GB2158321A - Arrangement for rotating television pictures - Google Patents

Abstract

An arrangement for rotating a television picture about an axis perpendicular to the plane of the picture and/or one or both of two axes in the plane of the picture comprises a control unit (1), which may contain selection switches, control potentiometer and analogue to digital converters, which produces control signals which are fed to a computer 2. The control unit (1) produces signals to specify the positions of the axes of rotation and the speed of rotation about those axes. The computer (2) together with a hardware calculator (3) produce control signals for a store write control unit (4) and a video signal interpolator (5). The interpolator (5) receives video signals from a picture source (6) and feeds interpolated signal samples to a field store (7) where the samples are inserted in the desired location by the write control unit (4). The output of the field store (7) is fed to an output (8) which may be connected to other units in an effects generator or signal routing system. The control unit (1) provides signals to the computer (2) which specify the point about which rotation is to take place (centre of rotation) and the point about which perspective is to take place (centre of perspective). Further the angles of rotation and perspective may be set by means of a conventional cursor control while the angles of rotation and perspective may be set by switches or, if continuous resolution is desired potentiometer control. <IMAGE>

Description

SPECIFICATION Arrangement for rotating television pictures The invention relates to an arrangement for rotating a television picture about an axis perpendicular to the plane of the picture and/or one or both of two axes in the plane of the picture.

In this specification a rotation of the picture about an axis perpendicular to the picture will be referred to as rotation and a rotation of the picture about an axis in the plane of the picture will be referred to as perspective.

It is an object of the invention to enable the provision of an arrangement capable of rotating a television picture about one or more axes in or perpendicular to the plane of the picture.

The invention provides an arrangement for rotating a television picture about an axis perpendicular to the picture and/or one or both of two axes in the plane of the picture comprising means for selecting the position of the axis perpendicular to the picture and/or the positions of the two axes in the plane of the picture, means for selecting angles of rotation about each of the axes, means for calculating the positions of the corners of the rotated picture, means for determining the length and gradient of the lines joining the corners of the rotated picture, a field store for storing a television field, a store write control including address counters, means for presetting the address counters at the start of each field so that a seiected starting address is obtained for the calculated position of one of the corners of the rotated picture, means for generating incrementing or decrementing instructions for the address counters from the length and gradient values of the lines joining the corners of the rotated pixture, and means for generating interpolation values from the length and gradient values of the lines joining the corners of the rotated picture.

The means for calculating the positions of the corners of the rotated picture may comprise means for multiplying the positions of the corners of the non rotated picture by one or both of the matrices.

cos A -sin A 0 0 sin A cos A 0 0 | 0 '0 0 1 CI Xr-XrCosA-YrsinA Yr+XrsinA-YrcosA 0 1 and cos C 0 sin C sin B sin C cos B -sin Bcos C CI -cos B sin C sin B cos Bcos C CI Xp-XpcosB-YpsinBsinC Yp-YpcosB YpsinBcosC-XpsinC 1 where A is the angle of rotation B is the angle of rotation about the horizontal axis (line scan direction) C is the angle of rotation about the vertical axis (field scan direction) Xr, Yr are the co-ordinates of the centre of rotation Xp, Yp are the co-ordinates of the centre of perspective.

The interpolator may be fed with values xlp = (y.Ax'V - #x'V-xF#y'V)/(#y'V.#x'H - x'V.Ay'H) and YIP - (y.Ax'H - xAy'H)/(Ay'V.Ax'H - X)V. L\ylH) where x,p and y,p are the distances of a point x,y in the output plane from a sample in the input plane in the line and field scan directions respectively, XF and y, are the distances of a sample in the input plane from the point x,y in the output plane, Ax'V and hy'V are the incremented positions of a picture element in line n of an input picture in the input plane, and AXFH and hy'H are the incremented positions of line m in the input picture plane.

so that interpolation co-efficients may be determined by the interpolator.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 shows in block schematic form an arrangement according to the invention for rotating a television picture about an axis perpendicular to the plane of the picture and/or one or both of two axes in the plane of the picture, Figure 2 shows the outline of a rotated picture on a television screen, Figure 3 shows on an enlarged scale a portion of the rotated picture of Fig. 2 to illustrate the location of samples in the input plane and a method of traversing the samples in the output plane, Figure 4 shows the superimposition of the input and output planes, Figure 5 illustrates assumptions made for interpolation of the input signal, Figure 6 illustrates in block schematic form the hardware calculator of Fig. 1, Figure 7 shows a part of the hardware calculator of Fig. 6 in greater detail, and Figure 8 shows a further part of the hardware calculator of Fig. 6 in greater detail.

