Note: Descriptions are shown in the official language in which they were submitted.
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96/11712
SUBSEQUI~NT FRAME VARL~BLE DATA RATE INDICATION
METMOD FOR VARIOUS VARL~BLE DATA RATE SYSTEMS
BACKGROUND OF THE INVE~NTION
The present invention relates generally to the field of data cc~....2...i~icqti~ n.c
and more specifirally, to the field of synchronous, fixed boundary, variable data
rate co~u~ication systems, such as code division multiple access (CDMA) North
~m~ni~.qn digital cellular telephone and personal co~~ ication systems.
Synchronous cn.. ~ catit)n systems which utilize fixed frame boundary
data frames including data at variable rates are known in the art. One example of
such a system is the CDMA North American digital cellular system, a well-known
class of mod.. lsti~2n using sper;sli7ed codes to provide multiple co.. i~i~ stion
~hsnnele in a deci~Dted seg... ~1 ofthe electromagnetic spectrum. Thus, the
30 definition of "synchronous" is understood to include all systems in which an
CA 02260974 1999-01-14
W O 98/02986 PCT~US96J11712
attempt is made in at least one tr~Ami~Qinn direction to synchro~ e system timing
(frame and bit timing are recoverable) between trPncmhtin~ and reGeiving .station~
The Teleco...~ c~,l;on~ Industry Association (TIA) has standardized a CDMA
~lf .. l ~tinn in the "Mobile Station-Base Station Compatibility Standard for
5 Dual-Mode Wi~lebqn~ Spread Spectrum Cellular System TIA/EIA/IS-95 Interim
Standard" (IS-95) and the "Speech Service Option Standard for Wi~le~ql~ Spread
Spectrum Digital Cellular System TLAIEL~IIS-96 Interim Standard" (IS-96).
Sections 6 - 6.2.4 and 7 - 7.2.4 of IS-95 and the entire IS-96 are particularly
relevant. In addition, updated versions ofthese standards, known as IS-95A and
10 IS-96A, are also available. Of particular note in these updated versions are
p~ es associated with a second rate set accommodating a higher speed
vocoder.
Another example of a variable data rate co.~...-...~icqtion system is the
CDMA personal co~ qtinn system described in the industry standard TIA
proposal no. 3384, published as J-STD-008, entitled "Personal Station Base
Station Corn~atibility Re4~ e~ nts for 1.8 to 2.0 GHz Code Division Multiple
Access (CDMA) Personal Co~ication Systems". While other sections are
also relevant to the present invention, sections 2.1.3.3 - 2.2.3 of J- STD-008 are
particularly relevant. As would be understood by one reasonqbly skilled in the art
20 ofthe present invention, the personal co-l..ll,lllication system (PCS) mobile and
base statinn~ of J-STD-008 are very similar to the mobile and base stations,
respectively, of IS-95A except for the operating frequencies, thus, unless
otherwise noted, the term "mobile stations" should be understood to refer to
cellular mobile stations and personal cc~ icatinn st~tion~
In the conv~tic~ CDMA digital cellular and personal co.~.. ----irati~ n
systems, variable data rates are utilized to reduce the data tr~ mir~ion rate during
times of reduced speech activity. This data rate reduction results both in a
reductinn of interference with other users (thereby increasing capacity in the
system) as well as in a reduction in average transmit power ofthe CDMA mobile
30 station (thereby increasing battery life). On the tr~n~mitter end (transmitting base
CA 02260974 l999-0l-l4
w 0 98/02986 PCTrUS96/11712
station or tr~ne~ ng mobile station), a vocoder (voice or speech
encoder/decoder) compares voice energy levels to adaptive thresholds based on
background noise levels to determine an appropliate data rate for each frame of
~ speech data, thereby ~ul)pressing background noise and providing good voice
5 ~ ceinn in noisy en~iro....-- .l e Using a code excited hnear pre~liction (CELP)
method, the vocoder receives pulse code modulated speech samples and reduces
the number of bits le~uir~d to l~present speech by exploiting the intrinsic
pro~,.lies of speech signals to remove redundancy. Subsequently, the speech
encoded data is convohltinn~lly encoded for forward error correction before being
10 interleaved and mo~ lqted for tr~n.cmiceion.
Since the data rate may change at each frame boundary, the CDMA
receiver must first det~rmine the data rate of each frame of data. The process by
which this is accomphshed in the conventional CDMA digital cellular and personalcn.. ~ ation systems is a source of wasted time and processing energy.
15 According to the convPntion~l systems, each data frame must be separately
processed at each ofthe various possible data rates (including convolutional
decoding) before a deçi~;~ n is made regardi~lg which data rate was utilized on the
tr~nemittP~r end. Since this method is clearlv inefficient, there is a need in the
industry for a new method for determini Ig the data rate of each frame of data in
20 the CDMA digital cellular and personal co~u~ication systems, as well as other systems using fLxed boundary frames with variable data rates.
One possible method of addressing this problem is the addition of a
conventional header before each frame of data. Such a header could include the
data rate of the corresponding frame to which it is attached. Unfo~ ately~ such a
25 header would also need error protection to reduce the likelihood oftr~nemieei~n
errors. In view ofthe relatively small size of each frame of data, the ad~lhion~l bits
required for an error protected header would certainly add substantial overhead
and undesirable complexity to the syste~
There is, therefore, a need in the industry for a system which addresses
30 these and other related, and unrelated, problems.
CA 02260974 1999-01-14
W O 98/02986 PCT~US96111712
SUMMARY OF THE INVENTION
Briefiy described, the present invention in~ des a subseq~nt (or "next")
frame variable data rate in(1i~.ation method whereby a tranQ-m:tt~ inserts into the
5 frame structure of a current frame an in~lic,ation ofthe data rate ofthe next frame.
According to a first prer~lled embodiment ofthe present invention, after the first
frame is received and conventionally processed at a receiver, the data rates of
subsequent frames are known before processing, thereby red~rir g proceseing load.
Furthermore, because the rate indication is inserted into the current frame to be
10 error protected along with the rest ofthe frame infc.~ ~~ion~ reliability is high,
while ~a(~ tion~l data overhead and complexity are very low.
According to the first l)lerelled embodiment, as applied to one
implemPntqsinn of a CDMA system (cellular or PCS), as a tr ~QmittinE station
mn~l~m (SM) ~located in either the mobile station or the base station) assembles a
15 current traffic channel frame for convolutional encoding, the trqncmitting SMinserts an in~licqti~m of the data rate of the subsequent channel frame of data. In
many cases, (e.g. primary traffic frames) a vocoder speech encodes PCM data for
the SM and notifies the transmitting SM lhrùu~ a central procesQing unit (CPU)
ofthe applup -~e data Tate for the subsequent frame, and in other cases, the CPU20 issues commqn~c to the SM and the vocoder to influence the selection ofthe data
rate.
The number of bits necessary to provide a one-to-one representation of the
various rates equals the .cmqllect integer greater or equal to the Log2 Of the total
number of possible rates, e.g., since the current CDMA system utilizes four
25 possible data rates, two bits are adequate to provide a one-to-one infli-a~ion of
each ofthe possible data rates, whereas three bit would be required tû similarlyepres~,nl five to eight possible rates, etc. ln the convPntinn~l CDMA IS-95 frame
structure (also similar to rate set 1 of IS-95A and J-STD-008), the two it~ catinn
bits are, for example, easily ~u~liLuled for two frame quahty in~iclqtinn bits for the
30 top two rates and for two i.lrul~lion bits ffir the lower two rates. Since the rate
CA 02260974 1999-01-14
W O 98/02986 PCTAUS96/11712
indication is eml)edlled in the structure ofthe frame itself~ the rate infliraLion bits
rece*e the same error protection (error correction and error detection) as the
other informqfion in the data frame. Consequently, the current inventive method
exhibits high reliability without the need for great cornplexity or expense.
