Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
71~i,~t.
METHOD OF REFORMATTING COLOR TELEVISION SI~NALS
~ACKGROUND OF THE INVENTION
This invention relates to the field of transmission and
recording of color television signals.
As it will shortly be necessary to refer to the
drawings, these will first be briefly described as
follows:
Figure 1 illustrates an NTSC composite color video
signal;
Figure 2 illustrates a typical signal where the
luminance and chrominance data have been compressed and
serially arranged;
Figure 3 illustrates a reformatted signal;
Figure 4 illustrates an embodiment of a system that
can perform a reformatting operation; and
Figure 5 illustrates a system used during playback
for changing the reformatted recorded signal to NTSC for
display on a television screen.
Figure 6 illustrates a system for converting from the
prior art NTSC signal to the reformatted MAC signal.
-la-
Video tape recorders are available that can directly
record television signals as they are normally broadcast.
These recorders must have very high bandwidth capabilities
to record such signals and are, therefore, very expensive.
The common video recorders that are available on the
market are less expensive and have much lower bandwidth
capabilities. Since these low cost recorders cannot
directly record televislon signals accurately, various
types of processing are performed on the signals to reduce
the bandwidth requirements. Standard television signals
contain composite chrominance and luminance information
broadcast during the active video portion of a conventional
video line. Figure 1 illustrates an NTSC composite color
video signal. The active video portion is approximately
52.S microseconds in duration with the entire video line
occupying 63.5 microseconds. The remaining 11 microseconds
is reserved for synchronizing pulses (H-SYNC), clamping,
transition times, and a color reference signal (BURST). A
typical prior art method of recording such a signal was to
first separate the luminance and chrominance information
and then record the separated signals on tape by different
techniques, most commonly by frequency modulating the
luminance component onto a carrier and amplitude modulating
the chrominance signal onto a lower-frequency subcarrier.
The signal containing the chrominance information is then
added to the frequency modulated luminance carrier and the
combined signal is recorded.
Z
The above described method is subject to a number of problems including
several sources of picture degradation such as a poor chrominance signal-to-noise
ratio, gain and delay inequalities between chrominance and luminance and distor-tion such as differential phase and gain. These problems become amplified when
several generations of copying occur.
Several methods have been introduced which attempt to alleviate some or
all of the above sources of degradation. One involves arranging the chrominance
and luminance information serially. A conventional composite video signal is
reformatted serially into a new signal containing lines of serially arranged
chrominance and luminance information. The chrominance and luminance com-
ponents àre compressed before they are arranged into the serial format. The
compression step is necessary since the serially reformatted signal also has a line
duration of 63.5 us and if the components are not compressed they cannot ~fit~
into this time period when they are arranged serially. Figure 2 illustrates a typi-
cal signal where the luminance and chrominance data have been compressed and
serially arranged. The lumlnance data has been compressed from its original 52.5microseconds, the length of the active line portion, to 46 microseconds with thechrominance being further compressed to 11.5 microseconds. The chrominance
data is usually time compressed more than the luminance data, typically by a
small integer such as 2, 3 or 4. (Chrominance data usually consists of two colordifference signal~s, here Cr and Cb, and is arranged on alternate lines with thecorresponding luminance intormation Y).
Thls type of serially formatted signal is similar to a Multiplexed Analog
Component (MAC) signal which is proposed for satellite and cable television
transmission. A MAC signal is a serially reformatted signal with a luminance
component that has been compressed from its original 52.5 microseconds to 35
microseconds and a chrominance component that is compressed to 1~.5
microseconds. Although the use of MAC signals has overcome many of the
7~1.6;Z
- 3
problems inherent in the use of NTSC signals, several new problems have
developed.
Television signals are transmitted along cable systems in a nominal 6 MHZ
band. The MAC carrier, by convention, must be 1.25 MHZ above the bottom end
of the band; and the system response falls off above about 5.25 MHZ above the
bottom OI the band which leaves a usable 4 MHZ of bandwidth. Using ordinary 3:2
compression for luminance, this means that after decompression, the luminance
signal will have a bandwidth of only 2/3 of 4.0 MHZ or 2.67 MHZ.