Fig. 1 shows in block schematic form an arrangement for rotating a television picture about an axis perpendicular to the picture and/or two axes parallel to the plane of the picture. The arrangement comprises a control unit 1 which may contain selection switches, control potentiometers, and analogue to digital converters to produce control signals which are fed to a computer 2. The control unit 1 produces signals to specify the positions of the axes of rotation and the speed of rotation about those axes. The computer 2 together with a hardware calculated 3 produce control signals for a store write control unit 4 and a video signal interpolator 5.The interpoiator 5 receives video signals from a picture source 6 and feeds interpoiated signal samples to a field store 7 where the samples are inserted in the desired location by the write control unit 4. The output of the field store 7 is fed to an output 8 which may be connected to other units in an effects generator or signal routing system. The term field store is intended to mean a store capable of storing one or more fields of a television picture and, in particular, includes a frame store which stores two fields of a television picture.

In this specification a rotation of the picture about an axis perpendicular to the plane of the picture will be referred to as rotation while a rotation of the picture about an axis parallel to the plane of the picture will be referred to as perspective. While, in principle the axes parallel to the plane of the picture may extend in any direction in the following description it will be assumed that these axes are horizontal and vertical.

The control unit 1 provides signals to the computer 2 which specify the point about which rotation is to take place (centre of rotation) and the point about which perspective is to take place (centre of perspetive). Further the angles of rotation and perspective are set by the control unit 1. The centres of rotation and pespective may be set by means of a conventional cursor control while the angles of rotation and perspective may set by switches or, if continuous resolution is desired potentiometer control. The form and construction of such control units is well known to those skilled in the art.

Fig. 2 shows a television picture screen 10 on which the outline of a picture 11 which has been rotated about axes parallel to the screen and, optionally, about an axis perpendicular to the screen is shown. The computer 2 determines from the centres of rotation and perspective and the three angles of rotation about the three axes the positions of the points a, b, c, and d which define the corners of the rotated picture by multiplying the positions of the corners of the screen by the matrices cos A -sin A o 0 sin A cos A 0 0 0 0 1 0 Xr-XrCosA-YrsinA Yr+XrsinA-YrcosA 0 1 in the case of rotation, and cos C 0 sin C 0 sin B sin C cos B -sin B cos C 0 -cos B sin C sin B cos B cos C 0 Xp-XpcosB-YpsinBsinC Yp-YpcosB YpsinBcosC-XpsinC 1 in the case of perspective, where A is the angle of rotation B is the angle of perspective about the horizontal axis, C is the angle of perspective about the vertical axis, Xr, Yr are the co-ordinates of the centre of rotation, and Xp, Yp are the co-ordinates of the centre of perspective.

If rotation and perspective are applied simultaneously the positions of the points a, b, c, and d are determined by multiplying the positions of the corners of the screen by one matrix and then multiplying the result by the other matrix.

The field store 7 is arranged as an analogue of the picture screen and contains 720 locations for storing picture samples of each of 576 lines, i.e. 720 x 576 locations thus any store location can be considered to have co-ordinates (x, y) defining the position at which that sample will be displayed and starting from (0, 0) in the top left hand corner and ranging to (719, 575) in the bottom right hand corner. These co-ordinates are referred to as being in the output plane.

The picture received from the picture source 6 will likewise consist of 720 samples per line and 576 lines per frame. When perspective is performed these samples have to be compressed to fit within a picture frame such as that shown as 1 1 in Fig. 2. Thus it is necessary to convert each input sample point within the picture frame 11 to co-ordinates relating to the output plane.

Referring to Figs. 2 and 3 and given the four points a(x1, y,), b(x2, y2), c(x3, y3), and d(x4, y4) defined by a perspective projection incremental terms AxH and AyH can be derived which allow the line ab (a horizontal input line) to be traversed in the output plane.

ThusAxH=(x2-x1)/m) 1 AyH = (Y2 - y1)/m) where m is the number of picture elements per line Similarly for line ad AxV = (x4 - x1)/n) 2 AyV = (y4 - y1)/n) where n is the number of lines per frame.

By considering the incremental values AxH' and AyH' along cd the change of increments for sequential horizontal lines can be derived as follows.