On the receiving end, rather than needing to process each frame of data
multiple times at each ofthe possible data rates, including convolntinnql decoding,
to ~etPrmine the a~prop-iale data rate for each frame of data, the receiving SM
discovers the data rate of each frame of data subsequent to the first frame of data
by analyzing the ilLformqtion contained in the immP~ --ely preceding frame of data.
In other words, after the very first frame is, in the conventional manner, processed
at each ofthe various rates to detennine the appropliate data rate for the firstframe, the receiving SM is able to determine the data rate ofthe second frame ofdata before needing to process the second frame of data. This process continues
so that the data rates of each of the subsequent data frames are determined in the
frames prece~ling each ofthe subsequent frames.
1iti~ 11y~ according to the first p~ert,led embodiment ofthe present
invention, in an effort to prevent tr~nemiC~ion errors from propagating through the
series of data frames, the rate selection process is ct)ntim.~lly examined, such as
through monitoring frame quality indicators, symbol error rates, and/or other
methods of de~ g rate selection integrity, such as using viterbi decoding
internal information to deterrnine rate selection accuracy. If FQI (frame quality
in.iication) checking fails, the symbol error rate is too high, or if other rateselection integrity methods indicate improper rate selection, the receiver method
further includes conventionally processing the frame at each ofthe re...~ g
25 possible rates to ensure accurate data rate det~nnin~tion for that particular frame,
after which rate det~rminatinne proceed according to the new method. If the rate- still cannot be determined after being processed at the various possible data rates,
the frame is cl~.c.eified as an erasure frarne, and the process contiuues by processing
the next frarne as the first frarne was processed. As should be evident, the
30 processing load on the receiving SM is greatly reduced by not needing to process
CA 02260974 l999-0l-l4
W O 98/02986 PCT~US96/11712
each frame at each ofthe various rates. Thus, for the mobile and base SMs, clearly
understood benefits can be realized in reduced power con~.u~lion and reduced
processing load.
The present invention also inr.lv~les a second pfer~led embo&ent whiGh
is very similar to the first pler~,led embodim~nt However, rather than requiringthe receiving station to convPntir~nally process the first frame to determine the
proper data rate, this second p~er~lled embodiment inc.l~l.les tranemi~ting the first
frame with data at a known data rate so that, at the receiving station, no frame is
convoh.finnally processed at various rates unless an error in tran~mie.~inn occurs,
at which point recovery processing proceeds conventionally as in the first pyerelled
emhodiment. ~ a~l-1itinn7 a second frame structure is utilized in conjunction with a
higher speed vocoder. The second frame structure is an adaptation ofthe "Rate
Set 2" frame structure disclosed in IS-95A and J-STD-008. The two next frame
indication bits are substituted for two ;~f(i. ~ I ion bits in the full, quarter, and
eighth rate frames and for two frame quality indication bits for the half rate frames.
Again, since the next frame rate indication is embedded in the structure of the
frame itself, the rate in~ tion bits receive the same error protection (error
correction and error detection) as the other inrol~lion in the data frame. For
mixed mode frames including only ci~lin~ and/or secondary traffic inrol~on
(i.e., no speech), an erasure bit in each frame is utilized by the mobile station to
request the base station to re-transmit an erroneous frame at a known rate so that,
with such "no speech" frames, no frames ever need to be convol--ti-n~lly
processed at various rates to ~etPrmi~e tranemi~;on rates ofthose types of mixedmode frames.
According to a third l)r~r~ d embodiment, the first frame of speech data
is immP.rli~tely preceded by a preamble frame encoded at a known rate. The
preamble frame does, however, include a data rate indication for the next frame,corresponding to the first frame of speech data. In this way, it is not necessary
that the first frame of speech encoded data be transmitted at a fixed rate. A fourth
~)r~r~lled embodiment ofthe present invention is also very similar to the second
CA 02260974 1999-01-14
- WO 98/02986 PCT/US96/11712
preferred embodiment. The primary di~ ces relate to the method of
transferring; r~ ion between the vocoder and the transmitting SM. Rather
than separately outputting the encoded data and then the subsequent frame rate
in(lir ~ion the i. ro~n~l;ol- is combined and relayed together to the transmitting
5 SM. In another (f~h) p~erelled embodiment ofthe present invention, the vocoderhas a process delay greater than the sampled time in a frame of data, thus the
vocoder speech encodes multiple frames of data ~imlllt~neously through a type ofparallel vocoder processing. Because ofthis time overlap, the vocoder is able todetermine a data rate of a subsequent frame of data before speech encoding îs
10 complete on the current frame of data. This subsequent rate in.lirqtil-n is output to
the tranemitting SM before the current frame of data is output to the tr~nem;tting
SM.
Other prere,led embo-limPnt.c include inserting subsequent frame rate
indications in other locations within the data frame or inserting inc,e~
15 subsequent frame rate in-lications which in~ic~te changes in rates (i.e., upward,
downward, no change, max~mum, minimum, etc. ) rather than providing one-to-one
inr~ tion.e ofthe rates. Yet other prerelled embo~im~nts include inserting
subse~uent frame rate indications only in selective frames, such as inserting
in~ir~qtil)nS only when a change in rates is about to occur or only for certain types
20 of data frames, such as when assumptions can be made about other types of frames
or when it is better to simply allow the other types of frames to be processed
convPntinnally. Still other ~lt~ te embo-limPnt.c including recei~qng and buffering
variable rate data frames from other sources besides the vocoder, such as external
variable data rate devices. ln still other prelèlled embo~limPnts ofthe present
25 invention, rate choice evaluations are employed only occaein ~lly under the
assumption that tr~nemie.ei--n errors are very rare. Additionally, rate choice
~ evaluations are omitted in other embodiments where periodic fLxed rate
tran.emie.einns are imposed to ~.~tomqticqlly reduce the potential for propagation of
rate determination errors. In other words? the tr~n.cm:~ting stations of such
30 embo~iimpnte periodically transmit frames at known rates according to periods
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96/11712
known by the receiving station so that any rate Aeterrnin-q-tion errors are
auln~ .qlly f~xed.
It is the.~,f~le an object ofthe present invention to provide a subsequent
frame variable data rate jnAi~lqtion method.
Another object of the present invention is to provide a radio telephone
operative to co.~ rqte subsequent frame variable rate information.
Yet another object ofthe present invention is to provide an appal~lus and a
method for in~lirpqting a data rate of a subsequent frame of data in a synchronous,
fL~ced frame boundary system including frames of data at variable data rates.
Still another object ofthe present invention is to provide an apparatus and
a method ffir inserting a subseqllent frame data rate indi~.qtiQn into a current frame
of data and subsequently error protecting the frame of data.
Another object ofthe present invention is to provide an apparatus and a
method for inserting a subsequent frame data rate inllicqltinn in a beginninE portion
of a current frame of data and subsequently error protecting the frame of data.
Yet another object ofthe present invention is to provide a memory ~lefining
a tr~ncmi~sion frame memory structure ir~ ling current frame speech ;..rol~lion
and a subsequent frame rate indication.
Yet another object of the present invention is to provide a memory defining
20 a convoh~tion~lly encoded tr:m.~mi~ion frame memoly structure including current
frame speech inro~ ';0l-, a subsequent frame rate indication, and a frame quality
indicator based upon the current frame speech i~fo~ a~ion and the subsequent
frame rate indication.
Still another object of the present invention is to provide an apparatus and
25 a method for in~ ting and determining subsequent frame data rates in a CDMA
digital cellular system.
Another object ofthe present invention is to provide an apparatus and a
method for in~ ting and determinin~ subsequent frame data rates in a CDMA
personal co,....-~ ir~tinn syste~
CA 02260974 1999-01-14
W O ~81'~9~~ PCT~US96/11712
Yet another object ofthe present invention is to provide an apparatus and a
method for generating a current frame of data, including determining a desired data
rate of a subsequent frame of data and inserting an indication of the subsequentframe data rate into the current frame of data.