A similar problem is encountered in the use of the video recording equip-
ment commonly sold on the market. As noted previously, a typical recorder has
bandwidth llmitations and the use oi compressed signals causes problems similar
to those encountered during cable transmission. U.S. Patent No. 4,335,393 to
Pearson, for example, discloses a method of recording a serially reformatted sig-
nal. Pearson compresses the luminance and chrominance data from its original
duration in the active video region to form a signal similar to the one illustrated
in fig. 2. This compressed signal is then recorded. When Pearson~s signal is
recovered for playback there will have been a sigaificant bandwidth loss due to
the compression step. Several systems have been introduced which attempt to
solve this problem.
U.S. Patent No. 4,467,368 to Horstmann discloses a method of serially
recording luminance and chrominance data. Horstmann time-expands the
luminance signal berore recording in order to avoid the bandwidth problems inher-
ent in a system such as Pearson's. Since Horstmann has time expanded his signal,however, the signal will not fit within the 63.5 us limit and Horstmann must usetwo separate channels when recording (see Figure 4). Each channel consists of a
time expanded luminance signal and a compressed chrominance signal arranged
serially. Since the recording is being performed in two channels, there is twice as
much recording time per line and the signal can therefore be time expanded. The
J~
,i. ~.~ .
716~
drawback to this method i8 that a much greater amount of
recording space is required.
U.S. Patent No. 3,781,463 to Van den Bussche is
directed to a method of recording luminance and
chrominance information serially in a single recording
channel without uslng a compressed luminance signal.
Since the luminance signal i8 not compressed, the
chrominance signal must be compre~sed to a much greater
degree than was necessary in oth~r prior art systems to
satisfy the time requirements. For example, the ratio
between the length of the active line time and the
length of the compressed chrominance signal is
approximately 8 to 1. Such a high compression ratio
means that there is also a high ratio between the time
occupied by the luminance and chrominance components
which will cause the signals to become degraded, with
the color having the tendency to ~mear over the
luminance.
SUMMARY OP THE INVXNTION
It is an ob;ect of an a~pect of the present invention to
reformat color television signals to reduce the
bandwidth requirements for transmission or recording.
Further, an ob~ect of an aspect of this invention
i8 to reformat color television signals in such a way
that luminance and chrominance data are accurately
serially recorded in a single channel.
An ob~ect o~ an aspect of the invention is to avoid
excessive compression of luminance and chromin~nce data
when transmitting or recording color television signals.
Conventional television receivers overscan the
screen horizontally by anywhere from 10%-18%. This
means that only 82%-90% of each line of the transmitted
picture signal is displayed. There is no need to
transmit or record the portion of the video signal which
is never displayed on the screen. Accordingly, a
portion of the 52.5 microsecond active video time of a
conventional signal can be truncated with no loss of
useful information.
, .~.,.. )i .
A t ~ mul~plexed signal or MAC signal is reformatted by
a method c~prising ~e steps of: (l) at le~t partially
deo~ressing ~e compressed l~nance c~nent; (2) bn~cating
both the dec ~ ressed l ~ nance and compressed ~ o~nance ca~nen~;
and (3) serially c ~ ining ~e ca~nen~ to form reformatted
video lin~. me order of th~e steps may be varied ~ indicated
more fully below. The b~ncation step involves the
elimination of signals that repr~aent portions of the
active line that can be eliminated without any loss of
useful information since television picture tubes
normally overscan and the left and right edges of the
picture are not seen. With the reduction in signal
length due to the truncation ~tep, the decompressed
luminance can be transmitted or recorded serially with a
chrominance signal that is compressed by a low ratio.
Various aspects of the invention are as follows:
A method of reformatting each line of a time
multiplexed color television signal such as a ~AC
signal, wherein each line comprises a serially combined
~0 compressed luminance and compressed chrominance
component, comprising the steps of:
at least partially decompressing the compressed
luminance component:
truncating a portion of both the luminance and
chrominance components that represents an overscanned
portion of the television signal: and
serially co~bining the components to form
reformatted video lines.
A method of reformatting each line of a standard
color television signal, wherein each line comprises
composite chrominance and luminance information,
comprising the steps of:
separating the composite signal into chrominance
and luminance component~:
truncating a portion of each of the chrominance and
luminance components that represents an overscanned
portion of the television signal;
compressing the chrominance component; and
5a
~.~ r3'716~
serially combining the luminance and compressed
chrominance components to form reformatted video lines.
