Description
Title of Invention
DECODING APPARATUS AND METHOD, AND PROGRAM
Tecl~nicaFl ield
[OOOI]
The present technology relates to a decoding apparatus, a decoding method
and a program, and, Inore particularly, a decoding apparatus, a decoding method and
10 a program which make it possible to obtain sound with higher quality.
Background Att
[0002]
In related art, in an audio coding technique of moving picture experts group
15 (MPEG) advanced audio coding (AAC) (1~011~~14496-%2001it ),i s possible to
record aaoiliaty information of do\vnti~ixingo r dinatnic range coinpressiotl (DRC) in
a bit stream and use the anxiliaty iafonnation at a reproduction side according to an
environment of the reproduction side (see, for example, Nan-Patent Literature I).
[0003]
20 Use of such auxilia~yit ifonnation enables an audio signal to be downmixed
at the reproduction side or volulne to be appropriately controlled through the DRC.
Citation List
Nan-Patent Literature 1: Infor~nation technology Coding of audiovisual
objects Part 3: Audio (ISOIIEC 14496-3:2001)
S~cmmaryo f Itiveation
Technical Problem
For example, with the above-described coding techniqne, it is possible to
designate DRC gain information for volume control as auxiliary information of DllC
in units of a frame of an audio signal, and, at a reproduction side, by correcting
volume of the andio signal based on this DRC gain information, it is possible to
5 obtain sound with appropriate volume.
[0006]
However, a gain indicated by such DRC gain information becomes the same
value for each sample \vitI~ino ne frame of the audio signal which is a temporal signal.
That is, all samples included in one frame are corrected with the same gain.
10 [O007]
Therefore, for exaniple, when a magnitude of the gain indicated by the DRC
gain infonnation largely changes between frames, portions of temporal waveforms of
the audio signal become discontinuous between the frames, which may cause
degradation in auditory terms.
The present technology has been made in view of such circutnstances, and is
directed to making it possible to obtain sound with higher qnality.
Solution to Problem
20 [0009]
A decoding apparatus according to a first aspect of the present technology
includes: a gain readout unit configured to read ont encoded gain values at at least
two gain sample positions of a time series signal; an interpolation infor~nation
readout unit configured to read out interpolation infornlation indicating whether the
25 gain value at each sample position of the time series signal is obtained through linear
interpolation or obtained through non-linear interpolation; and an interpolation
processing unit configured to obtain the gain value at each sample position located
behvcen the two gain sample positions of the time series signal based on the gain
values at the gain sample positions thro~~glhin ear interpolation or non-linear
30 interpolation according to the interpolation infomation.
[OO lo]
The gain readout unit can be caused to fi~rther read out gain inclination
values indicating inclination of the gain values at the gain sample positions. When
the gain value is obtained through non-linear interpolation, the interpolation
processing unit can be caused to obtain the gain valoe at each sample position
5 located between the hvo gain sample positions based on the gain values and the gain
inclination values at the gain sample positions.
[OOI I]
The decoding apparatus can further include: a litniting processing unit
configured to perform limiting processing on the gain valoe obtained through non-
10 linear interpolation so that the gain value becomes a value equal to or greater than a
predetermined lower limit or a value equal to or less than a predetermined upper
limit.
[OO 121
The limiting processing unit can be caused to perfonn limiting processing
-
15 using zero as the lower limit, limiting processing using one as the lower limit or
limiting processing using one as the upper limit.
[OOI 31
The decoding apparatus can further include: an operation unit configured to
obtain at the gain sample positions, straight lines having the gain values at the gain
20 sample positions and having inclination indicated by the gain inclination values at
the gain satnple positions, and obtain differences behveen a gain value at an
itlteisection of the straight lines obtained for the hvo gain sample positions and the
gain values at the hvo gain satnple positions. When the interpolation information is
information.h~dicati~tg?hatth e gain value is obtained through linear interpolation, the
25 interpolation processing unit can be caused to obtain the gain value through linear
interpolation, and, when the interpolation information is infonnation indicating that
the gain value is obtained through non-linear interpolation, the interpolation
processing unit can be caused to obtain the gain value through non-linear
interpolation or linear interpolation according to the differeuces.
30 [OO14]
A decoding method or a prograln according to the first aspect of the present
tcchnology includes the steps of: leading out encoded gain valoes at at least two gain
satnple positions of a time series signal; reading out interpolation infortnation
indicating \vhether the gain value at each sample position of the time scries signal is
obtained through linear interpolation or obtained through non-linear interpolation;
5 and obtaining the gain valne at each sample position located between the twvo gain
sample positions of the time series signal based on the gain values at the gain satnple
positions through linear interpolation or non-linear interpolation according to the
interpolation information.
[0015]
10 According to the first aspect of the present technology, encoded gain values
at at least two gain sarnple positions of a time series signal are read out.
Interpolation information indicating whether the gain value at each sample position
of the time series signal is obtained through linear interpolation or obtained through
non-linear interpolation is read out. The gain value at each sarnple position located
15 behveen the twvo gain sample positions of tile time series signal based on the gain
values at the gain sample positions is obtained through linear interpolation or nonlinear
linear interpolatiot~a ccording to the interpolation information.
[0016]
A decoding apparatus according to a second aspect of the present
20 teclut~ology includes: a gain readout unit configured to read out encoded gain values
at at least h o gain satnple positions of a time series signal and gain inclination
values indicating inclination of the gain values; an operation unit configured to
obtain at the gain satnple positions, straight lines having the gain values at the gain
satnple positions and having inclination indicated by the gain inclination values at
25 the gain satnple positions, and obtain differences beheen a gain value at an
intersection of the straight lines obtained for the two gain sample positions and the
gain values at the hvo gain satnple positions; and an interpolation processing ltnit
configured to obtain the gain value at each satnple position located behveen the h o
gain sample positions of the time series signal thro~tgh linear interpolation or non-
30 linear interpolation according to the differences.
[OO 1 71
A decoding method or a program according to the second aspect of the
present technology includes the steps of: reading out encoded gain values at at least
two gain satnple positions of a time series signal and gain inclination values
indicating inclination of the gain values; obtaining at the gain sample positions,
5 straight lines having the gain values at the gain sample positions and having
inclination indicated by the gain inclination values at the gain sample positions, and
obtaining differences between a gain value at an intersection of the straight lines
obtained for the two gain sample positions and the gain values at the two gain sample
positions; and obtaining the gain value at each sample position located betbveen the
10 two gain sample positions of the tirne series signal through linear interpolation or
non-linear interpolation according to the differences.
[0018]
According to the second aspect of the present technology, encoded gain
values at at least two gain sample positions of a time series signal and gain
-
15 inclination values indicating inclination of the gain values are read out. At the gain
sample positions, straight lines having the gain values at the gain satnple positions
and having inclination indicated by the gain inclination valoes at the gain sample
positions are obtained and differences behveen a gain value at an intersection of the
straight lines obtained for the hvo gain sample positions and the gain values at the
20 two gain sample positions are obtained. The gain value at each salnple position
located between the two gain sample positions of the time series signal is obtained
through lineal. interpolation or non-linear interpolation according to the differences.
Advantageous Effeets of Invention
25 [0019]
According to the first aspect and the secotld aspect of the present technology,
it is possible to obtain sound with higher qilality.
[0020]
Note that advantageous effects are not limited to the advantageous effect
30 described herein and may be any advantageous effects described in the present
Brief Description of Drawings
[0021]
[FIG. I ] FIG. I is a diagram for explaining linear interpolation of a gain according to
5 an embodiment ofthe present technology.
[FIG. 21 FIG. 2 is a diagram illustrating an exatnple of a gain wavefolrn according to
an embodiment of the present technology.
[FIG. 31 FIG. 3 is a diagratn for explaining non-linear interpolation of a gain
according to an einboditnent of the present technology.
10 [FIG. 41 FIG. 4 is a diagram illitstrating a configoration example of an encoding
apparatus according to an embodiment of the present technology.
[FIG. 51 FIG. 5 is a flowchart explaining encoding processing according to an
embodiment of the present technology.
[FIG. 61 FIG. 6 is a diagram illustrating DRC characteristics according to an
15 embodiment of the present technology.
[FIG. 71 FIG. 7 is a diagratn illustrating a configi~ration example of a decoding
apparatus according to an elnboditnent of the present technology.
[FIG. 81 FIG. 8 is a flowchart explaining decoding processing according to an
etnbodiment of the present technology.
20 [FIG. 91 FIG. 9 is a flowchart explaining gain decoding processing according to an
eotbodiment ofthe present technology.
[FIG. 101 FIG. 10 is a diagratn illustrating a configitration example of a decoding
apparatus according to an enlboditnettt of the present technology.
[FIG. 111 FK:-l.l'is a flowchart explaining gain decoding processing according to an
25 etnbodirnent of the present technology.
[FIG. 121 FIG. 12 is a diagram for explaining interpolation of a gain waveform
according to an embodiment of the present techttology.
[FIG. 131 FIG. 13 is a diagram for explaining interpolation of a gain wavefonn
according to an embodiment of the present technologp.
30 [FIG. 141 FIG. 14 is a diagram for explaining interpolation of a gain waveform
according to an entboditnent of the present techno log)^.
[FIG. 151 FIG. 15 is a diagram illustrating a configuration example of a decoding
apparatus according to an e~nbodimenot f the present technology,
[FIG. 161 FIG. 16 is a flowchart explaining gain decoding processing according to an
embodiment of the present technology.
5 [FIG. 171 FIG. 17 is a diagram explaining interpolation of a gain wavefor~n according
to an embodiment ofthe present technology.
[FIG. 181 FIG. 18 is a flowchart explaining gain decoding processing according to all
embodiment of the present technology.
[FIG. 191 FIG. 19 is a diagrarn illustrating a configuration example of a computer
10 according to an embodiment of the present technology.
Description of Enlboditnents
[0022]
Hereinafter, an embodiment to which the present technology is applied will
-
15 be described with reference to drawings.
100231

The present technology relates to a technique of encoding a gain value when
20 volume of an audio signal is corrected at a reproduction side, multiplexing a gain
code string obtained by encoding the gain value and a signal code string obtained by
encoding the audio signal and transmitting the multiplexed code string, and a
technique of decoding these gain code string and signal code string and correcting
volume of the audio signal.
25 [0024]
In the present technolog): by designating an arbitrary value for each sample
within a flame of the audio signal as a gain value for volu~nec orrection, it is possible
to obtain sound with a s~nootherte mporal wavefonn. By this means, it is possible
to obtain sound with higher quality which does not cause a feeling of strangeness.
30 Here, while the gain value for volume correction may be a dB value or a linear value,
description will be continued belo\v assuming that the gain value is a linear value.
[0025]
Further, \vhen the gain value is encoded, if a gain code string is obtained by
encoding only gain values at part of satnple positions such as, for example,
characteristic positions sach as inflection points of a gain waveform and gain values
5 which are arranged at predetermined intervals among gain values at respective
sample positions within a kame, it is also possible to reduce a code amount of the
gain code string.