Still another object ofthe present invention is to provide an apparatus and
a method for receiving a current frame of data and analyzing the current frame of
data to determine a data rate of a subsequent frame of data.
Other objects, features and advantages of the present invention will become
apparent upon reading and under~ g the present specification, when taken in
conjunction with the accompanying drawings.
BR~EF DESCRIPTION OF T~E DRAWINGS
FIG. I is a block diagram representation of circuital ~ " ~ls of a speech
path in a CDMA digital cellular telephone in accordance with a first l)lerel,ed
embodiment of the present invention.
FIG. 2 is a block diagram represP.ntatinn of circuital clc....~ c of a speech
path in a CDMA base station in accordance with the first plertlled embodiment ofthe present invention.
FIG. 3 is a block diagram representation of selected frame generation
functions provided by a vocoder~ a CPU, and an SM in accordance with the first
prere.l~d embodiment ofthe present invention.
FIG. 4 is a frame structure diagram for the CDMA traffic channel frames at
various rates before being convoh.ti( n~lly encoded in accordance with the firstInt;r~lled embodiment ofthe present invention.
FIG. 5 is a flow chart l~res~ ation of selectec~ frame generation steps
taken by a tr~ncm;~tin~ station vocoder in accordance with the first plert;lled
~ embodiment of the present invention.
FIG. 6 is a ilow chart r~les~r~ m of selecte~l frame generation steps
taken by the trancm;tting station SM and CPU in accordance with the first
prer~ ed embodiment of the present invention.
CA 02260974 1999-01-14
W O g8/02986 PCT~US96/11712
FIG. 7 is a flow chart r~le~A~ in-~ of sP1~cte(1 frame analysis steps taken
by the receiving station vocoder, SM and CPU in accordance with the first
prefe.led embodiment ofthe present invention.
FIG. 8 is a block diagram ~ se~ 1 ;nn of selected frame generation
5 fimctinns provided by a vocoder, a CPU, and an SM in accordance with a second
p~rwled embodirnent ofthe present invention.
FIG. 9 is a vocoder timing diagram in accordance with the second
prt;r~llt;d embo~limPnt ofthe present invention.
FIG. 10 is a frame structure diagram for the CDMA traffic channel frames
10 at various rates before being convolutionally encoded in accordance with the
second prerelled embodiment ofthe present invention.
FIG. 11 is a flow chart representation of seiected frame generation steps
taken by a tr~n~mitting station vocoder in accordance with the second pr~r~lled
embodiment of the present invention.
FIG. 12 is a flow chart repres~nt~tion of selected frame generation steps
taken by the transmitting station SM and CPU in accordance with the second
plere..ed embo&ent ofthe present invention.
FIG. 13 is a flow chart representation of selected frame analysis steps taken
by the receiving station vocoder, SM and CPU in accordance with the second
20 pler~.led embodiment ofthe present invention.
FIG. 14 is a fiow chart repres~nt~tinn of sPlected frame generation steps
taken by a tr~n~mitting station SM and CPU in accordance with a third
embodiment ofthe present invention.
FIG. 15 is a fiow chart repre~Pntation of selected frame analysis steps taken
25 by the receiving station vocoder, SM and CPU in accordance with the third
prc;l~..ed embodiment ofthe present invention.
FIG. 16 is a fiow chart representation of selected frame generation steps
taken by a tr~n~mitting station vocoder in accordance with a fourth pre~e.led
embodiment ofthe present invention.
CA 02260974 1999-01-14
W O 9X~0~6 PCTrUS96/11712
FlG. l 7 is a flow chart repre~entQtion of se1ected frame generation steps
taken by a vocoder in accordance with a fifth pref~.,.ed embodiment ofthe present
nvention.
FIG. l 8 is a frame structure La~ for the CDMA trafflc channel frames
at various rates before being conVQ1llt~ lly encoded in accordance with a sixth
rt;lled embodiment ofthe present invention
DETAILED DESCRIrrION OF THE PREFF,RR~ EMBODIMENTS
Referring now in greater detail to the drawings, FIGS. 1 and 2 are very
similar to each other since a CDMA mobile station l 0, circuital portions of which
are represented in FIG. l, and a CDMA base station 30, circuital portions of which
are represented in FIG. 2, both transmit and receive CDMA signals, jnch~/lin~
traffic channel frames of data. The term "mobile station" is understood to refer to
any type of cellular telephone. including units in~stall~d in vehicles and hand-held
units, including conventional cellular hand-held devices and PCS personal sta,tinnc
Both the CDMA mobile station l0 and the CDMA base station 30 include,
respect*ely, according to the pre~.led embodim~nts ofthe present invention, an
antPnna 12, 32, a radio frequency (RF) section 14, 34, a CDMA baseb~n~l
applicatio~ specific integrated circuit (BB ASIC) 16, 36, a station modem (SM)
18, 38, a central processing unit (CPU) 20~ 40, and a vocoder (voice or speech
encoder/decoder) 22, 42. The CDMA mobile station l0 further includes,
connected to the vocoder 22, an analog-to-digitaVdigital-to-analog (A-to-D/D-to-A) converter 24 connected to a llficlophone 26 and a speaker 28 for interaction
with a mobile station user. ~he CDMA base station 30 further includes a public
2~ switched t~1eph--ne network (PSTN) interface 44 for interaction with the PSTN, as
well as other conventio ~1 intçrf~ces In other words, the PSTN interface 44 is
~ understood to include a digital switch connected to both an interface to the PSTN
and to other CDMA base stat~ Thus, the vocoder 42 is understood to by a
pass-through to the PSTN interface 44 (i.e., avoid speech encoding and decoding)for signals to and from other CDMA base stations. 1~ additi.m, as would be
CA 02260974 1999-01-14
W O 981'~29.~ PCT~US96111712
m~lP.r~tood by one rea~ lna~ly skilled in the art, the al~t~nna 32 and CPU 40 ofthe
CDMA base station 30 actually r~.es~ ls multiple elçmPntc~ i.e., multiple
nnaC and multiple controllers are l~r~s ted by blocks 32 and 40.
According to the pl~,f~ d embo~ of the present invention, except
for the SMs 18, 38, CPUs 20, 40, and vocoders 22, 42 of FIGS. 1 and 2, the
remaining ~ s of the CDMA mobile station 10 and the CDMA base station
30 find acceptable examples in conventional el~mpnt~ and circuital combinatinnc
functi~ming as would be understood by those reqson~ly skilled in the art.
Furthermore, the new el~mP~c (the SMs 18, 38, CPUs 20, 40, and vocoders 22,
42) also maintain a large degree of ~ ,;lalily to conventional elements, differing
only to accommodate the teaGhing.c in this specification, as would be understoodby those re~acrn~ly skilled in the art after review ofthis spe~ifirr~ n In
accordance with the prer~lled embol1im~nts ofthe present invention, the BB
ASICs 16, 36 include customary means for providing baseband frequency analog
processing and conversion of signals to and from the digital domain for interfacing
with the SMs 18, 38. In particular, functions of the BB ASICs 16, 36 include
b~eb~n-l signal quadrature sphtting and com~ing, baseband analog-to-digital and
digital-to-analog conversion, b~seba ld direct current (DC) offset control, local
oscillator quadrature generation. Further in accordance with the prefelled
emboAim~nts ofthe present invention, the SMs 18, 38 conventionally provide the
rnajority of physiGal layer ~ ng through a demo~ tin~ unit, a decoding unit,
and an interleavingldeinterleaving unit. Among other functional el~mPnt.~, the
dem- ~nl~ting unit includes multiple path and searcl~g receivers along with a
signal combiner; the decoding unit includes a viterbi decoder and data quahty
VPrifir~ ~ion means; and the interleaving/~lPinterleaving unit includes a convolutional
encoder, an interleaver, a deinterleaver, a psuedo-random number (PN) sequence
*,r~ r, a data burst randornizer, and a f~nite impulse response (FIR) filter. Inad~hi. n to customary memory and support circuitry, acceptable examples ofthe
CPUs 20, 40 include conventional static CMOS (compl~ . y-symmetry metal-
oxide-semicnn~ ctor) high-integration Il~icloprocessors with general registers,
CA 02260974 1999-01-14
W O ~8~ 586 PCTrUS96/11712
segmPnt registers, base registers, index registers, status registers, and control
regi~lels. The vocoders 22, 42 provide the fi~nction of using a code excited hnear
pre~lirtion method to convert between pulse code modulated speech samples and
data with a reduced number of bits obtained by exploiting the intrinsic properties
5 of speech signals to remove redundancy.
