Apparatus and method for transmitting audio signals as part of a television video signal

Description Submit all comments and votes

BACKGROUND OF THE INVENTION

1. Field of the Invention

This invention relates to encoding audio frequency signals on a television video signal while maintaining said video signal in a standard form, thus eliminating audio to video delay problems and providing high quality audio frequency channels.

2. Description of the Prior Art

As television technology develops, several previously unforseen audio related problems are also developing. One of these problems pertains to properly maintaining the timing or synchronization relationship of audio and video signals. This problem was explained in U.S. Pat. No. 4,313,135 which may be referred to for further details. Another problem is that normal broadcast television audio is limited to one channel of 5 KHz bandwidth. With the drastic improvements in television video equipment over the past years, the image quality of television programs now far surpasses the audio quality. There is a great need for a device which can improve the audio quality of a television program, preferably to a stereo high fidelity level, and which can encode this audio in the video signal to prevent audio to video delay problems. Such an encoded audio system will have substantial cost benefits in the transmission of television programs by eliminating the need for a separate audio channel.

Several television systems exist which add digitized audio in one form or another in the blanking interval of the television video signal. This system has been used as a scrambling technique where it is desirable to prevent unauthorized viewing of the television program. Digitized audio requires a great deal of bandwidth, thus causing a substantial portion of the video blanking interval to be filled with digital data seriously affecting sync and burst. Unfortunately, this digital audio in blanking is not directly compatible with existing video systems and the digital audio conversion components, i.e. A-D and D-A, are fairly expensive. The digital audio in blanking does work well as a scrambled system because the digital information requires most sync information to be removed causing television receivers to malfunction.

SUMMARY OF THE INVENTION

The inventive concept herein disclosed provides an apparatus and method for encoding signals having a bandwidth lower than video signals which may or may not be program related, on a television vide signal. Said apparatus and method described herein provides a pulse modulation means responsive to a signal such as a television audio signal to provide an analog modulated signal, a combination circuit to combine said modulated signal with said video signal at predetermined locations in said video signal's waveform while allowing said video signal to remain in a form which can be passed by standard video equipment with little or no modification. Also described are an associated decoder having separation means responsive to aforementioned video signal containing modulated signal to recover said modulated signal from said video signal and filter means to reconstruct the audio frequency signal which was input to the above encoder. The ability of low cost equipment to encode one or more audio frequency signals on a video signal while maintaining said video in a standard form is a very important feature of this invention. If said audio frequency signal is chosen to be program related, such as timecode or program audio, audio to video timing relationships are inherently preserved since both signals are always passed through the same delay after their combination.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a drawing of a typical Horizontal Blanking Waveform of an NTSC or PAL video signal.

FIG. 2 is a drawing of the waveform of FIG. 1 with modulated pulses added after burst.

FIG. 3 is a drawing of a typical vertical interval showing audio samples for lines 1-9 combined in line 10.

FIG. 4 is a drawing of a pulse modulation signal system with typical waveforms.

FIG. 5 is a block diagram of a 4 cell FIFO.

FIG. 6 is a block diagram of an audio in video transmission system showing an encoder and a decoder.

FIG. 7 is a block diagram of a differential analog pulse modulation system.

DESCRIPTION OF THE PREFERRED EMBODIMENT

FIG. 1 is a drawing of a typical video signal horizontal waveform showing sync 1, color burst 2, back porch after burst 3, active video leading edge 4, active video trailing edge 5.

FIG. 2 is a drawing of the same waveform as FIG. 1 with modulated pulses 6, 7, 8 and 9 added in the back porch after burst.

FIG. 3 is a drawing of a typical video signal vertical waveform showing modulated pulses 10 on a video line with expanded drawing showing detail of 10 having pulses 11, 12, 13 and 14.

