« PreviousContinue »
METHOD AND APPARATUS FOR CORRECTING
TIMING ERRORS AS FOR A MULTI-PICTURE
BACKGROUND OF THE INVENTION
This invention relates to apparatus and a method for reducing the visibility of timing errors in for example, the inset image of a picture-in-a-picture (pix-in-pix) television display system. 10
In a pix-in-pix system, two images from possibly unrelated sources are displayed simultaneously as one image. The compound image includes a full size main image with an inset compressed auxiliary image. The subjective quality of the inset image may be affected by timing 15 errors in either the main signal or the auxiliary signal.
Timing errors relevant to the present invention may occur, for example, when either the main or auxiliary signal is a nonstandard signal. As used herein, the term nonstandard signal means a video signal having a hori- 20 zontal line period which may vary in length relative to the horizontal line period set by the signal standard to which the video signal nominally conforms (e.g. NTSC, PAL or SECAM). A noisy but otherwise standard signal may appear to be a nonstandard signal if the noise 25 is of sufficient amplitude to mask transitions of the horizontal line synchronization (horizontal sync) signal.
To understand how these timing errors may affect the inset image, it is helpful to know how the auxiliary signal is processed and displayed. In a conventional 30 pix-in-pix display system, the auxiliary signal is sampled at instants determined by a sampling clock signal which, desirably, bears a fixed relationship to the horizontal line scanning frequency of the auxiliary signal. To aid demodulation of the chrominance signal components of 35 color television signals, the sampling clock signal desirably has a frequency that is a multiple of the chrominance subcarrier frequency. If the multiple is an even number, e.g., 4, for standard signals, this is a suitable sampling signal since, under all major video signal stan- 40 dards, it produces an integer number of samples per line interval. Under the NTSC system, this sampling clock signal may be developed, for example, by a phase locked loop which produces a sampling signal having a frequency of 4/c, four times the frequency, fc, of the 45 color subcarrier signal, and which is locked in phase to color reference burst component of the auxiliary composite video signal.
The auxiliary video signal is separated into its component parts, generally a luminance signal and two color 50 difference signals. These component signals are then subsampled both horizontally and vertically to develop signals that represent a compressed image. The lines of samples taken during one field of the auxiliary signal are stored in a memory. These samples are read from the 55 memory for display using a clock signal that is desirably related to the horizontal line scanning frequency of the main video signal.
When the auxiliary signal originates from a noisy source or from a nonstandard source such as a video 60 tape recorder (VTR) or a video game, the frequency of the horizontal sync signal may appear to vary significantly from line to line while the frequency of the color subcarrier signal, and thus of the color reference burst signal, may seem relatively stable. This variation can be 65 caused by pickup head misalignment or by stretched tape in a VTR or by inaccuracies in the frequencies used by video game circuitry. Since, in the example set forth
above, the sampling clock signal is locked in phase to the color reference burst signal, corresponding samples on successive lines may be shifted or skewed relative to each other. When these lines of samples are displayed in synchronism with the main signal, the pixels produced by these corresponding samples may not line up vertically. Consequently, any vertical lines in the inset image may appear jagged (if the period of the horizontal sync signal changes randomly) or tilted (if there is a fixed error in the relative frequencies of the horizontal sync and color burst signals). The frequency and phase variations which cause this type of image distortion are known as timing errors or, alternatively, as skew errors.
One type of timing error, which is relevant to the present invention, results from frequency or phase variations between the main horizontal sync signal and a video display clock signal that is phase locked to the color reference burst component of the main signal. Errors of this type may randomly change the distance between the left side edge of the main image (defined by the horizontal sync pulses) and the beginning of lines of the inset image (defined by the display clock signal). Main signal timing errors of integral numbers of sampling clock periods may be compensated for in the phase locked loop circuitry which generates the horizontal sync signal. Skew errors which are a fraction of a sampling clock period may be more difficult to correct.
One method of correcting these types of timing errors is to use interpolation to develop sample values that are matched to the clock signal used to store or display them. Another method is to shift the phase of the clock signal used to display the sample values so that it is properly aligned to the horizontal sync signal. These methods are described in U.S. Pat. No. 4,638,360 entitled "Timing Correction for a Picture-In-Picture Television System" which is hereby incorporated by reference.
Skew errors may also be corrected by generating samples that represent component video signals in synchronism with a skew shifted line locked clock signal. These samples are then applied to clock transfer circuitry which aligns the samples with a line-locked clock signal that is not skew shifted. U.S. Pat. No. 4,782,391 entitled "Multiple Input Digital Video Features Processor for TV Signals," which is hereby incorporated by reference, relates to a system of this type.
The first two methods described above use two substantially independent clock signals. Aside from the extra circuitry used to generate an additional clock signal, systems which use multiple clock signals may need to be carefully shielded to prevent radio-frequency interference between the signals.
In the third method described above, the luminance and color difference signal components of the auxiliary signal are separated by analog circuitry and then digitized. A system using this method may be more complex than a system which digitizes the composite video signal and then separates it into its component parts. In addition this second method uses line-locked clock signals, so it may be difficult to encode the color information signals of the compressed video signal so that the two signals may be time-division multiplexed for display.