The following describes eYamples of acceptable P~ nts of gl i~1ance for at
least one ofthe ~ r~.ed embo~1imPnt~ ofthe present invention. Except for the
intP~nal configuration mnrlifi~atinn~ and other inventive functions tiicc~ ed herein
(prog~ .;.,g, etc.), prior art examples similar to those of at least one ofthe
pr~rt;lled embodimPntc ofthe present inventions for the CPUs 20, 40, SMs 18, 38,BB ASICs 16, 36, and vocoders 22, 42, are, respectively, the 80C186
microprocessor available from Advanced Micro Devices of Sunnyvale, CA, the
Q52501- 1 S2 MSM available from Qualcomrn, Inc. of San Diego, CA, the
Q53101- 1 S2 baseband ASIC also available from Qualcomm, Inc., and, also from
15 Qualcomm, Inc., both the QCELP variable rate CDMA vocoder (first ~le~ d
embodiment of the present invention) and the High Rate Speech Service Option
CDMA vocoder (13.8 kbps) (second prertlled embodiment ofthe present
invention).
Tran~mic.~inns from the CDMA base station 30 to the CDMA mobile
20 station 10 are often referred to as the fon,vard channel link~ whereas trancmi~ c
from the CDMA mobile station 10 to the CDMA base station 30 are often referred
to as the reverse channel link. Thus, frames of data generated by the CDMA base
station 30 and tran~mitted in the forward channel link between the base station
antPnna 32 and the mobile station antenna 12 are often referred to as folward
25 channel data frames, and frames of data generated by the CDMA mobile station 10
and tran~mitted in the reverse channel ~ink between the mobile station antP.nna 12
and the base station antenna 32 are often referred to as reverse channel data
frames. Since both the CDMA mobile station 10 and the CDMA base station 30
are transceivers capable of sending and receiving information, most of the elements
30 of the CDMA mobile station 10 and the CDMA base station 30 are capable of
CA 02260974 1999-01-14
W O 98,'~29n6 PCTrUS96/11712
14
performing transmitter and receiver functions, e.g., both the mobile SM 18 and the
base SM 38 are each capable of p~,.r~ ~ing tr~ncmitting and receiving functions.Regarding the general functions of each of the various Pl~mPnts shown in
FIGS. 1 and 2, the typical process of speech co.. ~ ti~m in the forward channel
link begins with the PSTN interface 44 receiving pulse code modulated (PCM)
speech data from the PSTN. For typical voice telephone calls, PCM speech data isdigital data represPnting digital samples of a user's voice. After this data is passed
through the PSTN interf~ce 44, the data arr*es at the base station vocoder at 64kbps (8kHz samples of ~-law 8 bits per sample). Conversely, in the reverse
channel hnk, speech is rece*ed into the rnicrophone 26 and supplied in analog
forrn to the A-to-D/D-to-A converter 24 which converts the speech into a digitalsignal which is similar to that supplied to the base station vocoder 42. Thus, in the
first pre~l,ed embodiment ofthe present invention, the typical input for both the
base station vocoder 42 and the mobile station vocoder 22 are streams of PCM
speech data. However, as ~iCc~sse~ above, the CDMA base station 30 is also
capable of receiving encoded signals from other base st~tinnc which simply pass
through the PSTN interface 44 and vocoder 42 to the CPU 40.
Subsequently, the transmitting functions for both the CDMA mobile station
10 and the CDMA base station 30 are relat*ely similar. On a high level, the
vocoders 22, 42, CPUs 20~ 40, and SMs 18, 38 cooperate to assemble channel
frames of data, as ~i.cc~cced in more detail below. Subsequent to the SMs 18, 38,
the channel frames of data are processed in a conventional manner by the BB
ASICs 16, 36 and RF sections 14, 34 to be converted to analog signals, modulatedand transmitted through the ~ntenn~s 12, 32. When receiving channel frames of
data, the CDMA mobile station 10 reverses the above-stated functions to finally
produce PCM speech data output from the vocoder 22 and then converted into
analog signals and output through the speaker 28. Likewise, the CDMA base
station 30 produces PCM speech data through the vocoder 42 and PSTN interface
44 for tr~ncmiccion on the PSTN and passes encoded data through to other mobile
StS~fi-m.c
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96/11712
Now, regarding a more specific deswil.Lion of the new functions of the
vocoders 22, 42, CPUs 20, 40, and SMs 18, 38, since the relevant process steps
are similar in both the forward and reverse hnks, the process will be described from
- the viewpoint ofthe CDMA mobile station 10, but it should be understood that the
5 process is also applicable to the CDMA base station 30. Refer now also to FIG. 3
which shows a block diagram l~res~ ;on of selected frame generation
function~ A vocode function 50 is shown ,~)lece~ .g a group of s~lected SM &
CPU functions 52. The PCM speech data is first vocoded (speech encoded) as
indic~ted by the vocode fil-~ctirn 50. The CPU 20 acts as an int~rfqce between the
vocoder 22 and SM 18. The slolected SM & CPU functions 52 include an add next
rate function 54, an add frame quality indicator (FQI) for full and half rates
function 56 (i.e., CRC for error detection), an add encoder tail function 58, a
convolutionally encode function 62 for forward error correction, a repeat symbols
for haLt; quarter, and eighth rates filncti~m 64, and a block interleave function 66
15 for co.l~ burst errors. Refer briefiy to FIG. 4 which shows a frame structurediagram for CDMA traffic channel frames at various rates as the frames e~ast
;,~"~ tçly before the convolutionally encode function 62 The frame structures
include a full rate frame structure 70, a halfrate frame structure 72, a quarter rate
frame structure 74, and an eighth rate frame structure 76. After the block
interleave function 66, as would be understood by one rea~onqbly skilled in the art,
other convpntionql SM functions, broadly termed "modulate" in FIG. 37 are also
performed by the SM 18, including 64-ary orthogonal mo(llllating data burst
randomiz;ing, long code generating, offset quadrature phase shift key mo~ lqtingfil~rin~, etc., as ~ cile~e~l above.
Referring back to FIG. 3, the vocode function 50 includes converting
frames of PCM speech data into frames of speech encoded data at variable data
rates to be included as inform~ti- n in sl~bsequ~ntly formed traffic channel frames of
data. Thus, the term "frame of data" can refer to a frame of PCM data, a frame of
speech encoded data and/or a channel frame of data (traffic channel) which
includes as i~o~lion a frame of speech encoded data. In a convPntin~l manner,
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96111712
16
the vocode fi~n~i~n 50 includes c~ voice energy levels to adaptive
thresholds based on ~tletectel1 background noise levels to det~nnin~ an appl~liale
data rate for each frame of speech encoded data, and, using a code excited linear
pre.1ictî~ n (CELP) method, removing intrinsic redundancies to reduce the number5 of bits l~uhed to represent the speech. Such rate deterination is, however,
subject to rate s~1ectil-n commands from the CPU 20. Thus, the conventional
vocode (speech encoding) fimction in~ des receiving PCM speech data and
uul~Jullil~g frames of speech encoded data at variable data rates. However, in aquite unconv~ntif -- ~l manner, the vocode function 50 of the present invention also
lO includes determining a data rate of a subsequent frame of speech encoded data and
ouL~ulli,-g an in~1icati~n ofthat rate for being included in the current channel frame
of data, as shown by the add next rate function 54. Thus, according to the firstplerelled embol1im~nt ofthe present invention, the vocoder outputs frames of
speech encoded data at, for example, 8600 bps, 4000 bps, 1900 bps, and 700 bps.
