U.S. patent application number 14/285026 was filed with the patent office on 2015-11-12 for spectrally weighted analog to digital conversion.
This patent application is currently assigned to CABLE TELEVISION LABORATORIES, INC.. The applicant listed for this patent is CABLE TELEVISION LABORATORIES, INC.. Invention is credited to Belal Hamzeh.
Application Number | 20150326237 14/285026 |
Document ID | / |
Family ID | 54368720 |
Filed Date | 2015-11-12 |
United States Patent
Application |
20150326237 |
Kind Code |
A1 |
Hamzeh; Belal |
November 12, 2015 |
SPECTRALLY WEIGHTED ANALOG TO DIGITAL CONVERSION
Abstract
Systems and methods presented herein provide for analog to
digital conversion with variable bit resolution. In one embodiment,
a system includes a processor and a multiplexer. The processor is
operable to receive an analog signal, to detect power spectral
densities in the analog signal, to segment the analog signal into
at least two frequency bands, to sample each of the frequency
bands, and to quantize each of the sampled frequency bands with bit
resolutions according to detected power spectral densities of the
frequency bands. The multiplexer is operable to multiplex the
quantized frequency bands into a data stream.
Inventors: |
Hamzeh; Belal; (Westminster,
CO) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
CABLE TELEVISION LABORATORIES, INC. |
Louisville |
CO |
US |
|
|
Assignee: |
CABLE TELEVISION LABORATORIES,
INC.
Louisville
CO
|
Family ID: |
54368720 |
Appl. No.: |
14/285026 |
Filed: |
May 22, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61991833 |
May 12, 2014 |
|
|
|
Current U.S.
Class: |
341/141 |
Current CPC
Class: |
H03M 1/122 20130101;
H03M 1/0854 20130101; H03M 1/121 20130101 |
International
Class: |
H03M 1/08 20060101
H03M001/08; H03M 1/12 20060101 H03M001/12 |
Claims
1. A system, comprising: a processor operable to receive an analog
signal, to detect power spectral densities in the analog signal, to
segment the analog signal into at least two frequency bands, to
sample each of the frequency bands, and to quantize each of the
sampled frequency bands with bit resolutions according to detected
power spectral densities of the frequency bands; and a multiplexer
operable to multiplex the quantized frequency bands into a data
stream.
2. The system of claim 1, wherein: the processor is further
operable to assign a lower bit resolution to a first of the at
least two frequency bands having a lower power spectral density to
reduce the overall bit rate of the data stream.
3. The system of claim 1, wherein: the processor is further
operable to assign a higher bit resolution to a first of the at
least two frequency bands having a higher power spectral density to
improve a signal to noise ratio of the first frequency band.
4. The system of claim 1, wherein: the processor comprises first
and second bandpass filters operable to respectively segment the
analog signal into first and second of the at least two frequency
bands; and the processor is further operable to sample the first
and second frequency bands at a first sampling rate, to quantize
the first frequency band with a first number of bits per sample,
and to quantize the second frequency band with a second number of
bits per sample, wherein the first and second numbers of bits per
sample are different.
5. The system of claim 1, wherein, for first and second of the at
least two frequency bands: a bit rate of a first frequency band is
about a sampling rate of the first frequency band times a number of
bits assigned per sample of the first frequency band; a bit rate of
a second frequency band is about a sampling rate of the second
frequency band times a number of bits assigned per sample of the
second frequency band; and a bit rate of the data stream is about
the sum of the bit rate of a first frequency band and the bit rate
of a second frequency band.
6. The system of claim 1, wherein, for first and second of the at
least two frequency bands: the processor is further operable to
sample the first frequency band at a first sampling rate, to assign
a first number of bits to each sample of the first frequency band,
to sample the second frequency band at a second sampling rate, and
to assign a first number of bits to each sample of the second
frequency band.
7. The system of claim 1, further comprising: a digital signal
processor operable to receive the data stream, to demultiplex the
data stream into separate substreams according to the bit
resolutions, and to convert the substreams into another analog
signal having a total bandwidth corresponding to a sum of the at
least two frequency bands.
8. A method, comprising: processing an analog signal; detecting
power spectral densities in the analog signal; segmenting the
analog signal into at least two frequency bands; sampling each of
the frequency bands; quantizing each of the sampled frequency bands
with bit resolutions according detected power spectral densities of
the frequency bands; and multiplexing the quantized frequency bands
into a data stream.