[0026]
In this case, a decoding side of the gain code string needs to obtain an
10 original gain waveform based on gain values at some satnplc positions obtained
through decoding of the gain code string.
[0027]
Here, as a method for obtaining the original gain waveform, for example,
there is a possible tnethod for obtaining gain values at sample positions which are not
- -
15 included in the gain code string by perfonning linear interpolation as illustrated in
FIG. 1.
[0028]
It should be noted that FIG. 1 indicates a gain value on a vertical axis and a
sample position within a fi.ame of a n a~idiosi gnal on a horizontal axis.
20 [0029]
Futthel; hereinafter, a sample position of an encoded gain value which is
included in a gain code string will be also specially referred to as a gain satnple
position. Still fi~lther,i n the following, a point on a gain \vavefornl expressed with
an encoded sample position and a gain value included in the gain code string will be
25 also simply referred to as a gain satnple position.
[0030]
In the example of FIG. 1 , infonnation of a gain sample position GI 1 ,and a
gain sample position GI2 is obtained through decoding ofthe gain code string.
[003 11
30 Here, a gain value at the k-th gain sample position within a fiame is set as
g[k], and a sample length (the nctntber of samples) in a sample axis direction fio111
the k-th gain sample position to the k+l-th gain sample position will be expressed as
TLkl.
[0032]
In this case, when it is assumed that a sample position of the k-th gain
5 sample position GI 1 is n = 0, the gain satllple position GI 1 is a point expressed with
a coordinate (0, g[k]), and the gain sainple position GI2 is a point expressed with a
coordinate (T[k], g[k+l]). Here, n is an index indicating the n-th sample position
from the head of the frame.
[0033]
10 Furthet; a gain waveform behveen the gain satnple position GI I and the
gain sample position G12 obtained through linear interpolation becomes a wavefonn
indicated with a straight line L11. That is, between the gain sample position G11
and the gain sample position G12, a gain value at each satnple position is obtained
through interpolation assunling that the gain value linearly changes.
--
15 [0034]
However, if the gain wavefonn is estitnated throogh linear interpolation, for
example, as indicated with a cutve C11 in FIG. 2, when a stnooth gain wavefom~ is
tried to be encoded, tlle number of points to be encoded in the gain waveform, that is,
the number of gain sample positions increases. It should be noted that FIG. 2
20 indicates a gain value on a vettical axis and a sarnple position within a fi.ame of an
audio signal on a horizontal axis.
[0035]
In this example, because the gain waveforrn indicated with the cutye C 11 is
a s~~loowtha vefornl, if the decoding side tries to reproduce the gain waveform with a
25 ccttain level of precision, it is necessary to encode gain values at a number of gain
sample positions. This will increase a code amount of a bit stream obtained by
multiplexing the gain code string and the signal code string, that is, increase a bit rate.
[0036]
Therefore, in the present technologp, in order to make it possible to obtain
30 sound with higher quality with a less code amount, non-linear interpolation is newly
pcrfornled as appropriate in addition to linear interpolation. That is, a gaiu
waveform is generated by performing interpolation processing using a more
appropriate method selected behw'een linear interpolation and non-linear interpolation.
It should be noted that non-linear interpolation can be, for example, interpolation
using a quadratic fi~nction or a cubic function.
5 [0037]
For example, when non-linear interpolation utilizing a cubic function is
performed, a waveform indicated with a curve C21 in FIG. 3 can be obtained as a
gain wavefonn behveen the gain sample position GI 1 and the gain sample position
GI2 illustrated in FIG. 1. It should be noted that FIG. 3 indicates a gain value on a
10 vel-tical axis and a sample position within a fi'ame of an audio sigt~alo n a horizontal
axis. Further, in FIG. 3, the same reference nu~neralsa s those in FIG. 1 are assigned
to portions corresponding to those in FIG. 1, and explanation thereof \\fill be omitted
as appropriate.
[0038]
15 In this example, the gain code string includes information indicating the
sample position, the gain value and a gain inclination value at the gain satnple
position G11, and information indicating the sample position, the gain value and a
gain inclination value at the gain sanlple position G12.
[0039]
20 Here, the gain inclination value is it~fonnation indicating inclination of the
original gain wavefonn at a gain satnple position. Hereinafter, a gain inclination
value at the k-th gain sample position will be expressed as s[k].
[0040]
In-FSCP. 3, an arrow Dl 1 indicates the gain inclination value s[k] at the gain
25 sample position GI 1, and an arro\v Dl2 indicates a gain inclination value s[k+l] at
the gain sample position G12.
[004 I ]
At the decoding side, the gain wavefornl between the gain sample position
GI1 and the gain sample position G12 is obtained through non-linear interpolation
30 utilizing a cubic filnction, and, as a result, the gain ~vavefortn indicated with the
curve C21 is obtained.
[0042]
The gain \\raveform indicated with the curve C21 is, for exatnple, a curve of
a cubic function \vhich passes through the gain sample position GI1 and the gain
sample position GI2 and whose inclination at the gain sample position GI 1 and the
5 gain sample position GI2 is respectively s[k] and s[k+l].
[0043]
In this mannet; by ntilizing non-linear interpolation as appropriate, even
when the gain waveform is a sn~oothw aveform, it is possible to reproduce the gain
wvaveform with high precision through encoding of less gain satnple positions, that is,
10 with a less code amount.
[0044]
In the present technology, for example, as a parameter for switching
between linear interpolation and non-linear interpolation, interpolation mode
information indicating an interpolation scheme using linear interpolation or an
- -
15 interpolation scheme using non-linear interpolation is included in the gain code string.
The decoding side switches behveen linear interpolation and non-linear interpolation
according to this interpolation mode inforination.
[0045]
Here, the interpolation mode information may be, for example, an index of
20 two bits for s~vitchinga mong linear interpolation, interpolation nsing a quadratic
function and interpolation using a cubic function, or may be a flag of one bit for
switching between linear interpolation and interpolation using a cubic function wvl~icl~
is non-linear interpolation. That is, any infortnation tnay be nsed as the
interpolation mode information if the information indicates a method for
25 interpolating a gain wavefoml.
[0046]
Fo~thet; in the present technology, when the interpolation mode information
is itlfor~uatio~ini dicating an interpolation scheme using non-lillear interpolation, in
addition to a gain value, a gain inclination value is included in the gain code string
30 for each gain sample position.
[0047]
Here, the gain inclination valtte s[k] indicates change ofa gain value per one
satnple. For example, the gain inclination value s[k] at the k-th gait1 satnple
position is inclination of a straight line which connects a point on the gain waveform
at the k-th gain sample position and a point on the gain waveforln at the next satnple
5 position of the k-th gain sample position. It should be noted that the gain
illclillation value may be obtained using any method if the gain inclination value
indicates inclination at the gain sample position on the gain waveform.
[0048]
It should be noted that the gain inclination value as is may be stored in the
10 gait1 code string, or a quantization value of the gain inclination value or an entropy
encoded value such as a Huffinat1 encoded value ofthe gain inclination value may be
stored in the gain code string.
[0049]

-
15 Further, specific examples of a method for performittg linear interpolatiotl
and a tnethod for performing non-linear interpolation on a gain value at each sample
position between two gain sample positions will be described. First, a method for
perfortning linear interpolation will be described.
[OOSO]
20 When linear interpolation is performed using the interpolation mode
infornlation, at the decoding side, a gain value is read out for each gain sample
position from the gain code string.
[OOS 11
Here:an index at the k-th gain sample position is set as k, and a gain value
25 at the k-th gain sample position read out fiom the gain code string is set as g[k].
Futthet; a sample length between the k-th gain sample position and the k+l-th gain
satnple position is set as T[k], and it is assumed that the satnple length T[k] is
included in the gain code string as information indicating the sample positiotl of the
k+l-th gain sa~nplep osition.
30 [0052]
It is no\\r assumed that the k-th gain sample position is a head position of a
frame, that is, the satnple position of n = 0. In such a case, a gain value
g-it~terpolated[n] of the sanlple n which is located between the k-th gain sample
positiotl aud the k+l-th gain sample position and which is the n-th (where 0 < n <
t[k]) sample from the head is calculated using the following equation (I).
5 [0053]
[Math. 11
g-i nterpo I ated [nl =a [kl x n+b [kl
(OIn
25 Subsequently, a case will be described where a gain value of the sample n
behveen the k-th gain sample position and the k+l-th gain sample position is
obtained through non-linear interpolation. Here, description will be continued with
a case where interpolation using a cubic function is perforrued as an example of nonlinear
intet.polation.
[OOSS]
When non-linear interpolation is performed according to the interpolation
mode information, at the decoding side, a gain value and a gain inclination value are
read out from the gain code string for each gain sample position.
5 [0059]
Here, in a similar manner to a case of linear interpolation, a gain value at the
k-tb gain satllple position is set as g[k], and a sample length between the k-th gain
satnple positiou and the k+l-th gain sample position is set as T[k]. Futthel; a gain
inclination value at the k-th gain satnple position is set as s[k].
10 [0060]
It is now assumed that the k-th gain sample position is a head position of the
fiatne, that is, a sample position of n = 0. In such a case, a gain value
g-interpolated[n] of the sample n which is located between the k-th gain sample
position and the k+l-tll gain sample positiotl and which is the n-th (where 0 i n <
- -
15 T[k]) satnple from the head is calculated using the following equation (4).
[0061]
[Math. 41
[0062]
20 It should be noted that, in eqltatio~i( 4), c[k], d[k], e[k] and flk] are values
respectively obtained using the following equation (5) to equation (8).
[0063] L *
[Math. 51
c [ k I = (1/T[kl) x I ( ~ [ k + l l + s [ k l ) / T [ k l
-2x ( g [ k + l l - g C k l ) / ( ~ [ k l ~ ) I - . (5)
25 [Math. 61
[Math. 71
e [kl =s [kl - - - ( 7 )
[Math. 81
f [kl =g[kl - - (8)
5 [0064]
In this example, as described with reference to FIG. 3, a gain value of each
sample n is obtained through non-linear it~terpolation, that is, interpolatiot~ using a
cubic function assuming that the gain value changes according to a cubic function
indicated in equation (4) between the k-th gain sample position and the k+l-th gain
10 sample position.
[0065]
As described above, by obtaining a gain value as appropriate through nonlinear
interpolation, a smooth gain waveform as illustrated in, for example, FIG. 2
can be encoded at a lower bit rate, so that it is possible to improve coding efficiency.
1 [0066]

Specific operation ofthe encoding apparatus 11 will be described next.
[0079]
When the input time series signal corresponding to one frarne is supplied,
the encoding apparatus 11 performs encoding processing of encoding the input time
15 series signal and outputting the output code string. Hereinafter, encoding
ptlocessing by the encoding apparatus I 1 will be described with reference to the
flowcha~ot f FIG. 5.