15 Af~er next frame data rate indicator bits, FQI bits, and encoder tail bits are added,
the frames r~lesellt 9600 bps, 4800 bps, 2400 bps, and 1200 bps as shown in
FIG. 4.
Refer now to FIG. 5 which, in accordance with the first ~ rt;l,ed
embodiment ofthe present invention, shows a flow chart repres~ontation of steps of
20 the vocode (speech encode) function 50 of FIG. 3 as performed by the mobile
station vocoder 22 (FIG. 1) in the reverse channel link (again underst~ inp; that
similar steps are taken by the base station vocoder 42 (FIG. 2) in the forward
channel link). A first step 100 includes receiving a first frame of PCM speech data
for processing into a first frame of speech encoded data (also referred to as a
25 s~eech encoded frame of data). Subseql~Pntly, in step 102, the vocoding (speech
encoding) process begins for the first frame, including an initial step of determining
a data rate for the first frame of data through the above~ cn~se~ adaptive
threshold method. Step 104 shows that an indication ofthe first data rate is then
output from the vocoder (transmitted to the SM 18 through the CPU 20). Speech
30 encoding contin-les in step 106 until complete, after which the current speech
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96/11712
enco~led data frame is output in step 108 (during the initial pass through the
vocode fi~n~tir.n 50, the "~ ,.,l" frame is equivalent to the "first" frame and the
"next" frame is the "second" frame). The next frame of PCM data is received in
step 110, and the data rate ofthe next frame is quickly determined in step 112.
Thus, unlike other speech encoding methods that determine data rates late in thespeech encoding process, the present method is one in which an in~licqti~ n ofthis
newly determined data rate of the next frame is generated early and then output
from the vocoder 22 in step 114. Also, even if, depending on imple~ ;OI~
choices. a slight delay in generating the traffic channel data frame is introduced
10 through the generation ofthe subsequent frame data rate in-lirq~i- n a redllcti-)n in
time required to determine the data rate on the receiving end ofthe trqn~ .Cion is
available. Subsequently, as inflic l~e~l at step 116, the process loops back to step
106 where yet another frame of PCM speech data is received, and the process
contin~les
The next frame rate in~lirq~ti~n consists oftwo bits in the first prereIled
embodiment ofthe present invention since two bits are adequate to provide a one-to-one representation of the four possible data rates. With any number of possible
rates, the number of bits necessary to provide a one-to-one repres~ntPti- n ofthe
various rates equals the cmqll~st integer greater or equal to the Log2 of the total
number of possible rates. Refer now also to FIG. 6 which shows a flow chart
repres~ntqtion of selected channel frame assembly steps 53 taken by the SM 18
and CPU 20 (and SM 38 and CPU 40). According to the first ple~ed
embodiment~ the f~rst frame data rate in~ir.lqtir)n (for the first pass, "current" is
equivalent to "first", and "subsequent" is equivalent to "second") is stored by the
SM 18 and CPU 20 (step 118) until the current frame of speech encoded data
(step 120) and the next frame data rate infliclqtion (step 122) arrive from the
vocoder 22, as explained above. Thus, when the SM 18 and CPU 20 have both
the subsequent frame data rate in-lirqtion and the current frame of speech encoded
data, both are combined into the beginnings of a current traffic Gh~nnP.l ~ame of
data (step 124), as inflicated by the add next rate function 54 (FIG. 3). Another
CA 02260974 1999-01-14
W O 98/02986 PCT~US96/11712
18
way of e,~res~g this combining function is that the subsequent frame data rate
liA.qtifm is embedded or inserted into the current channel frame of data which
contains the current frame of speech encoded data as the i~ol~lion portion of
the current r~ A1 frame of data. Furthermore, it is understood that the exact bits
for the subsequent frame data rate in-lis~ n received from the vocoder 22 need
not necess-. ily be used as the actual subsequent frame data rate jn~ -a~i-n rather
the SM 18 and CPU 20 are understood to generate and insert bits representative of
the subsequent frame data rate indication
Subsequently, for the full and half rates, a frame quality indicator is
l 0 computed and added to the current channel frame of data, as also indicated by the
add FQI function 56 of FIG. 3 . Then, encoder tail bits are added to the currentchannel frame of data, as shown by step 128 of FIG. 6 and the add encoder tail
function of 5 8. Thus, as shown in FIG. 4~ the pre-encoder frame structures of the
first prerel,ed embodiment differ from convention~l channel frame structures in
15 that the subsequent frame data rate indication is substituted for two FQI bits for
the full and halfrate structures 70, 72, and for two .,lrO....~, ;on bits in the quarter
and eighth rate structures 74, 76 (i.e., the conventional frame structures inc1ude 12
FQI bits for full rate, 8 FQI bits for half rate, 40 information bits for quarter rate,
and 16 information bits for eighth rate). As discussed below, this particular
20 pl~.cem~nt ofthe subsequent frame data rate indicatiQns, as well as the particular
forrnat ofthe rate in~ication~, are given only as acceptable examples ofthe
teaching ofthe present invention. A-~-lition~lly, since, in the first prerel,ed
embodirnent, the add FQI function 56 (step 126) includes cornputing the FQl
based upon the i~o.... ~;fm and subsequent frame data rate indication, additional
25 error detection capabilities are realized.
After the current channel frame of data is assembled in one of the rate
formats shown in FIG. 4, the current channel frame of data is convolutionally
encoded at the data rate for the current frame of data, as indicated by step 130 of
FIG. 6 and the convohltiQn~lly encode function 62 of FIG. 3. In this way, the
30 subsequent frame data rate indication is also encoded along with the i.,fol, lation
CA 02260974 1999-01-14
W O 98102986 PCTrUS96111712
19
bits to provide good error correction for the subsequent frame data rate in(lira~ n
without a~liti~ n~l overhead or complexity. Subsequent to convoh~ti( n~l
encoding, encoder symbols (representative of pre-encoder bits) are repeated
thlou~out the frame for rates lower than full rate (step 132 of FIG. 6, function 64
5 of FIG. 3), and block interleaving is used to further protect the integrity ofthe
current channel frame, including the inllicatif n ofthe next frame data rate (step
134, function 66). Both ofthese functions, as well as the le ..~ i..g steps
neces~ry for completing the proc~s~g (step 136), including mn~llllstion, etc., are
conv~ntinn~l steps as would be understood by those re~cQn~bly skilled in the art.
Finally, this process loops back through step 138 to step 120 as shown in FIG. 6for processing the next channel frame of data.
On the receiving end of a tr~ncmiccil~n of a channel frame of data, such as
the forward traffic channel hnk, (underst~n~lin~ that similar events occur in the
reverse link) the CDMA mobile station l O is able to easily determine the data rate
of the information contained in the next channel frame of data. Refer now to FIG.
7 which shows a flow chart f~presf ~ on of selected frame analysis steps taken by
the vocoder 22, CPU 20! and SM l 8 in accordance with the first p~ led
embodiment of the present invention. A first step includes receiving a first channel
frame of data (convohltion~lly encoded data) at one of the four potential data rates
(step 152). Subsequently, the SM 18 processesthe first channelframe of data at
all of the four possible data rates to ~lett~.nnine (through conventional analysis of
FQI bits, symbol error rates, and other means for determining whether the correct
rate has been chosen, etc.) the correct data rate ofthe first frame of data. Then, in
step 156, the data rate of the next channel frame of data is ~let~.~mined by isolating
and analyzing the subsequent frame data rate inrliratinn ofthe current channel
frame of data. Step 158 infliç~tes that processing ofthe first channel frame of data
is then completed, including speech decoding the information at the current datarate.