9. The method of claim 8, further comprising: assigning a lower bit
resolution to a first of the at least two frequency bands having a
lower power spectral density to reduce the overall bit rate of the
data stream.
10. The method of claim 8, further comprising: assigning a higher
bit resolution to a first of the at least two frequency bands
having a higher power spectral density to improve a signal to noise
ratio of the first frequency band.
11. The method of claim 8, further comprising: bandpass filtering
the analog signal to segment the analog signal into first and
second of the at least two frequency bands; sampling the first and
second frequency bands at a first sampling rate; quantizing the
first frequency band with a first number of bits per sample; and
quantizing the second frequency band with a second number of bits
per sample, wherein the first and second numbers of bits per sample
are different.
12. The method of claim 8, wherein, for first and second of the at
least two frequency bands: a bit rate of a first frequency band is
about a sampling rate of the first frequency band times a number of
bits assigned per sample of the first frequency band; a bit rate of
a second frequency band is about a sampling rate of the second
frequency band times a number of bits assigned per sample of the
second frequency band; and a bit rate of the data stream is about
the sum of the bit rate of a first frequency band and the bit rate
of a second frequency band.
13. The method of claim 8, further comprising, for first and second
of the at least two frequency bands: sampling the first frequency
band at a first sampling rate; assigning a first number of bits to
each sample of the first frequency band; sampling the second
frequency band at a second sampling rate; and assigning a first
number of bits to each sample of the second frequency band.
14. The method of claim 8, further comprising: processing the data
stream; demultiplexing the data stream into separate substreams
according to the bit resolutions; and converting the substreams
into another analog signal having a total bandwidth corresponding
to a sum of the at least two frequency bands.
15. A non-transitory computer readable medium comprising
instructions that, when executed by a processor, direct the
processor to: process an analog signal; detect power spectral
densities in the analog signal; segment the analog signal into at
least two frequency bands; sample each of the frequency bands;
quantize each of the sampled frequency bands with bit resolutions
according detected power spectral densities of the frequency bands;
and multiplex the quantized frequency bands into a data stream.
16. The computer readable medium of claim 15, further comprising
instructions that direct the processor to: assign a lower bit
resolution to a first of the at least two frequency bands having a
lower power spectral density to reduce the overall bit rate of the
data stream.
17. The computer readable medium of claim 15, further comprising
instructions that direct the processor to: assign a higher bit
resolution to a first of the at least two frequency bands having a
higher power spectral density to improve a signal to noise ratio of
the first frequency band.
18. The computer readable medium of claim 15, further comprising
instructions that direct the processor to: bandpass filter the
analog signal to segment the analog signal into first and second of
the at least two frequency bands; sample the first and second
frequency bands at a first sampling rate; quantize the first
frequency band with a first number of bits per sample; and quantize
the second frequency band with a second number of bits per sample,
wherein the first and second numbers of bits per sample are
different.
19. The computer readable medium of claim 15, wherein, for first
and second of the at least two frequency bands: a bit rate of a
first frequency band is about a sampling rate of the first
frequency band times a number of bits assigned per sample of the
first frequency band; a bit rate of a second frequency band is
about a sampling rate of the second frequency band times a number
of bits assigned per sample of the second frequency band; and a bit
rate of the data stream is about the sum of the bit rate of a first
frequency band and the bit rate of a second frequency band.
20. The computer readable medium of claim 15, further comprising
instructions that direct the processor to, for first and second of
the at least two frequency bands: sample the first frequency band
at a first sampling rate; assign a first number of bits to each
sample of the first frequency band; sample the second frequency
band at a second sampling rate; and assign a first number of bits
to each sample of the second frequency band.
Description
BACKGROUND
[0001] Digital signal processing (DSP) is the mathematical
manipulation of a signal to modify or improve the signal.
Generally, a continuous time analog signal is first converted to a
discrete time digital representation of the signal via the widely
known process of analog to digital (A/D) conversion. While DSP and
A/D are well known concepts that are used in many applications,
larger bandwidth signals can create problems. For example, a 1 MHz
bandwidth signal generally requires a 2 MHz sampling rate to avoid
aliasing. Each sample is represented digitally by some number of
bits. With more bits comes better resolution and better
signal-to-noise ratio (SNR) over quantization noise. So, if each
sample in the 2 MHz sampling rate was represented by 8 bits, the
bit rate of the sampled signal would 16 Megabits per second (Mbps).
