t
`
`PROVISIONAL APPLICATION COVER SHEET
`
`a request for filing a PROVISIONAL APPLICATION under 37 CFR 1.53(b)(2).
`
`LAST NAME
`Murdock
`Pearson
`Sajda
`
`FIRST NAME
`Michael
`John
`Paul
`
`Docket Number SAR 13807P
`
`Type a plus sign (+)
`inside this box
`
`INVENTOR(S)/APPLICANT(S)
`RESIDENCE (CITY AND EITHER STATE OR FOREIGN COUNTRY
`MIDDLE INITIAL
`East Windsor, NJ
`Lawrenceville, NJ
`Princeton, NJ
`
`TITLE OF THE INVENTION (280 characters max.)
`
`A SYSTEM AND METHOD FOR VOICE NAVIGATION FOR NETWORKED TELEVISION-CENTRIC INFORMATION
`SERVICES WITH BACK CHANNEL
`
`CORRESPONDENCE ADDRESS
`
`Sarnoff Corporation
`Director, Law and Patents
`201 Washington Road
`CN 5300
`Princeton
`
`STATE New Jersey
`
`ZIP CODE
`
`08543-5300
`
`COUNTRY
`
`US
`
`r Specification
`
`Drawing(s)
`
`Number of Pages
`
`Number of Sheets
`
`ENCLOSED APPLICATION PARTS (check all that apply)
`
`ri Small Entity Statement
`
`Other (specify)
`
`METHOD OF PAYMENT (check one)
`
`A check or money order is enclosed to cover the Provisional filing fees
`
`iq'The Commissioner is hereby authorized to charge
`
`filing fees and credit Deposit Account Number:
`
`2o- O78?
`
`PROVISIONAL
`FILING FEE
`AMOUNT(S)
`
`$150.00
`
`The invention was made by an agency of the United States Government or under a contract with an agency of the United States Government.
`
`RiNo.
`
`ri Yes, the name of the U.S. Government agency and the Government contract number are:
`
`
`
`Respectfully Submitted,
`
`SIGNATURE
`TYPE or
`PRINT NAME Kin-Wah Tong
`
`
`
`Date:
`
`REGISTRATION NO.
`
`39,400
`
`Additional inventors are being named on separately numbered sheets attached hereto.
`Comcast - Exhibit 1011, page 1
`
`Comcast - Exhibit 1011, page 1
`
`

`

`***EXPRESS MAIL CERTIFICATION***
`
`"Express Mail" mailing label number EL446266677US
`Date of deposit (1 /7 7 9
`
`
`
`I hereby certify that this paper or fee is being deposited with the United States Postal Service "Express
`Mail Post Office to Addressee" service under 37 CH( 1.10 on the date indicated above and is addressed to the
`t Commissioner for Patents, Box Provisional Patent Application, Washington, D.C. 20231.
`Assist
`
`C9-yt
`Signature of person mailing paper or fee
`
`/! nit
`0 3 0-4"
`Name of person mailing paper or fee
`
`i.4
`
`; A
`
`Comcast - Exhibit 1011, page 2
`
`Comcast - Exhibit 1011, page 2
`
`

`

`SAR 13807P
`
`-1-
`
`A System and Method for Voice Navigation for Networked Television-
`Centric Information Services with Back Channel
`
`The present invention relates to an apparatus, system and
`
`5 concomitant method for voice control of a networked television-centric
`information service that utilizes a backward channel capability.
`
`DETAILED DESCRIPTION
`While speech-controlled consumer appliances are attractive for
`10 many reasons, accurate and reliable speech recognition requires a fairly
`powerful computer. It is the first advantage of this invention to provide a
`
`central speech recognition server that can be used by many speech-
`controlled appliances that are connected by a network. This means that a
`speech-controlled appliance can be delivered in a much more economical
`
`15 manner than if it required a fast CPU with a large amount of memory. It
`
`is the second advantage of this invention to provide a mechanism, the
`network's backward channel, for transmitting voice control data to the
`network information source. The central speech recognition server is
`
`contained at the network information source. It processes speech control
`
`20 data arriving through the backward channel and then sends the
`
`appropriate programming modifications through the forward channel as
`
`requested by the viewer. One of the principal reasons for speech
`
`recognition inaccuracies is the background room noise that masks the
`
`25
`
`speech signal. Even a small amount of noise from other speakers can lead
`to a large number of speech errors. Additionally, noise in the speech
`signal can make it very difficult to find an economical code for
`transmission. It is the third advantage of this invention to provide local
`speech enhancement in order to achieve an optimal coding for economical
`transmission through the backward channel. Since multiple microphones
`30 can be used to accurately focus on the viewer's speech, it is the fourth
`advantage of this invention to use multiple microphones to acquire the
`speech signal and then use a local signal separation method in the home
`entertainment processor to extract the clean signal prior to speech coding.
`
`Comcast - Exhibit 1011, page 3
`
`Comcast - Exhibit 1011, page 3
`
`

