Document text
NASA-TM-87578 19860005008
NASA Technical Memorandum 87578
ANALOG FM/FM VS. DIGITAL COLOR TV TRANSMISSION
ABOARD SPACE STATION
MICHAEL M. HART
OCTOBER 1985
IWV5A
National Aeronautics and
Space Administration
Langley Research Center
Hampton, Virginia 23665
Summary
Langley Research Center is developing an integrated fault tolerant
network to support data, voice, and video communications aboard Space
Station. The question of transmitting the video data via dedicated analog
channels or converting it to the digital domain for consistency with the
rest of the data is addressed. The recommendations in this paper are based
on a comparison in the signal-to-noise ratio (SNR), the type of video
processing required aboard Space Station, the applicability to Space
Station, and how they integrate into the network.
Introduction
For Space Station rendezvous and proximity operations, a video image
system will provide man with enhanced and augmented capability for man-in-
the-loop manipulation and observations. As the Space Station matures, some
of these operations will become nearly autonomous with man monitoring and a
video system could be a supportive element for this autonomy trend. In
addition, the Orbital Maneuvering Vehicle (OMV) has a remotely piloted
vehicle (RPV) requirement implying potentially extensive video for viewing.
Obviously, full motion and high fidelity video images are required to
perform the above task.
Langley Research Center took on the task of developing a fault-tolerant
integrated network for Space Station. The transmission of video data to and
from points in the Space Station will be supported by the network.
Typically, the output of video cameras is an analog signal. Hence, the
transmission of these data can be accomplished via dedicated analog channels
or by first converting the data to the digital domain before transmitting it
over the network. That leads to a question to be resolved: Which method
will be most beneficial? Although most data transmissions are accomplished
2
in the digital domain, further analysis and trade off must be considered
before the choice is made. This paper compares the FM and the PCM in terms
of signal -to-noise ratio (SNR), their applicability to Space Station, which
is most suited for the type of video processing needed aboard Space Station,
and how they integrate to the rest of the network. Based on the above
criteria, a recommendation is presented.
Bandwidth and Signal-to-Noise Ratio
First, the signal -to-noise ratio will be derived by using an ideal
unrealizable communication system implied by the Shannon-Hartley law to
establish a figure of merit. Then a comparison between the practical PCM
and FM systems will be presented, particularly in the exchange of bandwidth
(BW) for signal-to-noise ratio (SNR). Suppose we have a communication
system (fig. 1) transmitting signal X(t) bandlimited to f x Hz. Further,
suppose that the system is ideal, the channel bandwidth is By, and the
noise power spectral density is r\/2. Also, let us assume that the average
signal power at the receiver is S r and that the desired value of the
output SNR is ( S/N )<-) . The channel capacity of such a system is given by
Shannon-Hartley law as
C = By 1 ogg [ 1 + (S/N) r ] 0)
where ( S/N ) r is the SNR at the receiver input. At the receiver output,
the information rate can be no greater than Rmax> where
Rmax 2. f x ^°92^ + ( S/N ) d ]
( 2 )
An optimum or ideal system is defined as one that is operating at its
capacity, with maximum output rate that is
**max = C
or B t log 2 Cl + (S/N ) r ] = f x log 2 Cl + (S/N) d ]
solving for (S/N) d we have
(S/N ) d = [1 + (S/N) r ] B T/f x - 1
* Cl + (S/N) r )] B T/f x
when the SNR is large. In eq. (3) the input SNR (S/N) r is given by
(S/N ) r = J_r
nBj
The ratio of By/f is called the bandwidth expansion ratio (or BW
X
compression ratio if the ratio is less than one).
Let H = By /f x
and a = Sr
nf x
then eq. (3) can be rewritten as
4
(S/N) d - [1 + - 1
( 5 )
if the SNR is 1 arge.
Eq. (5) shows that, in an ideal system, the SNR at the output and the
bandwidth are exponentially related. This means that doubling the
transmission BW of an ideal system squares the output SNR. Alternately,
since a = Sr/pf x is proportional to the transmitted power Sj, the
transmitted power can be reduced to the square root of its original value
without reducing ( S/N ) ^ by increasing the bandwidth by a factor of 2.