This bit rate is easily obtained with current state of the art
processing. But, if the signal is a 100 MHz active Radio Frequency
(RF) signal and the resolution is 16 bits, then the bit rate of the
signal is 3.2 Gigabits per second (Gbps) without even considering
the effects of additional data required for error correction in
real time processing. These larger bandwidth signals require much
faster bit rates that are simply too difficult to process in real
time.
SUMMARY
[0002] Systems and methods presented herein provide for analog to
digital conversion with variable bit resolution. In one embodiment,
a system includes a processor and a multiplexer. The processor is
operable to receive an analog signal, to detect power spectral
densities in the analog signal, to segment the analog signal into
at least two frequency bands, to sample each of the frequency
bands, and to quantize each of the sampled frequency bands with bit
resolutions according to detected power spectral densities of the
frequency bands. The multiplexer is operable to multiplex the
quantized frequency bands into a data stream.
[0003] The various embodiments disclosed herein may be implemented
in a variety of ways as a matter of design choice. For example,
some embodiments herein are implemented in hardware whereas other
embodiments may include processes that are operable to implement
and/or operate the hardware. Other exemplary embodiments, including
software and firmware, are described below.
BRIEF DESCRIPTION OF THE FIGURES
[0004] Some embodiments of the present invention are now described,
by way of example only, and with reference to the accompanying
drawings. The same reference number represents the same element or
the same type of element on all drawings.
[0005] FIG. 1 is a block diagram of an exemplary A/D processing
system.
[0006] FIG. 2 is a flowchart illustrating an exemplary process of
the A/D processing system of FIG. 1.
[0007] FIG. 3 is a graph of an exemplary frequency domain.
[0008] FIG. 4 is a block diagram of another exemplary A/D
processing system.
[0009] FIG. 5 is a block diagram of another exemplary A/D
processing system.
[0010] FIG. 6 is a block diagram of one exemplary communication
system employing the A/D processing system.
[0011] FIG. 7 is a block diagram of an exemplary computing system
in which a computer readable medium provides instructions for
performing methods herein.
DETAILED DESCRIPTION OF THE FIGURES
[0012] The figures and the following description illustrate
specific exemplary embodiments of the invention. It will thus be
appreciated that those skilled in the art will be able to devise
various arrangements that, although not explicitly described or
shown herein, embody the principles of the invention and are
included within the scope of the invention. Furthermore, any
examples described herein are intended to aid in understanding the
principles of the invention and are to be construed as being
without limitation to such specifically recited examples and
conditions. As a result, the invention is not limited to the
specific embodiments or examples described below.
[0013] FIG. 1 is a block diagram of an exemplary A/D processing
system 100. The processing system 100 includes a processor 101 and
a multiplexer 102. The processor 101 is any system, device,
software, or combination thereof operable to sample an analog
signal x(t), determine power spectral densities across the
frequency domain of the analog signal x(t), and assign bit
resolutions to frequency sub bands according to those power
spectral densities. The processor 101 then outputs a plurality of
discrete variable bit rate signals x.sub.1(n)-x.sub.N(n) (where "N"
is merely intended to represent an integer greater than 1 and not
necessarily equal to any other N reference numeral herein)
corresponding to the assigned bit resolution/power spectral
densities. The multiplexer 102 is any system, device, software, or
combination thereof operable to combine the discrete signals
x.sub.1(n)-x.sub.N(n) from the processor 101 into a data stream.
Additional details regarding the A/D processing system 100 are
shown and described with respect to the process 200 of FIG. 2.
[0014] FIG. 2 is a flowchart illustrating an exemplary process 200
of the A/D processing system 100. The processor 100 receives the
analog signal x(t) to initiate processing on the analog signal, in
the process element 201. From there, the processor 100 detects
power spectral densities in the analog signal x(t), in the process
element 202. The processor 100 segments the analog signal x(t) into
at least two frequency bands, in the process element 203, and
samples those bands, in the process element 204. This generally
produces multiple streams of sampled data which the processor 101
quantizes with different bit resolutions. In other words, the
processor 101 may quantize each of the sampled frequency bands with
bit resolutions according to detected power spectral densities of
the frequency bands, in the process element 205, and in doing so,
the processor 101 assigns a particular number of bits for each
sample of a certain frequency band.