FIG. 4 is a drawing of a pulse modulation signal system showing signal input 15, sample input 16, switch 17, transmission channel 18, switch 19, hold capacitor 20, low pass filter 21, signal output 22, buffer amplifiers 23a, b and c, input waveform 24 corresponding to 15, sample waveform 25 corresponding to 18, sample and hold waveform 26 corresponding to 20 and output waveform 27 corresponding to 22.

FIG. 5 is a drawing of a first in first out memory (FIFO) showing audio input 28, buffer amplifiers 23d and e, analog multiplexers 29a and b hold capacitors 30a-d input clock input 31 output clock input 32 counters 33a and b and set input 34.

FIG. 6 is a drawing of the preferred embodiment of the invention as used in a system having an encoder made of combiner 46 and modulator 47 and having a decoder made of a separator 48 and demodulator 49. Combiner 46 is composed of buffer 23f having video input 35, adder 36, sync stripper 38a, phase lock loop 39a, timing generator 40a, and switch 41. Modulator 47 is composed of FIFOs 43a and b having audio inputs 42a and b respectively. The decoder is composed of Separator 48 and Demodulator 49. Separator 48 is composed of buffer 23g having video output 37, sync stripper 38b, phase locked loop 39b, and timing generator 40b. Demodulator 49 is composed of FIFOs. 43c and d, low pass filters 44a and b having outputs 45a and b respectively. Encoded video and audio from aforementioned encoder pass thru transmission channel 45 to the decoder.

FIG. 7 is a differential pulse amplitude system having difference amplifier 50, sample and hold circuits 51a and b, integrators 52a and b, transmission channel 53, low pass filter 54, resampler 55 and showing optional connections 56a and b used when 55 and 51b are deleted.

In the preferred form of the invention, a relatively low bandwidth signal, which may be a time code signal, control signal, chroma signal, program audio or other low frequency signal, but which signal will be called an audio signal in this description, is sampled periodically as in a pulse amplitude modulation system and the pulses from the sampler are combined with and transmitted as part of the video signal. At the receiving end the pulses are recovered from the video by resampling by a sample and hold which drives a low pass filter to recover the audio signal. An example of such a sampled signal system without the video combination is shown in FIG. 4. Waveform 24 shows the analog signal input at AF input 15 in the block diagram. Waveform 25 shows the output of sample switch 17 which is caused to sample the buffered AF waveform by an external oscillator (not shown) which sampled waveform is transmitted via channel 18 to buffer amplifier 23b. The output of buffer amplifier 23b is resampled at an appropriate time, related to sample signal 16, by switch 19 which samples are held during open periods of 19 by hold capacitor 20 which waveform is shown by 26. The hold waveform on 20 is buffered by 23c and passed through low pass filter 21 to AF output 22 which is shown by 27 which output is essentially the same as the input waveform 24. The above described circuit is a variation of textbook examples of sampled signal systems. Further information of design criteria for these systems relating to signal to noise, bandwidths, spectrum and alias considerations for various pulse modulation techniques which would be suitable for use in this invention may be found in "Reference Data For Radio Engineers", Howard W. Sams & Co., Indianapolis, Ind. 46268; .COPYRGT. 1975. Chapter 23 is particularly useful and several further references are given. Such pulse modulation systems include but are not limited to pulse amplitude modulation, pulse position modulation, pulse phase modulation, pulse duration modulation and multilevel pulse amplitude modulation. Also of interest are delta modulation and the family of differential encoding techniques which transmit binary digits in response to the change in an analog signal from one sample time to the next. These systems are not limited in dynamic range as are pulse modulation and binary pulse amplitude modulation systems. Since in high quality audio systems a dynamic range in excess of 80 db is desirable, differential encoding techniques become attractive although they are generally more difficult to implement than analog pulse amplitude modulation systems. Further information on delta modulation and differential encoding may be found in "Principles of Pulse Code Modulation" by K. W. Cattermole, American Elsevier Publishing Co., New York, N.Y. 10017, pages 198-241. In all of the above listed pulse modulation techniques some analog parameter of a pulse such as amplitude, width, phase or position with respect to a reference; are caused to vary in response to a modulating signal. Of particular interest are pulse amplitude modulation such as described with FIG. 4 and multilevel pulse amplitude modulation because of their relative simplicity and low cost. In the former system the pulse amplitude is caused to vary in a linear fashion in response to the amplitude of the input modulating signal. This is the system which is the preferred system for the invention described herein due to its simplicity and low cost. The major disadvantage of this modulation system is that it will have a dynamic range limited by the video channel to around 60 db which is somewhat lower that desirable. If the limited dynamic range becomes a problem in a particular application it can be considerably improved by use of differential pulse modulation such as the differential pulse amplitude system, believed to be previously unknown, such as shown in FIG. 7 which will be discussed later.

