Television – Image signal processing circuitry specific to television – Noise or undesired signal reduction
Utility Patent
1997-05-22
2001-01-02
Peng, John K. (Department: 2714)
Television
Image signal processing circuitry specific to television
Noise or undesired signal reduction
C348S607000, C348S612000, C348S618000, C348S620000, C382S275000
Utility Patent
active
06169583
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a method and a circuit to extract from a signal, in particular a digitized video signal, a value that corresponds to the noise in a signal.
Noise values that correspond to the noise in a signal can be used to control noise reduction filters as well as other digital improvement circuits, in particular picture improvement circuits. This is of special interest for digital signal processing in modern television receivers. The invention helps to use picture improvement algorithms that exist in modern digital television receivers (e.g. noise reduction, peaking, CTI, contour correction) more efficiently. To obtain a meaningful noise value the used methods and means have to discriminate between the noise and the signal.
2. Description of Related Art
As described in “Automatic Assessment of Signal-to-thermal noise ratio of Television Images” by Qui Zhang and Rabab Ward published in IEEE Transactions on Consumer Electronics, Vol. 41, No. 1 on February 1995, cable providers measure in the frequency domain the amount of energy in the guard band, wherein the guard band is the portion of the bandwidth between two consecutive channels within a continuous frequency range. The measured energy is assumed to correspond to the noise in the picture. The noise in the guard band does not necessarily reflect the noise in the picture. Additionally, the extreme expensive use of a spectrum analyzer which is necessary for such a method is a disadvantage.
Modern television receivers are using in the front-end a so-called ‘automatic gain control’ (AGC) which measures the amplitudes of the sync or burst signal. The amplitudes are assumed to correspond to the noise in the signal. But since this is not a real measurement of the noise even a noisy picture which has been recorded by using a bad camera or a cheap VCR can contain a correct sync amplitude.
In another method, particularly used by broadcasters and cable providers, test signals are inserted in the field retrace time or during the first lines of a field. Thus, the signal is known in advance and it is easy to separate the noise from the signal. But these test signals cannot be used in television receivers because each cable provider and broadcaster uses different test signals and positions in the CVBS signal to check the manner of transmission (from antenna via air, cable, satellite). This technique cannot assess the noise in the picture but only the amount of noise added during transmission.
SUMMARY OF THE INVENTION
It is a general object of the present invention to provide a method and a circuit that can extract from a signal a value that corresponds to the noise in the signal.
Preferred embodiments of the present invention aim to provide a noise measurement method and circuit that can extract from a digitized video signal a value that corresponds to the noise in the video signal to control digitized picture improvement circuits. With such a noise value picture improvement, algorithms can perform more efficiently. The improvement algorithms may, for example, be noise reduction algorithms, peaking algorithms, CTI algorithms, or contour correction algorithms.
Noise reduction algorithms can reduce the noise level in a picture to a level which is more pleasing to the viewer. Peaking algorithms can emphasize the high frequency portion of the video signal to improve the impression of sharpness. But if the video signal has a high level of noise, the emphasis of a high frequency portion would as well lead to an undesirably increased visibility of noise. The application of peaking algorithms can be made more efficient by the use of noise values. The same is true for the application of CTI algorithms. CTI algorithms (Color Transient Improvement) produce a sharper impression at color transitions. But to apply such algorithms first the color transitions have to be detected. The lower the noise level is in a video signal, the better algorithms for detecting color transitions perform and the more the detected color transitions are reliable. Since the reliability of the transition detection depends to a high degree on the noise level in the signal, CTI applications controlled by a noise value could be applied more efficiently.
Modern television receivers usually comprise means for automatic search tuning. Since a broadcasting station in general uses different frequencies to cover a larger area by automatic search tuning the same station will be found at different frequencies. With the help ofthe Videotext service broadcast with the video signal it is possible to decide automatically whether two stations found at different frequencies are the same or different ones. To decide automatically which of the found frequencies for one station should be used, the measured noise level is an appropriate criterion since the frequency having the lower noise level and correspondingly being more pleasing to the viewer should be selected.
According to a first aspect ofthe present invention, the noise measurement method and circuit evaluate only the absolute differences between corresponding lines in subsequent fields or frames. The evaluated pixels in the corresponding lines have a constant value from frame to frame or field to field in the noise free signal. This makes a noise measurement possible which takes place as close as possible to the real video information because it is insensitive to signals inserted in these lines as long as these signals are not varied in temporal respect. Therefore, the measured noise values can be more accurate than it has been possible with prior art techniques.
According to another aspect of the present invention, the measured noise values are filtered temporally. This ensures that short distortions in the signal, especially in the video signal, do not lead to wrong settings in the controlled features (e.g. noise reduction). Additionally, this results in a smooth change of the controlled features when the noise level changes.
According to another aspect of the present invention, extreme peak values of the measured differences are limited to make limited hardware complexity possible.
According to another aspect of the present invention, a reliability test is introduced to make a self detection of wrong measurements possible. Since extreme peak values of the differences are usually not due to noise but to other distortions or to an inserted time-varying signal, these values are not evaluated for the calculation of the noise value. This ensures an accurate noise value even if unpredictable time-variant signals occur in the evaluated line.
A noise measurement circuit according to the present invention makes a more efficient use of digital image processing algorithms possible. Such a circuit has a low hardware complexity but provides a more accurate noise measurement compared with the prior art. Since the noise measurement circuit is intended for use in digital television receivers, it works with different television standards (e.g. NTSC, PAL, PALplus, SECAM), with different sources (e.g. television signal via tuner, VCR, laser disk, set top box) and with existing and future services in vertical blanking.
REFERENCES:
patent: 4249210 (1981-02-01), Storey et al.
patent: 4796088 (1989-01-01), Nishimura et al.
patent: 4873574 (1989-10-01), Darby
patent: 5025316 (1991-06-01), Darby
patent: 5105207 (1992-04-01), Iizuka et al.
patent: 5105275 (1992-04-01), Iizuka et al.
patent: 5185664 (1993-02-01), Darby
patent: 5241370 (1993-08-01), Desor
patent: 5268760 (1993-12-01), Suyama
patent: 5294979 (1994-03-01), Patel et al.
patent: 5387946 (1995-02-01), Knee
patent: 5657401 (1997-08-01), De Haan et al.
patent: 5684720 (1997-11-01), Hein
patent: 5742355 (1998-04-01), De Haan et al.
patent: 5867228 (1999-02-01), Miki et al.
patent: 31 21 611 (1982-12-01), None
patent: 43 41 760 (1995-06-01), None
patent: 54-118727 (1979-09-01), None
patent: 58-24275 (1983-02-01), None
patent: 59-45770 (1984-03-01), None
patent: 62-1379 (1987-01-01), None
patent: 62-172879 (1987-07-01), None
patent: 3-7
Merkl Bernd
Saenger Jens
Matsushita Electric - Industrial Co., Ltd.
Natnael Paulos M.
Peng John K.
Wenderoth , Lind & Ponack, L.L.P.
LandOfFree
Method and circuit to determine a noise value that... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and circuit to determine a noise value that..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and circuit to determine a noise value that... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2523413