A method of reformatting each line of a time
multiplexed color television signal such as a MAC
signal, wherein each line comprises serially combined
compressed luminance and compressed chrominance
components, co~prising the steps of:
decompressing the compressed luminance component;
truncating a portion of both the decompressed
luminance component and the compressed chrominance
component, ~aid portion representing an overscanned
portion of the television signal; and
serially combining the components for recording and
playback on a single recording channel with a
chrominance compression ratio of less than or equal
to 9 to 2.
An apparatus for reformatting each line of a time
multiplexed color television signal such a MAC signal,
wherein each line comprises a serially combined
compressed luminance and compressed chrominance
component, said apparatus comprising:
decompressing means for at least partially
decompressing the compressed luminance component:
truncating means for truncating a portion of both
the luminance and chrominance components that represents
an overscanned portion of the television signal; and
combining means for serially combining the
components to form reformatted video lines.
An apparatus for reformatting each line of a
standard color television signal, wherein each line
comprises composite chrominance and luminance
information, said apparatus comprising:
separating means for separating the composition
signal into chrominance and luminance components;
truncating means for truncating a portion of each
of the chrominance and luminance components that
represents a~ overscanned portion of the television
signal;
5b~ 716~
compressing means for compressing each chrominance
component; and
combining means for serially combining the
components to form reformatted video lines.
DE~AILED DESCRIPTION OF ~ PREYEF~ED E~BQDr~E~
Standard television picture tubes overscan and the
left and right edges of the picture are never seen. By
eliminating this unnecessary in~ormation, television
signals can be formatted in a serial manner with
significantly les~ compression than that performed by
prior art system~. Approximately 10-18% of the actual
line time of 52.5 microseconds contains information that
the picture tube does not display. The elimination of
this portion of the signal requires less information to
be transmitted or recorded. The luminance and
chrom~nance signals can now be combined with much less
compression than that which was necessary when the
entire active vldeo portion was used.
When reformatting a signal such as the one
illustrated in Figure 2 for recording on conventional
video recorder~, the luminance signal is decompressed to
restore it to its original 52.5 microseconds. The first
2.5 microseconds and the last 2.5 microseconds of the
signal are then eliminated resulting in a 47.5
microsecond signal. The compressed chrominance data is
also truncated by an amount corresponding to five
microseconds of the active video region. If the
chrominance signal had been compressed by a 4 to 1
(or reduced to 13.125 microsecondsj ratio the 5.0/4.0
= 1.25 microseconds would be eliminated resulting in an
11.875 microsecond chrominance
3716~
signal. The combined luminance and chrominance signals now only occupy a total
time of 59.375 microseconds which allows the signaLs to be serially recorded
within the 63.5 microsecond limit. The chrominance signal can be compresssed by
other appropriate ratios such as 15:4 which would result in a 12.66~ microsecondchrominance signal. Figure 3 illustrates such a reformatted signal.
The elimination of the unnecessary portion of the signal has allowed the
signal to be refomatted into a serial combination of luminance and chrominance
without c~pressing the l~n~nance and with only m~ni~nal c~ression of
chrominance. Since the luminance signal is not compressed there will be no
reduction in bandwidth on playback and, therefore, the entire available bandwidth
of 4.0 MHZ w~ll be utilized. Thls is a signi~icant improvement over systems which
record compressed limlnance data.
1~ lt is desired to reformat a standard television signal such as the signal
illustrated in Figure 1, the method would comprise the following steps:
(1) separating the luminance and chrominance components; (2) truncating 2.5
mlcroseconds from the beginning and end of both the luminance and
chrominancne signals; (3) compressing the chrominance signal by an appropriate
ar.:ount such as 4 to l and (4) serially combining the two signals.
Figure 4 shows one embodiment of a system that can perform such a
reformatting operation. For illustrative purposes, the system will be described
operating on a standard MAC signal. As previously described, such a signal uses a
3 to 2 compression for luminance and a 3 ~o 1 compression ror chrominance.