[0015] Since there are variable numbers of bits per sample on a
frequency band by frequency band basis, there may be different
rates among the frequency bands. Accordingly, the multiplexer 102
multiplexes the quantized streams from the processor 101 into a
single datastream for real-time processing by a receiving end, in
the process element 206.
[0016] To illustrate, assume the analog signal has a frequency
bandwidth of 20 MHz and that bandwidth is segmented into two 10 MHz
frequency bands. Now, assume a common sampling rate of 20 MHz. If
one of the two frequency bands is designated with two bits per
sample (e.g., because it has a lower power spectral density than
the other frequency band), then the bit rate from the processor 101
is about 40 Mbps for the quantized lower spectral density frequency
band to process in real time at a receiving end (i.e., assuming no
other bits for error correction and the like). And if the other
frequency band is designated with four bits per sample, then the
bit rate of the processor 101 for that quantized higher spectral
density frequency band is about 80 Mbps. Thus, the multiplexer 102
interleaves the two bit streams associated with the two frequency
bands into a single datastream of about 120 Mbps such that the
entire 20 MHz spectrum of the analog signal can be
analyzed/observed in real time at the receiving end.
[0017] As one can see from this example, if the entire 20 MHz
spectrum of the analog signal was sampled at 40 MHz and a common
bit resolution of four bits was applied to each sample, then the
total bit rate of the quantized signal would be about 160 Mbps to
process in real time. Accordingly, the embodiments herein provide a
selective quantization/bit resolution based on power spectral
density across the frequency spectrum of the analog signal. And,
this selective quantization process means that bit rates for
real-time processing at a receiving end can be reduced.
[0018] Alternatively, more bits can be selectively assigned to
certain frequency bands to enhance the signal-to-noise ratio in
those bands. For example, if an increase in the SNR on the higher
power spectral density frequency band in the above example was
desired, more bits could be assigned to the quantization of that
frequency band to increase the SNR over the quantization noise. To
illustrate, assuming the same scenario above except that the higher
power spectral density frequency band is assigned six bits per
sample instead of the previous four bits per sample. Then, the bit
rate of that frequency band would be about 120 Mbps. When
multiplexed with the quantized signal of the lower power spectral
density frequency band, the overall bit rate returns to about 160
Mbps. But, the SNR improves significantly in the higher power
spectral density band because dB increases proportionally to the
number of bits of resolution--by about 6.02 times per bit. The
embodiments below provide additional details regarding this
process.
[0019] FIG. 3 is a graph of an exemplary frequency domain 300 of an
analog signal 301 (e.g., an RF signal). The frequency spectrum of
interest of the signal 301, in this example, is from 10 MHz to 100
MHz with the frequency spectrum of the signal 301 being segmented
into 10 MHz chunks. FIG. 4 is a block diagram of another exemplary
A/D processing system 350. The frequency domain 300 will now be
discussed in the context of the A/D processing system 350
processing the analog signal 301.
[0020] In this embodiment, the A/D processing system 350 as a
plurality of bandpass filters 351-1-351-N. Each bandpass filter 351
is operable to receive the analog signal x(t) and filter out a 10
MHz portion. For example, the band pass filter 351-1 filters around
the 10 MHz frequency band x.sub.1, the bandpass filter 351-2
filters around the 10 MHz frequency band x.sub.2, the bandpass
filter 351-3 filters around the 10 MHz frequency band x.sub.3, etc.
Each of these 10 MHz frequency bands is then sampled at, in this
embodiment, a common sampling rate of .DELTA.f (e.g., 20 MHz
Nyquist), thereby producing a corresponding number of sampled
sub-signals x.sub.1, x.sub.2, . . . , x.sub.9.
