This application claims the benefit, under 35 U.S.C. §365 of International Application PCT/EP2003/009260, filed Aug. 21, 2003, which was published in accordance with PCT Article 21(2) on Apr. 8, 2004 in German and which claims the benefit of German patent application No. 10242037.8, filed Sep. 11, 2002.
The invention relates to an arrangement for correcting colour video signals, in particular colour video signals generated by a film scanner, with a matrix, through which the colour video signals pass and which can be used to control the proportions of three primary colours in matrixed colour value signals, provision being made of means for controlling the matrix in a manner dependent on the hue which the colour video signals respectively represent.
When generating colour video signals from optical originals, for example when scanning films, colour errors occur, inter alia, by virtue of the fact that the filter curves during colour splitting do not correspond to the ideal curves. These errors can largely be corrected by matrixing the colour value signals. In this case, the coefficients of the matrix are set manually whilst observing the reproduced image.
The remaining, generally slight colour corruptions are conspicuous, however, particularly when the same original is scanned by two different recording apparatuses, particularly when the same film is reproduced by two different film scanners.
U.S. Pat. No. 5,668,596 discloses an arrangement for correcting colour video signals generated by an image sensor. In the known arrangement, provision is made of means for controlling a matrix in a manner dependent on the image sensor. Coefficients that serve for setting the matrix in a manner dependent on the image sensor are stored in a memory. The hue respectively represented by the colour video signals depends on the image sensor.
The arrangement according to the invention is characterized in that provision is made of means for reducing the effect of the correction in the case of low colour saturation. The arrangement according to the invention enables such a precise correction of the colour video signals that the same film produces an identical colour impression when scanned by different apparatuses. In this way, noise influences and exaggerated corrections of pixels with low colour saturation are furthermore avoided.
An advantageous refinement of the arrangement according to the invention consists in the fact that provision is made of memories for storing coefficients of the matrix or correction values for the coefficients of the matrix that are set previously in a manner dependent on the hue.
A further advantageous refinement of the arrangement according to the invention consists in the fact that a converter for generating a hue signal from the colour video signals is connected by its output to address inputs of memories for a respective correction value to be fed to the matrix. In this case, a reduction of the effect of the control in the case of low colour saturation may be achieved by virtue of the fact that the converter has a further output, which carries a colour saturation signal and is connected to multipliers located in the supply lines of the correction values to the matrix. The colour video signals are generally present as colour value signals, for which purpose, in the case of the arrangement according to the invention, it may be provided that the converter comprises a converter matrix for generating colour difference signals and a coordinate converter.
The colour video signals are generally present as colour value signals, for which purpose, in the case of the arrangement according to the invention, it may be provided that the converter comprises a converter matrix for generating colour difference signals and a coordinate converter.
Another advantageous refinement of the arrangement according to the invention consists in the fact that the matrix comprises nine further multipliers and three adders, in each case three further multipliers being connected to inputs of an adder and having three colour video signals—fed as colour value signals—applied to them, and in that a correction value/coefficient can be fed to a respective one of the further multipliers from one of the memories.
For setting the corrections, it is preferably provided that the correction values can be loaded into the memories from a computer, and in that the computer has a program for setting the coefficients. The computer may be provided with corresponding input devices, for example rotary regulators, which facilitate a selection of the hue that is respectively to be corrected and the setting of the magnitude and direction of the correction.
This development may be embodied in such a way that provision is made of a manual setting and/or an automatic determination of the correction values by scanning of a test film and comparison of the scanned values with desired values.
In another development, it is possible to take account of the non-linear film density by virtue of the fact that logarithmizers are connected upstream of the matrix and delogarithmizers are connected downstream of the matrix.
An exemplary embodiment of the invention is illustrated in the drawing on the basis of a plurality of figures and is explained in more detail in the description below. In the figures:
In the case of the representation of colours that is known per se in accordance with
Colour value signals R, G, B are fed to the arrangement according to
Since signals that are greater than the predetermined range of values of the digital video signals may arise during the correction, limiters 23, 24, 25 are furthermore connected downstream of the matrix 16 and limit the video signals to a maximum value governed by the respective quantization. The corrected colour value signals can be taken from outputs 26, 27, 28. The matrix 16 maps the following system of equations:
Rk=R.Krr+G.Kgr+B.Kbr
Gk=R.Krg+G.Kgg+B.Kbg
Bk=R.Krb+G.Kgb+B.Kbb
Correction values krr, krg, krb, kgr, kgg, kgb, kbr, kbg and kbb for the coefficients are stored in nine memories 31. The outputs of the memories 31 are connected via a respective one of the multipliers 32 to inputs of the multipliers 4 to 12 of the matrix 16.
In order that the correction values can be read from the memories 31 in a manner dependent on the hue, the colour value signals R, G, B are firstly converted into colour difference signals CB, CR in a converter matrix 33. The said colour difference signals are then converted from Cartesian coordinates into polar coordinates in a suitable converter 34. In this case, the angle represents the hue H and the magnitude represents the colour saturation S. The hue H is fed as a signal having a width of 12 bits to the address inputs of the memories 31. The colour saturation signal—likewise having a width of 12 bits—is forwarded via a switch 35 to the multipliers 32, with the result that the correction values read from the memories 31 are reduced in the case of low colour saturation.
The correction values read from the memories 31 have a width of 10 bits and can assume positive and negative values. The output signals of the multipliers 12 have a width of 12 bits, the sign being continued. The value 2048 is in each case added to the correction values krr, kgg and kbb after multiplication at 36, 37 and 38 for the purpose of forming the coefficients, while the remaining correction values serve directly as coefficients. This ensures that the matrix 16 is operated as a unit matrix if the correction values themselves are zero or are multiplied by zero. The colour value signals are then conducted unchanged through the matrix.
Number | Date | Country | Kind |
---|---|---|---|
102 42 037 | Sep 2002 | DE | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP03/09260 | 8/21/2003 | WO | 00 | 3/11/2005 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2004/030370 | 4/8/2004 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
4418358 | Poetsch | Nov 1983 | A |
4962418 | Kamaga | Oct 1990 | A |
5001663 | Parulski et al. | Mar 1991 | A |
5406325 | Parulski | Apr 1995 | A |
5596427 | Honma et al. | Jan 1997 | A |
5668596 | Vogel | Sep 1997 | A |
5943143 | Kawai et al. | Aug 1999 | A |
6337692 | Rai et al. | Jan 2002 | B1 |
6433898 | Bestmann | Aug 2002 | B1 |
6477271 | Cooper et al. | Nov 2002 | B1 |
6563531 | Matama | May 2003 | B1 |
20010008428 | Oh | Jul 2001 | A1 |
20020110376 | MacLean et al. | Aug 2002 | A1 |
Number | Date | Country |
---|---|---|
2191360 | Dec 1987 | GB |
11-136533 | May 1999 | JP |
2000-101766 | Apr 2000 | JP |
Number | Date | Country | |
---|---|---|---|
20060072172 A1 | Apr 2006 | US |