The MAC signal is received by a satellite dish or other means and is
supplied to tlme gate 5. Gate S allocates the signal betweesl three possible paths,
and will be in the upper position, supplying line store 10, during that portion of
the line that represents the luminance information that will be recorded. Gate Swill not supply all of the luminance information to line store lO but only the por-
tion that will be displayed by the television. For example, if 5.0us of the original
~.2~3716X
active line time is to be truncated, then 2/3 of 5.0us of the 35us luminance period
or 3.334us of this signal will not be supplied to line store 10. The truncated por-
tion of course, will be taken in equal increments from both sides of the active
video portion. Similarly, gate S will be in the lowermost position in Figure 4, sup-
plying line store 11 during the portion of each video line that represents desirable
chrominance information. According to our example, 5.0us of the active line timeor 1/3 of 5.0us of the compressed signal will be truncated. This 1.67us will also be
eliminated in equal amounts from each end of the chrominance portion (0.835us
from each end). During those portions, gate 5 is ln the center position thus
deliverlng data to delay 9.
Line stores 10 and 11 read the truncated information Into memory. The
speed at which the data is read into the memory and subsequently written out of
the memory is controlled by clocks 12 and 13. By varying the ratio of the read in
to write out rates, the data can be decompressed or compressed by any desired
amount. For example, if luminance data is read into line store 10 at a freguencyof Fl, it can be completely decompressed by writing it out of line store 10 at afrequency equal to (2/3) Fl. In this manner the original 3 to 2 compression of the
luminance data is eliminated. The compression of the chrominance data is simi-
larly controlled by line store 11 and clock 13. If it is desireable not to change the
compression of any data, it would simply be read and written at the same rate.
Time gate 15 wlll be operative to supply the truncated signals to an output.
There are three positions for gate 15. In the uppermost position, the truncated,decompressed luminance data is supplied to the output while in the lower position
the truncated chrominance data is supplied to the output. As noted earlier, tllechrominance data may or may not have been compressed or decompressed,
depending upon the difference between its initial compression rate and the rate
desired for the output. The 3 to 1 compression of chrominance contained in a
standard MAC signal is undesirable since the remaining time per line af ter
~ ~f~716~
-- 8 --
decompressing the luminance data will be insufficient ~or sync, clamping and
transition times. To obtain the desirable signal shown in Flgure 3, the
chrominance compression should be read into lire store 11 at trequency Fl, and
read out at (5/4) F1. This would result in a total compression ot (3/1)(5/4), or 15 to
4. Delay 9 functions to assure tha the sync pulses andother data is delivered tothe output at the appropriate time. Other delays may be used as necessary. The
information should ~ecome available to gate 15 in a sequential manner in order to
form the reformatte~d video llnes. In other words, for a particular line, line
store 11 would ~irSt write out its chrominance data with gate 15 set on the lower
level. Next, line store 10 would read out its data and then delay 9 would (ollow.
The order of these operations is unimportant and may be varied as desired. The
re~ormatted MAC slgnal can then be supplied to a standard VCR i'or recording.
F~gure S shows a system used during playback ror changing the reformatted
recorded signal to NTSC lor display on a television screen. The signal is played by
VCR 20 and is supplied to a time ga~e 2S. Gate 26 has two possible pQsitions. The
lower position supplies chromlnance informatlon to a line store 22. By controlling
the read and write rates ot clock 26 as previously discussed, the chrominance
in~ormation is decompressed. The decompressed signals are supplied to
demultiplexer 24 whlch separates the Cr chrominance trom the Cb chrominance.
These slgnals are then supplled to a NTSC encoder. During that portion of each
line that does not contain chrominance information, gate 26 is in the upper posi-
tion and supplies the rest ot the signal, Including the luminance data, to a time
delay 21. Delay 21 assures that luminance and chrominance intormation trom the
same line arrive at the NTSC encoder 25 simultaneously. The encoder 25 recei~lesthe three signals and produces a standard NTSC output. Such encoders are com-
mon and encoder 25 is not, there~ore, shown ~n greater detail.