From dc AxH' = (x3 - x4)/n AyH' = (y3 - y4)/n Thus A2xH = (xH - AxH')/n) 3 A2yH = (AyH - AyH')/n) where A2xH and A2yH are the changes in increments #xH and AyH with sequential input lines.

From which the increments to traverse any horizontal line Ax'H = AxH + Ynli2xH) Ay'H = AyH + yA2yH) where y, is the horizontal line number starting from zero at line ab.

Similarly by considering the incremental values along bc the change of increments for picture elements along the horizontal line A2xV and E2yV can be derived ç2xV = (#xV' - AxV)/m) A2yV = (AyV' - hyV)/m) Hence #x'V = AxV + xnA2xV) 6 #y'V = AyV + xA2yV) where xn is the picture element number starting from zero at line ad.

By combining equations 4) and 6) with the starting point x1, y, the values of co-ordination x', y' of any input picture element in the output plane can be determined.

x' = x, + ynAxV + xn (AxH + YnlS2xH) y' = y, + yAyV + xn (AyH + yA2yH).

Fig. 3 illustrates a number of picture elements 12 of the input picture signal in the rotated input plane starting from the point a and the incremental additions to define their co-ordinates in the output plane.

Consider now the integer part of x' as x' and the integer part of y' as y'. These represent x' and y' truncated to fit the possible output plane positions. The integer parts x' and y' give the address for the field store while the fractional remainders x, and y, give the distance from x', y' of the addressed point in the output plane. Fig. 4 illustrates this construction with the input plane 30 shown by chain dotted lines and the output plane 31 by solid lines. In order to provide the interpolator 5 with the necessary co-efficients it is necessary to convert from output plane co-ordinates to input plane co-ordinates.To achieve this it is assumed that over the small area over which interpolation takes places adjacent lines are parallel and adjacent sample positions are parallel, i.e. that four adjacent sample positions in the input plane are at the corners of a parallelogram. Fig. 5 illustrates the assumptions made in deriving the equations. In practice the angle between two adjacent lines is always less than +,. Thus the interpolator positions x,p and y,p are defined by x,, = (y,. bx'V - XF. Ay'V)/(Ay'V . Ax'H - Ax'V. Ay'H) Yip = (y.Ax'H - XF.YIH)/(ylV. L\x'H - Ax'V.Ay'H) Since the transformation of input points to the output plane in areas of compression is being considered values of x,p and y,p greater than one will be obtained in some instances.These values indicate that these positions are redundant and they are consequently discarded.

Thus the computer 2 produces once per frame ten parameters x1, y1, AxH, hyH, A2xH, A2yH, AxV, AyV, A2xV, and A2yV and supplies them to the hardware calcultor 3. The calculator 3 then provides at the picture element rate x' and y' to the store write control unit and x,p and y,p to the interpolator 5.

The hardware calculator 3 is shown in block schematic form in Fig. 6 and comprises first and second identical blocks 60 and 61, an inverter 62 and a third block 63. The block 60 has five inputs 70 to 74 to which the parameters AxH, A2xH, A2xV, AxV and x, respectively are applied by the computer 2 while the block 61 has five inputs 80 to 84 to which the parameters AyH, A2yH, A2yV, AyV, and y, respectively are applied by the computer 2.The block 60 has five outputs 75 to 79 which provide the calculated values x', x,, Ax'H, Ax'V, and Ax'V. lvy'H, while the block 61 has five outputs 85 to 89 which provide the calculated values y', y,, hy'H, hy'V, and Ax'H. hy'V. The block 60 has a further input 50 to which the calculated value of hy'H is applied while the block 61 has a further input 51 to which the calculated value of Ax'H is applied.The block 60 has a further output 52 which provides an instruction to the sample address counter in the store write control 4 to increment or decrement at the sample rate while the block 61 has a further output 53 which provides an instruction for the line address counter to increment ot decrement at the sample rate. It should be noted that the outputs x' and y' set the starting points for the address counters at the start of each line.

The block 63 has eight inputs 90 to 97 to which are connected x,, Ax'H, Ax'V, the output of inverter 62 (x'V. Ay'H), Ax'H.Ay'V, y,, by'H, and hy'V respectively. The block 63 has two outputs 98 and 99 at which the calculated values of x,p and Y,p respectively are available.

Fig. 7 shows an embodiment of the block 60. The block 61 is identical to the block 60 and thus will not be separately described, it being sufficient to note that block 60 calculates the x co-ordinates while block 61 calculates the y co-ordinates.

