kHz, but mobile telephone stations are being developed that will use a sampling rate of 16 kHz. According to the Nyquist theorem, a sampling rate of 16 kHz can represent speech in the frequency band 0-8 kHz. The sampled speech is then coded for communication by a transmitter, and then decoded by a receiver. Speech coding of speech sampled using a sampling rate of 16 kHz is called wideband speech coding.
When the sampling rate of speech is increased, coding complexity also increases. With some algorithms, as the sampling rate increases, coding complexity can even increase exponentially. Therefore, coding complexity is often a limiting factor in determining an algorithm for wideband speech coding. This is especially true, for example, with mobile telephone stations where power consumption, available processing power, and memory requirements critically affect the applicability of algorithms.
Sometimes in speech coding, a procedure known as decimation is used to reduce the complexity of the coding. Decimation reduces the original sampling rate for a sequence to a lower rate. It is the opposite of a procedure known as interpolation. The decimation process filters the input data with a low-pass filter and then re-samples the resulting smoothed signal at a lower rate. Interpolation increases the original sampling rate for a sequence to a higher rate. Interpolation inserts zeros into the original sequence and then applies a special low-pass filter to replace the zero values with interpolated values. The number of samples is thus increased.
Another prior-art wideband speech codec limits complexity by using sub-band coding. In such a sub-band coding approach, before encoding a wideband signal, it is divided into two signals, a lower band signal and a higher band signal. Both signals are then coded, independently of the other. In the decoder, in a synthesizing process, the two signals are recombined. Such an approach decreases coding complexity in those parts of the coding algorithm (such as the search for the innovative codebook) where complexity increases exponentially as a function of the sampling rate. However, in the parts where the complexity increases linearly, such an approach does not decrease the complexity.
The coding complexity of the above sub-band coding prior-art solution can be further decreased by ignoring the analysis of the higher band in the encoder and by replacing it with filtered white noise, or filtered pseudo-random noise, in the decoder, as shown in FIG.
2
. The analysis of the higher band can be ignored because human hearing is not sensitive to the phase response of the high frequency band but only to the amplitude response. The other reason is that only noise-like unvoiced phonemes contain energy in the higher band, whereas the voiced signal, for which phase is important, does not have significant energy in the higher band. In this approach, the spectrum of the higher band is estimated with an LP filter that has been generated from the lower band LP filter. Thus, no knowledge of the higher frequency band contents is sent over the transmission channel, and the generation of hi
Mikkola Hannu
Rotola-Pukkila Jani
Vainio Janne
Azad Abul K.
Dorvil Richemond
Nokia Mobile Phones Ltd.
Ware Fressola Van Der Sluys & Adolphson LLP
LandOfFree
Method and system for estimating artificial high band signal... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for estimating artificial high band signal..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for estimating artificial high band signal... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3329251