Next, equipped with an expectatir n of the data rate ofthe next channel
frame of data, the SM 18 receives the next channel frame of data in step 160, at
CA 02260974 1999-01-14
WO 981'~2586 PCTrUS96/11712
which point that "next" becomes "current". Then, the frame of data is processed at
the expected data rate, i-~t~ling reversing functions 66, 64, and 62 of FIG. 3. In
an effort to prevent tran~s~on errors from propag~ting through the series of
data frames, the validity of the choice of rate chosen for each processed frame is
5 evahuated at decision block 164 lhrou~ as an example, an FQI analysis and a
symbol error rate analysis. For example, for the full and half rates, if FQI checking
passes, and for the quarter and eighth rates, if the symbol error rate is below its
corresponding rate-related threshold, the rate is ~letermitled to be valid, and the
operation proceeds through the Y~S branch to step 166. In nd~liti~.n~ the scope of
10 the present invention is understood to include other known methods of
determining whether the choice of chosen rate is correct, such as using viterbi
decoding intPn.~ inform~tion to determine rate selP.cti- n accuracy. At that point,
the current frame is analyzed to isolate the subsequent frame data rate in~licatinn
and determine the data rate of the next channel frame of data. Subsequently, in
15 step 168, proces~ing ofthe current frame of data is contim.ed until complete, and
the process loops back to step 160 to continue proce~Qi~ g If the data rate was not
found to be valid at de~;cinn block 164, conventional processing is utilized in step
l 70 to determine the app~ ;ate data rate for the current frame and then, in step
172, determine the data rate ofthe next frame of data from the subsequent frame
20 data rate in~lic~tinn before co~.~;l.;g with step 168 as shown. Also, though not
shown, in FIG. 7, if the rate still cannot be detPrmined after being processed at the
various possible data rates, the frame is classified as an erasure frame, and the
process cnntimles by processing the next frame as the first frame was processed in
step 152.
The present invention also includes a second pl~r~led embodiment which,
in many respects, is very similar to the first prer~l.ed embodiment. Refer now to
FIG. 8 which shows a block diagram representation of selected frame generation
functions. A vocode function 50' is shown preceding a group of s~ cte~l SM &
CPU functions 52'. The s~1~cte~1 SM & CPU functions 52' include an add next ratefunction 54', an add erasure/reserved (E/R) bit function 265, an add frame quality
CA 02260974 1999-01-14
W O 98~'~29~6 PCTAUSg6/11712
infli~2tQr (FQI) fimr,tinn 56', an add encoder tail fimctinn 58', a convohltinn~lly
encode fimction 62' for forward error correction, a repeat symbols for hal~,
quarter, and eighth rates function 64', and a block interleave function 66' for
comhatin~ burst errors. FIG. 9 is a vocoder timing diagram in accordance with the
second ~)rer~l-ed embodiment ofthe present invention. As the vocoder 22' (a
variation of vocoder 22 of FIG. 1 adapted forthe second pl~r~.,ed embodiment)
receives a continual supply of PCM samples, (bit-by-bit or sub-frame bursts) thedata can be divided into 20 ms frames as shown, and the vocoder 22' is configured
to double-buffer the PCM data. The known data rate of the first frame is
determined by the vocoder 22' (for example, responsive to a full rate Gontrol
command from the CPU 20', a variation of CPU 20 of FIG. 1 adapted for the
second ple~.led embodiment) and then made available to be output early in the
speech encoding processing ofthe first frame of PCM sarnples, as indicated at time
"A". Then, at some point up until time "B"~ the vocoder 22' finishes encoding the
first frame of PCM data and makes it available to the CPU 20'. Subsequently, thedata rate ofthe second frame is calc~llated and made available at time "C" (shortly
afler time "B", and approximately 20 ms after time "A"), and the process cnntim~es
such that the second fr~me of encoded data is made available at some point up
until time "D" (appro~i~tely 20 ms afler time "B"). Further ~ c~s~inn of the
vocoder function 50' and the selected SM & CPU function 52' is given below.
FIG. 10 shows a frame structure diagram for CDMA traffic channel frames at
various rates as the frames exist ;"~ r~ e1y before the convolutionally encode
function 62', in accordance with the second plere,led embodiment ofthe present
invention. The frame structures include a full rate frame structure 270 (14,400
bps), a half rate frame structure 272 (7,200 bps), a quarter rate frame structure 274
(3,600 bps), and an eighth rate frame structure 276 ( 1,800 bps).
Refer now also to FIG. 11 which, in accordance with the second ,olef~lled
embodiment ofthe present invention, shows a ~low chart represP.nt~tinn of steps of
the vocode (speech encode) function 50' of FIG. 8 as pelroll.led by the mobile
30 station vocoder 22' (FIG. 1) in the reverse channel hnk, again under~ ing that
CA 02260974 1999-01-14
W O 98/02986 PCT~US96/11712
similar steps are taken by the base station vocoder 42' (a variation ofthe vocoder
42 of FIG. 2 adapted for the seGond pl~,f~led embodiment) in the forward channellink. A first step 300 i ~hldes be~nnin~ the process of receiving contim~al PCM
speech data. The first 20 ms of PCM data received will be processed into a first5 frame of speech encoded data. Thus, in step 302, the vocoding (speech encoding)
process begins for the first frame, inr~ ing an initial step of determining a speech
encoding data rate for the first frame of data. According to this second plt;r~lled
embo~1im~.nt the first frame data rate is required through instructions from theCPU 20' to the vocoder 22' to be a known full rate. Step 304 shows that an
10 indic2tion of the first data rate is then made available for output to the CPU 20'
(time "A" in FIG. 9). Speech encoding continues in step 306 until complete, after
which the current (first) speech encoded data frame is made avai1able for output in
step 308 (a point in t~me up until time "B" in FIG. 9). The data rate ofthe next(second) frame is quickly ~let~ ned in step 312, and an indi~atirm is made
available for output to the CPU 20' at step 314 (time "C" in FIG. 9). Be~inning
with the second frame, (step 312, first pass) the above~ ed adaptive
threshold method is utilized to determine the speech encoding rate, subject to other
conventi~nal CPU 20' rate control commands. Furth~ re because ofthe
operation of the conv~ntir -- ~1 Hamming window technique, a small portion of
PCM data from the subsequent next frame (e.g., the third frame) is also exa_ined(step 312) in determining the best speech encoding data rate for the second and
subsequent frames, as would be understood by one reasonably skilled in the art, an
examp}e of which is described IS-96, section 2.4.3.2.2. Subsequently, as in~licated
at step 316, the process loops back to step 306, and the process continlles
Refer now also to FIG. 12 which shows a flow chart repres~ntatic-n of
selected channel frame assembly steps 53'. In accordance with the second
plerelled embodiment ofthe present invention, as shown in steps 317 - 324, a first
frame is generated and output at a standard full rate, e.g., 288 bits at 14,400 bps.
The known first frame data rate in~ tion~ first frame speech of encoded data, _nd
second frame data rate indication are received as shown in steps 317 - 319. In step
CA 02260974 1999-01-14
W O 98,~2g~ PCTAUS96/11712
320, the first frame speech encoded data is combined with the second frame data
rate in.licati..n in a known full rate frame structure. In step 322, an E/R bit is
col~,uled and added (the function of which is explained in detail below), and a
frame quality inllicnor is computed and added to the traffic channel frame along5 with encoder tail bits to produce a first traffic channel frame at the full rate frame
structure. No symbol repetition is necessary since the frame is a full rate frame.
Finally, the first traffic channel frame is encoded and block interleaved before the
processing is further fin~li7~d in step 324 and output. It is understood that other
embo-limPnts ofthe present invention omit steps 304 of FIG. 11 and 317 of FIG.