As shown in Fig. 7 the inputs 70 to 74 are connected to respective 1 2 bit latches 100 to 104, these latches having tri-state outputs. The input 74 is also connected to pre-setting inputs of an up/down counter 105. The information at inputs 70 to 74 is clocked into the latches and counter during the field blanking interval. The information presented by the computer 2 at the input 74 is split into an integer portion which is fed to the counter 105 and a fractional portion which is fed to the latch 104. Alternatively separate inputs may be provided for the integer and fractional portions of x,. The output of the latch 100 is connected to a first input of an adder 106 while the output of the latch 101 is connected to a latch 107 whose output is connected to a second input of the adder 106.The latch 107 is resetable to zero output by the application of a reset signal. The output of the adder 106 is connected to the input of latch 108 which has a tri-state output which is connected to the first input of the adder 106 and to the output 77. For the first line the output of latch 100 is enabled and that of latch 107 set to zero and hence the contents of latch 100 are read into latch 1 08. The output of latch 100 is then disabled until the start of the next field while the output of latch 108 is enabled to produce the value axlH at output 77 and at the first input of the adder 106. The contents of latch 101 is read into the latch 107 and A2xH is added to the contents of latch 108 at the line rate by appropriate clocking of the inputs of latch 108. A similar arrangement is formed by latches 102 and 103 together with an adder 109, a latch 110 which has outputs pre-settable to zero, and a latch 111 having tri-state outputs. This produces the value Ax'V at output 78 but in this case the latch 111 is clocked at the picture element rate. The outputs of latches 102 and 111 are also connected to a first input of a multiplier 112 whose second input is connected to input 50 which receives the value aylH from the unit 61. The output of the multiplier 112 is accumulated in an arrangement comprising an adder 113 and a latch 114 whose output is presettable to zero and which is clocked at the picture element rate so that the latch 114 contains the value Qx'V. Ay'H, the output of latch 114 being connected to output 79.

The output of latches 103 and 104 are connected to a first input of an adder 115 which together with a latch 116 having its output presettable to zero forms an accumulator. The fractional part of x, is entered into the acumulator at the start of a field and the value of AxV is added at the line rate. The contents are monitored 117 to determine when the added fractional values add to the next integer (up or down) and an output is produced to increment or decrement the counter 105 which sets the position from which the sample counter in the store write control unit 4 starts each line.

The outputs of latches 108 and 116 are fed to first and second inputs of a multiplexer 118 which selects the output of latch 116 to feed to an accumulator which comprises an adder 119 and a latch 120 which has outputs resettable to zero at the start of each line. The multiplexer then connects the output of latch 108 to the accumulator which is clocked at the picture element rate. The output of the accumulator is monitored 121 and the output 52 activated to present increment or decrement signals to the sample address counter in the store write control unit 4. The output of the latch 120 is fed to output 76 as the value XF.

It should be noted that the values AxH, A2xH, AxV, A2xV, and of may be positive or negative and hence the accumulators may increment or decrement. Thus the output values of outputs 75 and 52 may be such that the address counters in the store write control unit are caused to increment or decrement.

The latches 100, 101, 102, 103 108, and 111 may be formed by standard TTL type 74LS374 integrated circuits, the input values being in 12 bit parallel form, while the latches 107, 110, 114, 116, and 120 may be type 74LS273 integrated circuits and the counter 105 a type 74LS 1 91 integrated circuit. A suitable circuit for the multiplier 112 is that sold by TRW Inc under the type reference MPY1 12K while the adders by be formed from type 74LS283 integrated circuits.

Fig. 8 shows an embodiment of the block 63 which calculates x,p and y,p from inputs xp, #x'V,yF, hy'H, hy'V, Ax'V.Ay'H and AxH . by'V. The value x, is applied to input 90 and inverted by an inverter 140 whose output is connected to a first input of each of two multipliers 141 and 142. The value dy'V is applied to the second input of multiplier 141 via input 97 while the value of hytH is applied to the second input of multiplier 142 via input 96. The multiplier 141 produces the value -x.Ay'V at its output and this is applied to a first input of an adder 143.The multiplier 142 produces the value -xF - x.Ay'H at its output and this applied to a first input of an adder 144.

