Speech band division decoder

Details Speech band division decoder Speech band division decoder

1998-11-05

2004-04-06

Knepper, David D. (Department: 2654)

Data processing: speech signal processing, linguistics, language

Speech signal processing

Psychoacoustic

C704S200100

Reexamination Certificate

active

06718295

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates to speech band division decoders and, more particularly, to speech band division decoders having a tone quality control function.
In the field of signal processing, digital processing of signals has been replacing conventional analog processing owing to the appearance of CPU (central processing unit) and DSP (digital signal processor, which are fast and highly accurate in operation.
In addition, with performance improvement of CPU and DSP, coding and decoding of images and speeches which have been thought to be relatively heavy processes, have been being progressively realized by digital processing. Particularly, in the processing of speeches the tone quality control has become realized digitally.
For example, Japanese Patent Laid-Open Publication No. 59-178808 shows a technique, which uses FFT (Fast Fourier Transform) as digital filter to replace an equalizer which were heretofore constructed as a combination of analog filter, thus dispensing with a number of analog filters heretofore necessary for fine coefficient setting, increasing the available dynamic range which is not so broad due to the characteristics of analog filters and permitting gain setting in correspondence to finer frequency ranges.
FIG. 6
outlines a prior art decoder according to the above technique. In this case, a digital equalizer
55
comprising a decomposing filter
52
, a gain controller
53
and a synthesis filter
54
, is constructed separately from and combined with a coded data decoder
51
as speech data decoder.
As another technique, Japanese Patent Laid-Open Publication No. 6-348295 shows a commonly termed MPEG/audio decoder prescribed in speech coding/decoding standards ISO/IFC 11172-3, specifically a method of realizing a tone control function of that decoder in conformity to the user's taste with a simple construction by utilizing the algorithmic character of the decoder.
The outline of the above speech coding/decoding standards will now be summarized. The standards prescribes dividing a band division coded signal, obtained by dividing the frequency band of a speech signal into a plurality of sub-bands, coding the individual sub-band signals and combining the coded sub-band signals, into a plurality of sub-bands, changing the scale factor of a predetermined sub-band coded signal, combining the individual coded signals and then analog converting the resultant signal to obtain a speech signal having a predetermined frequency characteristic.
FIG. 7
outlines the disclosed MPEG/audio decoder. The decoder will now be described with reference to the Figure.
An inverse quantizer
61
converts band division coded (or sub-band coded) data to speech data of each sub-band. Then, a gain control circuit
62
controls the gain of each sub-band speech data. A side-band synthesis filter bank
63
then synthesizes the resultant individual side-band speech data. The tone control is executed by duplicating the MPEG/audio algorithm other than the gain control.
The publication noted above also teaches that when each sub-band speech data from the inverse quantizer
61
includes range data representing the sound volume, the tone quality may be controlled by controlling the range data. In the MPEG/audio decoder, the tone control may thus be made by controlling scale factor data representing the playback scale factor of each sub-band.
The construction and operation of an MPEG 1/Audio layer
1
decoder with a tone control function will now be described.
FIG. 8
is a block diagram showing a prior art MPEG 1/Audio layer
1
decoder. The decoder is basically a band division coding (or sub-band coding) system, and thus executes the process in units of sub-bands having an equal bandwidth, for instance of 32 sub-bands.
The MPEG 1/Audio layer
1
decoder
71
comprises a bit stream decomposer
72
, an inverse quantizer
73
, a sub-band synthesis filter bank
74
and a side-band data decoder
75
.
The bit stream decomposer
72
decomposes an MPEG 1/audio bit stream signal supplied to the decoder
71
into a 32-sub-band quantized signal and side-band data as attribute data.
The side-band data decoder
75
decodes the side-band data to restore bit assignment data representing the number of bits assigned to each sub-band (the number of quantization bits being determined by utilizing human's acoustical character, a greater number of bits being assigned to a sub-band containing part important for human's hearing sense, a less number of bits being assigned otherwise) and scale factor data representing the playback scale factor of each sub-band.
The inverse quantizer
73
generates 32-sub-band equal bandwidth sub-band signals from individual signals obtained as a result of the bit stream decomposition. The sub-band synthesis filter bank
74
synthesizes the sub-band signals thus obtained, and thus supplies linear PCM audio data (LPCM). The frequency characteristic of the LPCM that is supplied can be controlled by controlling the playback scale factor of each sub-band.
However, when the scale factor data representing the playback scale factor in the increasing direction, it may happen that a number of bits are necessary, which exceeds a range that the LPCM as the output data of the decoder (for instance 16 bits) can be expressed.
To avoid this, the frequency characteristic of the LPCM is controlled in the direction of reducing the scale factor data, i.e., in the direction of reducing the playback scale factor. When a certain frequency band is to be emphasized, the scale factor data of the other bands is reduced. In this case, however, the gain of the signal of the sub-band to be emphasized is not changed. Therefore, a sufficient sense of emphasis cannot be obtained in this operation alone.
SUMMARY OF THE INVENTION
An object of the present invention, accordingly, is to provide a speech band division decoder, which permits a sense of emphasis of a desired frequency band to be reliably obtained.
According to the present invention, there is provided a speech band division decoder for decomposing a band division coded signal, obtained by dividing a speech signal into a plurality of frequency sub-band signals, coding these signals and synthesizing the coded signals, into a plurality of sub-bands, updating a playback scale factor of a predetermined sub-band coded signal, synthesizing the resultant coded signals and analog converting the resultant signal to obtain a speech signal having a predetermined frequency characteristic, the speech band division decoder further comprising speech signal gain control means for controlling the gain of the analog converted speech signal in the direction of canceling a change in the playback scale factor.
The updating of the playback scale factor reduces the playback scale factor of coded signals of sub-bands adjacent to the predetermined sub-band. The updating of the playback scale factor is executed by updating a scale factor value provided in correspondence to the playback scale factor. The scale factor value is set such that it is reduced by increasing the playback scale factor. The speech band division decoder further comprises scale factor value control means for setting the scale factor to the upper limit when a scale factor value exceeding the upper limit is inputted. The coded signals obtained after the division of the band division code signal into a plurality of sub-bands further includes bit assignment data, the bit assignment being determined according to the hearing characteristic of human. The speech band division decoder further comprises playback scale factor input means or externally updating the playback scale factor.
According to another aspect of the present invention, there is provided a speech band division decoder for synthesizing coded signals of a plurality of frequency bands obtained after changing a playback scale factor of a predetermined frequency band to obtain a speech signal having a predetermined frequency characteristic, wherein the gain of output of the synthesized coded signal is controlled in th

Affiliated with

Also associated with

Rating

Speech band division decoder does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Speech band division decoder, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speech band division decoder will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3231891