GB2243515A - Digital video signal processing using lookup tables - Google Patents

Abstract

In a digital video signal processor, two lockup tables 46, 48 with 16-bit address inputs receive 8 bits from a respective video signal input 42, 44 and 8 bits from control inputs 50, 52 connected to a control signal input 54 which may be common to the two tables. The outputs of the lookup tables are added in an adder 60. By appropriately loading the lockup table 5, any desired signal processing operation from a large range of transformations may be applied, e.g. fading, mixing, production of a negative image, "solarisation" (omitting the least significant bits), brightening or darkening by a constant amount. In a colouriser, a predetermined colour in the input image may be changed whenever it occurs. <IMAGE>

Description

DIGITAL VIDEO SIGNAL PROCESSING Background of the Invention This invention relates to the processing of digital video signals, and particularly but not exclusively to the mixing or fading and to the colourising of digital video signals.

A conventional video fader and colourising circuit 10 is shown in block diagram form in Figure 1 of the drawings. The circuit has two video signal inputs 12 and 14 for receiving video signal 1 and video signal 2 respectively. Each of these inputs is connected to a respective multiplier 16,18 which receives at its other input a respective fader control signal from an input 20,22. The outputs of the multipliers 16,18 are applied to an adder 24. A colouriser 26 is connected to the output of the adder1 and receives a colouriser control signal at an input 28.

The output 30 of the circuit then provides a faded colourised video signal.

Such a circuit has a number of disadvantages. It includes two multipliers, which are relatively complex circuits, and a separate colouriser. Normally colourisers either have a limited number of possible colour transformations which they can effect, or require a large amount of special purpose circuitry.

Also, the mixing provided by the fader is linear due to the multiplier operation, so that gamma correction leads to an incorrect mix. Finally, three control signals are required to mix and colourise two signals.

Summary of the Invention The invention in its various aspects is defined in the appended claims, to which reference should now be made.

A preferred embodiment of the invention, which is described in more detail below with reference to the drawings, can carry out for example a combined mixing (or fading) and colourising operation. Two lookup tables constituted by static RAM with 16-bit address inputs receive 8 bits from a respective video signal input and 8 control bits, which are preferably the same for the two lookup tables. The outputs of the two lookup tables are added together.

The digital signals are preferably in accordance with CCIR Recommendations 601 and 656. In the preferred embodiment the video signal is controlled by a control signal of a similar character to the video signal, or at least one component of it.

A bank of high speed static memory provides a mapping of the input luminance and chrominance to the output, controlled and selected by the control signal. This means that each pixel on the screen can have its value (chroma and luma) transformed in one of a large number of arbitrary ways. Examples of these transformations are fading, colourising and inverting.

Brief Description of the Drawings The preferred embodiment of the invention will be described in more detail below by way of example, with reference to the drawings, in which: Figure 1 (referred to above) is a block diagram of a known type of fader and colourising circuit for video signals; Figure 2 is a block circuit diagram of a circuit embodying the invention illustrating its use in a fading operation; Figure 3 is a more detailed circuit diagram of the circuit of Figure 2; and Figure 4 is a diagram of the circuit of Figure 2, illustrating its use in a colourising operation.

Detailed Description of the Preferred Embodiment The preferred embodiment of the invention illustrated in Figure 2 comprises a simple circuit which is usable with great flexibility in digital video signal processing, particularly in fading and colourising operations. The circuit is easy to construct, and yet can provide a much greater variety of transfer functions than the known circuit of Figure 1.

The circuit 40 of Figure 2 comprises two video inputs 42,44 for receiving respectively a video signal 1 and a video signal 2 which are to be mixed. These video signals each comprise a digital signal with 8-bit samples. Video input 42 is connected to 8 of the address inputs of a random access memory 46 and video input 44 is connected to 8 of the address inputs of a random access memory 48. Each of the random access memories has 65,536 (64K) locations each capable of storing an 8 bit number, and can be constituted by a static random access memory (RAM) such as Hitachi type number HM6208P-35. Such a 64K memory will have 16 address bits in total. Eight of these are connected to the respective video inputs, as noted above. The other eight bits constitute a control input 50,52 to each of the SRAMs 46,48.The two control inputs 50,52 are connected to a control signal input 54 of the circuit, which can receive an 8 bit control signal for application to the two SRAMs.

When a 16-bit address is applied to one of the SRAMs, one of its 64K locations is addressed and the 8 bit contents thereof are applied to the SRAM data output 56,58. These outputs are then applied to a combining circuit 60 in the form of an adder, the output 62 of which constitutes the output of the circuit 40.