The multilevel pulse amplitude system previously mentioned is similar to the pulse amplitude system shown in FIG. 4 in that the pulse amplitude changes in response to the modulating signal however the pulse amplitude is allowed to assume only a limited number of discrete steps. This multilevel system allows a much better signal to noise ratio and dynamic range than the pulse amplitude system at a cost in complexity and bandwidth. For quality audio signals it is required to transmit a number of multilevel pulses for each sample of the modulating signal allowing several hundred possible levels for each sample, which results in poorer utilization of the transmitting channel in order to achieve a better signal to noise ratio. Further discussion of alternate modulation schemes will not be given as this area is well known in the electronics industry and one skilled in the art could use any of several modulation schemes in this invention.

The inventive part of this apparatus and method involves taking the periodically generated pulses output from the sample switch (or modulator) and inserting them in a predetermined position in the video waveform. Since analog pulses or samples are inserted in the video much less space is needed than with digital or multilevel analog systems, thus very little disturbance to the video is generated. FIG. 1 shows a typical television waveform having horizontal sync 1, color burst 2, back porch after burst 3, leading active video 4 and trailing active video 5. FIG. 2 shows the same waveform with four sample pulses 6-9 added in the back porch. This location is the preferred location for adding sample pulses in horizontal blanking since it allows the video signal to be processed by most standard video equipment with little or no modification. Other locations, such as the front porch in the video waveform may be used, however, care must be used so that the pulses do not cause disturbances to the active video, by interfering with synchronization pulses or color burst. It is worthwhile to note here that it is not practical to use only this horizontal back porch area for encoding bits in normal digital audio applications due to the large number of digital bits required. For a typical video signal bandwidth of 5 MHZ it would only be possible to insert around 8 digital bits per horizontal line in the back porch, an insufficient number of bits for most audio signals. Typical state of the art digital audio systems utilize the entire blanking interval for digital bits causing a great deal of interference with, or removal of, sync or burst. Six amplitude modulated pulses per line will however be quite adequate for two high fidelity audio signals and will fit in the 1.4 ms space available on the back porch. Inserting the amplitude modulated pulses into the predetermined location of the video waveform is not a trivial problem. With digitized audio the digital bits are simply stored in a RAM as they are developed and then read out at the appropriate time. There is, however, no analog equivalent of a digital RAM available as a manufactured device, only analog delay lines which could be used but are fairly expensive. The circuit invented to perform the necessary time compression function in a low cost fashion is an analog first in first out memory shown in FIG. 5.