0 12 since the CPU 20'iS already knowledgeable of the known first frame data rate.
ln step 326, the CPU 22' receives the current frame of speech encoded data
(at this point, the "second" frame of speech encoded data) and subsequently
receives the in~i~sti- n ofthe next frame data rate in step 328. ln step 330, asshown in FIG. 8, assembly of the current traffic channel frame begins by adding the
15 next rate indicr~ n to the current frame of speech encoded data. Then, depending
on the data rate of the current speech encoded data, the trafflc channel frame will
be formed in step 332 according to one ofthe structures shown in FIG. 10 before
being finali7ed and output in step 334. The pre-encoding frame structures ofthe
second plerelled embodiment differ from conv~nti.mal CDMA higher rate
(standard Rate Set 2) traffic channel frame structures in that the subsequent firame
data rate indication is substituted for two information bits for the full, quarter and
eighth rate structures 270,274, and 276, and for two FQI bits in the half rate
structure 272.
On the receiving end of a trancmiC~i-m of a traffic channel frame of data,
25 such as the for,vard traffic channel link, (underctan~lin~ that similar events occur in
the reverse link) data rate determination is simplified. Refer now to FIG. 13 which
shows a flow chart reprecentatinn of sPlected frame analysis steps taken by the
vocoder 22', CPU 20', and the SM 18' in accordance with the second prerelled
embodiment ofthe present invention. FIG. 13 is very sirnilar to FIG. 7, thus thefirst and second plertlled emboflil-.. l.~i ofthe present invention are very similar to
CA 02260974 1999-01-14
PCT/US96/1 1712
WO ~1&.2986
24
each other with respect to the op~rati~ne represented by FIGS. 7 and 13. The
primary di~el~ces between FIGS. 7 and 13 relate to steps 352 and 354 where it isshown that the very first frame is received and convolutionally decoded at the
known full rate. The first frame's decoded data is then analyzed to determine the
5 data rate ofthe next traffic cllannel frame (step 356), correspondiug to the second
frame of speech encoded data, before processing cnntinlles in step 358. Then,
with step 360, operation proceeds as in the first pl~r~lr~d embodiment.
Up until this point, the diagrams and ~i.c&llc.cions regarding the present
invention have referred e~.sPnti~lly to primary traffic frame structures which do not
10 include .~raling or secondafy trafflc i~fo,~Lion. The scope ofthe present
invention is certainly int~Pnded to extend to such "mixed-mode" frame structureswhich include .ci~ling andlor secondary traffic i~u~ ;on Any necessary
modification.c to the diagrams would be understood by those reasonably skilled in
the art. Of particular note would be, for example, for the second l,ref~lled
embo~limPnt adding before steps 320 and 330 steps which include adding into the
trafflc channel the .si~n~ling or secondary traffic data and bits identifying the
structure of the frame. ~n addition, it would be necessary to adapt the next frame
rate indication in anticipatiûn ofthe overall frame data rate ofthe mixed mode
frame. ~n other words, as is understood by those reasonably skilled in the art, the
20 primary speech data may be speech encoded at, for example, a halfrate and
combined with ci~ling i~ollllalion into a full rate frame structure. This
technique is also used in other embo~1imPnt.c of the present inventiûn where frames
are required to be tran.cnlitted at a known rate, such as the second ple~elled
embodiment where the first frame is transmitted at a known rate, yet the vocoder25 22' is allowed to determine the speech encoding rate. In other words, the CPU 20'
and SM 18' would utilize mixed mode frames to accommodate speech encoding
rates which are ~etP~mined by the vocoder 22' to be less than full rate.
Furthermore, for any ofthe frame structures of the present invention conceptually
formed by ex~ anging frame quality indicator bits ofthe conventional frame
30 structures for the next frame data rate in-lications, the frame structures ofthe
, . . .
CA 02260974 1999-01-14
W O 98/02986 PCTrUS96/11712
information bits ofthose mixed-mode frames structures would be similar to the
conv~ntinn~l infnrrnqtion bit frame structures. On the other hand, for any of the
frame structures of the present invention conceptually formed by exchanging
inf rm~tion bits ofthe con~ 1 frame structures for the next frame data rate
indi.a irm~, the frame structures ofthe information bits ofthose mixed-mode
frames structures would change to msint~in a con~i~tPnt llu~bel of primary traffic
bits.
Furthermore"n accordance with the second plefelled embodiment ofthe
present invention, the E/R bit is used as an erasure bit in the reverse trafflc channel
link and a reserved bit in the foTward traffic channel link. Other embo~liments of
the present invention certainly include ~leei~ting an erasure bit for both directions
oftrrlcmieeion. According to the second plere~-ed embodiment, for frame types
utili7ing the erasure bit method (mi?ced-mode frame structures which do not
include any speech data, i.e., purely .ei~nali~lg or secondary traffic), the step of
convolutionally decoding at various rates to deterrni~le the correct rate (step 370 in
FIG. 13) is replaced (or, in other embo(limPnte supplemP.nted) by steps of usingthe erasure bit to notify the base SM 38' to retransmit the erasure frame at a
known full rate which can then be processed at that known rate when received by
the mobile station 10'.
Refer now also to FIG. 14 which shows a flow chart repre~ePntation of
selected channel frame assembly steps 53". In accordance with a third preferred
embodiment of the present invention, as shown in steps 374 - 377, a preamble
frame encoded at a known rate is first generated and output. This third
embodiment of the present invention is very similar to the second l)rere.led
embodiment in utili7.ing the vocode function 50' (FIG. 11) ofthe second prer~ ;dembodiment. However, the first frame speech encoding data rate is not required to
be predeterrnined. The preamble frame includes either blank speech information or
.ci~ information, depending on the need for .cj~n~ g i~ol~lion at that time,
and an indication ofthe data rate ofthe next frame, i.e., the data rate ofthe f~rst
frame of speech encoded data. Accordingly, the first frame data rate in~licati. n is
CA 02260974 1999-01-14
W 098/02986 PCTAUS96/11712
26
received (step 374) and combined in the preamble frame structure (step 375). In
step 376, an EIR bit is computed and added, and a frame quality indicator is
computed and added to the trafflc channel frame along with encoder tail bits to
produce a preamble traffic channel frame. Finally, the IJlea~le traffic channel
5 frame is encoded and block il~ttllca~ed before the processing is further fin~li7ed in
step 377 and output. Steps 374 - 377 take place at some point between times "A"
and "B" on FIG. 9. In step 378, the CPU 22" receives the current frame of speechencoded data (at this point, the "first" frame of speech encoded data) and
subsequently receives the in~ie~tinn ofthe next frame data rate in step 379. In
10 step 380, assembly ofthe current traffic channel frame begins by adding the next
rate indication to the current frame of speech encoded data. Then, depending on
the data rate ofthe current speech encoded data, the traffic channel frame will be
formed in step 381 according to one ofthe structures shown in FIG. 10 before
being finali7ed and output in step 382.
On the receiving end of a tr~n.cmic~ion of a traffic channel frame of data,
such as the forward traffic chalmel link, (again underst~n.ling that similar events
occur in the reverse link) data rate determination is again sirn~hfied. Refer now to
FIG. 15, which shows a flow chart repres~nt~tiQn of selected frame analysis steps
taken by the vocoder 22",CPU 20", and the SM 18" in accordance with the third
20 pl~;r~lled embodiment ofthe present invention. The very first frame to be received
will be the preamble frame encoded at the known rate. Thus, after receiving the
preamble frame in step 384, the SM 18" convohltinn~lly decodes the preamble
frame at the known rate in step 385 before analyzing the decoded data to
determine the data rate ofthe next traffic channel frame (step 386), which
25 corresponds to the first frame of speech encoded data. Then, with step 387,
operation proceeds as in the second pl~,relled embodiment.