[0021] A controller 352 may determine the power spectral densities
of the signals from the bandpass filters 351-1-351-N so as to
direct bit resolution assignments to the quantizer 353-1-353-N. For
example, any frequency band having an average power spectral
density less than 10 dB may receive two bits of resolution, a
frequency band having an average power spectral density between 10
and 20 dB may receive three bits of resolution, a frequency band
having an average power spectral density between 20 and 30 dB may
receive four bits of resolution, and any frequency band having an
average power spectral density between 30 and 40 dB may receive
five bits of resolution. Thus, for the purposes of this
illustration, the controller 352 directs the quantizer 353-1 to
assign three bits of resolution to the x.sub.1 frequency band
signal, five bits of resolution to the x.sub.2, x.sub.3, and
x.sub.4 frequency band signals, three bits of resolution to the
x.sub.5 frequency band signal, two bits of resolution to the
x.sub.6 and x.sub.7 frequency band signals, and three bits of
resolution to the x.sub.8 and x.sub.9 frequency band signals.
[0022] After quantization by the quantizer 353-1-353-N, the
frequency band signals, or "sub-streams", x.sub.1-x.sub.9 are now
represented by digital streams of varying bit rates. For example,
based on the assigned bit resolutions by the controller 352 to the
quantizers 353-1-353-N, the x.sub.1 frequency band signal is
represented by a data stream having a bit rate of 40 Mbps (i.e.,
due to .DELTA.f equaling 20 MHz with 2 bits of resolution, or 20
Mhz times 2). The x.sub.2, x.sub.3, and x.sub.4 frequency band
signals are similarly represented by datastreams having bit rates
of 100 Mbps, with the x.sub.5 frequency band signal having a bit
rate of 60 Mbps, the x.sub.6 and x.sub.7 frequency band signals
having bit rates of 40 Mbps, the x.sub.8 and x.sub.9 frequency band
signals having bit rates of 60 Mbps. The multiplexer 354 then
combines the sub-streams of each of the frequency bands
x.sub.1-x.sub.9 into a single output datastream having an overall
bit rate of 620 Mbps (i.e., 60 Mbps+100 Mbps+100 Mbps+100 Mbps+60
Mbps+40 Mbps+40 Mbps+60 Mbps+60 Mbps), neglecting effects of error
correction, quantization errors, and the like.
[0023] A more detailed and mathematical discussion of the above is
now presented. Nyquist's theorem for alias-free signal sampling
states that, for a baseband signal with a maximum frequency
f.sub.m, the sampling rate is at least 2f.sub.m. And, for a
bandpass signal with a bandwidth of B, the sampling rate is an
integer multiple of B.
[0024] After a signal is sampled at a rate off samples per second,
the sampled values are digitized/quantized to discrete levels based
on the number of bits per sample b. Thus, the overall bit rate of a
sampled signal r.sub.b is given by r.sub.b=f.sub.sb bits per
second. The reconstructed signal integrity, assuming alias-free
sampling, is directly related to the quantization error during
digitization, which is a function of the number of bits per sample
b. The RMS (Root Mean Square) quantization error QE is a function
of the least significant bit (LSB) in the bit resolution and is
generally given to be
Q E = ( 1 12 ) L S B volts . ##EQU00001##
[0025] The signal x(t) being sampled can be decomposed into
multiple sub-signals which when added together form the overall
x(t) such as:
x(t)=x.sub.0(t)+x.sub.1(t)+ . . . +x.sub.L-1(t).
Once sampled, that signal becomes:
x(n)=x.sub.0(n)+x.sub.1(n)+ . . . +x.sub.L-1(n), which is the sum
of the sampling of the sub-signals.
[0026] The contributions of the sub-signals to x(t) can be based on
various metrics, one of which being the relative sub-signal power.
In other words, the sub-signal with the highest relative power is
the signal that most contributes to x(t). With this in mind, an
alternate way of looking at the signal integrity is that the
sub-signals contributing the most to x(t) should be provided a
higher degree of fidelity and protection from errors/quantization
noise because errors in quantization noise in those sub signals
reflect more on the overall signal x(t) than an error in other sub
signals with lower power contributions to the signal x(t).
[0027] Accordingly, the bits per sample given to each sampled sub
signal can be based on the relative contribution of a sub-signal to
the overall signal x(t). And, based on how the bits per sample are
assigned to the sub-signals, two advantages arise: 1. The overall
number of bits per sample of x(t) can be reduced which reduces the
overall bit rate for real-time processing; and 2. The quantization
error observed in the reconstructed signal can be reduced in
certain desired bands.
[0028] So, assuming each sub-signal is digitized at b bits per
sample, the output bit rate (i.e., bits per second) is:
r.sub.b
sub-signal=.DELTA.f.sub.samples/secondb.sub.bits/samplek.sub.num-
ber of sub-signals.