Those skilled in the art will appreciate that the SRAMs 46,48 are being used as lookup tables. Prior to operation, the SRAM is loaded with the lookup table required for the specific application in which it is to be used. It will be seen that the video signal applied to each lookup table can have 256 values and the control signal applied to the lookup tables can have 256 values. Thus the lookup tables can be considered as containing 256 separate ways of transforming the 256 possible video values to output values, i.e. it contains 256 (the number of possible control signal values) smaller lookup tables each having 256 (the number of possible video signal values) entries.

If, say, the control signal were a 6-bit signal and the video signal a 10-bit signal, the lookup table would contain 64 separate ways of transforming the 1024 possible video values, i.e. it would contain 64 smaller lookup tables each having 1024 entries.

Depending upon the nature of the data loaded into the lookup tables, the circuit 40 of Figure 2 can be used in a multitude of different ways to effect signal processing on the applied digital video signals. Some examples of the way in which the circuit can be used will now be described. The following description describes what happens for each pixel of the incoming video signals. If the control signal changes for each pixel, then each pixel can have its own individual programmed effects.

(1) Fader/Mixer To mix two video signals, each signal must be reduced by a predetermined factor (faded), then these two faded signals are added together to generate a mix of the two signals. Control signals determine the amount that each signal is reduced so as to produce the desired mix. The control input represents a factor between say, zero and unity to yield output video levels between black and the input video level. In known multiplier-based faders that give a linear mix as shown in Figure 1, the control signal to one video signal is the complement of the other. On a scale of 0 to 1, where a and b are the control signals, b = (1 -a).If in the circuit of Figure 2 the lookup table 46 for video signal 1 is programmed with a linear scale such that it acts as a multiplier, the other lookup table 48 can be programmed to be the complement of this, so that only one control signal is required.

When the transfer function of the lookup tables are programmed in a non-linear fashion (i.e. they no longer act as linear multipliers), a number of fading and mixing effects can be provided. With digital signals conforming to CCIR Recommendations 601 and 656, the incoming digital video is already gamma-corrected, so a linear transfer function is undesirable. The transfer function can correct for gamma correction, so that one control signal can generate a mix which provides the same apparent birghtness for all mixes of the two video signals. The gain and offset of the virtual multiplier constsituted by the lookup table can be adjusted to give delayed, or more rapid or more sensitive, mix transitions.

(2) Colouriser In an alternative application, a video signal can be colourised by the lookup table. It is assumed first that only one video input signal is present and only the lookup table 46 is therefore active. For each value of the control signal, the lookup table is programmed with a control value table of all output pixel values, one per possible input pixel value Depending on the values programmed into the control value table in the lookup table for this control value, each unique input pixel value can result in any of the possible output pixel values. Note that for the moment the separation of luma and chroma is ignored.This transformation can be any desired function, such as an inverse (where black and white are flipped), a brightening or darkening (where a constant value is added to or subtracted from the pixel value), or solarising, where the least significant bits of the video signal are thrown away, or even a random transformation. There are as many control value tables in the lookup table as there are possible values of the control signal. If the range of the control value tables is programmed appropriately, a large number of effects can be produced. The control signal can be considered to be a colourise selector, in which case each value of the control signal selects a different unrelated effect. Or the control signal could be considered to be a range, say from black to pass-through, or pass-through to invert, and so on.

(3) Combinations Any combination of the above functions (1) and (2) can be performed at one time, under the control of one control signal.

For example, a key control signal could be used so that most of the picture is monochrome, but areas determined by the key are coloured negatively, and yet other areas are blacked out so that a second video picture can be mixed in.

The preferred embodiment of the present invention provides arbitrary, user-alterable, real-time transformation of digital video signals. The arbitrary transformation is accomplished with 65,536 words of high speed static transform RAM. Thus, the preferred embodiment achieves real-time video speed operation with eight bits of control input producing 256 levels of control or different transfer functions.

Figure 3 is an overall block diagram of the digital fader of the preferred embodiment showing one of the main digital video paths and the control path. The transform block consists of a bank of 65,536 words of 24 bit wide static random access memory.

Note that this block diagram shows only the luma portion of the circuit. The chroma portions are substantially identical, with the only difference being some multiplexing and demultiplexing of the two chroma signals, as would be understood by those skilled in the art.

Incoming digital video consists of 16 simultaneous data bits representing luminance (8 bits) and chrominance (8 bits) components. The representation for the luma is different to that for the chroma. Luma is represented in unsigned positive binary form while chroma is represented in 8-bit offset binary form. For this reason different transfer functions are necessary for luma and chroma.

The digital fader is split into two sections one for luma and one for chroma. CCIR 601/656 digital video comprises two luma samples and two chroma samples in each unit sample or pixel time. The two chroma samples comprise a blue-minus-luma sample and a red-minus-luma sample which are associated with the first luma sample. The second luma sample has no associated chroma samples.