[OOSO]
In step SlI, the sound pressure level calculating circuit 21 calculates a
20 representative sound pressure level of the input time series signal based on the
supplied inpot time series signal and supplies the representative sound pressure level
to the gain calculating circuit 22.
[OOSI ]
Specifically, the sound pressure level calculating circuit 21 calculates sound
25 pressure levels of respective channels constituting the input time series signal and
sets a represetltative value of the sound pressure levels of these channels as a
representative sound pressure level.
[0082]
For example, in a method fo~.calculatinga sound pressure level, a maximum
30 value, a root Inean square (RMS), or the like, of fiatnes of an audio signal of
channels constituting the itlput time series signal is used, aud a sound pressure level
is obtained for each of the channels constituting the input time series sigual for the
frames of the input titne series signal.
[0083]
Fu~thet; as a method for calculating a representative value as the
5 representative soi~~lpdre ssure level, for example, a method in which a ~naximum
value among sound pressure levels of the channels in the same frame is set as the
representative value, a method in which one representative value is calculated using a
specific calculation formula from the sound pressure levels of the channels, or the
like, can be used. Specifically, for example, it is possible to calculate the
10 representative value using a loudness calculation formula described in ITU-R
BS. 1770-2(03/20 1 I).
[0084]
In step S12, the gain calculating circuit 22 calculates a gain value based on
the representative sound pressure level supplied from the sound pressure level
15 calculating circuit 21 and supplies tl~ega in value to the gain encoding circuit 23.
[OOSS]
For example, the gain calculating circuit 22 calculates the gain value
according to DRC characteristics designated by a higher-order control apparatus.
[O086]
20 The DRC characteristics designated by the higher-order control apparatus
can be DRC characteristics as illustrated in, for example, FIG. 6. It should be noted
that FIG. 6 indicates an input sound pressure level (dBFS), that is, a representative
sound pressure level on a horizontal axis, and indicates an output sound pressure
level (dBFS), that is, a corrected sound pressure level when the sound pressure level
25 (volome) of the input time series signal is corrected, on a vertical axis.
[0087]
A broken line L3 1 and a broken line L32 respectively indicate relationship
of the input and output sound pressure levels. For exau~plea, ccording to the DRC
characteristic indicated with the broken line L31, when there is an input of the
30 representative sout~tlp ressure level of 0 dBFS, volume is corrected so that the sound
pressure level of the inpot time series signal becomes -27 dBFS.
On the other hand, for example, according to the DRC characteristics
indicated with the broken line L32, when there is an input ofthe representative sound
pressure level of 0 dBFS, volu~ne is corrected so that the sound pressure level of the
5 inpot titne series signal becomes -21 dBFS.
[0089]
The gain calc~~latincgir cuit 22 deteunines a gain value according to the
DRC characteristics indicated with such a broken line L31 and a broken line L32.
This gain value is outputted as a gain waveform synchronized with the frame at the
10 signal encoding circuit 24. That is, the gain calculating circuit 22 calculates a gain
value for each of samples constituting a frame wvhich is to be processed of the input
time series signal.
[0090]
More specifically, for example, the gain calculating circuit 22 obtains a gain
--
15 waveforin g(J, n) in a frame J by perfortning calculation of the following equation (9).
[0091]
[Math. 91
g (J, n) =A x G t (J) + (1 -A) x g (J, n-1) (9)
100921
20 It should be noted that, in equation (9), n indicates a position of a satnple
which takes values from 0 to N-1 when a frame length is set as N, Gt(J) indicates the
above-described DRC characteristics, that is, a target gain in the frame J determined
by the input sound pressure level and the output sound pressure level.
100931
25 Further, A in equation (9) is a value determined by the follo\ving equation
(10).
[0094]
[Math. 101
A=l-exp(-1/(2xFs~Tc(J))) - - (10)
30 [0095]
In equation (lo), Fs indicates a sampling frequency (Hz), Tc(J) irtdicates a
time constant in the frame J, and exp(x) indicates an exponent fiinction. Further, in
equation (9), a gain value of the last sample in a frame imelediately before the frame
is used as a gain \vaveform g(J, n-I) wllen n = 0.
[0096]
5 Returning to explanation of the flowchart in FIG. 5, in step S13, the gain
encoding circuit 23 performs gain encoding processing to encode the gain value
supplied fro111 the gain calculating circuit 22. The gain encoding circuit 23 then
supplies the gain code string obtained through the gain encodi~lg processing to the
multiplexing circuit 25.
10 [0097]
For example, the gain encoding circuit 23 extracts a gain sample position to
be encoded from the gain value at each sample position supplied from the gain
calculating circuit 22, that is, a gain waveform of the frame to be processed. For
example, characteristic samples such as inflection points it1 the gain \vavefortn may
-
15 be used as the gain sample positions, or samples ar~angeda t predetermined intervals
may be used as the gain sample positions.
[0098]
The gait1 encoding circuit 23 generates interpolation mode infonnation and
gain information for each of the gain sample positions extracted in this manner.
20 [0099]
For example, the gain encoding circuit 23 generates the interpolation mode
i~lforlnatiotbl y performing so-called local decoding.
[O 1 001
That is, the gain encoding circuit 23 generates a gain waveform between
25 hvo gain sartlple positions adjacent to each other through itlterpolation for linear
interpolation and non-linear interpolation, and calculates a difference between the
gain waveform and an actual gain wavefonn. The gain encoding circuit 23 then
generates inforillation indicatiug an interpolation scheme in which the obtained
difference is smaller as the interpolation mode information.
30 [OlOl]
It should be noted that whether linear interpolation is perfomled or nonlinear
interpolation is performed tnay be detertnined using any other mcthod. For
example, it is also possible to detenninc that linear interpolation is performed when
the gain value is the same behveen the gain sample position to be processed and the
gain sa~nplep osition immediately before the gain sample position to be processed,
5 and the gain inclination value of the gain sample position immediately before the
gain sample position to be processed is 0, and, that non-linear interpolation is
performed in other cases. Alternativel); it is also possible to einploy a
configitration where a higher-order control appal-atus designates linear interpolation
or non-linear interpolation.
10 [0102]
Further, the gain encoditig circuit 23 encodes a satnple length T[k], a gain
value g[k] and a gain inclination value s[k] indicating the sample position as
appropriate for each gain sample position to obtain gain information. It should be
noted that, when the interpolation [node information is information indicating an
- -
15 interpolation scheme using linear interpolation, gain info~~natioonn ly including the
sample length and the gain value and not including the gain inclination value is
generated.
[O 1031
The gain encoding circuit 23 supplies the gain code string including the gain
20 iufonnation of each gain sample position and the interpolation mode information
obtained in this mallnet to the ~nultiplexingc ircuit 25.
[O 1041
In step S14, the signal encoding circuit 24 encodes the supplied input time
series signal according to a predetermined encoding scheme atld supplies the signal
25 code string obtained as a result of encoding to the multiplexing circuit 25.
[0 1 051
In step S15, the multiplexing circuit 25 moltiplexes the gain code string
supplied from the gaiu encoding circuit 23 and the sigrlal code string supplied from
the signal encoding circuit 24 and ontputs the output code string obtained as a result
30 of tnultiplexing. When the output code string corresponding to one frame is
outputted as a bit stream in this manner, encoding processing ends. Then, encoding
processing of the next kame is performed.
[O 1 061
As described above, the encoding apparatus 11 obtains the gain value for
each sample within a fiatne of the input time series signal to extract the gain sample
5 position and generates the gain code string cotlstituted with gain information of each
gain sample position and interpolation [node information.
[0107]
By the gain value for each sample within a frame being determined in this
manner, at the decoding side, temporal waveforms behveen frames of the audio
10 signal are smoothly connected, so that it is possible to obtain sound with higher
qoality. Moreovet; by the interpolation mode information being iticluded in the
gain code string, it is possible to reproduce a gain waveform wit11 high precision with
a less code amount by utilizing non-linear interpolation as appropriate.
[0 1081
15
A decoding apparatus which receives the output code string outputted from
the encoding apparatus I I as at1 input code string and decodes the input code string
will be described next.
[0109]
20 FIG. 7 is a diagram illustrating a configuration example of an etnbodiment
of the decoding apparatus to which the present technology is applied.
[OIIO]
The decoding apparatus 51 illastrated in FIG. 7 has a demultiplexing circuit
61, a signal decoding circuit 62, a gain decoding circuit 63 and a gain applying
25 circuit 64.
[Olll]
The demultiplexing circuit 61 demultiplexes the supplied input code string,
that is, the output code string received from the encoding apparatus 11 and supplies
the signal code string obtained as a result of de~noltiplexing to the signal decoding
30 circuit 62, while supplying the gain code string to the gain decoding circuit 63.
[0112]
The signal decoding circuit 62 decodes the signal code string supplied from
the demultiplexing circuit 61 and supplies a time series sigtlal obtained as a result of
decoding to the gait1 applying circuit 64. Here, the ti~ne series sigtlal is, for
example, an audio signal of 11.1 ch or 7.1 ch, and an audio signal of channels
5 cotlstituting the time series signal is set as a pulse code modulation (PCM) signal.
[0113]
The gain decoding circuit 63 decodes the gaiu code string supplied froin the
demultiplexitig circuit 61 and supplies a gain value obtained as a result of decoding
to the gain applying circuit 64. The gain decoding circuit 63 has an interpolation
10 processing unit 71, which calculates a gain value at each satnple position of the time
series sigual through linear interpolation or non-linear interpolation based on the gain
information and the interpolatiotl mode infortnation obtained from the gain code
string.
[OI 141
15 The gain applying circuit 64 corrects volurne of the time series signal by
adjusting a gain of the time series signal supplied from the sigtial decoding circuit 62
based on the gain value supplied from the gain decoding circuit 63 and outputs an
output time series signal obtained as a result of volume correction.
[0115]
20
Subsequently, operation of the decoding apparatus 51 will be described.
[OI 161
When the input code string corresponding to one frame is supplied, the
decoding apparatus 51 performs decoding processing of decoding the input code
25 string aud outputting all output time series signal. The decodiug processing by the
decoding apparatus 51 will be described below \vith reference to the flowchart of FIG.
8.
[0117]
In step S41, the de~nultiplexi~lcgir cuit 61 receives the input code string
30 transmitted from the encoding apparatus 11 and demultiplexes ttie input code string,
and supplies a signal code string obtained as a result of demoltiplexiog to the signal
decoding circuit 62, while sopplying the gain code string to the gain decoding circuit
63.
[0118]
In step S42, the signal decoding circuit 62 decodes the signal code string
5 supplied fiom the demultiplexing circnit 61 and supplies a time series signal obtained
as a result of decoding to the gain applying circnit 64.
[0119]
111 step S43, the gain decoding circuit 63 perforn~sg ain decoding processing
to decode the gain code string supplied from the demultiplexing circuit 61 and
10 supplies a gain value at each sample position of a fiame to be processed obtained as a
result of decoding to the gain applying circuit 64. It should be noted that details of
the gain decoding processing will be described later.