The primary goal of the ideal system is to achieve reliable information
transfer in the context of an information theoretic approach. In analog
systems it is very difficult to assess the information rate. The primary
concern in such applications might be SNR, threshold power, and BW
requirement rather than channel capacity and its utilization. In fig. (1),
if we assume that the system is operating with large SNR, the FM and PCM
schemes are above threshold, and the message signal is normalized to
E[X 2 (t)] = E[X 2 (t)] = 1/2; then
( S/N ) d = 3
inf x
( 6 )
where (f^/f x ) is the deviation ratio
and S x is the power of X ( t )
*
f
0
5
The PCM SNR at the receiver output is given by
2 2n
{S/N) d = ~ ojT
l+4Pe2 2N
(7)
where N is the word length
and P e is the probability of a bit error.
Equation 7 is valid for all values of input signal-to-noise ratio.
However, notice when P e gets very small, i.e., the detection threshold is
reached, the value 1+4 P e 2 2 N tends to 1 and
(S/N) d - 2 2 N (8)
One can conclude from eq. 8 that increase in the input power above
threshold level yields no improvement in the value of (S/N)^.
The performance of the PCM operating above threshold is limited by
quantizing noise, and the SNR is given by
( S/N ) d = Q 2 (9)
where Q is the number of quantizer levels.
Now if we assume a binary PCM, and if the sampling rate is f s = 2f x , then
the transmission bandwidth By is given by
B T s f x i°g 2 Q (10)
Solving for Q we obtain
q = = 2^
( 11 )
6
For a PCM that is operating above threshold, the value of Q in eq. 9
can be substituted by the value of Q in eq. 11 and we have
( S/N ) d = 22H (12)
Eq. (6) shows that the FM does not have an exponential power dependence,
while the PCM does have an exponential power-bandwidth relationship.
In order to compare ( S/N )cj to ( S/N ) r = a and compare a to H, we must
derive an expression for a for the PCM. In order to derive the a for
the PCM, the following assumptions are made: the noise has a Gaussian
distribution; the (S/N)^ needed to be produced = 50 dB; and a binary PCM is
used. Hence, we need to derive the minimum ( S/N ) r = a needed to produce
(S/N)^ = 50 dB. To calculate the (S/N) r a threshold point must be defined.
The point at which symbol error due to channel noise occur? with probability
P e < 10"4 is chosen as the threshold.
For PCM, P e is obtained by
with r s = 2f x log2 Q
then for P e < 10"^
*G(Z 0 ) is the area under a normal Pdf.
7
if Z 0 satisfies G(Z 0 ) = 10"4 then we have
2
a > (log 2 q) Z 0
since (S/N)^ = Q 2
then for ( S/N ) ^ = 10 5
we need Q = 316.
Knowing the value of Q we can compute H. Fig. (2) indicates that
increasing the transmitted power for the PCM beyond threshold yields no
further improvement in ( S/N ) ^ since the limiting value of ( S/N ) ^ is
determined by quantization. Fig. (3) indicates that the power-bandwidth
exchanges in PCM is considerably better than the FM. The PCM system
requires about 6 dB more power than the ideal system. In summary, we can
say that FM and PCM offer wideband noise reduction and PCM is better than FM
at low input SNR. In addition, the exchange of bandwidth for power is
easier to accomplish in PCM as has been shown. Since PCM can be easily time
scaled, time can also be exchanged for signal power. Thus, the
communications system designer has added flexibility to meet a given
performance criteria.
Suitability for Image Processing
The following candidate image processing functions may be required of
the video system aboard the Space Station: Smoothing, enhancement,
* restoration and filtering, data compression, feature extrapolation detection
and identification, interpolation/extrapolation, spectral estimation,
spectral factorization, and synthesis.
8
These capabilities are much easier done in the digital domain than in
the analog domain in terms of computation, speed, and hardware complexity.
The reasons for the above is the availability of integrated circuits that
are more reliable and stable than the analog circuits. In addition, in the
analog domain a circuit that does a particular function does not lend itself
to doing another function unless the computations are quite similar to each
other, while in the digital domain the flexibility is a lot greater than the
analog. In other words, in order to do the previously mentioned processing
in the analog domain, we might end up with a dedicated circuit for each
function. In addition, it is easy to store and time scale PCM signals.