Thus, the sum of the sub-signal bit rates is the same as sampling
the overall signal x(t) at f.sub.s and assigning a general bit
resolution to each sample if the sub signals were indeed assigned
the same bit resolution. But, if one assumes that each sub-signal
of x(t) contributes a total power of P.sub.1 Watts to the original
signal x(t), then the relative contribution of each sub signal to
the original signal x(t) can be used to either reduce the total bit
rate or improve the reconstructed signal quality by reducing the
impact of quantization noise.
[0029] To achieve a bit rate reduction, each sub signal is assigned
a number of bits per sample b.sub.i based on its relative
contribution to the overall signal x(t) such that:
.SIGMA..sub.i=1.sup.kb.sub.i<kb.
[0030] Thus, one possible approach to assigning b.sub.i is to find
an x.sub.i with the maximum P.sub.i (x.sub.max and P.sub.max) and
assign it b bits per sample. Then, the remaining x.sub.i
sub-signals can be assigned b.sub.i according to:
b i = [ P i P max b ] . ##EQU00002##
The resulting values of b.sub.i should then satisfy
.SIGMA..sub.i=1.sup.kb.sub.i<kb.
[0031] Alternatively, a threshold value S can be defined for
P.sub.i that defines major contributing frequency bands to the
overall signal x(t). Thus, b.sub.i can be algorithmically assigned
according to:
if P i .gtoreq. S .fwdarw. b i = b , and ##EQU00003## if P i < S
.fwdarw. b i = [ P i P max b ] . ##EQU00003.2##
The resulting values of b.sub.i should also then satisfy
.SIGMA..sub.i=1.sup.kb.sub.i<kb.
[0032] To improve signal quality (e.g., by reducing the impact of
quantization noise), each sub-signal is assigned its per sample
b.sub.i based on the relative contribution to the overall signal
x(t) such that .SIGMA..sub.i=1.sup.kb.sub.i<kb. In doing so, one
might sort the P.sub.i values of the frequency bands in descending
order and then find the median P.sub.i (i.e., P.sub.i=P.sub.i,
median). From there, one could assign b.sub.i=b for x.sub.i that is
associated with P.sub.i, median. Then, one could choose a step size
for P.sub.i in terms of order and for every step greater than
P.sub.i, and the associated b.sub.i is increased by j bits. And,
for every step less than P.sub.i, the associated b.sub.i is
decreased by j bits.
[0033] An example of such is now presented using 11 sub-signals and
a b value of 10 bits.
TABLE-US-00001 P.sub.i P.sub.4 P.sub.2 P.sub.10 P.sub.7 P.sub.9
P.sub.1 P.sub.3 P.sub.5 P.sub.8 P.sub.11 P.sub.6 b.sub.i 13 12 12
11 11 10 9 9 8 8 7 median
Now, assume that P.sub.4/P equals 32%, and P.sub.6/P equals 2%, and
assuming a 10 volt P-P A/D design. Traditionally, the quantization
error observed by the frequency bands x.sub.4 and x.sub.6 would be
approximately 3 mV and the relative contributions would be 1.05 mV.
Now, however, the quantization error observed by x.sub.4 and
x.sub.6 would be about 0.35 mV and 22 mV, respectively, and the
relative contribution would be about 0.55 mV.
[0034] As illustrated in FIG. 4, one possible manner in which the
concepts herein may be implemented is the through the use of analog
band pass filters 351. Alternatively, however, the process may be
implemented in the digital domain as follows:
X ( k ) = n = 0 N - 1 x ( n ) - 2.pi. nk N ; ##EQU00004## and x ( n
) = n = 0 N - 1 X ( k ) 2.pi. nk N , ##EQU00004.2##
where
x I ( n ) = k = k i , start k i , end X ( k ) 2.pi. nk N ,
##EQU00005##
Thus, to perform signal decomposition of the digital domain as a
function of spectrum occupancy (i.e., the digital version of FIG.
4)
k = N f f s . ##EQU00006##
where the values of k.sub.i,start and k.sub.i,end are determined
from
k = N f f s . ##EQU00007##
The samples of x.sub.i(n) are then quantized with b.sub.i bids
according to the targeted criteria as described above.
[0035] FIG. 5 is a block diagram of another exemplary A/D
processing system 350. In this embodiment, the controller 352 is
also operable to control the sampling rates of the frequency bands
(i.e., or sub-signals). This process may be implemented in addition
to the selectively assigned bit resolutions discussed above. For
example, assuming that a particular frequency band was contributing
more to the overall signal x(t), then the controller 352, in
addition to assigning a higher bit resolution to that frequency
band, it would also direct the sampler to oversample that frequency
band to improve resolution.
[0036] FIG. 6 is a block diagram of an exemplary communication
system 400 employing the A/D concepts described herein. For
example, the A/D concepts disclosed herein may be implemented in a
cable television communication system that employs RF signaling
techniques across a substantial amount of RF spectrum. An upstream
link of the cable television communication system, in this
embodiment, provides high speed data services being delivered over
devices conforming to the Data Over Cable Service Interface
Specification (DOCSIS) specification. The communication system 400
includes a headend 401 configured with an upstream hub 420. The hub
420 is coupled to a downstream node 421 via optical communication
links 405 and 406.
[0037] The hub 420 includes a Cable Modem Termination System (CMTS)
402, an electrical to optical converter 403, and an optical to
electrical converter 404. The node 421 is similarly configured with
an optical to electrical converter 408 and an electrical to optical
converter 407. The A/D conversion concepts herein would generally
be configured with the electrical to optical converters 403 and
407. Thus, digital to analog (D/A) operations would be performed by
the optical to electrical converters 404 and 408 where the
sampled/quantized/multiplexed sub-signals of previously analog x(t)
signals are received.
[0038] To ensure that the sub-signals are properly converted, the
D/A operations may be configured with demultiplexers that extract
the sub signals from the overall datastream. To do so, the
demultiplexer may require information pertaining to the exact
structure of the datastream. For example, with one sub-signal
having a bit rate of 100 Mbps and another sub-signal having a bit
rate of 200 Mbps, then the overall bit rate of the datastream is
300 Mbps with one bit of the 100 Mbps sub-signal being interleaved
with every two bits of the 200 Mbps sub-signal. As there may be
several more frequency bands and a variety of bit
resolutions/sub-signal bit rates, the multiplexed datastream can
rapidly increase in complexity. Accordingly, a signaling technique
using extra bits within the data stream may be used to flag the
demultiplexer in the optical to electrical converters 404 and 408
and indicate which bit belongs to which sub-signal. However, such
an implementation is a matter of design choice. It should also be
noted that such design choices may include the use of extra bits in
the datastream for error correction and other data transmission
features (e.g., specific communication protocols, etc.).
[0039] With respect to the remaining features of FIG. 6, the
headend 401 is generally the source for various television signals.
Antennas may receive television signals that are converted as
necessary and transmitted over fiber optic cables 405 to the hub
420. Several hubs may be connected to a single headend 401 and the
hub 420 may be connected to several nodes 421 by fiber optic cable
links 405 and 406. The CMTS 402 may be configured in the headend
401 or in the hub 420. The fiber optic links 405 and 406 are
typically driven by laser diodes, such as Fabry Perot and
distributed feedback laser diodes.
[0040] Downstream, in homes/businesses are devices called the Cable
Modems (CM; not shown). A CM acts as a host for an Internet
Protocol (IP) device such as personal computer. Transmissions from
the CMTS 402 to the CM are carried over the downstream portion of
the cable television communication system generally from 54 to 860
MHz. Downstream digital transmissions are continuous and are
typically monitored by many CMs. Upstream transmissions from the
CMs to the CMTS 402 are typically carried in the 5-42 MHz frequency
band, the upstream bandwidth being shared by the CMs that are
on-line. However, with greater demands for data, additional
frequency bands and bandwidths are continuously being considered
and tested, including those frequency bands used in the downstream
paths.
[0041] The CMTS 402 connects the local CM network to the Internet
backbone. The CMTS 402 connects to the downstream path through the
electrical to optical converter 404 that is connected to the fiber
optic cable 406, which in turn, is connected to the optical to
electrical converter 408 at the node 421. The signal is transmitted
to a diplexer 409 that combines the upstream and downstream signals
onto a single cable. The diplexer 409 allows the different
frequency bands to be combined onto the same cable. The downstream
channel width in the United States is generally 6 megahertz with
the downstream signals being transmitted in the 54 to 860 MHz band.
Upstream signals are presently transmitted between 5 and 42 MHz,
but again other larger bands are being considered to provide
increased capacity. So, the variably assigned bit resolution
concepts herein may be particularly advantageous. However, the
invention is not intended to be limited to any particular form of
communication system.
[0042] After the downstream signal leaves the node 421, the signal
is typically carried by a coaxial cable 430. At various stages, a
power inserter 410 may be used to power the coaxial line equipment,
such as amplifiers or other equipment. The signal may be split with
a splitter 411 to branch the signal. Further, at various locations,
bi-directional amplifiers 412 may boost and even split the signal.
Taps 413 along branches provide connections to subscriber's homes
414 and businesses.
[0043] Upstream transmissions from subscribers to the hub
420/headend 401 occur by passing through the same coaxial cable 430
as the downstream signals, in the opposite direction on a different
frequency band. The upstream signals are sent typically utilizing
Quadrature Amplitude Modulation (QAM) with forward error
correction. The upstream signals can employ any level of QAM,
including 8 QAM, 32 QAM, 64 QAM, 128 QAM, and 256 QAM. Modulation
techniques such as Synchronous Code Division Multiple Access
(S-CDMA) and Orthogonal Frequency Division Multiple Access (OFDMA)
can also be used. Of course, any type of modulation technique can
be used, as desired.
[0044] Transmissions, in this embodiment, are typically sent in a
frequency/time division multiplexing access (FDMA/TDMA) scheme, as
specified in the DOCSIS standards. The diplexer 409 splits the
lower frequency signals from the higher frequency signals so that
the lower frequency, upstream signals can be applied to the
electrical to optical converter 407 in the upstream path. The
electrical to optical converter 407 converts the upstream
electrical signals to light waves which are sent through fiber
optic cable 405 and received by optical to electrical converter 403
in the node 420.
[0045] Those skilled in the art should readily recognize that the
invention is not intended to be limited to the examples disclosed
herein. For example, the invention should not be limited to any
particular number of frequency bands segmented, any number of bits
of resolution during quantization, and/or any frequency bandwidth
of an analog signal. Nor should the invention be limited to any
particular form of analog signal. That is, the inventive concepts
disclosed herein may be used in a variety of communication systems
regardless of bandwidth considerations.
[0046] Additionally, the invention can take the form of an entirely
hardware embodiment, an entirely software embodiment or an
embodiment containing both hardware and software elements. In one
embodiment, the invention is implemented in software, which
includes but is not limited to firmware, resident software,
microcode, etc. FIG. 7 illustrates a computing system 500 in which
a computer readable medium 506 may provide instructions for
performing any of the methods disclosed herein.
[0047] Furthermore, the invention can take the form of a computer
program product accessible from the computer readable medium 506
providing program code for use by or in connection with a computer
or any instruction execution system. For the purposes of this
description, the computer readable medium 506 can be any apparatus
that can tangibly store the program for use by or in connection
with the instruction execution system, apparatus, or device,
including the computer system 500.
[0048] The medium 506 can be any tangible electronic, magnetic,
optical, electromagnetic, infrared, or semiconductor system (or
apparatus or device). Examples of a computer readable medium 506
include a semiconductor or solid state memory, magnetic tape, a
removable computer diskette, a random access memory (RAM), a
read-only memory (ROM), a rigid magnetic disk and an optical disk.
Some examples of optical disks include compact disk-read only
memory (CD-ROM), compact disk-read/write (CD-R/W) and DVD.
[0049] The computing system 500, suitable for storing and/or
executing program code, can include one or more processors 502
coupled directly or indirectly to memory 508 through a system bus
510. The memory 508 can include local memory employed during actual
execution of the program code, bulk storage, and cache memories
which provide temporary storage of at least some program code in
order to reduce the number of times code is retrieved from bulk
storage during execution. Input/output or I/O devices 504
(including but not limited to keyboards, displays, pointing
devices, etc.) can be coupled to the system either directly or
through intervening I/O controllers. Network adapters may also be
coupled to the system to enable the computing system 500 to become
coupled to other data processing systems, such as through host
systems interfaces 512, or remote printers or storage devices
through intervening private or public networks. Modems, cable modem
and Ethernet cards are just a few of the currently available types
of network adapters.
* * * * *