[O 1201
In step S44, the gain applying circuit 6h adjusts a gain of the time series
15 signal supplied from thesignal decoding circuit 62 based on the gain value supplied
from the gain decoding circuit 63 and outputs the obtained output time series signal.
That is, each sample of the titne series signal is multiplied by the gain value to be
made an output time series signal with appropriate volume.
[0121]
20 When the output time series signal is outputted, the decoding processing
ends.
[O 1221
As described above, the decoding apparatns 51 decodes the gain code string,
and applies the obtained gain value at each sample position to the time series signal
25 to adjust a gain (volume) in a tirne domain. By adjusting a gain with a gait1 value
detennined for each sample position in this tnannet; it is possible to smoothly
connect tirne wavefonus between frames of the output time series signal, so that it is
possible to obtain sound with higher quality.
[0 1231
30 Moreovel; because a gain wavefol.m is obtained by utilizing non-linear
interpolation as appropriate, even tvhen the gain wa\refonn is a smooth waveform, it
is possible to reproduce the gain wvavefortn with high precision with a less code
aulount.
[0124]

5 Farthet; gain decoding processing corresponding to processing in step S43
of FIG, 8 will be described with reference to the flowvchatt of FIG. 9.
[0125]
In step S7l, the gain decoding circuit 63 reads out gain infonnation at a gain
sample position to be processed from the gait1 code string supplied fro111 the
10 demultiplexing circuit 61 and decodes a sample length T[k], a gain value g[k] and a
gain inclination value s[k] included as the gain infomiation as necessary. It should
be noted that when an interpolation scheme indicated by the interpolation mode
information is an interpolation scheme using linear interpolation, the gain inclination
value is not included in the gain information.
- -
15 [0126]
For exatnple, it1 the gain code string, gain information and interpolation
mode information at each gain satnple position are stored while being arranged in
ascending order of a distance from the head of the frame. Because the gain
decoding circuit 63 seq~tentiallyr eads out the gain infonnation and the interpolation
20 mode information from tlie gain code string, a gain satnple position is set as a gain
sample position to be processed in ascending order of a distance frotn the head of the
frame.
[0 1271
In steps67-2, the gain decoditig circuit 63 reads out the interpolation inode
25 information at the gain satuple position to be processed frotn the gain code string.
[0 1281
It should be noted that while an exatnple where the itlterpolation mode
illformation is included in the gain code string will be described here, tlie
interpolation mode itiformation may be included in a header, or tlie like, of a bit
30 stream it1 which ail itlput code string of each frame is included, or the interpolation
mode infortnation may be acquired fiotn a higher-order control apparatus, or the like.
[O 1291
In step S73, the interpolation processing unit 71 determines whether or not
the interpolation schelne indicated by the read-out interpolation mode information is
a scheme using linear interpolation.
5 [0130]
In step S73, when it is determined that the interpolation scheme is a scheme
using linear interpolation, in step S74, the interpolation processing unit 71 performs
linear interpolation to generate a gain wavefortn.
[0131]
10 Specifically, the interpolation processing unit 71 performs the same
calculation as that of the above-described equation ( I ) based on the gain value g[k]
and the sample length T[k-1] at the gain sample position to be processed, and a gain
value and a sample position at a gain sample position one position closer to the head
of the frame from the gain sample position to be processed to generate a gain
-
15 wvaveforn~b ehveen the gain sample positions. That is, a gain value at each sample
position located behveen hvo gain sample positions of the time series signal is
calculated, and a wavefonn constituted with gain values at the sample positions is set
as a gain waveform.
[0 1321
20 When the gain waveform behveen two adjacent gain sample positons is
obtained in this manner, the processing proceeds to step S76.
[O 1331
On the other hand, when it is determined in step 573 that the scheme is a
scheme which does not use linear interpolation, that is, a scheme using non-linear
25 interpolation, in step S75, the interpolation processing nnit 71 perfonns non-linear
interpolation to generate a gain wavefonn.
[O 1 341
Specifically, the interpolation processing unit 71 perfonns the same
calculation as that of the above-described eqaation (4) based on the gain value g[k],
30 the satnple length T[k-I] and the gain inclination value s[k] at the gain sample
position to be processed and a gain value, a sa~nple position and a gain inclination
value at a gain satnple position one position closer to the head of the frame fro111 the
gain sample position to be processed to generate a gain waveform between the gain
satnple positions. That is, a gain value at each sample position located between two
gain sa~nple positions of the time series signal is calcolated, and a waveform
5 co~lstitntedw ith gain values ofthe sample positions is set as a gain wavefornl.
[O 1351
Wheu the gain wavefor~n between two adjacent gain sample positions is
obtained in this manner, the processing proceeds to step S76.
[0136]
10 When the gait1 waveforn~ between the gait1 sample positons is obtained
through interpolation in step S74 or step S75, in step S76, the gain decoding circuit
63 determines whether or not processing is performed for all the gain satnple
positions.
[0137]
- -
15 When it is determined in step S76 that not all of the gain sample positions
are processed yet, the processing returns to step S71, and the above-described
processing is repeated. That is, the next gain sarnple position is selected as a
processing target, and a gait1 waveform is obtained through interpolation.
[0138]
20 011 the other hand, when it is detennined in step S76 that all of the gain
sample positions are processed, the gain decoding circuit 63 sopplies a gain
\vaveforrn corresponding to one frame constituted wit11 gain values at the sample
positions obtained through the processing so far to the gain applying circuit 64, and
the gain decoding processing ends. When the gain decoding processing ends, then,
25 the processing proceeds to step S44 in FIG. 8.
[0139]
The decoding apparatus 51 obtains a gain waveform through linear
interpolation or non-linear interpolation according to the interpolation mode
information as described above. By obtaining a gain waveform thro~~gnho n-linear
30 interpolation as appropriate according to the interpolation mode information in this
manna; it is possible to reproduce a gain waveform with high precision with a less
code amount.
[0 1 401
It should be noted that \\~liilea n example has been described above where
the interpolation mode information is generated for each gain sample position, and
5 the interpolation scheme is switched between linear interpolation and non-linear
interpolation, one piece of the interpolatioti mode information may be generated for
each frame. In this case, the interpolation scheme is switched between linear
interpolation and non-linear interpolation in units of a fianie.
[0141]
10 FUI-thel; the interpolation schetne may be switched between linear
interpolation and non-linear interpolation in units of a plurality of frames or in units
of a file. For example, when the interpolation scheme is switched in units of a file,
for example, one piece of interpolation mode information is stored in a header of the
bit stream. The interpolation processing unit 71 performs interpolation processing
--
15 of each kame using an interpolation scheme indicated by the interpolation mode
infonnation, that is, either the scheme using linear interpolation or the scheme using
non-linear interpolation to obtaiti a gain waveform corresponding to one file.
[0 1421

20
By the way, the gain waveform obtained through non-linear interpolatioti is
different from tlie gaiu ~vaveform obtained through linear interpolation, and there is a
case where a gain value at a sample position behveen two gain sample positions Itlay
be greater or smaller than gain values at hvo gain sa~ttplep ositions included in the
25 gain code string.
[0 1431
For example, in the example illustrated in FIG. 3, in part of the gain
\vavefonn indicated with the curve C21, \\~hicli is obtained through non-linear
interpolation, there is a portion where the gain value becomes sntaller than tlie gain
30 value g[k] at the gain sample position GI 1. Further, in part of the gain waveform
indicated with the curve C21, there is also a portion \\there the gain value becottles
greater than the gain value g[k+l] at the gain saulple position (312.
[0 1441
Therefore, there is a case where the gain value obtained through non-linear
interpolation becomes a negative (minus) value which is inappropriate as the gain
5 value. Therefore, in order to prevent the gain value obtained through interpolation
fro111 becotnitlg an inappropriate value, it is also possible to perfo1.m limiting on the
gain value using zero as a lower limit by performing calculation of the follo\ving
equation (1 1).
[0145]
10 [Math. 111
g-interpol ated En1 =max (0, g-i nterpo I ated [nl) - - - (11)
[0 1461
In equation (ll), between the gain value g_interpolated[tl] obtained through
interpolation and zero, a greater one is made a final gain value g-interoplated[n].
If, Accordingly, the final gain value is equal to or greater than zero, and the gain value
does not become a negative value.
[O 1471
Fullher, there is a case where it is desired to boost (amplify) the time series
signal and a case where it is desired to compress (suppress) the time series signal
20 through gain adjustment (volume correction).
[0148]
For example, when it is desired to boost the time series signal, if the gain
value is smaller than one, the gain value becomes an inappropriate value. Therefore,
when the time series signal is boosted, it is also possible to perform limiting on the
25 gain value using one as a lower litnit by perfomling calculation of the following
equation (12).
[0 1491
[Math. 121
g-interpoIated[nl =inax (1, g-i nterpoIated[nl) - - - (12)
30 [0150]
In equation (12), beheen the gain value g_interpolated[n] obtaiued through
interpolation and one, a greater one is made a final gain value g_interpolated[n].
Accordingly, the gain value does not become a value less than one. In other words,
the gain value is always equal to or greater than one which is the lower limit.
[0151]
5 Fmther, for example, when it is desired to compress the time series signal, if
the gain value is greater than one, the gain value becomes an inappropriate value.
Therefore, when the time series signal is compressed, it is also possible to perform
limiting on the gain value using one as an upper limit by performing calculation of
the following equation (13).
10 [0152]
[Math. 131
g-interpoIated[nl =mi n(1, g-i nterpoIated[nl) - - (13)
[0 1531
In equation (13), between the gain value g_interpolated[n] obtained through
~-
15 interpolation and one, a smaller one is made a final gain value g_interpolated[n].
Accordingly, the gain value does not become a value greater than one. In other
words, the gain value is always equal to or smaller that^ one which is the upper limit.
[0154]
When limiting processing as indicated in equation (12) or equation (13) is
20 perfonned, it is only necessary to provide limiting information indicating wlietlier the
gain wavefonn is used for boosting or used for compression to the gain decoding
circuit 63 as informati011 regarding the encoded gain wavefonn. For example, the
limiting information may be supplied from a higher-order control apparatus to the
gaiu decoding-oimuit 63, or the limiting information may be included in the gait1
25 eode string, tlie header of the bit stream, or the like.
[0155]
In the following, description will be continued assuming that tlie limitit~g
information is included in the gain code string. 111 this case, in the processing of
step S13 in FIG. 15, the gain eode string including the limiting information is
30 generated.
[0156]
By perfomling the limiting processing on the gain value as described above,
it is possible to obtain a more appropriate gain value. By this means, it is possible
to perfonn more appropriate gain adjustment (volume control), and, as a result, it is
possible to obtain sout~dw ith higher quality.
5 [0157j
iconfiguration Exatnple of Decoding Apparatus>
When limiting processing is perfonned on the gait1 valoe, the decoding
apparatus 51 is configured as illustrated in, for example, FIG. 10. It should be noted
that, in FIG. 10, the same reference uutnerals as those it1 FIG. 7 are assigned to
10 portions corresponditlg to those in FIG. 7, and explanation thereof will be omitted as
appropriate.
[0158]
The decoding apparatus 51 illustrated in FIG. LO has a different
configuration from that of the decoding apparatus 51 in FIG. 7 in that a litniting
p~
~~-p
15 processing unit 101 is newly provided at the gain decoding circuit 63, and has the
same configuration as that of the decoding apparatus 5 1 in FIG. 7 in other points.
[0 1591
The linliting processing unit 101 performs limiting processing on the gain
value calculated through non-linear interpolation perfonned by the interpolation
20 processing unit 71 to obtain a final gain value.
[0 1601

Gain decoding processing performed in the case where the decoding
apparatus 51 has the configuration illustrated in FIG. 10 will be described next.
25 [Ol6lj
For example, at the decoding apparatus 51, the decoding processing
described with reference to FIG. 8 is perfonned. Ho\vevel; it1 gain decoding
processing corresponding to step S43, the gain decoding processing illustrated in FIG.
I I is performed. The gain decoding processing by the decoding apparatus 51 in
30 FIG. 10 \\rill be described below with refereuce to the flowchart of FIG. 11.
[0 1621
It shoold be noted that the processing from step S l Ol to step S 105 is the
same as processitig fron~s tep S71 to step S75 it1 FIG. 9, explanation thereof will be
omitted.
[0 1631
5 In step S106, the limiting processing unit 101 changes the gain value as
appropriate so that the gain value does not become a negative value by performing
calculatio~i of the above-described equation (11) on the gain value at each sample
position obtained through processing in step S105.
[0164]
10 Furthei; the limiting processing unit 101 obtains a final gain value by fi1121ier
perfornliag calculatiotl of either equation (12) or equation (13) on the gain value
limited through calculation of equation (11) according to the limiting information
included in the gain code string.
[0 1651
-
15 Specifically, when the limiting information included in the gain code string
indicates that the gain wavefonn is to be used for boosting, the litniting ptocessing
unit 101 performs calculation of equation (12) so that the gain value does not become
a value less than one.
[0 1661
20 On the other hand, when the limiting information included in the gain code
string indicates that the gain wavefonn is to be used for compression, the litniting
processing unit 101 performs calculation of equatio~l (13) so that the gain value does
not becotne a value greater than one.
[O167] ..-
26 When the gain wavefornl is generated through linear interpolation in step
S104 or limiting processing is performed in step S106, the processing in step S107 is
performed, atid the gain decoding processing ends. Because the processing in step
S 107 is tlie same as tlie processing in step S76 in FIG. 9, explanation thereof will be
omitted.
30 [0168]
As described above, the decoding apparatus 5 1 performs limiting processing
on the gain value obtained through non-linear interpolation. By this means, it is
possible to perform gain adjustment (volume correction) with a more appropriate
gain value. It is therefore possible to obtain sound with higher quality.
[O 1691
5 e h i r d Embodiment>

Fu~thet; while, in the above description, an example has been described
where the gain waveform is obtained while the interpolation scheme for interpolating
the gain value is switched between linear itlterpolation and non-linear interpolation
10 for each gain sample position, it is also possible to employ a configuration where
non-linear interpolation is basically perfonned, and linear interpolation is perfonned
only under specific conditions.
[0170]
For example, a case will be studied where the gain \\laveform indicated with
15 a broken line LA1 illustrated in FIG. 12 is encoded, and the gain waveform is
obtained at the decoding side through non-linear interpolation. It should be noted
that FIG. 12 indicates a gain value on a vertical axis and a satnple position on a
horizontal axis.
[0171]
20 It is assumed that, at the encoding apparatus 11, the k-th gain sample
position G21 and the k+l-th gain satnple position G22 are extracted, and the gain
code string including gain values, sainple lengths and gain iilclination values at these
gain sample positions is obtained.
[O 1721
25 Here, an arrow 021 indicates a gain inclination value s[k] at the gain sample
position G21, and an arrow D22 indicates a gain inclination value s[k+l] at the gain
sample position G22.
[0 1731
It is now assulned that noti-linear interpolation using a cubic function is
30 performed at the decoding apparatus 51 based on the gain values, the sample lengths
and tile gain inclination values included in the gain code string, and a gain wa\~eform
indicated with a curve C31 is obtained.
[0 1741
In this example, a difference behveen the gain wavefonn indicated with tlie
curve C31 obtained through non-linear interpolation and the gain waveform
5 indicated with a broken line L41 becomes large.
[O 1751
In a scl~emefo r obtaining a gain waveform through non-linear interpolation,
wlien a gain wavefortii whose gain value linearly changes is encoded as wit11 this
example, a difference between the original gain waveform and a gain wavefonn
10 obtained through non-linear interpolation upon decoding becomes large.
[0 1 761
To make this difference small, it is necessary to perform processing (local
decoding) of adjusting a gain value and a gain inclinatioti value to be encoded at the
encoding apparatus I I by calculating the gain waveform obtained tlirough non-linear
--
15 interpolation, which increases a processing amount of encoding.
[O 1771
Therefore, in tlie present technology, when non-linear interpolation is
performed at the decoding apparatus 51, by allowing linear interpolation to be
performed ortder specific cotlditions, a gain waveform is reproduced with high
20 precision with a less processing amount of encoding.
[0178]
Specifically, wlien, for example, a gaiii value at a sample position between
the k-th gain sample position and the k+l-tli gain sample position is obtained through
interpolation, an intersection X[k, k+l] of two straight line l[k] and straight line
25 l[k+l] is obtained from gain values and gain inclination values at these gain sample
positions.
[0179]
Were, tlie straight line l[k] is a straight line which passes through tlie k-tli
gain sample position (point) on the gain waveform atid which has inclination
30 indicated with the gain incli~iation value s[k]. That is, he hen a value of tlie
coordinate in the sample axis direction is the same as a value of the k-tli gain sample
position, the straight line l[k] is a straight line which has a gain value g[k] at the k-th
gain sample position as the value o f the coordinate in the gain axis direction and
which has inclination indicated with the gain inclination value s[k].
[O 1 801
5 In a similar mannet; the straight line l[k+l] is a straight line whiel~ passes
through the k+l-t11 gain sample position and which has inclination indicated with the
gain inclination value s[k+l].
[0181]
Further, it is determined whether or not a distance between either the k-th
10 gain satnple position or the k+l-th gain satnple position and the obtained intersection
X[k, k+l] is equal to or less than a predetermined threshold. 111 the determination
here, it is determined whether, for exatnple, the following equation (14) holds true.
[0 1821
[Math. 141
-
((d-samp I e [ k l I thre-samp I e) && (d-ga i nCk1 I thre-ga i n))
[ I ((d-sampIeCk+ll~thre_sampIe)
&&(d&gai n[k+lI Ithre-gain)) . . (14)
15
[0183]
It shonld be noted that, in equation (14), d-sample[k] and d_sample[k+l]
respectively indicate distances from the k-th gain sample position and the k+l-th
gain sample position to the intersection X[k, k+l] in the satnple axis direction.
20 Further, d_gain[k] and d_gain[k+l] respectively indicate distances fiotn the k-th gain
sample position.and.the k+l-th gain sample position to the intersection X[k, k+l] in
the gain axis direction, that is, differences of the gain values.
[0184]
Further, thre-sample and thre_gain respectively indicate a threshold of a
25 distance in the sample axis direction and a threshold of a distance in the gait1 axis
direction.
[0 1851
Therefore, in equation (14), when the distance d-sample[k] is equal to or
less than tl~re~samplaen,d the distance d-gain[k] is equal to or less than tl~te~gai0n1. ,
when the dista~lce dsample[k+l] is equal to or less than thre-sample and the
distance d-gain[k+l] is equal to or less than the threshold thre-gain, a distance fiom
the gain sample position to the intersection X[k, kt11 is equal to or less tllan a
5 threshold.
[O 1 861
For example, when the k-th gain sample position is a head position of the
frame, that is, a sample position of n = 0, the distance d_sample[k], the distance
d_gain[k], the distance d_sample[k+l] and the distance d_gain[k+l] in equation (14)
10 are respectively obtained using the following equation (15) to equation (18).
Furthe]; the threshold thre-sample and the threshold thre-gain are, for example, the
threshold thresample = 32 and the thresl~oldt hresain = 0.01.
[O 1871
[Math. IS]
d~sample~kl=abs((g[k~ll-g[kl-s[k+llxT[kl)
15 /(s[kI-s[k+lI)) . (15)
[Math. 161
d-gain[kl=abs(s[kl x (g[k+ll-g[kl-s[k+llxT[kl)
/ ( ~ [ k l- s[k+lI)) - - - (16)
[Math. 171
d-sample[k+ll =abs((g[k+ll-g[kl -s[k+ll x T[kl)
/ ( ~ [ k l- s[k+ll) -T[kl) - = . (17)
20 [Math. 181
d-gain[k+13 =abs (s[kl x (g[k+l] -g[k] -s[k+l] x T[k])
/(s[kI -s[k+ll) +g[kI--g[k+ll) . . . (18)
[O 1881
It should be noted that, in equation (15) to equatio~i (18), abs(x) indicates
that an absolute value of x is obtained.
25 [0189]
When it is determined that such a conditional expression indicated \\lit11
equation (14) holds true, a gain wa\'efonll is obtained through linear interpolation,
that is, tl~rot~gcahl culation of the above-described equation (I). On the other hand,
when the conditional expression indicated with equation (14) does not hold true, a
gain wavefonn is obtained through non-linear interpolation, that is, through
calculation of the above-described equation (4).
5 [O19O]
For example, as illustrated in FIG. 13, when a gain value at each satnple
position between the gain sample position G3 1 and the gain sa~nple position G32 is
obtained through interpolation, whether or not the conditional expression indicated
with equation (14) holds true is detennined by specifying which of a region TRI I
10 and a region TR12 the intersection CPll is included in. It should be noted that FIG.
13 indicates a gain value on a vertical axis and a sample position within a fiatne of
the time series signal on a horizontal axis.
[O191]
In FIG. 13, the gain sample position G31 indicates the k-t11 gain sample
15 position, and the arrow D31 itldicates the gain inclination value s[k] at the gait1
sample position G31. Thet.efore, the straight line L51 is a straight line l[k].
[0 1 921
In a similar manner, the gain sample position G32 indicates the k+l-th gain
sample position, and the arrow D32 indicates the gain inclination value s[k+l] at the
20 gain sample position G32. Therefore, the straight line L52 is a straight line l[k+l].
The intetsection CPl I which is an intersection of the straight line L51 and the
straight line L52 is an intersection X[k, k+l].
[0 1931
It isonow assutned that the region TRI 1 has the gain satnple position G31 in
25 the center and has a length in a vertical direction of 2 x thre-gain, and a length in a
horizontal direction of 2 x thre-sample in the drawing. In a similar mannet; it is
assutned that the region TR12 has the gain sample position G21 in the center and,
has a length in a vettical direction of 2 x thre-gain and a length in a horizontal
direction of 2 x thre_satnple in the drawing.
30 [O194]
In this case, when tl~ein tersection CPll is located \vitt~int he region TR1 1 or
the intersection CPll is located within the region TR12, the conditional expression
i~tdicated with equation (14) holds true. In the example of FIG. 13, because the
intersection CPll is located within the region TR12, the conditional expression
indicated with equation (14) holds true.
5 [0195]
In the example illustrated in FIG. 13, the origiual gain wavefortn to be
reproduced (restored) should have been a wavefornl close to a waveform constituted
with the straight line L51 and the straight line L52. That is, in more detail, the
wavefonn should have been close to the straight line L5 1 fro111 the gain sample
10 position G31 to the iute~section CPl1, and should have been close to the straight line
L52 from the intersection CPll to the gait1 sample position (332.
[0196]
Ho\vevel; because, in this example, the intersection CPI I is located within
the region TR12, and a distance between the intersection CPl l to the gain sample
15 position G32 is sufficiently short, it is possible to detennine that the original gain
waveform is approxitnated as a straigl~tli ne connecting the gain sample position G3 I
and the gain sample position G32.
[0197]
In this case, because at the gain waveform between the gain satnple position
20 G31 and the gain satnple position G32, the gain value can sobstatltially linearly
change, it is possible to reproduce the gain waveform with higher precision by
obtaining the gain waveform through linear interpolation rather than obtaining the
gain \\raveform through non-linear interpolation. Therefore, in the present
tecl~nology;~whet~h~e conditional expression indicated with the above-described
25 equation (14) holds true, the gain waveform is obtained through linear interpolation.
[0198]
Accordingly, in the example of FIG. 13, a gain value at each sample positiorl
between the gain sample position G31 and the gain sample position G32 is obtained
through linear interpolation, at~d, by this means, for example, a gain \\ravefonn
30 illustrated in FIG. 14 can he obtained. It should be tloted that, in FIG. 14, the satne
reference numerals as those in FIG. 13 are assigned to portions corresponding to
those in FIG. 13, and explanation thereof will be omitted as appropriate.
[0 1991
In FIG. 14, a straight line L61 connecting the gain sample position G31 and
the gain sample position G32 is obtained as a gain waveform between the gain
6 sa~nplep osition G31 and the gain sample position G32.
[0200]
Frnthel; for example, also in the above-described exa~npleil lustrated in FIG.
12, because the conditional expression indicated with equation (14) holds true, the
gain \vavefonn is obtained through linear interpolation.
10 [0201]
In the example of FIG. 12, becanse the intersection X[k, k+l] is located at
the gain sample position G22, equation (14) holds true, and a straight line connecting
the gain sample position G21 and the gain sarnple position G22 is set as a gain
wavefonn beheen the gain sample positions. Therefore, in this example, the
-
15 original gain wavefonn is accurately reproduced.
[0202]
As described above, when non-linear interpolation is basically performed
while linear interpolation is perfol~ned under specific conditions, it is possible to
make a difference between the original gain wavefonn and the decoded gain
20 waveform smaller without increasing a processing amount of encoding.
[0203]
Moreover, by employing such a decoding scheme, because both linear
interpolation and non-linear interpolation can be performed only with a scheme in
which non-linear interpolation is perfonned, it becomes unnecessary to include the
25 interpolation mode informatiotl in the gain code string, so that it is possible to lower
a bit rate of the output code string. That is, it is possible to reduce a code amount of
the output code string.
[0204]

30 When linear interpolation is performed under specific conditions, the
decoding apparatus 51 is configored as illustrated in, for example, FIG. 15. It
should be noted that, in FIG. 15, the same reference numerals as those in FIG. 7 are
assigned to portions corresponding to those in FIG. 7, and explanation thereof will be
omitted as appropriate.
[0205]
5 The decoding apparatus 51 illustrated in FIG. 15 has a different
configuration from that of the decoding apparatus 51 in FIG. 7 in that an operation
unit 131 is newly provided at the gain decoding circuit 63 and has the same
configuration as that ofthe decoding apparatus 51 in FIG. 7 in other points.
[0206]
10 The operation unit 131 performs calculation of the above-described
conditional expression indicated with equation (14).
[0207]

The gain decoding processing performed when the decoding apparatus 51 is
-
15 configured as illustrated in FIG. 15 will be described next.
[0208]
For example, while, at the encoding apparatus 11, the encoding processing
described with reference to FIG. 5 is performed, in the gain encoding processing in
step S13, the gain code string only including the gain infor~nation and not including
20 the interpolation mode infornlation is generated, and the ontpot code string obtained
through ~nultiplexingis outputted. Further, in this case, the gain information always
includes the gain inclination value.
[0209]
At the decoding apparatus 51, the decoding processing described with
25 reference to FIG. 8 is perfomled. However, in the gain decoding processing
corresponding to step S43, the gain decoding processing illustrated in FIG. I6 is
performed. The gain decoding processing by the decoding apparatus 51 in FIG. 15
will be described below with reference to the flowchart of FIG. 16.
[02 101
30 It should be noted that because processing in step S131 is the same as the
processing in step S71 in FIG. 9, explanation thereofwill be o~nitted.
[021 I]
In step S132, the operation unit 131 calculates the conditional expression
indicated with equation (I 4) based on the read-out gain information.
[02 121
5 That is, the operatio11 unit 13 1 performs the same calculation as the abovedescribed
equation (15) to equation (18) based on the gain value, the sample length
and the gain inclination value at the gain sa~nple position read out as the gain
information. The operation unit 131 then perfonns calculation of equatiotl (14)
based on a distance from the gain sa~nple position obtained as a result of the
10 calculation to the intersection X[k, k+l].
[02 1 31
This calculation of equation (15) to equatiotl (18) is equivalent to obtaining
the straight line l[k] and the straight line l[k+l] and obtaining the intersection X[k,
k+l] of these straight lines and, further, obtaining differences between the gain
15 values at the k-th gain sample position and the k+l-th gain sample position and the
gain value at the intersection X[k, k+l]. Further, calculation of the conditional
expression of equation (14) is equivalent to determining whether or not the
differences between the gain values at the gain sample positions and the gain value at
the intersection X[k, k+l] are equal to or less than a predeterniined threshold.
20 [0214]
Therefot~a, t the decoding apparatus 51, it is possible to obtain a gain value
at each sample position between the two gain sample positions through linear
interpolation or non-linear interpolation according to the differences between the
gain values..atthe-gaiwsample positions and the gain value at the i~ltersectionX [k,
25 k+l].
[02 151
In step S133, the interpolation processing unit 71 determines whether or not
linear interpolation is perfonned based on a calculation result of the conditional
expression in step S132. For example, when the conditional expression indicated
30 with equation (14) holds true, it is determined that linear interpolation is performed.
[02 161
When it is determined in step S133 that linear interpolation is performed, in
step S134, the interpolation processing unit 71 perforn~s linear interpolation to
generate a gain waveform, and, then, the processing proceeds to step S136. In step
S 134, the same processing as the processing in step S74 in FIG. 9 is perfor~ned.
5 [0217]
On the other hand, when it is determined in step S133 that linear
interpolation is not perfonned, in step S135, the interpolation processing unit 71
performs non-linear interpolation to generate a gain waveform, and, then, the
processing proceeds to step S136. It should be noted that, in step S135, the same
10 processing as the processing in step S75 in FIG. 9 is performed.
[02 181
When the gait1 waveforln is generated in step S134 or step S135, the
processing in step S136 is perfonned, and the gain decoding processing ends.
Because the processing in step S136 is the same as the processing in step S76 in FIG.
-
15 9, explanation thereof will be omitted.
[02 191
As described above, the decoding apparatus 51 generates a gain waveforin
through linear interpolation under specific conditions. By this means, it is possible
to obtain the original gain waveform with higher precision with a less processing
20 amount and it is possible to reduce a code amount of the output code string.
[0220]

It should be noted that, while, in the third embodiment, a case has been
25 described where linear interpolation is performed under specific conditions, it is also
possible to perform linear interpolation on the gain value by utilizing the gait1 sample
positions and the intersection.
[0221]
That is, in the third embodiment, a gain value at each sample position
30 between two gain sample positions is calculated through linear interpolation itsing
equation (1). In the present embodiment, instead, a wavcforn~ constituted with
straight lines respectively connecting the intersection CPI 1 of the twvo straight litle
L51 and straight line L52 as illustrated in FIG. 17 and two gain sample positions is
set as a gain wavefonn obtained through linear interpolation. It should be noted
that, in FIG. 17, the same reference numerals as those in FIG. 13 are assigned to
5 portions corresponding to those in FIG. 13, and explanatiotl thereof \\rill be omitted as
appropriate.
[0222]
In this exatnple, a broken line L71 constituted \writ11 a straight line
connecting the gain satnple position G31 and the intersection CPII and a straight
10 line connecting the gait1 satnple position G32 and the intersection CPI I is set as a
gain waveform behveen the gain sample position G31 and the gait1 satnple position
G32.
[0223]
In the example of the gain waveform illustrated in FIG. 17, when linear
-
15 interpolation is performed with the straight line connecting the hvo gain sample
positions, in order to reproduce the gait1 wavefotln more accurately, it is necessary to
set three or Inore gain sample positions in a section behveen the gain satnple position
G31 and the gain satnple position G32 upon eticoditlg of the gain wavefonn.
[0224]
20 That is, if satnple positions of the gait1 satnple position G3 1, the intersection
CPll and the gain sample position G32 are set as the gain satnple positions upon
encoding of the gain wavefonn, a difference (error) occuts behveen the gain
waveform before encoded atld the gain wavefortn obtained through decoding.
[0225] * - % -
25 On the other hand, when the broken line L71 is set as the gain waveform by
utilizing the intersection CP11, by only setting hvo positions of the gain sample
position G31 and the gain sample position G32 as the gain satnple positions, it is
possible to make a difference between the gain wavefonn before encoded and the
gain \vavefonn obtained through decoding smaller. Therefore, in a tnethod in
30 which linear interpolatiotl is performed while utilizing the intersection, it is possible
to reduce the number of gain sample positions, and thereby it is possible to suppress
a bit rate ofthe output code string and improve coding efficiency
[0226]
It should be noted that, when linear interpolation is perfommed while
utilizing the intersection, the intersection of the two straight lines has to be located
5 between the two gain sample positions.
[0227]
For example, in the example of FIG. 17, the intersection CPll has to be
located between the gain sample position G31 and the gain sample position G32 in
the sample axis direction.
10 [0228]
Therefore, a region used for determining whether non-linear interpolation is
performed or linear interpolation is performed is different between the exatnple in
FIG. 13 and the exatnple in FIG. 17. 111 the example in FIG. 17, when the
intersection CPI I is included in either the region TR21 or the region TR22, linear
- -
15 intelpolation utilizing the intersection is performed.
LO2291
Here, the region TR21 is a right half region of the region TRI 1 illustrated in
FIG. 13, that is, a region at the right side of the region TRll with respect to the gain
sample position G31 in FIG. 13. In a similar mannet; the region TR22 is a left half
20 region of the region TR12 illustrated in FIG. 13, that is, a region at the left side of the
region TR12 with respect to the gain sample position G32 in FIG. 13.
[0230]
In this mannet; when linear interpolation utilizing the intersection is
performed, the conditional expression corresponding to equation (14) is as indicated
25 with the following equation (19). That is, when the following equation (19) holds
true, linear interpolation utilizing the intersection is perfonned.
[023 I ]
[Math. 191
[0232]
In equation (19), when the distance dsample[k] is greater than zero, and
equal to or less than thresample, and the distance d-gain[k] is equal to or less than
5 the threshold thre-gain, or when the distance d-sa~nple[k+l] is greater than zero and
equal to or less than the threshold thre-sample, and the distance d_gain[k+l] is equal
to or less than the threshold thre-gain, the distance h t n the gain sample position lo
the intetsection X[k, k+l] is equal to or less than a threshold.
[0233]
10 For example, when the k-th gain sample position is the head position of the
fmme, that is, a satnple of n = 0, the distance d-sample[k], the distance d_gain[k],
the distance d-satnple[k+l] and the distance d-gain[k+l] in equation (19) are
respectively obtained using the following equation (20) to eqllation (23).
[0234]
15 [Math. 201
d-sample[kl=(g[k+ll-g[kl-s[k+ll xT[kl)
/ ( ~ [ k l - s [ k + l I ) . . (20)
[Math. 211
d-gain[kl=abs(s[kl x (g[k+ll -g[kl -s[k+11 x T[kl)
/(s[kI -s[k+lI)) = • (21)
[Math. 221
d~sampIe[k+1l=T[kl-(g~k+1]-g[k]-s[k+1lxT[kl)
20 / ( ~ [ k l - s [ k + l I ) . - (22)
[Math. 231
d-gain[k+l]=abs(s[k] X (g[k+1] -g[kl-s[k+l] X T[kl)
/(s[kI -s[k+ll) +g[kl -g[k+ll) . . . (23)
[0235]
When it is detemiined that the conditional expression indicated with
equation (19) does not hold true, a gain waveforai is obtained throogh non-linear
interpolation, that is, the above-described calculation of equation (4).
[0236]
5 On the other hand, when it is deterniined that the conditional expression
indicated with equation (19) holds true, a gain waveform is obtained through linear
interpolation.
[0237]
For example, when the k-th gain satliple position is tlie head position of the
10 frame, that is, a satnple position of n = 0, when the sample position of the
intersectiotl X[k, kill, that is, a sample length from the k-th gain satnple position to
the intersection X[k, k+l] is set as T'[k], the sample position T'[k] is obtained from
the following equation (24).
[0238]
- ~
15 [Math. 241
T'[kl= (g[k+ll -g[kl - s [ k + l l x T [ k I ) / ( s [ k l - s [ k + l l )
. (24)
[0239]
Further, a gain value g-interpolated[n] of tlle satnple n which is located
between the k-th gain sample position and the intersection X[k, k i l l and which is
20 the n-th (where 0 5 n < TT'[k])s ample fio~itih e head of tlie fianie is calculated froni
tlie following equation (25).
[0240]
[Math. 251 <.-
25 [0241]
It should be noted that, in equation (25), al[k] and bl[k] are values
respectively obtained from the follo\\ring equation (26) and equation (27).
[0242]
[Math. 261
a1 [kl =s [kl . . (26)
[Math. 271
bl [kl =g[kl - - . (27)
5 [0243]
al[k] and bl[k] indicate illclinatior~ and intercept of the straigbt line
connecting the k-th gain sample position and the intersection X[k, k+l]. Therefore,
in this example, as described with reference to FIG. 17, it is determined that the gait1
value linearly changes between the k-th gait1 satnple position G31 and the
10 intersection CPI1, and a gain value of each sa~llple n is obtained through linear
interpolatiotl.
I02441
Further, a gain value gginterpolated[n] of the sample n which is located
between the it~tersectionX [k, k+l] and the k+l-th gain satnple position and which is
15 the n-th (where T'[k] 5 n < T[k]) satnple fromlhe head of the frame is calculated
from the following equation (28).
[0245]
[Math. 281
g-i nterpol ated [nl =a2[kl x n+b2[k]
(T'[klln

Farthel; in the third embodiment and ~nodified example I of the third
embodiment, a case has been described where the interpolation mode information is
not included in the gain code string, and non-linear interpolation is basically
10 perfonned.
102551
Howvever, while the interpolation mode information is included in the gain
code string, and the gain wavefor111 is basically obtained using an interpolation
scheme indicated in the interpolation mode information, when the interpolation
~p
15 . scheme indicated in the interpolation mode infonnation is a scheme using non-linear
interpolation, linear interpolation may be performed under specific conditions.
102561
In such a case, at the decoding apparatus 51, the decoding processing
described with refe~ence to FIG. 8 is perfornled. However, in the gain decoding
20 processing corresponding to step S43, the gain decoding processing illustrated in FIG.
18 is performed. The gain decoding processi~~bgy the decoding apparatus 51 in
FIG. 15 will be described below with reference to the flowchart of FIG. 18.
[0257]
It should be noted that because the processing in step S161 to step S164 is
25 the same as the processing in step S71 to step S74 in FIG. 9, explanation thereof will
be omitted.
[0258]
In step S163, when it is determined that the sche~neis a scheme usiug nonlinear
interpolation, in step S165, the operation unit 131 perfom~s calculation of the
30 conditional expression indicated with equation (14) based on the read-out gain
infonnation.
[0259]
Then, while the processing in step SI66 to step S168 is performed, because
these processing is the same as the processing in stcp S133 to step S135 in FIG. 16,
explanation thereof will be omitted. It sl~ouldb e noted that, from step S165 to step
5 S168, the pmcessitig described in the third e~nboditnent may be performed, or
processing described in modified exatnple 1 of the third eeiboditnent may be
perfonned. Fut-thet; when non-linear interpolation is perfortned, limiting processing
may be perfonned.
[0260]
10 When a gait1 waveform is generated through interpolation in step S164, step
S167 or step S168, the processing then proceeds to step S169.
[0261]
In step S169, the gain decoding circuit 63 determities whether or not
processing is performed for all the gain satnple positions.
- -
15 102621
When it is detertiiined in step S169 that not all of the gain satnple positions
is processed, the processing returns to step S161, and the above-described processing
is repeated.
[0263]
20 On the other hand, whet1 it is determined in step S169 that all of the gain
satnple positions are processed, the gain decoding circuit 63 supplies the gain
wavefonn corresponding to one fratne constituted with gain values at the sample
positions obtained through the processing so far to the gain applying circuit 64, and
the gain decoding processing ends. When the gain decoding processing ends, the
25 processing then proceeds to step S44 in FIG. 8.
[0264]
As described above, the decoding apparatus 51 obtains a gain wavefort11
through linear interpolation or non-linear interpolation according to the itlterpolation
mode infonnation. By obtaining the gain waveform tlirough non-linear
30 interpolation as appropriate it1 this manner according to the interpolation mode
infonnation, it is possible to reproduce a gain waveform with high precision witli a
less code amount.
[0265]
Moreover, even when the interpolation scheme indicated in the interpolation
mode information is a scheme using non-linear interpolation, by performing linear
5 interpolation under specific conditions, it is possible to reproduce the original gain
\\laveform with higher precision with a less processing amount of encoding. Futthel;
it is possible to reduce a code atnount of the output code string.
[0266]
The series of processes described above can be executed by hardware bat
10 can also be executed by sofhvare. When the series of processes is executed by
soft\\~area, progratn that constructs such sofhvare is installed into a cotnputer. Here,
the expression "con~puter" includes a computer in which dedicated hardware is
incorporated and a general-purpose personal computer or the like that is capable of
executing various functions when various progratns are installed.
--
15 [0267]
FIG. 19 is a block diagram showing a hardware configuration example of a
cotnputer that perfonns the above-described series of processing using a progratn.
[0268]
In such computer, a CPU (Central Processing Unit) 501, a ROM (Read Only
20 Memory) 502, and a RAM (Randorn Access Metnory) 503 are connected to one
another by a bus 504.
[0269]
An inputloutpot interface 505 is also connected to the bus 504. An input
unit 506, an output unit 507, a recording unit 508, a co~l~rnunicatiount~it 509, and a
25 drive 510 are connected to the input/outpot interface 505.
[0270]
The input unit 506 is configured h t n a keyboard, a mouse, a microphone,
an imaging device or the like. The output unit 507 is configured from a display, a
speaker or the like. The recording unit 508 is configured fi.otn a hard disk, a non-
30 volatile tnemory or the like. The cotntnunication unit 509 is configured fro111 a
network interface or the like. The drive 510 drives a removable ~nediutn 511 such
as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor meltlory
or the like.
[0271]
In the compoter configured as described above, as one example the CPU
5 501 loads a program recorded in the recording unit 508 via the inputloutput interface
505 and the bus 504 into the RAM 503 and executes the program to carry out the
series of processes described earlier.
[0272]
Programs to be executed by the computer (the CPU 501) are provided being
10 recorded in the re~novable medium 51 1 which is a packaged ~nedii~o~r nth e like.
Also, programs may be provided via a wired or wireless transmission medium, such
as a local area nehvork, the Internet or digital satellite broadcasting.
[0273]
In the compoter, by loading tlie removable recording medium 51 1 into the
p~
15 drive 510, the program can be illstalled into the recording unit 508 via the
inputloutput interface 505. It is also possible to receive tlie program fio~na wired
or wireless transfer n~edium using the co~ntnl~nicatiounn it 509 and install the
program into the recording unit 508. As another alternative, the program can be
installed in advance into the ROM 502 or the recording unit 508.
20 [0274]
It should be noted that the program executed by a conlputer may be a
program that is processed in time series according to the sequence described in this
specification or a program that is processed in parallel or at necessary timing such as
upon calling.
25 [0275]
An embodiment of the present technology is not li~nitedto the etnbodi~netlts
described above, and various changes and modifications tnay be made without
departing from the scope of the present tech~lology.
[0276]
30 For example, the present technology can adopt a configuration of cloi~d
computing which processes by allocating and connecting one function by a plurality
of apparatnscs through a network.
[0277]
Futthel; each step described by the above ~nentionedf low cha~tsc an be
executed by one apparatus or by allocating a plurality ofapparatnses.
5 [0278]
In addition, in the case where a plurality of processes is included in one step,
the plurality of processes included in this one step can be executed by one apparatus
01. by allocating a plurality of apparatuses.
-
[0279]
10 The advantageous effects described herein are not limited, but merely
examples. Any other advantageous effects may also be attained.
[0280]
Additionally, the present technology may also be configured as below.
[0281]
- -
15 (1)
A decoding apparatus including:
a gain readout unit configured to read out encoded gain values at at least
t\vo gain sample positions of a time series signal;
an interpolation information readout unit configured to read out
20 interpolation information indicating whether the gain value at each sample position
of the time series signal is obtained through linear interpolation or obtained throngh
non-linear interpolation; and
an interpolation processing unit configured to obtain the gain value at each
sample position.-located between the two gain sample positions of the time series
25 signal based on the gain values at the gain sample positions through linear
interpolation or non-linear interpolation according to the interpolation infonnation.
(2)
The decoding apparatus according to (I),
wherein the gain readout nnit fin-ther reads out gain inclination values
30 indicating inclination ofthe gain values at the gain sample positions, and
wherein, when the gain value is obtained tluough non-linear interpolation,
the interpolation processing unit obtains the gain value at each satnple position
located between the h o gain sample positions based on the gain values and the gain
inclination values at the gain sample positions.
(3)
The decoding apparatns according to (I) or (2), fnrther including:
a limiting processing unit confignred to perform litniting processing on the
gain value obtained throngh non-linear interpolation so that the gain value becomes a
value equal to or greater than a predeternmined lower limit or a value equal to or less
than a predetermined upper limit.
10 (4)
The decoding apparatus according to (3),
wherein the limiting processing nnit perfonns limiting processing using zero
as the lower limit, litnititig p!.ocessing using one as the lower limit or litniting
processing using one as the upper limit.
15 (5)
The decoding apparatus according to any one of (2) to (4), further including:
an operation unit configured to obtain at the gain satnple positions, straight
lines having the gain values at the gain sample positions and having inclination
indicated by the gain inclination values at the gain sarnple positions, and obtain
20 differences between a gain value at an intctsection of the straight lines obtained for
the two gain sample positions and the gain values at the two gain sample positions,
wherein, when the interpolation information is information indicating that
the gain value is obtained through linear interpolation, the interpolation processing
unit obtains the gain value through linear interpolation, and, when the interpolation
25 information is infortnation indicating that the gain value is obtained through nonlinear
interpolation, the interpolation processing unit obtains the gain value through
non-linear interpolation or linear interpolation according to the differences.
(6)
A decoding nlethod including the steps of:
30 reading out encoded gain values at at least hvo gain satnple positions of a
time series signal;
reading out interpolation information indicating whether the gain value at
each san~plep osition of the tinle series signal is obtained through linear interpolation
or obtained through non-linear interpolation; and
obtaining the gain value at each sample position located between the two
5 gain sarilple positions of the time series signal based on the gain values at the gain
sanlple positions through linear interpolation or non-linear interpolation according to
the interpolation information.
(7)
A program causing a colnpater to execute processing including the steps of:
10 reading out encoded gain values at at least two gain satnple positions of a
time series signal;
reading out interpolation information indicating whether the gain value at
each sample position of the time series signal is obtained through linear interpolation
or obtained through non-linear interpolation; and
-
15 obtaining the gain value at each sample position located between the two
gain sample positions of the tinie series signal based on the gain values at the gain
sample positions through linear interpolation or non-linear interpolation according to
the interpolation information.
(8)
20 A decoding apparatus including:
a gain readout unit configured to read out encoded gain values at at least
h a gain sample positions of a time series signal and gain inclination values
indicating inclination of the gain values;
an operation unit configured to obtain at the gain sample positions, straight
25 lines having the gain values at the gain sample positions and having inclination
indicated by the gain inclination values at the gain sample positions, and obtain
differences behveen a gain value at an intersection of the straight lines obtained for
the two gain sample positions and the gain values at the two gain sample positions;
and
30 an interpolation processing unit configured to obtain the gain value at each
satnple position located between the hvo gain san~ple positions of the time series
signal through linear interpolation or non-linear interpolation according to the
differences.
(9)
A decoding method including the steps of:
5 reading out encoded gain values at at least hvo gain sample positions o f a
time series signal and gain inclination values indicating inclination of the gain
values;
obtaining at the gain sample positions, straight lines having the gain values
at the gain salnple positions and having inclination indicated by the gain inclination
10 values at the gain sample positions, and obtaining differences between a gain value at
an intersection of the straight lines obtained for the hvo gain salnple positions and the
gain values at the two gain satnple positions; and
obtaining the gain value at each sample position located between the hvo
gain satnple positions of the time series signal through linear interpolation or non-
- -
15 linear interpolation according to the differences.
(10)
A progranl causing a cotnputer to execute processing including the steps of:
reading out encoded gain values at at least hvo gain san~ple positions of a
time series signal and gain inclination values indicating inclination of the gain
20 values;
obtaining at the gain sample positions, straight lines having the gain values
at the gain sample positions and having inclination indicated by the gain inclination
values at the gain satnple positions and obtaining differences behveen a gain value at
an intersection ofthe straight lines obtained for the hvo gain sample positions and the
25 gain values at the hvo gain sample positions; and
obtaining the gain value at each satnple position located between the h o
gain satnple positions of the time series signal through linear interpolation or nonlinear
interpolation according to the differences.
Reference Signs List
decoding apparatus
demultiplexing circuit
signal decoding circuit
gain decoding circait
gain applying circuit
interpolation processing unit
litniting processing unit
operation onit
CLAIMS
Claim 1
A decoding apparatus comprising:
a gain readout unit configured to read out encoded gain values at at least
5 two gain sattlple positions of a time series signal;
an interpolation information readout unit configured to read out
itlterpolation information indicating whether the gain value at each sample position
of the titlie series signal is obtained through linear interpolation or obtained through
non-linear interpolation; and
10 an interpolation processing unit configured to obtain the gain value at each
satnple position located between the two gain sample positions of the time series
signal based on the gain values at the gain sample positions through linear
interpolation or non-linear interpolation according to the interpolation infor~nation.
15 Claim7
The decoding apparatus according to claim 1,
wherein the gain readout unit fi~~therera ds out gain inclination values
indicating inclination of the gain values at the gain sample positions, and
wherein, when the gain value is obtained thmugh non-linear interpolation,
20 the interpolation processing unit obtains tlle gain valoe at each sample position
located between the two gain sample positions based on the gain values and the gain
inclination values at the gain sample positions.
Claim 3 . .,
25 The decoding apparatus according to claim 1, further comprising:
a limiting processing unit configwed to perfonn limiting processing on the
gait1 value obtained through non-linear interpolation so that the gain value becomes a
value equal to or greater than a predetermined lower litnit or a value equal to or less
than a predetermined npper limit.
30
Claim 4
The decoding apparatus according to clainl 3,
whel.ein the linliting processing unit performs limiting processing wing zero
as the lowcr limit, limiting processing nsing one as the lower limit or limiting
processing using one as the upper limit.
5
Claim 5
The decoding apparatus according to clairll2, finther comprising:
an operation unit confignred to obtain at the gain sample positions, straight
lines having the gain values at the gain sample positions and having inclination
10 indicated by the gain inclination values at the gain sanlple positions, and obtain
differences behveen a gain value at an intersection o f the straight lines obtained for
the two gain sample positions and the gain valaes at the hvo gain sample positions,
wherein, when the interpolation information is information indicating that
the gain value is obtained through linear interpolation, the interpolation processing
-
15 unit obtains the gain value throngh linear interpolation, and, when the interpolation
information is infonnation indicating that the gain value is obtained through nonlinear
inte~.polation, the interpolation processitlg unit obtains the gain value through
non-linear interpolation or linear interpolation according to the differences.
20 Claim 6
A decoding method comprising the steps of:
reading out encoded gain values at at least two gain satnple positions of a
time series signal;
reading out interpolation information indicating whether the gain value at
25 each sample position of the time series sigual is obtained through linear interpolation
or obtained through non-linear interpolation; and
obtaining the gain value at each sample position located behveen the hvo
gain sample positions of the time series signal based on the gain values at the gain
sample positions through linear interpolation or non-linear interpolation according to
30 the interpolation information.
Claim 7
A program causing a computer to execute processit~g including the steps of:
reading out encoded gain values at at least h o gain satnple positions of a
time series signal;
5 reading out interpolation information indicating whether the gain value at
each sample position of the tinle series signal is obtained through linear interpolatiotl
or obtained through not]-linear interpolation; and
obtaining the gain value at each sample position located behveen the twvo
gain sample positions of the titne series signal based on the gain values at the gain
10 sample positions through linear interpolation or non-linear interpolation according to
the interpolation information.
Claim 8
A decoding apparatus comprising:
-
15 a gain readout unit configured to read out encoded gain values at at least
two gain satnple positions of a titne series signal and gain inclination values
indicating inclination of the gain values;
an opetation wit configured to obtain at the gain sample positions, straight
lines having the gain values at the gain sample positions and having inclination
20 indicated by the gain inclination values at the gain sample positions, and obtain
differences between a gain value at an intersection of the straight lines obtained for
the two gain satnple positions and the gain valoes at the hvo gain sample positions;
and
an interpolation processing unit configured to obtain the gain value at each
25 sample position located between the twvo gain sample positions of the time series
signal through linear interpolation or non-linear interpolation according to the
differences.
Claim 9
30 A decoding method cotnprising the steps of:
reading out encoded gain values at at least two gain sample positions of a
time series signal and gain i~~clinatiovna lues indicating inclination of the gain
values;
obtaining at the gain sample positious, st~aiglltl ines having the gain values
at the gain sanlple positions and Ilaving i~~clinatioind icated by the gain inclination
6 values at the gain sample positions, and obtaining differences between a g a i ~va~lu e at
an intersection of the straight lines obtained for the two gain samplc positions ant1 tlle
gain values at the two gain sample positions; and
obtaining the gain value at each sample position located between the hvo
gain sample positiol~so f the ti~nes eries signal throug11 lit~eari nterpolation or non-
10 linear interpolation according to the differe~~ces.
Claim 10
A progratn causing a coinpotel. to execute processing including the steps of:
reading out cr~coded gait1 values at at least I\vo gain sample positions of a
-
15 time series signal and gain incli~~atiovna lues indicating inclination of the gain
values; .
obtaining at tl~eg ain sample positiorls, straight lines having the gain values
at the gain sa~nl~ploes itions and having inclination indicated by the gain inclination
values at the gain sample positions and obtaining differences between a gain value at
20 all intersection of the stsaigl~tli nes obtained for the t\vo gain sample positions and tlie
gain values at the two gain sample positions; and
obtaining the gain value at cacl~s ample position located bet\veen the two
gain sample positions of the tirne series sigilal through linear interpolatiot~ or lionlinear
interpolation according to the differences