The first in first out analog memory or FIFO of FIG. 5 is constructed of two buffer amplifiers 23d and e, two 4 position analog multiplexers 29a and b which are driven from four bit digital counters 33a and b, and four hold capacitors 30a-d. To understand the operation of the FIFO, assume a short pulse at set input 34, sets counters 33a to count 1 and 33b to count 3. Both counters count 1,2,3,4,1 etc. Audio is input to buffer 23d at 28 and the buffer in turn charges capacitor 30a (because 33a is on count 1) with the audio signal. Next, assume a clock pulse arrives at 31 causing counter 33a to count to 2 allowing buffer 23d to charge capicator 30b. Capacitor 30a will now be holding the charge that was placed on it by the input audio signal at the instant before counter 33a changed count. At each count, the audio signal will charge another capacitor in sequence in effect creating a multiplexed sample and hold. Alternately, for low clock frequencies it may be desired to cause counter circuit 33a to output only a short pulse rather than a count length pulse as described above. This short pulse would effect a more precise sample and hold. The short pulse could be simply effected by differentiating the counter output with an RC Network. Now assume multiplexer 29b is allowing each stored charge in turn to be fed into buffer amplifier 23e in response to output clock input to counter 33b at 32. The output clock does not need to be synchronous with the input clock, only the same frequency. The output to input clock relationship must be such that a new sample is not put on a capacitor until the old has been read out, and a new sample cannot be read out of a capacitor until it has been put on. One skilled in the art will see that by increasing the number of hold sections in this scheme a quite large variation of input clock to output clock phase may be handled providing that the average frequency of the two clocks is exactly the same, i.e., one sample is clocked out for every one clocked in. This will provide for temporary storage of samples, allowing samples to be generated at a steady rate, stored and read out at an intermittant rate. This FIFO memory system is used to generate and store analog samples of an audio input and read out those samples in correct sequence when necessary to insert them in the proper location in the video waveform. This FIFO can be easily built by one skilled in the art using standard integrated circuits. The buffer amplifiers 23d and e could be a National Semiconductor LF347, the counters 33a and b a type 74LS163 the analog multiplexers 29a and b could be a National Semiconductor LF11509. If large enough volume were anticipated it would be possible to build an analog FIFO integrated circuit using MOS charge coupled device technology.

FIG. 6 shows the preferred embodiment of the invention in a system configuration. An encoder made up of a modulator 47 and combiner 46 has as its input two audio channels and an associated video signal. The signals applied at the audio channels need not be program audio but any signal such as those previously described. It should be noted that in this example the audio signals are D.C. offset to an amount corresponding to 1/2 of the peak positive video voltage to ensure that no negative pulses are added to the video signal. Audio input on 42a and b is applied to FIFOs 43a and b respectively. A clocking signal at exactly twice the video horizontal rate is also input to said FIFO's input clock causing said FIFOs to periodically sample the audio signals, thus effecting pulse amplitude modulation. The clocking signal may be derived from an oscillator utilized in the timing of the input video signal or from a phase locked loop responsive to said input video. At the appropriate time during horizontal blanking, as determined by timing generator 40a, and corresponding to positions 6-9 of FIG. 2, the previously stored samples or modulated pulses are clocked out of said FIFO's 43a and b. Said pulses clocked out are time multiplexed by switch 41 in response to timing generator 40a which time multiplexed pulses are added to aforementioned input video which has been buffered by 23f by adder 36. When pulses are not being inserted switch 41 selects ground, allowing video to pass through adder 36 unchanged. The combination of said multiplexed pulses and buffered video will appear typically as shown in FIG. 2, which combination is output from aforementioned encoder and passed through transmission channel 45 to the decoder. Phase lock loop 39a and timing generator 40a are driven from sync stripper 38a which recovers information from buffered video signal output from 23f. Phase lock loop 39a generates a clocking signal exactly twice the horizontal sync rate of the video signal which provides a precise frequency for the pulse modulation. Alternately a precision oscillator or external reference could be used, and frequencies other than twice horizontal may be used. Timing generator 40a ensures that exactly the same number of pulses are taken out of said FIFOs as are input. This is a relatively simple process during active video lines, however during the vertical sync it is undesirable to insert pulses in vertical sync. The timing generator 40a must therefore not insert pulses during vertical which causes each FIFO 43a and b to store the 18 pulses (2 pulses per H.times.9 H) generated in each audio channel until after the vertical sync, where all of the stored pulses are then inserted in video as is shown in FIG. 3. It is also possible to store all samples made during the video frame or field and combine those samples with video only in the vertical interval. The need for adding pulses in horizontal blanking is eliminated, thus absolutely no changes are made to the horizontal blanking interval. The main disadvantage to this system is the large FIFO requir