According to a fourth prere~l~d embodiment ofthe present invention, as
represented in FIG. 16, vocoder steps ofthe second embodiment are colllbii~cd sothat the vocoder outputs only one package of data per frame to the CPU. Ai?~er
30 the first frame, such an output to the CPU and SM would include the current
CA 02260974 l999-0l-l4
W O 98/02986 PCTfUS96/11712
speech encoded data and the in-lica~ion ofthe subsequent frame data rate, as
in~lic~ted in step 414 of FIG. 16 and the o,..lssion of a step corresponding to step
308 in FIG. 11. Others steps ofthe fou~th prerelred embo~1im~t are similar to
steps ofthe second pler~led embodiment as r~les~ ed in FIG. 11.
According to a fi~h prer~led embodiment ofthe present invention, as
represented by an ~h~rn~te vocode filnc~ion 50"' in FIG. 17, after procee~lin~ in a
manner sirnilar to the first p~ ed embodiment of the present invention, PCM
data for the next frame of speech encoded data is received in step 504, before the
speech encoding processing on the first frame of data is complete. This is due to
the vocode function 50"' requiring more time than (have a process delay greater
than) the amount oftime represented by one frame of data, e.g., 20 rns. Thus, the
vocoder of this aher~ e embodiment processes mu1tiple frames of data
llt~neously in a parallel procç~r;ng arrangement (steps 502, 503, 510, and 512
related to one processor, and steps 506 and 508 related to a second), as evidenced
by step 506 where the speech encoding process also begins on the next
(subsequent) frame of data. Because ofthis time overlap, the ~ltern~te vocoder is
able to determine and output a data rate ofthe next frame of data before speech
encoding is complete on the current frame of data, as evidence by steps 508 - 512.
Accordingly, the corresponding selecte(l SM & CPU functions (represented in
FIG. 6 for the first prerelled embodiment) would be changed by reversing the
order between steps 120 and 122 since the next frame data rate would arrive at the
SM before the current frame of speech encoded data.
FIG. 1 X shows a frame structure diagram for the CDMA traffic channel
frames at various rates before being convolutionally encoded in accordance with a
sixth prc;re"ed embodiment ofthe present invention. This sixth embodiment is
ntical to the first embodiment of the present invention except for distinctions
related to the frame structure shown in FIG. 18. Rather than using information or
FQI bits for the subsequent frame data rate in~lic~tiQn bits, two tail bits from each
rate of the conventional frame structure are used, and the subse~uent frame data30 rate h~ro-l ;on is placed at the beginni ng of the frame. To accomplish such a
CA 02260974 1999-01-14
W O 9~ 6 PCT~US96111712
re~r,ti. n in tail size, the convohltifmql encoding method utilizes the conv~nti-mql
tailbiting "unknown tail" method whereby starting and ending states are the samefor the coded message. This method is understood by those skilled in the art, asfliecl~eee~l in "An F.ffif~ nt Adaptive Circular Viterbi Algoli~ for Deco~ling
S Generahzed Tqilhiting Conventional Codes", IEEE TranQac~i- ne on Vehicular
Technology, Vol. 43, No. 1, February 1994, pages 57 - 68.
The present invention also includes various other preL4led embodiments,
especially those formed by combi~g the various disclosed pl~r~ cd
embo~im~ntc In one such embodiment, depending on the amount oftime required
O by the vocoder for any particular frame, data (speech data or rate inflicqti~m data)
is output to the SM and CPU whenever available. Since vocoders often take more
or less time depending on the rates used, the next frame data rate in~lic~tiQn may
be available and output before or after the current frame speech encoded data.
In another class of prere~rcd embodim~nte ofthe present invention, rate
15 choice ev~hl~tif)n steps are performed only occasionally under the understan-ling
that tr ~ cmic~eion errors are normally very rare. Ad~1ition~11y, rate choice
evah.~tione are omitted in other embo~im~nts where periodic fL~ced rate
tranemie.eion.C are imposed to autom~tic~lly reduce the potential for propag~tion of
rate det~ ifm errors. In other words, the tr~n.em;tting stations of such
20 embo~ nl s periodically transmit frames at known rates according to periods
known by the receiving station so that rate determination errors are addressed.
Yet other pr~relled embo-lim~nts include inserting the subsequent frame
data rate indications in altemate locations within the data frame or inserting
in~;f~ nl~l subsequent frame data rate in~ ti-)ne which in~lisate changes in rates
2~ (i.e., upward, downward, no change, ma~mum"..;..i....l..., etc.) rather than
providing one-to-one indications ofthe rates. Still other ~lcrcllcd embollimt-nt.e of
the present invention include inserting subsequent frame data rate indications only
in selective frames, such as inserting in~lications only when a change in rates is
about to occur or for certain types of data frames, such as when assumptions can30 be made about other types of frames or when it is better to simply allow the other
CA 02260974 1999-01-14
- W 098t02986 PCTrUS96/11712
29
types of frames to be processed conventionally. Similarly, a system in which only
one direction of co.. ~~ .icati~ n utili7i~ subsequent frame rate id~ntific~tinnc is
also contemplated. In one example of such a system, forward traffic channel
frames tr~n~mitte~l by the base station, for example~ would include subsequent
5 frame rate iA~ntifications, yet reverse traffic channel frames tr~n~mitted by the
mobile station would not include such inAil~.ation.C Such a system would be
applicable when base station receiving resources are freely available for
convqntion~l rate determin~ti~nc and/or when it is advantageous not to use mobile
station tr~n~mittin~ resources to include such subsequent frame rate inAi~tinn.cStill other pr~r~l.ed embo-Aimf nt~ include speech encoding processes which
receive PCM data in other formats and at other rates, such as linear PCM as
opposed to ~-law PCM, as well as oull uLl ng frames of alternate l~n~thi, such as
those encompassing 10 ms of sampled speech. Other l)ler~lled methods include
variable data rate co~unication systems other than the CDMA digital cellular
15 systems and PCS systems. Furthermore, other p~er~.led embodimf~ntc include
receiving and buffering variable rate data frames from other sources besides thevocoder, such as external data devices co.. ic~tin~ at variable data rates. Still
other prere..ed embo~Aim~nte include lltili7ing ~ltPrn~te error protection (error
detectinn and error correction) methods, such as various block encoding methods
20 as opposed to the convolutional encoding method disclosed. Finally, as would be
understood by one rea.cnn:.hly skilled in the art, many ofthe clf ~ ls ofthe
various prere~ed emboAim~nts ofthe present invention can easily be split into
combinations of more discrete Pl~mf~ts or co~ ...cd into fewer, more complex
elements, as well as combined and substituted for functional clcm.,llls of between
25 the various emboflimPnts Thus~ the scope of the present invention certainly
includes any such increase or decrease in the number and complexity of el~- ... ~Is
necessary to perform the described functions, as well as combin~ti~ n~ ofthe
various embo-iim~ontc One particular combination of note includes modifying the
second plel~-led embodiment ofthe present invention to utitize the slower rate
30 frame structures of the first pref~l ~ed embodiment. Furthermore7 all of the rate
CA 02260974 1999-01-14
- W O 98l02986 PCTtUS96tll712
frame structures ofthe ,~ rt;lred embo~lim~tc, as well as others c~ f~mplated
herein, can easily be used with any ofthe various methods ofthe pl~re .ed
embo-lim~nts taught or suggested herein.
While the embodiments of the present invention which have been disclosed
5 herein are the prert;~l~d forms, other embo~1im~nts ofthe present invention will
suggest th~mee1ves to persons skilled in the art in view ofthis disclosure.
Therefore~ it will be unders ood that v~ and modifications can be effected
within the spirit and scope ofthe invention and that the scope ofthe present
invention should only be limited by the claims below. Furthermore, the equivalents
10 of all means-or-step-plus-function ~ U ,l~l e in the claims below are intended to
include any structure, materiaL or acts for pelr"~g the function as specificallyc1~imecl and as would be understood as acceptable substitutes by persons skilled in
the art.
We cla~m:
