Embodiment
Below, with reference to the description of drawings embodiments of the invention.
(embodiment 1)
Fig. 1 is the block diagram of the picture signal mixed washing apparatus of the embodiment of the invention 1.
In Fig. 1, from the data image signal input store 18 of input terminal 10 inputs.
Blocking address generating unit 13 generates the address that is used for the input data are divided into piece on memory.Arrange and change the address that generating unit 14 generations in address are used on memory piece being changed arrangement.
Memory writes control part 12 control blocking address generating units 13 and arranges and changes address generating unit 14 according to importing the unique address of data on the position calculation memory on the picture and will importing the address that data write appointment.
Memory is read controller 15 and is utilized according to the data on channel address generating unit 17 readout memories 18 of the generation channel address of the sector address generating unit 16 of the data section of making on the memory 18 and deferent segment and to the channel of appointment and export.
Fig. 2 is the figure that writes to memory of explanation embodiment 1.Below, illustrate input signal with Horizontal number of pixels Xp, vertical line number Yp to write Horizontal number of pixels Ap ((situation of the memory area of Bp>Yp) (makes Xp * Yp=Ap * Bp) here, for Ap>Xp), vertical line number Bp.
In Fig. 2, solid line is represented the pixel region of input signal, and dotted line is represented the zone of memory 18.Make picture and memory one by one to the memory that can not write direct at once, so, arrange to change address generating unit 14 and generate the address that zone that zone with the Xp on the picture * (Yp-Bp) moves to (the Ap-Xp) * Bp on the memory writes.
In addition, above-mentioned arrangement changes and the moving process such as formation of section all are that least unit is carried out with the piece.
Now, establish this piece by horizontal Kp pixel * vertical capable formation of Kp.Here, the Kp of the capable least unit with respect to piece of vertical Y p, when being in the concerning of Yp=Kp * Lp (Lp is a natural number)+Kp/2, blocking address generating unit 13 takes place with the level (address of the piece write memory of 2 * Kp) pixels * vertical (Kp/2) row for the capable portion of Kp/2.
Memory is read controller 15 and is used sector address generating unit 16 and channel address generating unit 17 these 2 kinds of address generating units to carry out reading from memory.The generation of reading the address from memory is at first by the number corresponding address of sector address generating unit 16 generations with the section of exporting to channel.Be that the simple method that deviant with the level of calculating each channel that will be generated by channel address generating unit 17, vertical direction is added to the value on this address is carried out.
This be because, when forming section, and though be positioned at mutually on the picture away from the memory of position on piece how read the address channel, all be identical deviant in level and vertical direction, different in each channel, only be the address on the memory that begins.
According to the address of reading of decision in the above described manner, by reading a plurality of sections of formation.
In addition, address, sector address, the channel address of the address of blocking, arrangement change are arbitrarily.
(embodiment 2)
Fig. 3 is the block diagram of the picture signal mixed washing apparatus of the embodiment of the invention 2.
In Fig. 3, be marked with identical symbol for piece, and omit its explanation with Fig. 1 same action.
In Fig. 3, from the picture signal of input terminal 10 inputs, to format converter 11 outputs.
11 pairs of input signals of format converter are frequency band limits in addition, carries out the pixel transformation of variable of input signal simultaneously.
Here, suppose that input signal is that Horizontal number of pixels is that Mp, vertical line number are the signal of Np.
Wherein, suppose that Mp, Np are in the relation of Mp 〉=Xp, Np 〉=Yp for the Horizontal number of pixels Xp of input signal, the signal of vertical line number Yp.
11 pairs of format converters input pixel is in addition after the frequency band limits, takes out pixel, thereby input signal is transformed to the signal of Horizontal number of pixels Xp, vertical line number Yp and to memory 18 outputs.
As mentioned above, the picture signal mixed washing apparatus of embodiment 2 makes input signal become the form consistent with shuffling figure by increase format converter on the basis of embodiment 1, just can make its signal with a lot of specifications corresponding.
Format conversion can be horizontal direction or vertical direction, also can not carry out frequency band limits and only reduces pixel.In a word, as long as make the conversion consistent of input pixel count with the pixel count that shuffles.
(embodiment 3)
Fig. 4 is the figure that is used to illustrate the embodiment of the invention 3.
The block diagram of the picture signal mixed washing apparatus of present embodiment is identical with Fig. 3.
Below, use Fig. 3 and Fig. 4 that the action of the picture signal mixed washing apparatus of present embodiment is described.
Effectively line number be 1080 row, luminance signal (below, table is for Y-signal) the valid pixel number (the valid pixel number of per 1 row) of horizontal direction be that the picture signal (referring to Fig. 4 (a)) that the valid pixel number of the horizontal direction of 1920 pixels, 2 color difference signals (below, table is Cr, Cb signal respectively) is respectively 960 pixels is input on the terminal 10.
11 pairs of input signals of format converter are frequency band limits in addition, and the Horizontal number of pixels of Y-signal is transformed to 1280 pixels, and the Horizontal number of pixels of Cr, Cb signal is transformed to 640 pixels (referring to Fig. 4 (b)) respectively.By the zone 1 of 1072 row in 13 pairs of effective line numbers 1080 of blocking address generating unit, Y-signal generates the blocking address with the piece of level 16 pixels * vertical 16 row, and Cr, Cb signal generate the blocking address with the piece of level 8 pixels * vertical 16 row.
For the zone 2 of remaining 8 row, Y-signal generates the blocking address with the piece of level 32 pixels * vertical 8 row, and Cr, Cb signal generate the blocking address with the piece of level 16 pixels * vertical 8 row.
In Fig. 4 (c), in order to understand easily the relation with pixel, expression be the situation that the identical block size of piece of the piece utilization in zone 2 and Fig. 1 is divided into each piece.
On picture, will be positioned at 1 Y-signal piece of same position and Cr, Cb block as 1 macro block (below, show to be MB), shown in Fig. 4 (d) like that, change and be arranged in horizontal direction 90MB, vertical 60MB.
Fig. 5 is the figure that the arrangement of the piece of explanation present embodiment changes.
As shown in Figure 5, do not arrange change, be aligned to right-hand member and will part up and down change for the vertical 60MB of central authorities.
The horizontal direction 80MB on top, the zone of vertical direction 4MB are that 1 unit is divided into 8 zones from U0 to U7 with horizontal direction 10MB.
Arrange to change and do not change the interior MB arrangement in zone, and just move.As a result, just become the zone of horizontal 10MB, vertical 32MB.
The zone of the horizontal 80MB of bottom, vertical 3MB, the same with the situation on top, the zone of D0-D7 moving like that as shown by arrows, thus become the zone of horizontal direction 10MB, vertical direction 24MB.The zone of the D8 of remaining horizontal direction 40MB, vertical direction 1MB also is divided into 4 zones of horizontal direction 10MB and moves, thereby becomes the zone of horizontal direction 10MB, vertical direction 4MB.
As mentioned above, by arranging the zone that changes the horizontal direction 90MB * vertical direction 60MB (referring to Fig. 4 (d)) that forms, 1MB unit is cut apart by vertical direction, shown in Fig. 4 (e) like that, be divided into 2 zones of horizontal direction 90MB, vertical direction 30 (Ns=30) MB.In addition, these 2 zones are cut apart with 9MB unit 90MB in the horizontal direction.Form 1 zone by zone, just form 4 zones (referring to Fig. 4 (e)) with the every 9MB of this horizontal direction.That is, formation is by 4 zones of the macro block formation of horizontal direction 45MB (Ms=45), vertical direction 30MB (Ns=30).
In the present embodiment, get Lh=5, Lv=5, constitute horizontal direction 9MB (Ms/Lh=9), vertical direction 6MB (Ns/Lv=6) in the dotted line in Fig. 4 (e).
These 4 zones and 4 output channel A, B, C, D are corresponding.
Little square in Fig. 4 (e) is represented the piece of reading beginning of each channel.
As shown in FIG., according to the address of reading of different starting position, per 4 zones, what will have identical value in all zones forms section to 5 pieces combinations of the deviant of the appointment of level and vertical direction.
With being formed on of such section vertical and horizontal direction move the position of each piece repeatedly and form 1 frame section.
Other MB meet the relation of (1), (2) formula with respect to the deviant of the position of 1 MB of 5 MB that form section.
Hoff=MOD(k×Ms/Lh,Ms) (1)
Voff=MOD(p×k×Ns/Lv,Ns) (2)
MOD (p, Lv) ≠ 0, and MOD (Lv, p) ≠ 0.
k=1、2、…、Lh-1
Wherein, and MOD (a, b)=a-b * INT (a/b)
(a, b are integer, and INT (a/b) is the following function of decimal point of giving up the operation result of undependent variable)
In the present embodiment, the value of p adopts 3, and concrete deviant is the relation that meets (3), (4) formula as position Hn, the Vn of the level of the beginning piece of individual section of the n (being positive integer) of (9,18), (18,6), (27,24), (36,12) and vertical direction.
Hn=MOD(H1+INT((n-1)/Lv),Ms/Lh) (3)
Vn=MOD(V1+Ns/Lv×MOD(n-1,Lv)
+INT((n-1)/Lv/(Ms/Lh)),Ns) (4)
Wherein, H1, V1 are the level of beginning piece and the position of vertical direction.
In the present embodiment, in 1 zone, form (the section of Ms/Lh * Ns)=270.
Particularly, in Fig. 4 (e), in the horizontal direction in the 30MB of dotted line 5 row of the 9MB of dotted line 1 row and vertical direction,
In step 1: every 6MB is scanned in vertical direction
In step 2: the every scanning of vertical direction 5 times, move in the horizontal direction 1MB, and execution in step 1
In step 3: every scanning 45 times, move 1MB in vertical direction
Execution in step 1,2,3 repeatedly, form 270 sections.
(embodiment 4)
Fig. 6 is the figure that is used to illustrate the embodiment of the invention 4.
The block diagram of the mixed washing apparatus of present embodiment is identical with Fig. 3.
Below, use Fig. 3 and Fig. 6 that the action of present embodiment is described.
The picture signal that is input on the terminal 10 is identical with embodiment 3.
In the present embodiment, the picture signal of input is by format converter 11 restricted band, and about the pixel count of horizontal direction, Y-signal is transformed to 960 pixels, and Cr, Cb are transformed to 320 pixels (referring to Fig. 6 (b)) respectively.
In addition, effective line number of Cr, Cb is transformed to 540 row.
Blocking address generating unit 13 is for the Y-signal in the zone 1 of 1072 row in effective line number 1080, and the piece generation blocking address with level 24 pixels * vertical 16 row generates the blocking address for Cr, Cb signal with the piece of level 8 pixels * vertical 8 row.
For the Y-signal in remaining 8 zone 2 of going, the pieces generation blocking address with level 48 pixels * vertical 8 row generates the blocking address for Cr, Cb signal with level 16 pixels * vertical 4 pieces of going.
In Fig. 6 (c), in order to understand easily the relation with pixel, expression be with the piece in zone 2 situation with the block size block identical with regional 1 piece.The piece that makes like this is used in 1 Y-signal piece being positioned at same position on the picture and Cr, Cb block as 1MB, shown in Fig. 6 (d) like that, arrange and change into horizontal direction 45MB, vertical direction 60MB.
In Fig. 6 (d), be used in 1 Y-signal piece being positioned at same position on the picture and Cr, Cb block and constitute 1 macro block (below, show to be MB).
Fig. 7 is the figure that the arrangement of the piece of explanation present embodiment changes.As shown in Figure 7, the number of the MB of horizontal direction is half of embodiment 3.Therefore, arranging the detailed description that changes is omitted.
As mentioned above, the MB (referring to Fig. 6 (d)) that arranges after changing is cut apart 60MB by 3 block units in vertical direction, gather alternately with the zone that 3MB unit segments, shown in Fig. 6 (e) like that, be divided into 2 zones that constitute by level 45 (Ms=45) MB, vertical 30 (Ns=30) MB.
The section that makes in these two zones is respectively to A channel and the output of B channel.
In the present embodiment, get Lh=5, Lv=10, each zone in the dotted line of Fig. 6 (e) forms 9MB (Ms/Lh=9) in the horizontal direction, forms 3MB (Ns/Lv=3) in vertical direction.
Blockage in Fig. 6 (e) is represented the piece of reading beginning of each channel.
In the present embodiment, two channels all will have 5 MB sections of being combined to form to the deviant of the appointment of the level of identical value and vertical direction from identical starting position.The deviant of the position of 5 MB of the formation section in the same area meets the formula (1) of embodiment 3, the relation of (2).
In the present embodiment, the p value adopts 6, and concrete deviant is (9,18), (18,6), (27,24), (36,12).
The level of the piece of the beginning that n (n is a natural number) is individual section and position Hn, the Vn of vertical direction meet the relation of (5), (6) formula.
Hn=MOD(H1+INT((n-1)/(Ns/Lv)),Ms/Lh)
(5)
Vn=(Ns/Lv)×MOD(INT(V1/(Ns/Lv))
+INT((n-1)/(Ms/Lh)/(Ns/Lv),Lv)
+MOD(Hn+1,2)×MOD(V1+(n-1),(Ns/Lv))
+MOD(Hn,2)×MOD((Ms/Lh)×(Ns/Lv)
-MOD(Ci+n,(Ms/Lh)/(Ns/Lv)),(Ns/Lv))
(6)
Wherein, Ci=(Ns/Lv)-MOD (V1, (Ns/Lv)), the position of the vertical and horizontal direction of the MB when V1 is n=1.
Particularly, the piece in the dotted line is scanned in vertical direction, when the scanning of finishing vertical direction, move in the horizontal direction 1MB, scan in vertical direction repeatedly then.In addition, whenever mobile in the horizontal direction 1MB makes vertical scanning direction counter-rotating 1 time.
During the end of scan of the MB in 1 dotted line, the MB in the next dotted line is scanned, thereby in each zone, form 270 sections in vertical direction.
(embodiment 5)
Fig. 8 is the block diagram of the mixed washing apparatus of embodiment 5.
In Fig. 8, it is substantially the same that memory writes the generating unit 13 of the action of blocking address generating unit 23 of controller 22 and embodiment 1, so explanation is omitted.
Memory is read controller 24 and is used sector address generating units 25 and channel address generating unit 26 these two kinds of address generating units to carry out the processing of reading from memory.
For the zone of the central authorities on the picture, carry out the processing the same with embodiment 1.For in addition zone then only to for example channel A output of specific channel.
In the present embodiment, following straightforward procedure is carried out in use, generate and number corresponding address the level to each channel, the deviant of vertical direction and the value after this address addition that calculating will be generated by channel address generating unit 26 to the section of each channel A, B, C, D output by sector address generating unit 25 from the generation of reading the address of the memory of picture middle section.
As mentioned above, the picture signal mixed washing apparatus of embodiments of the invention 5 does not carry out the arrangement carried out and changes in embodiment 1, just can be applied to the Hi-Vision signal with shuffling and exports to a plurality of channels.In addition, whole channels are not used in the zone beyond the middle section, so, can import other data to obsolete channel.
In addition, in the present embodiment, address, sector address, channel address that the address of blocking, arrangement change are arbitrarily.
(embodiment 6)
Fig. 9 is the block diagram of the mixed washing apparatus of the embodiment of the invention 6.
In Fig. 9, be marked with identical symbol for piece with Fig. 8 same action.
In Fig. 9, export to format converter 21 from the picture signal of input terminal 20 inputs.Format converter 21 carries out conversion with embodiment 2 identical resolutions to picture signal, and data are exported to memory 27.
Memory writes controller 22, blocking address generating unit 23, memory, and to read the action of controller 24, sector address generating unit 25 and channel address generating unit 26 identical with embodiment 5.
As mentioned above, the picture signal mixed washing apparatus of embodiment 6 is by increasing format converter 21 on the basis of embodiment 5, just input signal can be transformed to shuffle figure consistent.
Therefore, can make it corresponding with the received image signal of more specifications.
The format conversion of present embodiment can be only in the horizontal direction or vertical direction carry out, can not carry out frequency band limits yet and only reduce pixel.
In a word, be transformed to consistent as long as will import pixel count with the pixel count that shuffles.
(embodiment 7)
Figure 10 is the figure that is used to illustrate the embodiment of the invention 7.
The block diagram of the mixed washing apparatus of present embodiment is identical with Fig. 9.
Effectively line number is that the pixel count of the horizontal direction of 1080 row, Y-signal is that the picture signal (referring to Figure 10 (a)) that the valid pixel number of the horizontal direction of 1920 pixels, Cr, Cb signal is respectively 960 pixels is input on the terminal 20.
21 pairs of input signals of format converter are frequency band limits in addition, and the Horizontal number of pixels of Y-signal is transformed to 1440 pixels, and the Horizontal number of pixels of Cr, Cb signal is transformed to 720 pixels (referring to Figure 10 (b)) respectively.
Blocking address generating unit 23 generates the blocking address of cutting apart with the piece of level 16 pixels * vertical 16 row with the Y-signal in the zone 1 of 1056 row in effective line number 1080 row, Cr, Cb signal is generated the blocking address of cutting apart with the piece of level 8 pixels * vertical 16 row.
As a result, just to make be 90 in the horizontal direction, be the Nb=66 piece in vertical direction.
Regional 2a that is made of 16 row and the regional 2b that is made of 8 row are divided in the zone 2 of remaining 24 row.
The Y-signal of regional 2a is generated the blocking address with level 16 pixels * vertical 16 pieces of going, Cr, Cb signal are generated the blocking address with the pieces of level 16 pixels * vertical 8 row.That is, make 45 of horizontal directions, 1 of vertical direction.
In Figure 10 (c), (d), (e), in order to understand easily the relation with pixel, expression be that the piece of regional 2v is used the line of pixels column split identical with the piece in zone 1 is the situation of piece.
The piece that makes like this will be positioned at 1 Y-signal piece of same position and Cr, Cb block as 1MB on picture, shown in Figure 10 (d), constitute horizontal 90MB, vertical 67.5MB.
Horizontal direction 90MB, the vertical direction 66MB in the zone 1 of Figure 10 (d) are cut apart by vertical direction 1MB unit, carry out after divided area gathers with 1MB unit, shown in Figure 10 (e) like that, be divided into 2 zones of horizontal 90MB, vertical 33MB and the zone of zone 2 horizontal 90MB, vertical 1.5MB.
Below, the processing of reading from zone 1 memory is described.
2 zones of above-mentioned zone 1 and then with 9MB unit 90MB is cut apart in the horizontal direction.Make 1 zone by the zone at the every 9MB of this horizontal direction, the result just forms 4 zones (referring to Figure 10 (e)).In the dotted line in Figure 10 (e), constitute 9MB in the horizontal direction, constitute 3MB in vertical direction.
These 4 zones and output channel A, B, C, D are corresponding.
Blockage in Figure 10 (e) is represented the piece of reading beginning of each channel.As shown in the figure, according to the address of reading of the different starting position of each channel, will in all channels, have 5 MB sections of being combined to form to the deviant of the appointment of the level of identical value and vertical direction.
In the present embodiment, the deviant of the position of 5 MB of the formation section in the same area is (9,18), (18,6), (27,24), (36,12).
The level of the beginning piece that n (n is a positive integer) is individual section is the same with embodiment 3 with position Hn, the Vn of vertical direction, meets the relation of formula (3), (4).
In the present embodiment, in formula (3), (4), Ms=45, Ns=33, Lh=5, Lv=11.
As a result, in each 4 zone of per 1 frame, make 297 sections.
The MB in zone 2 amounts to 135MB.In the present embodiment, from top to the bottom, from left to right order is by per 17 MB sections of formation on picture.
That is, the section that begins most is that 0,27,54,81,108 MB makes by number, and next section is that 1,28,55,82,109 MB makes by number, and by that analogy, order constitutes section.With 27 sections forming like this shown in Figure 10 (e) like that, export to the A channel.
By the explanation of the foregoing description 1~embodiment 7 as can be known, picture signal mixed washing apparatus of the present invention, memory write is fashionable carrying out, by utilizing the blocking address to be divided into piece, and generate as required to arrange and change the address, the arrangement of piece is changed, carrying out memory when reading, by utilizing sector address and channel address calculated address, just can be applied to the Hi-Vision signal, to a plurality of channel outputs with shuffling.
(embodiment 8)
Figure 11 is the block diagram of the picture signal mixed washing apparatus of the embodiment of the invention 8.
In Figure 11, the progressive image signal that is input to input terminal 30 is to memory 37 inputs.
Memory as cutting unit writes controller 32 from the position of input data on picture, the input data is written to the address on the memory that is calculated by blocking address generating unit 33.
The memory that makes the unit as section is read controller 34 and is used sector address generating units 35 and regional address generating unit 36 these 2 kinds of address generating units to generate to read the address, and from memory 37 sense data.
The zonule that regional address generating unit 36 will be made of a plurality of on the memory is as 1 unit, and the unit piece that will be kept at the image signal level direction on the memory is divided into 2 * Lh (Lh=2,3,4 frame by frame ...) individual.The even number zoning of the horizontal direction in the 2 * Lh that will cut apart piece is the 1st zone, is the 2nd zone with the odd number zoning.
In addition, 1st zone corresponding with the 1st frame in continuous 2 frames of input progressive signal is called the 1, the 2nd zone, zone and is called zone 2; Will 1st zone corresponding be called the 3, the 2nd zone, zone and be called zone 4, judge the data of reading from memory are arranged in which zone of zone 1 to zone 4 with the 2nd frame, and OPADD.
From the generation of reading the address of memory 37, at first generate number corresponding address with the section of exporting in each zone by sector address generating unit 35.
Secondly, by being added on the address that generates by generating unit 35, generate and read the address by the deviant that regional address generating unit 36 generates to each regional level, vertical direction.
When forming section, carry out be positioned on the picture mutually away from the memory 37 of position on the reading of piece, no matter the zone how, the skew of level and vertical direction is all identical, in each zone the address on the different memory that just begins to carry out, so, just can generate with above-mentioned simple method and to read the address.
As mentioned above, the picture signal mixed washing apparatus of the embodiment of the invention 8 can be divided into a plurality of zones with the input progressive signal, and uses shuffling method.
(embodiment 9)
Figure 12 is the block diagram of the picture signal mixed washing apparatus of embodiment 9.
In Figure 12, be marked with identical symbol for piece, and omit its explanation with Figure 11 same action.Export to format converter 31 from the progressive picture information of input terminal 30 inputs.31 pairs of input signals of format converter are frequency band limits in addition, carries out the pixel transformation of variable of input signal simultaneously, then to memory 37 inputs.
Here, when input signal was the high image of input signal resolution than embodiment 8, format converter 31 just carried out taking out processing between frequency band limits and pixel, exported after being transformed to the pixel size identical with the input signal of embodiment 8.
As mentioned above, the picture signal mixed washing apparatus of embodiment 9 can be applied to a greater variety of progressive image signals than embodiment 8 by increase format converter 31 on the basis of embodiment 8.
The format conversion of present embodiment not only can reduce pixel count, and can increase empty data by filling, interpolation.In addition, also can not carry out frequency band limits.In a word, as long as make the input pixel count consistent with the pixel count that shuffles just passable.
(embodiment 10)
Figure 13 is the figure that is used to illustrate the embodiment of the invention 10.
The block diagram of the picture signal mixed washing apparatus of present embodiment is identical with Figure 12.
In Figure 12, Figure 13, effectively the progressive image signal (referring to Figure 13 (a)) of several 640 pixels of horizontal valid pixel of several 1280 pixels of horizontal valid pixel of line number 720 row, Y-signal, Cr, Cb signal is input on the input terminal 30.
31 pairs of received image signals of format converter are frequency band limits in addition, shown in Figure 13 (b) like that, the horizontal direction pixel count of Y-signal is transformed to 960 pixels, the horizontal direction pixel count of Cr, Cb signal is transformed to 480 pixels.
Blocking address generating unit 34 generates the blocking address of cutting apart with the piece of level 16 pixels * vertical 16 row for Y-signal, generates the blocking address of cutting apart with the piece of level 8 pixels * vertical 16 row for Cr, Cb signal.
As a result, shown in Figure 13 (c), Y-signal forms 60 * 45, and Cr, Cb signal form 60 * 45 respectively.
The piece of the Y-signal by utilizing same position on picture, Cr, Cb signal constitutes MB, shown in Figure 13 (d), and formation level 60 * vertical 45MB.
Concentrate continuous 2 frames (referring to Figure 13 (e)) to carry out later processing the piece of Figure 13 (d) of above-mentioned blocking.
The MB of each frame is divided into 10 (2Lh=10) five equilibrium in the horizontal direction, is divided into 5 five equilibriums in vertical direction.
With being carried out of horizontal direction the zone of 2Lh five equilibrium be divided into the zone of odd and even number, be zone 1 with the even number zoning of the 1st frame, be zone 2 with the odd number zoning, be regional 3 with the even number zoning of the 2nd frame, be regional 4 with the odd number zoning.
Dotted line in Figure 13 (e) represents to cut apart the image of above-mentioned arrangement, and each zone in the dotted line is made of horizontal 6MB, vertical 9MB.
Blockage in Figure 13 (e) is represented the MB that reads beginning in each zone.As shown in the figure, according in the address of reading of each different starting position, zone, will in all zones, have 5 MB sections of being combined to form to the deviant of the appointment of identical value level and vertical direction.
Particularly, the level of 5 MB of formation section, the deviant of vertical direction meet the relation of the formula (1) of embodiment 3, formula (2) in the same area.
In the present embodiment, at k=1,2,3,4 o'clock, Ms=30, Ns=45, p=3, Lv=5, Lh=5, deviant is (6,27), (12,9), (18,36), (24,18).
With such section vertical and horizontal direction move repeatedly each piece section, form the section of 2 frames.
That is, the level of the beginning piece that the n in each zone (n is a positive integer) is individual section is the same with embodiment 3 with position Hn, the Vn of vertical direction, meets the relation of formula (3), formula (4).
The situation of present embodiment, in formula (3), formula (4), Ms=30, Ns=45, Lh=5, Lv=5.
As a result, in each zone, just form 270 sections.
(embodiment 11)
Figure 14 is the figure that is used to illustrate the embodiment of the invention 11.
The block diagram of the picture signal mixed washing apparatus of present embodiment is identical with Figure 11.
Below, use Figure 11 and Figure 14 that the picture signal mixed washing apparatus of present embodiment is described.
Shown in Figure 14 (a), effectively line number is that the progressive image signal that the horizontal valid pixel number that 480 horizontal valid pixel numbers capable, Y-signal are 720 pixels, Cr and Cb signal is 360 pixels is input on the input terminal 30.
Blocking address generating unit 32 generate with Y-signal with the pieces of level 16 pixels * vertical 16 row cut apart, with Cr and Cb signal with blocking address that level 8 pixels * vertical 16 capable pieces are cut apart.
That is, shown in Figure 14 (b), Y, Cr, Cb signal become 45 * 30 piece.
Constitute macro block by Y, the Cr of the same position on picture, the piece of Cb signal.That is, shown in Figure 14 (c) like that, constitute the MB of horizontal 45MB * vertical 30MB.
Concentrate continuous 2 frames (referring to Figure 14 (d)) to carry out following processing the MB of Figure 14 (c) of above-mentioned blocking.
The piece of each frame is divided into level 5 five equilibriums, vertical 5 five equilibriums, and the slip gauge of the 1st frame is decided to be zone 1, the slip gauge of the 2nd frame is decided to be zone 2.
Dotted line in Figure 14 (d) is represented image that above-mentioned arrangement cut apart to become horizontal 9MB, vertical 6MB in the dotted line.
Blockage in Figure 14 (d) is represented the piece of reading beginning in each zone.As shown in the figure, according in the address of reading of each identical starting position, zone, will in all zones, have 5 piece sections of being combined to form to the deviant of the appointment of the level of identical value and vertical direction.
The deviant of the position of 5 MB of formation section meets the formula (1) of embodiment 3, the relation of formula (2) in the same area.
In the present embodiment, k=1,2,3,4 o'clock, Ms=45, Ns=30, p=3, Lv=5, Lh=5 were so deviant is (9,18), (18,6), (27,24), (36,12).
The level of the beginning piece that n (n is a positive integer) is individual section is the same with embodiment 3 with position Hn, the Vn of vertical direction, meets the relation of formula (3), formula (4).
The situation of present embodiment, in formula (3), formula (4), Ms=45, Ns=30, Lh=5, Lv=5.
By making the position that being formed on of such section is vertical and horizontal direction moves each piece repeatedly, form 270 sections of 2 frames.
As mentioned above, the picture signal mixed washing apparatus of present embodiment is that 480 progressive signals of going carry out can shuffling after the Region Segmentation to effective line number.
(embodiment 12)
Figure 15 is the figure that is used to illustrate the embodiment of the invention 12.
The block diagram of the picture signal mixed washing apparatus of present embodiment is identical with Figure 12.Below, use Figure 12 and Figure 15 that the picture signal mixed washing apparatus of present embodiment is described.
Shown in Figure 15 (a), effectively line number is that the horizontal valid pixel numbers of 720 row, Y-signal are that the horizontal valid pixel number of 1280 pixels, Cr and Cb signal is that 640 progressive image signal is input on the input terminal 30.
Format converter 31 shown in Figure 15 (b) like that be 960 pixels with the horizontal valid pixel transformation of variables of Y-signal, be 240 pixels with the horizontal valid pixel transformation of variables of Cr and Cb signal.
Blocking address generating unit 34 generate with Y-signal with the pieces of level 32 pixels * vertical 8 row cut apart, with Cr and Cb signal with blocking address that level 8 pixels * vertical 8 capable pieces are cut apart.
That is, shown in Figure 15 (c) like that, Y, Cr, Cb signal become 30 * 9 respectively.
Figure 15 (d) is the figure that the piece of Y, the Cr of same position on picture, Cb signal is represented as macro block.
Shown in Figure 15 (e), the MB of Figure 15 (d) is divided into level 5 five equilibriums, vertical 5 five equilibriums.The BOB(beginning of block) of reading beginning each zone shown in the blockage in Figure 15 (e) calls over, and forms 540 sections.
As mentioned above, the picture signal mixed washing apparatus of present embodiment carries out the format conversion different with embodiment 10, can be that the progressive signals of 720 row carry out shuffle process to effective line number.
In addition, the part of the foregoing description is carried out following change, also can obtain the effect identical with embodiment 12.
The horizontal valid pixel transformation of variables of Cr, Cb signal is 480 pixels, vertical effectively line number is transformed to 360 row.Cut apart with level 16 pixels * vertical 16 row.As a result, just constitute horizontal 60MB, vertical 45MB, shown in Figure 15 (e) like that, level and vertical 5 five equilibriums that all are divided into.
As mentioned above, the part of change embodiment 12 is handled, and forms 540 sections.
By the explanation of the foregoing description 8~embodiment 11 as can be known, the memory that the memory that picture signal mixed washing apparatus of the present invention has a blocking address generating unit by formation writes controller and has sector address generating unit and a regional address generating unit is read controller 104,1 frame of input progressive image signal is divided into 1 zone at least, with 1 frame or 2 continuous frames serves as to handle unit, the hop count that per 1 zone is made can be divided into the hop count identical with existing TV signal, thereby can use shuffle process the input progressive signal.
Though it is undeclared in the above-described embodiments,, to also being easily, and be also contained among the present invention with the regional corresponding output channel deferent segment that makes by the regional address generating unit.
A kind of record is used for also belonging to the present invention by all or part of functional programs of all or part of key element of each picture signal mixed washing apparatus of computer execution the foregoing description 1~12 and/or the program record medium of data, it is characterized in that: can utilize computer to read, the collaborative above-mentioned functions of carrying out of said procedure that reads and/or above-mentioned data and aforementioned calculation machine.
Below, the embodiment about encoding amount control of the present invention is described.
The quantizer and the relative importance value of following examples at first, are described.
Each quantizer is set the quantification number of regulation quantization step.Set each the quantizer affix that quantizes number and quantized number, show into quantizer 0, quantizer 1, quantizer 2 ...It is big more to quantize number, and quantization step is more little.(quantization step is more little, just means that the generation encoding amount is many more)
In addition, relative importance value affix relative importance value number, show into relative importance value 0, relative importance value 1, relative importance value 2 ...And, set the deviant corresponding with the relative importance value number.
When given deviant, quantizer just adds that by the quantification number of giving self the quantification number after the deviant carries out quantification treatment.
Quantification number after the addition is sometimes outside the scope of the quantification number that device is prepared.At this moment, when the quantification number of the maximum that surpasses device, just adopt maximum quantification number.Though do not adopt in the present embodiment,, in deviant with negative value definition and the quantification number of the minimum that the quantification number ratio device after adding deviant is prepared hour just adopts minimum quantification number.
(embodiment 13)
Figure 16 is the block diagram of the code device of the embodiment of the invention 13.
In Figure 16, import section from the picture signal of input terminal 40 inputs and make portion 41.Duan Zuocheng portion 41 is the piece that is made of 8 * 8 pixels with the image segmentation of input, concentrates 20 pieces to constitute the section of controlling unit as encoding amount.
Each section is transfused to quadrature transformer 42 and the 1st relative importance value calculating part 43.
The section of 42 pairs of inputs of quadrature transformer is that unit carries out orthogonal transform by piece, and the output orthogonal transform data.
The orthogonal transform data are imported the 1st relative importance value calculating part the 43, the 1st encoding section 44, and the 45, the 2nd encoding section 47 and 49 and encoding section 51.
The 1st encoding amount calculating part 55 is made of the 1st encoding section 44 and 45.The 2nd encoding amount calculating part is made of the 2nd encoding section 47 and 49.
The 1st relative importance value calculating part 43 calculates the relative importance value of each piece according to the block message before and after the orthogonal transform.In the present embodiment, relative importance value is 0 and 1 these two kinds.The 1st relative importance value of each piece is imported the 1st encoding section 44 and the 45, the 2nd relative importance value calculating part 46 and 48 and quantizer determination section 50.
The specific relative importance value of the 1st relative importance value change of 46 pairs of inputs of the 2nd relative importance value calculating part, and set the 2nd relative importance value.The 2nd relative importance value is imported the 2nd encoding section 47 and quantizer determination section 50.
The specific relative importance value of the 1st relative importance value change of 48 pairs of inputs of another the 2nd relative importance value calculating part, and set the 2nd relative importance value.The 2nd relative importance value is imported the 2nd encoding section 49 and quantizer determination section 50.
The 1st encoding section 44 that is made of quantizer 0 (not shown) and the 1st encoding section 45 that is made of quantizer 1 (not shown) are respectively by carrying out quantification treatment to the orthogonal transform data of input by the quantification number after the quantification number addition of given deviant of the 1st relative importance value and quantizer.
Deviant is intrinsic and certain to relative importance value, and in the present embodiment, the deviant of relative importance value 0 is 1, and the deviant of relative importance value 1 is 0.
For example, a certain relative importance value is 1 o'clock, and quantizer 0 carries out quantification treatment by quantizing number 1 (result of the quantification number addition of deviant and quantizer is limited by the maximum quantizer number of the 1st encoding amount calculating part) to this piece by quantizing number 1, quantizer 1.Relative importance value is 0 o'clock, and quantizer 0 carries out quantification treatment by quantification number 0, quantizer 1 by quantizing 1 pair of this piece of number.
Encoding section 44 and 45 and then carry out Variable Length Code, and calculation code amount.
20 pieces in the section are carried out same processing, and the encoding amount of compute segment.
The encoding amount input quantizer determination section 50 of the section of calculating by the 1st encoding section 44 and 45.
The 1st relative importance value of 46,48 pairs of inputs of the 2nd relative importance value calculating part changes specific relative importance value respectively, and sets the 2nd relative importance value.Import the 2nd encoding section 47 and quantizer determination section 50 by the 2nd relative importance value that the 2nd relative importance value calculating part 46 is set.
Import the 2nd encoding section 49 and quantizer determination section 50 by the 2nd relative importance value that the 2nd relative importance value calculating part 48 is set.
The 2nd encoding section 47 that is made of quantizer 0 (not shown) is the same with encoding section 44 and 45 with the 2nd encoding section 49 that is made of quantizer 1 (not shown), will be by the given deviant of the 2nd relative importance value and the quantification number addition of quantizer, and this piece carried out quantification treatment.Encoding section 47 and 49 and then carry out Variable Length Code, and calculate the encoding amount of this piece.
20 pieces in the section are carried out same processing, and calculating with the section is the encoding amount of unit.Import quantizer determination section 50 by each encoding amount that the 2nd encoding section 47 and 49 is calculated.
Quantizer determination section 50 determines final encoding amount according to the encoding amount and the target code amount of each section of the 1st encoding section of importing 44 and the 45 and the 2nd encoding section 47 and 49, and will be final quantizer and final relative importance value with this final encoding amount corresponding quantitative device and relative importance value decision respectively.
For example, when the encoding amount of the 1st encoding section 45 is final encoding amounts, just quantizer 1 decision is final quantizer, the decision of the 1st relative importance value is final relative importance value.
When total encoding amount of the 2nd encoding section 47 is final encoding amounts, just quantizer 0 decision is final quantizer, the 2nd relative importance value decision of the 2nd relative importance value calculating part is final relative importance value.
Final quantizer (the quantification number of final quantizer) and final relative importance value input coding portion 51.
Encoding section 51 is by adding that the quantification number by after the given deviant of final quantizer and final relative importance value carries out quantification treatment to the orthogonal transform data of input, and then carries out Variable Length Code, and exports the coded data of this section.Coded data input and output terminal 52.Surpass at 1 o'clock by final quantizer and the given deviant of relative importance value, selected quantizer just becomes quantizer 1.
Lead-out terminal 52 with the coded data of input to record and output such as transmission medium.
Figure 17 is explanation determines final quantizer and final relative importance value to the section S0 of present embodiment figure.
In Figure 17, encoding amount 60 is encoding amounts of encoding section 44, and encoding amount 61 is encoding amounts of encoding section 45, and encoding amount 62 is encoding amounts of encoding section 47, and encoding amount 63 is encoding amounts of encoding section 49.
Encoding amount beyond the encoding amount 62 has surpassed target code amount 64.At this moment, just become final quantizer with encoding section 47 corresponding quantitative devices 0, the 2nd relative importance value that the 2nd relative importance value calculating part 46 calculates just becomes final relative importance value.
Figure 18 is used to illustrate that present embodiment is to the final quantizer of a certain section S1 and the figure of the decision of final relative importance value.
In Figure 18, encoding amount 65 is encoding amounts of encoding section 44, and encoding amount 66 is encoding amounts of encoding section 45, and encoding amount 67 is encoding amounts of encoding section 47, and encoding amount 68 is encoding amounts of encoding section 49.
Have only encoding amount 66 to surpass target code amount 69.
At this moment, just become final quantizer with encoding section 44 corresponding quantitative devices 0, the 1st relative importance value just becomes final relative importance value.
Though the encoding amount of encoding section 49 has the encoding amount that does not surpass the maximum of target code amount,, more preferential in order to make the 1st encoding section than the 2nd encoding section, so encoding section 44 becomes final encoding amount.
(embodiment 14)
The block diagram of the image signal coding device of present embodiment is identical with Figure 16.
In Figure 16, basic structure is the same with explanation in embodiment 13 with action.
In the present embodiment, the 2nd relative importance value calculating part 46 is decided to be minimum relative importance value 1 with whole relative importance values, and the 2nd encoding section 47 uses the quantizer 0 of quantization step maximum to encode.At this moment, the generation encoding amount of encoding section 47 becomes minimum generation encoding amount.
Like this, become minimum relative importance value by setting encoding amount, and in the 2nd encoding amount calculating part 57, select encoding amount to become minimum quantizer, just can avoid overflowing by the 2nd relative importance value calculating part 46.
(embodiment 15)
The block diagram of the image signal coding device of present embodiment is identical with Figure 16.
In Figure 16, basic structure is the same with action and explanation in embodiment 13.
In the present embodiment, the 1st encoding section 44 and 45 orthogonal transform data for input, the data after relative importance value is respectively whole AC coefficients to be divided by with 2 in 1 o'clock are as new orthogonal transform data.
And each quantizer is by carrying out quantification treatment by given deviant of the 1st relative importance value and the quantification number that quantizes the value (addition result be just be taken as 1) after the number addition at 2 o'clock.
In addition, encoding section 44 and 45 is carried out Variable Length Code respectively, and calculates the encoding amount of this piece.20 pieces in the section are carried out same processing, and calculate encoding amount by the section unit.
Import quantizer determination section 50 by the 1st encoding section 44 and 45 encoding amounts that calculate.
The 2nd encoding section 47 and 49 is 1 o'clock orthogonal transform data for input at relative importance value, and the data after whole AC coefficients are divided by with 2 are as new orthogonal transform data.
And each quantizer is by carrying out quantification treatment by given deviant of the 2nd relative importance value and the quantification number that quantizes the value (addition result be just be taken as 1) after the number addition at 2 o'clock.
In addition, encoding section 47 and 49 is carried out Variable Length Code respectively, and calculates the encoding amount of this piece.20 pieces in the section are carried out same processing, and calculating is the encoding amount of unit with the section.
Import quantizer determination section 50 by the 2nd encoding section 47 and 49 encoding amounts that calculate.
Encoding section 51 is 1 o'clock orthogonal transform data for input at relative importance value, and the data after whole AC coefficients are divided by with 2 are as new orthogonal transform data.
And encoding section 51 is by adding that the quantification number by after the given deviant of final quantizer and final relative importance value carries out quantification treatment to the orthogonal transform data of input, and and then carries out Variable Length Code.
As a result, just export the coded data of this section.Coded data is to efferent 52 outputs.Surpassing at 1 o'clock by final quantizer and the given deviant of final relative importance value, selected quantizer just becomes quantizer 1.
Figure 19 is owing to illustrate that present embodiment is to the final quantizer of a certain section S0 and the figure of the decision of final relative importance value.
In Figure 19, encoding amount 70 is encoding amounts of encoding section 44, and encoding amount 71 is encoding amounts of encoding section 45, and encoding amount 72 is encoding amounts of encoding section 47, and encoding amount 73 is encoding amounts of encoding section 49.
Encoding amount 70,71,72,73 is all less than target code amount 74.
The situation of present embodiment is decided to be final quantizer with encoding section 49 corresponding quantitative devices 1, and the 2nd relative importance value that the 2nd relative importance value calculating part 48 calculates is decided to be final relative importance value.
(embodiment 16)
The block diagram of the image signal coding device of present embodiment is identical with Figure 16.
In Figure 16, basic structure and action with in embodiment 13, illustrate the same.
In the present embodiment, the 2nd relative importance value calculating part 46 is minimum relative importance value 1 with whole the 2nd relative importance value decisions, and the quantizer 0 of the 2nd encoding section 47 usefulness quantization step maximums is encoded.
The 2nd relative importance value calculating part 48 is decided to be the highest relative importance value 0 with whole the 2nd relative importance values, and the quantizer 1 of the 2nd encoding section 49 usefulness quantization step minimums is encoded.In addition, identical with embodiment 15.
The situation of present embodiment, the generation encoding amount of encoding section 47 become minimum generation encoding amount, and the generation encoding amount of encoding section 49 becomes maximum generation encoding amount.
Figure 20 is used for the final quantizer of S0 and the figure of the decision of final relative importance value.
In Figure 20, encoding amount 75 is encoding amounts of encoding section 44, and encoding amount 76 is encoding amounts of encoding section 45, and encoding amount 77 is encoding amounts of encoding section 47, and encoding amount 78 is encoding amounts of encoding section 49.
Encoding amount 75,76,77,78 is all less than the target code amount.
At this moment, just become final quantizer with encoding section 49 corresponding quantitative devices 1, the 2nd relative importance value that the 2nd relative importance value calculating part 48 calculates just becomes final relative importance value.
(embodiment 17)
The block diagram of the image signal coding device of present embodiment is identical with Figure 16, in Figure 16, basic structure and action with in embodiment 13, illustrate the same.
To be quantizer determination section 50 will not surpass the encoding amount of maximum of target code amount as final encoding amount to present embodiment in the encoding amount of encoding section 44,45,47,49.
Below, with reference to the decision of Figure 18 explanation to final quantizer and the final relative importance value of a certain section S0.
In the present embodiment, it is final encoding amount that encoding amount 68 takes place, and becomes final quantizer with the 2nd encoding section 49 corresponding quantitative devices 1, and the relative importance value that the 2nd relative importance value calculating part 48 calculates becomes final relative importance value.
By the explanation of the foregoing description 13~17 as can be known, the code device of the control about encoding amount of the present invention, from the quantizer that can select, select L quantizer, the 1st relative importance value and L quantizer that use is set according to the feature of piece carry out quantification treatment to the orthogonal transform data of each piece, calculate L encoding amount, and then by using 2nd relative importance value different to set new T encoding amount calculating part of the calculating of carrying out encoding amount with the 1st relative importance value, just can avoid overflowing, thereby stable image quality can be provided.In addition, a kind of record is used for also belonging to the present invention by all or part of functional programs of all or part of key element of each image signal coding device of computer execution the foregoing description 13~17 and/or the program record medium of data, it is characterized in that: can utilize computer to read, the collaborative above-mentioned functions of carrying out of said procedure that reads and/or above-mentioned data and aforementioned calculation machine.
(embodiment 18)
Figure 22 is the block diagram of the image signal coding device of the embodiment of the invention 18.
In Figure 22, received image signal is transformed to the picture signal of low resolution by the 1st format converter 101.
Motion detector 102 will be as the picture signal record of the output of format converter 101 in video memory 104.In addition, detector 102 is according to coded frame be recorded in the picture signal of the reference frame in the video memory 104 and the decoding picture signal of the reference frame of coding and decoding has detected macro block (in luminance signal, the piece that 16 pixels in picture * 16 are capable) unit activity.
Motion compensator 103 is exported with the differential signal of the decoding picture signal of the reference frame that is detected by motion detector 102 picture signal of coded frame to macro block (MB).
105 pairs of each pieces of DCT device carry out DCT (discrete cosine transform) with the output signal of motion compensator 103 to be handled.
106 pairs of DCT coefficients of quantizer carry out quantification treatment.
The coefficient that 107 pairs of inverse quantizers have carried out quantification treatment by quantizer 106 carries out re-quantization to be handled.
The output of 108 pairs of inverse quantizers 107 of IDCT device is carried out inverse DCT and is handled.
Motion compensator 109 with the output of IDCT device 108 with carried out the decoding picture signal addition of the reference frame of motion compensation by motion compensator 103, generate the decoding picture signal, and store in the video memory 104.
The output of 110 pairs of quantizers 106 of variable length coder and the sign that comprises the appointment of motion vector carry out Variable Length Code.
111 pairs of decoding picture signals as the output of motion compensator 109 of the 2nd format converter are carried out format conversion, are transformed to the resolution identical with received image signal.
Differential signal maker 112 generates differential signal according to the output of received image signal and the 2nd format converter 111.Motion vector converter 113 uses the motion vector as the output of motion detector 102, according to the decoding differential signal of the reference frame in the video memory 114 of coded frame and coding and decoding by macro block unit's output differential signal use with reference to image.Motion compensator 115 is pressed the differential signal of macro block unit's output encoder frame and the differential signal of the signal of exporting from video memory 114.
116 pairs of each pieces of DCT device carry out the DCT processing with the differential signal of the output of motion compensator 115.
117 pairs of DCT coefficients of quantizer carry out quantification treatment.
The coefficient that 118 pairs of inverse quantizers have carried out quantification treatment by quantizer 117 carries out re-quantization to be handled.
The output of 119 pairs of inverse quantizers 118 of IDCT device is carried out inverse DCT and is handled.
Motion compensator 120 with the output of IDCT device 118 with carried out the decoding differential signal addition of the reference frame of motion compensation by motion compensator 115, generate the decoding differential signal, and store in the video memory 114.
The output of 121 pairs of quantizers 117 of variable length coder and the sign that does not comprise the appointment of motion vector carry out Variable Length Code.
As an example, the valid pixel numbers of considering to establish received image signal and be effective line number and be 1080 row, 1 row be the progressive signal of 1920 pixels (below, table is for the 1080p signal), be that effective line number is that the pixel counts of 720 row, 1 row are the situation of the progressive signal of 1280 pixels (below, show the signal into 720p) by the 1st format converter 101 from the picture signal of the low resolution of received image signal conversion.
At this moment, to not additional movement of 1080p signal vector, to 720p signal additional movement vector.
As long as only the 720p signal is had motion detector, can encode to 720p signal and 1080p signal.
In addition, motion vector converter 113 is according to the sampling rate of 720p signal and 1080p signal, will be as the motion vector of the output of motion detector 102 in level and vertical direction all with 3/2 times of use.
In addition, if lift an example of the 1st compressive flow, be exactly the primary image stream of MPEG.
For the 2nd compressive flow, grammer is identical with said primary image stream in MPEG, but content is the composition that is not included in the high image resolution picture signal in the 1st compressive flow (a high image resolution picture signal with will the decoding of the 1st compressive flow then carried out the differential signal of the signal of format conversion).
In addition, the 1st format converter 101 is in the relation of inverse transformation mutually with the 2nd format converter 111, can change the characteristic of employed filter.
In addition, obviously, the received image signal form of the 1st format converter is identical with the output image signal form of the 2nd format converter with the output image signal form of the 1st format converter and the received image signal form of the 2nd format converter.
(embodiment 19)
Figure 23 is the block diagram of the image signal coding device of the embodiment of the invention 19.
In Figure 23, be marked with identical symbol for moving identical piece with Figure 22.
In Figure 23, received image signal is transformed to the picture signal of low resolution from received image signal by the 1st format converter 101.Here, for convenience of explanation, horizontal direction and vertical direction all be transformed to 1/2 resolution.
Motion detector 102 records the output image signal of format converter 101 in the video memory 104, and, according to coded frame be recorded in the picture signal of the reference frame in the video memory 104 and the decoding picture signal of the reference frame of coding and decoding has detected macro block unit's (in luminance signal, the piece of 16 pixels in picture * 16 row) activity.
105 pairs every of DCT device (pieces of 8 pixels in picture * 8 row) carries out DCT with the differential signal of the output of motion compensator 103 and handles and export this signal.
Motion vector converter 113 makes the motion vector as the output of motion detector 102 all become 2 times of uses afterwards with vertical direction in the horizontal direction, according to the decoding differential signal of the reference frame in the video memory 114 with macro block unit's (in luminance signal, the piece of 32 pixels in picture * 32 row) output differential signal use with reference to image.
116 pairs every of DCT device (pieces of 16 pixels * 16 row) carries out DCT with the differential signal of the output of motion compensator 115 and handles and export this signal.
Motion compensator 120 with the output of IDCT device 203 with carried out the decoding differential signal addition of the reference frame of motion compensation by motion compensator 115, generate the decoding differential signal, and store in the video memory 114.
(in Figure 23, the action of above-mentioned piece in addition is referring to embodiment 18).
Here, though the format conversion of horizontal vertical is decided to be 1/2 than all,, level, Vertical Resolution can be carried out conversion with ratio arbitrarily independently.
Usually, level, vertical format conversion ratio are taken as N2/N1, L2/L1 (N1, N2, L1, L2 are natural numbers) respectively by DCT and inverse transformation thereof that DCT device 105 and IDCT device 108 carry out, are defined as the capable unit of M pixel * M.
At this moment, DCT and the inverse transformation of being undertaken by DCT device 116 and IDCT device 203 thereof just becomes that (pixel of M * N1/N2) * (M * L1/L2) goes unit.
And the size of macro block is the capable integral multiple of M pixel * M in the 1st high efficiency coding unit 181, is (the pixel of the M * N1/N2) * (integral multiple of row of M * L1/L2) in the 2nd high efficiency coding unit 182.
As object lesson, be the 1080p signal at received image signal, when low resolution picture signal is the 720p signal, then N1=3, N2=2, L1=3, L2=2.
Motion vector converter 113 makes the motion vector as the output of motion detector 102 all become 3/2 use afterwards in the horizontal vertical direction.
DCT device 105 and IDCT device 108 are so that 8 pixels * when 8 row carried out DCT and inverse transformation thereof, DCT device 116 and IDCT device 203 just all carried out DCT and inverse transformation thereof with 12 pixels * 12 row.
(embodiment 20)
Figure 24 is the block diagram of the image signal coding device of the embodiment of the invention 20.
In Figure 24, be marked with identical symbol for moving identical piece with Figure 22.
In Figure 24, received image signal is transformed to the picture signal of low resolution from received image signal by the 1st format converter 101.Here, for convenience of explanation, all be transformed to 1/2 resolution in the horizontal direction with vertical direction.
Motion detector 102 just as the picture signal record of the output of format converter 101 in video memory 104.In addition, detector 102 is according to coded frame and be recorded in the picture signal of the reference frame in the video memory 104 and the activity of the decoding picture signal detection macro block of the reference frame of coding and decoding (in luminance signal, the piece of 16 pixels in picture * 16 row) unit.
Motion compensator 103 is with the differential signal of picture signal with the decoding picture signal of the reference frame that is detected by motion detector 102 of macro block unit's output encoder frame.105 pairs every of DCT device (pieces of 8 pixels in picture * 8 row) carries out DCT with the differential signal of the output of motion compensator 103 to be handled.
The output of 108 pairs of inverse quantizers 107 of IDCT device is carried out inverse DCT and is handled this signal of back output.
Motion detector 201 records the picture signal and the differential signal of the coded frame of received image signal in the video memory 114, and detect the activity of macro block (in luminance signal, the piece of 32 pixels in picture * 32 row) according to coded frame and the picture signal that is recorded in the reference frame in the video memory 104.
116 pairs every of DCT device (pieces of 16 pixels in picture * 16 row) carries out DCT with the differential signal of the output of motion compensator 115 and handles and export this signal.
117 pairs of DCT coefficients as the output of DCT device 116 of quantizer carry out quantification treatment.
The output of 203 pairs of inverse quantizers 118 of IDCT device is carried out inverse DCT and is handled.
Motion compensator 120 with the output of IDCT device 203 with carried out the decoding differential signal addition of the reference frame of motion compensation by motion compensator 115, generate the decoding differential signal, and store in the video memory 114.
(in Figure 24, the action of above-mentioned piece in addition is referring to embodiment 18).
The 1st format converter is to be transformed to 1/2 with the format conversion of horizontal vertical than all, still, level, Vertical Resolution can be carried out conversion with ratio arbitrarily independently.
Usually, level, vertical format conversion ratio are taken as N2/N1, L2/L1 (N1, N2, L1, L2 are natural numbers) respectively by DCT and inverse transformation thereof that DCT device 105 and IDCT device 108 carry out, are defined as the capable unit of M pixel * M.
At this moment, DCT and the inverse transformation of being undertaken by DCT device 116 and IDCT device 203 thereof just becomes that (pixel of M * N1/N2) * (M * L1/L2) goes unit.
And the size of macro block can be the capable integral multiple of M pixel * M in the 1st high efficiency coding unit 191, can be (the pixel of the M * N1/N2) * (integral multiple of row of M * L1/L2) in the 2nd high efficiency coding unit 192.
As object lesson, be the 1080p signal at received image signal, when low resolution picture signal is the 720p signal, then N1=3, N2=2, L1=3, L2=2.
DCT device 105 and IDCT device are so that for example 8 pixels * when 8 row carried out DCT and inverse transformation thereof, then DCT device 116 and IDCT device 203 just carried out DCT and inverse transformation thereof with 12 pixels * 12 row.
(embodiment 21)
Figure 25 is the block diagram of the picture signal code translator of the embodiment of the invention 21.
Present embodiment is and a for example embodiment of the corresponding picture signal code translator of code device of the present invention of explanation in embodiment 18,19.
In Figure 25,401 pairs the 1st compressive flows of variable length decoder carry out the decoding of appointment, and output movement vector and decoded signal.
Signal after inverse quantizer 402 will be deciphered by decoder 401 carries out the re-quantization processing by the quantization step of appointment, and output DCT coefficient.
403 pairs of DCT coefficients of IDCT device carry out IDCT to be handled.
Motion compensator 404 is the output of IDCT device and decoding picture signal addition by the reference frame of motion vector appointment, and stores in the video memory 405.
The 2nd format converter 406 will be deciphered picture signal and carry out format conversion and export this signal.407 pairs the 2nd compressive flows of variable length decoder carry out the decoding of appointment.
Signal after inverse quantizer 408 will be deciphered by decoder 407 carries out the re-quantization processing by the quantization step of appointment, and output DCT coefficient.
409 pairs of DCT coefficients of IDCT device carry out IDCT to be handled.
Motion vector decoder 412 is deciphered motion vector from the 1st compressive flow, and carries out the conversion of appointment, exports the motion vector that the 2nd compressive flow is used then.
Motion compensator 410 with the output of IDCT device 409 with by decoding differential signal addition as the reference frame of the motion vector appointment of the output of motion vector decoder 412, generate the decoding differential signal, and store in the video memory 411.
Adder 413 generates the decoding picture signal of the 1st picture signal with the output signal of compensator 410 and the output signal addition of the 2nd format converter 406.
At level, vertical resolution ratio is N2/N1, L2/L1 (N1, N2, L1, L2 are natural numbers), IDCT device 403 when handling for carry out IDCT with M * M, and then IDCT device 209 is preferably with (M * N1/N2) * (M * L1/L2) carry out IDCT to handle.
As object lesson, be 1080p signal and when being the 720p signal in high image resolution picture signal by the low resolution picture signal of the 1st decoding unit 471 decodings by the decoding of the 2nd decoding unit 472, then N1=3, N2=2, L1=3, L2=2.
Motion vector decoder 412 uses the motion vector in the 1st compressive flow after level and vertical direction all become 3/2 times.
When IDCT device 403 carried out the inverse transformation of 8 * 8 DCT, then IDCT device 409 just can carry out the inverse transformation of 12 * 12 DCT.
(embodiment 22)
Figure 26 is the block diagram of the picture signal code translator of the embodiment of the invention 22.
In Figure 26, be marked with identical symbol for moving identical piece with Figure 25.
Present embodiment is and a for example embodiment of the corresponding picture signal code translator of code device of the present invention of explanation in embodiment 18,19.
Motion vector converter 501 will as the motion vector of the output of variable length decoder 401 by the conversion of appointment after the motion vector used of output the 2nd compressive flow.
Motion compensator 410 with the output of IDCT device 409 with by decoding differential signal addition as the reference frame of the motion vector appointment of the output of motion vector converter 501, generate the decoding differential signal, and store in the video memory 411.
Differential signal after adder 413 will be deciphered and decoding picture signal addition as the output of the 2nd format converter 406 generate the decoding picture signal of the 1st picture signal.
(about the action of the piece beyond above-mentioned, referring to embodiment 21)
Being N2/N1, L2/L1 (N1, N2, L1, L2 are natural number), IDCT device 403 at level, vertical format conversion ratio carries out DCT when handling with M * M, in IDCT device 409, preferably with (M * N1/N2) * (M * L1/L2) carry out IDCT to handle.
As object lesson, be the 1080p signal at received image signal, when low resolution picture signal is the 720p signal, then N1=3, N2=2, L1=3, L2=2.
Motion vector converter 501 uses the motion vector as the output of variable length decoder 401 after level, vertical direction all become 3/2 times.
When IDCT device 403 carried out the inverse transformation of 8 * 8 DCT, then IDCT device 409 just can carry out the inverse transformation of 12 * 12 DCT.
(embodiment 23)
Figure 27 is the block diagram of the picture signal code translator of the embodiment of the invention 23.
In Figure 27, be marked with identical symbol for moving identical piece with Figure 25.
Present embodiment is and a for example embodiment of the corresponding picture signal code translator of code device of explanation in embodiment 20.
403 pairs of DCT coefficients of IDCT device carry out the inverse transformation of 8 * 8 DCT.
The 2nd format converter 406 will be deciphered picture signal and carry out format conversion and output.Here, for convenience of explanation, all be transformed to 2 times resolution in the horizontal direction with vertical direction.
Variable length decoder 407 carries out the decoding of appointment and output movement vector and decoded signal with the 2nd compressive flow.
Inverse quantizer 408 carries out re-quantization with decoded signal by the quantization step of appointment to be handled, and output DCT coefficient.
IDCT device 409 carries out the DCT data inverse transformation of 16 * 16 DCT.
Motion compensator 410 with the output of IDCT device 409 with by decoding differential signal addition as the reference frame of the motion vector appointment of the output of variable length decoder, generate the decoding differential signal, and store in the video memory 411.
Differential signal that adder 413 will have been deciphered and decoding picture signal addition as the output of the 2nd format converter 406 generate the decoded signal of the 1st picture signal.
Here, though level is taken as 1/2 with vertical format conversion than all,, level, vertical transformation ratio can be ratios arbitrarily.
Usually, level, vertical ratio are taken as N2/N1, L2/L1 (N1, N2, L, L2 are natural numbers) respectively, when M * M unit was adopted in the inverse transformation of the DCT that is undertaken by IDCT device 403, then the inverse transformation of the DCT that is undertaken by IDCT device 409 just became the (M * N1/N2) * (unit of M * L1/L2).
As object lesson, be the 1080p signal at received image signal, when low resolution picture signal is the 720p signal, then N1=3, N2=2, L1=3, L2=2.
When IDCT device 403 adopted the inverse transformation of 8 * 8 DCT, then IDCT device 409 just became the inverse transformation of 12 * 12 DCT.
By the foregoing description 18~23 as can be known, of the present inventionly picture signal is carried out level encoded image signal code device or its code translator has following function:
(1) when low resolution picture signal is encoded, the motion vector that is detected is used in the coding of the differential signal of high image resolution, so 1 of motion detector is just much of that.By in the 2nd compressive flow, not comprising the information of motion vector, just comparablely more effectively picture signal was carried out the stratification coding in the past.
(2) use and the consistent orthogonal transform of resolution ratio of hanging down the resolution picture signal by coding the differential signal of high image resolution, on picture, the high image resolution picture signal is consistent with the section of low resolution picture signal, so, can prevent that the compression distortion from intersecting, thereby can carry out effective stratification coding with high image quality.
(3) by having the motion vector decoder, can decipher the compressive flow of having encoded with simple structure.
(4) by having the motion vector converter, can decipher the compressive flow of having encoded with simple structure.
(5) by having IDCT device 409, can decipher the compressive flow of having encoded.
In addition, all or part of of each image signal coding device of the foregoing description 18~23 or each structural element of code translator can be hardware, also can be the software that has with the function identical functions of this hardware.
In addition, a kind of record is used for also belonging to the present invention by all or part of functional programs of all or part of key element of image signal coding device of computer execution the foregoing description 18~23 or code translator and/or the program record medium of data, it is characterized in that: can be read by computer, said procedure that reads and/or above-mentioned data and aforementioned calculation machine be executable operations synergistically.