As discussed above, a standard television signal such as the
signal illustrated in Figure 1 can also be reformatted in accordance with
the method of this invention. Figure 6 sh~s a system that can perform
such a reformattin~ operation. me system illustrated in Figure 6 is
substantially identical to the system of Figure 4, the only difference being
the addition of composite signal separator 32. Separator 32 will receive
a standard NTSC signal and will serially reformat the signal in a manner
similar to a MAC signal. The serially reformatted signal is then supp1ied
to time gate 5 which will operate in conjunction with the remainder of the
circuit in exactly the same manner as discussed above in relation to
Figure 4. There have been many devices disclosed in the prior art which
function to convert NTSC signals to serially formatted signals, and it is
therefore unnecessary to specifically describe the operation of composite
signal separator 32.
Alternatively, the Cr, Cb and Y outputs could be supplied to a matrix which
would produce R, C and B outputs. These outputs could then be supplied to
approp.iate inputs on the consumer's television.
~ 2~3716~
When a signal is to be reformatted for transmission purposes the same prin-
ciple applies although a transmitted signal must keep the video portion within the
original 52.5us active line time. In other words, when the signal is being
reformatted for recording purposes a larger portion of the 63.5us line time can be
used to carry the video information. In transmission, however, their is no extratime available and the luminance and chrominance data is confined to 52.5us.
Approximately 10% of the active video line of 52.5us will be truncated. A
truncation leaving 90.4% o~ the active line has achieved the best results. The
actlve line of 52.Sus is thereby reduced to 47.46us.
It is possible to re~ormat the signal so that the luminance component is
completely decompressed as was done when re~ormatting i'or record~ng purposes.
This would leave 5.û4us to carry the chrominance data (52.5-47.46) . The
chrominance data would have to be compressed by a ratio o~ more than 9 to 1
which ls undesirable slnce it would severely reduce the available bandwidth for
chrominance. In the preterred embodiment the chrominance is compressed by a 6
to 1 ratlo thus reduclng the 47.46us chrominar.ce signal to 7.91us. This leaves
44.59us to carry the luminance data. It is undesirable to utilize all ol this time
since a small amount ot time is required tor transition tlmes, etc. The luminance
signal is therefore compressed trom its original 47.46us (after truncation) to
44.33us. Thls is a compression ratlo of approximately 1.0706 to 1. It may be
desirable to low pass l'ilter the chrominance signal prior to compressing the signal
in order to avoid problems with aliasing.
By using such a low compression ratio for luminance, a signal can be trans-
mitted down a cable system utilizing most of the available 4.0 MHZ o~ bandwidth.Specil'ically, by compressing the signal by a 1.0706 to 1 ratio the utilized bandwith
Is 3.74 MHZ (4.0/1.0706).
In re~ormatting television signals, either for recording or transmitting pur-
po~ " the present invention relies on the basic premise that there is no need to
~ 2~3'71~,X
- 10 --
transmit or record data which is not displayed on the television screen. The
amount of truncation that has been performed in the illustrative embodiments
contained herein do not eliminate all of the overscan since different televisions
overscan by varying degrees. Some of the overscan should therefore be left in
order to insure that the edges of the picture can never be seen on the screen. It
is, however, possible that some consumers may have televisions that overscan by
less than a ~normal~ amount and picture edges will be displayed.
In order to correct for the above unlikely although possible occurence, the
displayed picture can be ad~usted by the consumer. A slight expansion of the dis-
played picture can be performed in the same way that other expansions and com-
pressions have been effectuated in the above illustrative embodiments. Prior to
display, the signal can be read into a line store at a certain frequency and subse-
quently read out a slightly lower frequency, thus slightly expanding the picture.
This could be easily performed using the components in a standard subscriber
decoder. A typical decoder receives a MAC signal and reads it into line stores at
the standard rate (the same rate that it is received). Decompressions are formedby clocking the data out at a slower rate. A variable switch could be provided so
that a subscriber could slightly ad~ust the read out rate, thus changing the size of
the displayed picture.
The specific embodiments recited above are given only for illustrative pur-
poses and it should be clearly understood that various modifications are possible.
The precise amount of truncation can be varied, for example, as can the specificcompression ratios used. The specific order of the steps performed by the present
invention may also be modified. For example, the truncation can be performed
before or after decompression or compression. Similarly, various modifications of
the components, parts etc. may be employed without departing from spirit of the
invention or the scope of the appended claims.