SUMMARY OF THE INVENTION
The present invention is embodied in a system which compensates for timing errors in a first video signal relative to a second stored video signal. This system 5 includes a first clock signal for retrieving samples of the second stored video signal from memory. A signal phase alignment circuit shifts the phase of the first clock signal to generate a second clock signal that is synchronized to the horizontal scanning signal derived from the 10 first video signal. A clock transfer circuit, responsive to the second clock signal, aligns retrieved sample synchronous with the first clock signal, to a predetermined phase relationship with the phase shifted clock signal.
BRIEF DESCRIPTION OF THE DRAWINGS 15
FIG. 1 is a block diagram of a television receiver which includes an embodiment of the present invention.
FIG. la is a block diagram of exemplary circuitry for generating a subsampling clock signal used in the televi- 20 sion receiver shown in FIG. 1.
FIG. 2 is a block diagram of a clock phase shifter suitable for use in the television receiver shown in FIG. 1.
FIG. 3 is a block diagram of circuitry which illus- 25 trates the operation of the signal phase alignment circuitry shown in FIG. 2.
FIG. 4 is a block diagram of clock transfer circuitry suitable for use in the television receiver shown in FIG. 1. 30
FIG. 5 is a timing diagram which is useful for describing the operation of the clock phase shifter circuitry shown in FIG. 2.
The present invention is described in the context of digital circuitry which implements, for example, a pixin-pix feature for a consumer television receiver. It is contemplated, however, that this invention has broader application. It may be used in other systems where two 40 images or portions of two images are displayed concurrently (e.g., side by side or one over the other) and it may employ analog circuitry, such as charge-coupled devices, in place of the digital memory circuitry.
THEORY OF OPERATION 45
In the television system described below, a main video signal is processed by conventional analog circuitry to produce a full-size image. An auxiliary signal is received, digitized and processed by digital circuitry to 50 produce a luminance signal and two quadrature phase related color difference signals. These separated signals are subsampled to develop signals representing a compressed image. The subsampled signals are stored in a memory which holds one field interval of the com- 55 pressed signal. When the compressed image is to be displayed, the stored signals are retrieved from the memory and encoded into a composite video signal. This composite video signal is substituted for a portion of the main composite video signal to generate a com- 60 pound signal which is processed by the analog circuitry to display a compound image. This compound image includes a full-size main image with a compressed auxiliary image displayed as an inset.
A composite video signal includes three component 65 signals, a luminance signal, Y, and two color difference signals, for example, (R-Y) and (B-Y). The two color difference signals modulate respective quadrature phase
related color subcarrier signals to produce a chrominance signal which is additively combined with the baseband luminance signal to generate the composite video signal. Conventional analog techniques for decoding a composite video signal include low-pass filtering to recover the luminance signal, Y, and band-pass filtering to recover the chrominance band signals. The chrominance band signals are then synchronously demodulated using a regenerated color subcarrier signal.
In general, when digital processing techniques are used, a composite video signal is first sampled and digitized. The sampling clock signal used to develop these samples is typically locked in phase to the color burst signal of the composite video signal. This sampling signal may aid in the demodulation of the chrominance signal. For example, if the selected sampling clock signal has a frequency of 4fc, four times the frequency, fc, of the color subcarrier signal, successive samples of the separated chrominance signal may be represented by the sequence (R-Y), (B-Y), -(R-Y), -(B-Y), (R-Y), etc. where the minus signs indicate sampling phase and not necessarily sample polarity. The (R-Y) and (B-Y) color difference signals may be recovered from this sequence by a process of demultiplexing and selective polarity inversion.
Thus, if these conventional techniques are used to decode the chrominance signal components of two independent composite video signals, it is desirable to generate two oscillatory signals related to the respective color subcarrier signals of the two composite video signals. Using two clock signals may complicate the design of the receiver, since electromagnetic shielding may be needed, to limit interference between the two signals.
An alternative method for demodulating two video signals is to generate only one oscillatory signal, for example, the main signal color subcarrier signal. This signal is then used to generate a sampling clock signal for digital circuitry which processes the auxiliary signal. However, since the chrominance signal phases of the main and auxiliary signals may be different, it may be desirable to include circuitry which corrects the phase of the digitized decoded color difference samples based on the color reference burst component of the auxiliary signal.
The choice of a sampling clock signal is also a factor in determining what type of skew-error compensation circuitry is to be used in the system. Since a clock signal that is locked to the main burst signal may be less closely aligned to the auxiliary horizontal sync signal than a clock signal that is locked to the auxiliary burst, the possibility of skew errors when the compressed auxiliary signal is stored in the memory is increased. Moreover, if the sampling clock signal is locked to the main signal burst, skew errors caused by noise in the main signal or by variations in the relative frequencies of the main signal horizontal sync and color subcarrier signals are not reduced. A sampling clock signal that is locked to the horizontal synchronizing component of the auxiliary signal may also be used.
In the television receiver described below, skew errors of the first type are reduced to a maximum of onethird of one clock period (e.g. to a maximum of 23 ns for a sampled data NTSC signal having a sampling frequency of 4fc) and skew errors of the second type are substantially eliminated.
In the system described below, the luminance signal component of each horizontal line interval of the auxil