The luma fader takes incoming luma, a sample at a time, along with the incoming 8-bit digital control signal. These t6 bits form the address inputs of a 65,536 byte block of high speed static random access memory (RAM). The eight bit output then is the transformation of the luma input. The chroma fader takes the incoming chroma samples two at a time and does a similar mapping with the incoming digital control signal. The difference in representation is accommodated by different values written into the map RAM. The chroma map is twice as large as the luma map, requiring separate maps for each chroma component (red and blue). The output transformed chroma signals are recombined into the 8-bit time-multiplexed chroma signal.

The digital fader is pipelined to allow full video speed operation with off-the-shelf components.

Figure 3 shows an input latch 70 and an output latch 72 for latching the input to and output from the SRAM 46. The output 74 of latch 72 represents the circuit output. The figure also illustrates the circuitry necessary to load the SRAM. When in a load mode, control is taken over by a host processor which applies an enabling signal to an input 78 of synchronisation and control circuitry 80. This disables latches 70 and 72, and enables a latch 82 which receives an address at an input 84 from the host processor and applies it to the address input of the SRAM, and also enables a latch 86 which receives data for that address from an input 88 from the host processor.Finally the circuitry 80 enables the chip enable (CE) and write enable (WE) inputs of the SRAM. In this way new data can be loaded into successive addresses of the SRAM. The memory contents are filled, during vertical blanking of the video signal, with values representing the desired transfer function. The chroma lookup tables are filled in the same way as described for the luma lookup table 46.

A preferred application of the system is in a digital video and special effects studio, wherein various fading and colourising effects are to be implemented as part of the repertoire of effects for the studio.

Figure 4 is a block diagram showing one possible application of the circuit of Figure 2. In this example, it is desired to show a figure of a news reader (who is filmed in front of a chroma key background) in front of a pre-recorded backdrop, and at the same time make the eyes (a dull grey colour) appear bright blue. The control signal is a type of matte signal, coded with the chroma key information in the following way: 255 of the possible 256 control values correspond to the foreground chroma key (e.g. 0 refers to full background, while 254 corresponds to full foreground with no background, while 127 is half foreground, half background). The other value (say 255) is present when that pixel has the particular grey of the news reader's eyes.

Lookup table B is programmed so that at a control value of 0, the output is the same as the input, and at a control value of 254 or 255, the output is black. The values in between are an interpolation of this.

Lookup table A is programmed so that at a control value of 0, the output is black, and at a control value of 254, the output is the same as the input. The values in between are an interpolation of this. However, at a control value of 255, the (R-Y) component is the same as the (R-Y) input, but the (B-Y) component has a number added to it and the luma component has another number added to it such that the grey of the news reader's eyes becomes bright blue. This does not make the eyes all the same blue colour, but merely shifts the colour of each pixel of the eyes towards light blue from grey.

Claims (11)

1. A digital video signal processor, comprising: a first video input for receiving a first multibit digital video signal; a second video input for receiving a second multibit digital video signal; control signal input means; a first lookup table having a multibit address input coupled to the first video input and the control signal input means such as to receive the first multibit video signal and a multibit control signal and to be addressed by the resultant combination of bits; a second lookup table having a multibit address input coupled to the second video input and the control signal input means such as to receive the second multibit video signal and a multibit control signal and to be addressed by the resultant combination of bits; and combining means coupled to the outputs of the first and second lookup tables to combine the outputs thereof to form a combined output digital video signal.

2. A signal processor according to claim 1, in which the control signal input means comprises a single control signal input coupled to the address inputs of both lookup tables.

3. A signal processor according to claim 1 or 2, in which the number of bits of the video signal and the number of bits of the control signal are substantially equal.

4. A signal processor according to claim 1, 2 or 3, in which the lookup table provides both a mixing/fading and a colourising operation.

5. A signal processor according to any preceding claim, in which the lookup tables are comprised in random-access-memory.

6. The use of a signal processor in accordance with any preceding claim to provide both a mixing or fading function and a colourising function to an input video signal.

7. A method of processing a digital video signal, comprising the steps of: providing a lookup table having a multibit address input; applying a multibit digital video signal to some of the address input bits of the lookup table; applying a multibit control signal to others of the address input bits of the lookup table; and applying the contents of the thus-addressed location in the lookup table to an output.

8. A method according to claim 7, in which the number of bits of the video signal and the number of bits of the control signal are substantially the same.

9. A method according to claim 7 or 8, in which the number of bits of the video signal is at least 8.

10. A method according to claim 7, 8 or 9, in which the number of bits of the control signal is at least 6.

11. A method of mixing two digital video signals, comprising the steps of combining each of the digital video signals with a single, common multibit control signal, to provide two combined signals, addressing each of two pre-programmed lookup tables with the two combined signals respectively to provide two data outputs, and combining the data outputs to form a mixed digital video output signal.