Digital memories can accomplish storage and retrieving a lot more
efficiently than analog. PCM signals can be time-dimension multiplexed a
lot easier than analog. With PCM systems, source coding and channel coding
can be used to reduce the redundancy in messages and to reduce the effects
of noise and interference.
Applicability to Space Station
One of the design considerations in the network is to use intelligent
nodes, where the nodes make the routing decision in cooperation with other
nodes. The connection among these nodes will be fiber optic buses. Hence,
there will be constant conversion from electric to optic and vice versa at
each node. That will result in a considerable amount of loss and the need
to regenerate the signal. PCM signals can be completely regenerated at each
repeater station if the repeater spacing is such that the magnitude of the
noise is less than 1/2 the separation between levels (with a high
probability). With the exception of occasional errors, a noise and
9
distortion free signal is transmitted at each repeater. Furthermore, the
effect of noise does not accumulate and in designing repeaters one needs to
be concerned only about the effects of channel noise and signal loss between
repeater stations. Repeaters for analog modulation schemes consist of
amplifiers that raise the signal level at each transmitting station. While
raising the signal level, the amplifier also raises the level of the noise.
However, the above issue might not be of great importance, since utilizing
FM means a dedicated video channel where the loss and noise might not be of
considerable magnitude. Some might argue that FM is more applicable to
Space Station since the complexity of a PCM system is greater than that
required for the FM. However, the complexity varies little as the number of
channels is increased. Hence, PCM can compare quite favorable when the
number of channels, is large, which might be the case in Space Station.
Integrability to the Network
Since all other components on the network require a digital communi-
cation channel, the use of FM will require separate analog channels. In
that case, if the video data are to be used in any other purpose other than
display, the data must be converted into the digital domain. Using the PCM
does not mean that it eliminates the choice of having separate video
channels. It will add the flexibility of either using dedicated channels or
integrating the video channels with the rest of the network.
Concluding Remarks
It has been shown that the PCM has a better flexibility in power
bandwidth trade-off, is better in image processing and is integrable to the
10
network. Therefore, it is obvious that the PCM will be a more beneficial
scheme to use in transmitting the video data. The only exception where the
FM will be a better choice is in case the video data will not require any
processing and only be used in display.
<
!
11
References
1) Shanmugan, K. S.: "Digital and Analog Communication Systems," Published
by John Wiley and Sons, 1979.
(S/N) H {dB)
13
oi(dB)
Fig. 2- SNR in Communication System vs a = (S/N) r and H = 8.
a(dB)
1. Report No. 2. Government Accession No.
NASA TM-87578
3. Recipient's Catalog No.
4. Title and Subtitle
"Analog FM/FM vs. Digital Color TV Transmission Aboard
Space Station"
5. Report Date
October 1985
6. Performing Organization Code
506-58-13-02
7. Author(s)
Michael M. Hart
8. Performing Organization Report No.
10. Work Unit No.
9. Performing Organization Name and Address
NASA Langley Research tenter
Hampton, VA 23665
11. Contract or Grant No.
13. Type of Report and Period Covered
Technical Memorandum
12. Sponsoring Agency Name and Address
National Aeronautics and Space Administration
Washington, D.C. 20546
14. Sponsoring Agency Code
15. Supplementary Notes
16. Abstract
Langley Research Center is developing an integrated fault tolerant network to
support data, voice, and video communications aboard Space Station. The question
of transmitting the video data via dedicated analog channels or converting it to
the digital domain for consistancy with the rest of the data is addressed. The
recommendations in this paper are based on a comparison in the signal-to-noise
ratio (SNR), the type of video processing required aboard Space Station, the
applicability to Space Station, and how they integrate into the network.
17. Key Words (Suggested by Author(s)) 18. Distribution Statement
Signal-to-Noise Ratio (SNR), Applicability
of PCM and FM to Space Station, PCM and FN Unclassified - Unlimited
Integrabil ity to the rest of the Network
Aboard Space Station Subject Category (32)
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified 15 A02
N-305
For sale by the National Technical Information Service, Springfield, Virginia 22161