Similarly the value YF is applied via input 95 to a first input of each of two multipliers 145 and 146. The value of Ax'H is applied via input 91 to the second input of multiplier 145 while the value of bx'V is applied via input 92 to the second input of multiplier 146. The multiplier 145 produces the value y.Ax'H at its output and this is applied to the second input of the adder 144. Similarly the multiplier 146 produces the value y.Ax'V at its output and this is applied to the second input of the adder 143.

The value Ax'V. Ay'H is fed via input 93 to the first input of an adder 147 while the value y'V.Ax'H is fed via input 94 to the second input of the adder 147. The adder produces the value Ay'V.Ax'H - Ax- 'V.Ay'H at its output and this is fed to a latch 148 whose output addresses a read only memory 149 which is programmed to convert the value #y'V.#x'H-#x'V - Ax'V.Ay'H to 1 /(Ay'V - Ax- 'V.Ay'H). This value is fed via a latch 150 to a first input of a multiplier 151 and to a first input of a multiplier 152.The output of the adder 144 is fed to the second input of the multiplier 151 while the output of the adder 143 is fed to the second input of the multiplier 1 52. The output of multiplier 152 which is the value (y . Ax'V - XF. Ay'Y)/(Ay'V. Ax'H - # Ax'V. Ay'H) is applied to output 98 as the value x,p, while the output of the multiplier 151 is inverted in an inverter 153 to produce a value -(yF.Ax'H - x.Ay'H)/(Ay'V.Ax'H - Ax'V.Ay'H) which is applied to output 99 as the value Yip The multipliers 141, 142, 145, 146, 1 51 and 1 52 may be those sold by TRW Inc. under the type reference MPY08HUJ since the values are truncated to 8 bits before multiplication while the latches 148 and 1 50 may be formed from TTL type 74LS273 integrated circuits and the adders 143 and 144 may be formed from TTL type 74LS283 integrated circuits.

The values x,p and Y,p are fed to the interpolator 5 to enable the signal from the picture source to be interpolated as required before being applied to the field store 7.

Claims (4)

1. An arrangement for rotating a television picture about an axis perpendicular to the picture and/or one or both of two axes in the plane of the picture comprising means for selecting the position of the axis perpendicular to the picture and/or the positions of the two axes in the plane of the picture, means for selecting angles of rotation about each of the axes, means for calculating the positions of the corners of the rotated picture, means for determining the length and gradient of the lines joining the corners ofthe rotated picture, a field store for storing a television field, a store write control including address counters, means for presetting the address counters at the start of each field so that a selected starting address is obtained for the calculated position of one of the corners of the rotated picture, means for generating incrementing or decrementing instructions for the address counters from the length and gradient values of the lines joining the corners of the rotated picture, and means for generating interpolation values from the length and gradient values of the lines joining the corners of the rotated picture.

2. An arrangement as claimed in Claim 1, in which the means for calculating the positions of the corners of the rotated picture may comprise means for multiplying the positions of the corners of the non rotated picture by one or both of the matrices cos A -sin A C 0 sin A cos A 0 0 O 0 1 0 Xr-XrCosA-YrsinA Yr+XrsinA-YrcosA C 1 and cos cos C 0 sin C C sin sin B sin C cos B -sin Bcos C 0 -cos B sin C sin B cos Bcos C 0 Xp-Xp.cosB-YpsinBsinC YpYpcosB YpsinBcosC-XpsinC 1 where A is the angle of rotation B is the angle of rotation about the horizontal axis (line scan direction) C is the angle of rotation about the vertical axis (field scan direction) Xr, Yr are the co-ordinates of the centre of rotation XP, Yp are the co-ordinates of the centre of perspective.

3. An arrangement as claimed in Claim 1 or Claim 2, in which the interpolator is fed with values x,p = (yF.Ax'V - xFAy'V)/(Ay'V.Ax'H - L\x'V.y'H) and Yip = - (y.Ax'H - xAy'H)/(Ay'V.Ax'H - Ax'V. Ay'H).

where x,p and y,p are the distances of a point x, y in the output plane from a sample in the input plane in the line and field scan directions respectively, x, and y, are the distances of a sample in the input plane from the point x, y in the output plane, Ax'V and Ay'V are the incremented positions of a picture element in line n of an input picture in the input plane, and Ax'H and Ay'H are the incremented positions of line m in the input picture plane.

so that interpolation co-efficients may be determined by the interpolator.

4. An arrangement for rotating a television picture about an axis perpendicular to the picture and/or one or both of two axes in the plane of the picture substantially as described herein with reference to the accompanying drawings.