`

`SAR 13807P
`
`System Description
`
`-2-
`
`The following is a brief definition of each of the important terms in
`this invention. The Home Entertainment Processor (HEP) is the core
`
`5
`
`component of a home entertainment system which is primarily a
`television viewing system. The HEP is an electronic device for receiving
`network programming content through the forward channel and
`
`transmitting speech control data through the backward channel. The
`forward channel is the channel in which the programming content from
`the information source is transmitted into the HEP. The back channel is
`
`the channel in which the Voice Control Data from the HEP is transmitted
`to the information source. In this invention, the HEP is controlled by a
`speech recognition engine that resides in the Information Source. The
`content from the information source could be web content, video-on-
`demand, cable television, or a local video cassette player.
`
`10
`
`15
`
`A Hand-Held Controller (HHC) is used by the viewer in a manner that is
`similar to the way current television remote controls are used with the
`exception that the viewer makes selections by speaking into the HHC
`rather than pushing buttons. The viewer's speech command ("Show me
`20 all of the NCAA football games on this afternoon") is acquired from the
`HHC with one or more microphones, which is then transmitted to the
`
`HEP. It is also in the current embodiment to acquire a video signal of the
`speaker using a video camera in the Hand-Held Controller. This would
`also be transmitted to the HEP and processed in conjunction with the
`25 speech signal. The speech signal is then used to determine the speaker's
`
`identification. The speech signal is enhanced to reduce noise in a manner
`
`that improves the speech recognition accuracy and the efficiency of the
`speech coder. Finally, the speech signal is coded for efficient transport
`over the back channel and sent out to the Back Channel Multiplexor.
`30 It should be noted that the HEP could also be connected to the Home
`Compute Server, which is a computer device for controlling non-
`entertainment compute devices in the home (e.g. communications,
`security, personal computing, kitchen appliances, etc.). Figure 1
`illustrates a block diagram of an illustrative embodiment of the present
`
`Comcast - Exhibit 1011, page 4
`
`Comcast - Exhibit 1011, page 4
`
`

`

`SAR 13807P
`
`-3-
`
`invention. Each of the components is described in more detail in the
`following section.
`
`Component Descriptions.
`The functionality of each of the components of this invention in one
`
`5
`
`embodiment is as follows:
`
`Hand-Held Controller - This component is a small electronic device that is
`held by the viewer and used to navigate through the programming
`10 content. It contains one or more microphones, a video camera and a radio-
`frequency transmitter for sending the speech and video data to the HEP.
`The current embodiment of this invention contains push button controls
`for viewers that do not wish to use speech control.
`
`15 Home Entertainment Processor - This component is an electronic device
`for receiving network content through the network's forward channel. It
`contains software modules that perform speaker identification, speech
`enhancement and coding and video tracking. The video component of the
`system is used to extract visual information (such as lip location) from the
`20 viewer that will render the speech processing more accurate. The speaker
`identification is used by the speech server to perform a more accurate
`recognition by employing a user profile that contains a statistical model of
`an individual's preferences and command patterns. The coded speech
`commands are sent out to the network's backward channel in order that
`the speech commands can be processed by a speech recognition engine at
`the information source.
`
`25
`
`30
`
`Backward Channel Multiplexor - This component mixes the backward
`channel content received from several HEPs into one signal that is sent to
`the Information source. In situations in which a large number of viewers
`are likely to speaking simultaneously, the BCM can be used to delay and
`stagger the content, thus avoiding channel overload.
`
`Comcast - Exhibit 1011, page 5
`
`Comcast - Exhibit 1011, page 5
`
`

`

`SAR 13807P
`
`-4-
`
`Information Source - This component is located at the site of the
`information provider. The information could be web content, video-on-
`demand, or cable television. However, in any case, this component
`contains the central speech recognition server. The speech recognition
`5 server is connected to electronic programming guides, vocabulary lists
`and other meta-data. Also, an inference engine for processing user profile
`information is used by the Information Source to help guide the speech
`recognition engine.
`
`15
`
`10 Conclusion
`It is the first advantage of this invention to provide a central speech
`recognition server that can be used by many speech-controlled appliances
`that are connected by a network. This means that a speech-controlled
`appliance can be delivered in a much more economical manner than if it
`required a fast CPU with a large amount of memory.
`It is the second advantage of this invention to provide a mechanism,
`the network's backward channel, for transmitting voice control data to the
`network information source. The central speech recognition server is
`contained at the network information source. It processes speech control
`20 data arriving through the backward channel and then sends the
`appropriate programming modifications through the forward channel as
`requested by the viewer. One of the principal reasons for speech
`recognition inaccuracies is the background room noise that masks the
`speech signal. Even a small amount of noise from other speakers can lead
`to a large number of speech errors. Additionally, noise in the speech
`signal can make it very difficult to find an economical code for
`transmission.
`It is the third advantage of this invention to provide local speech
`enhancement in order to achieve an optimal coding for economical
`transmission through the backward channel. The speech enhancement
`also eliminates noise from the speech signal so that the recognition is
`much more accurate.
`It is the fourth advantage of this invention to use multiple
`microphones to acquire the speech signal. Multiple microphones can be
`
`25
`
`30
`
`Comcast - Exhibit 1011, page 6
`
`Comcast - Exhibit 1011, page 6
`
`

`

`SAR 13807P
`
`-5-
`used to accurately focus a beam on the viewer's speech A local signal
`separation method in the home entertainment processor is used to extract
`the clean signal prior to speech coding.
`
`Comcast - Exhibit 1011, page 7
`
`Comcast - Exhibit 1011, page 7
`
`

`

`SAR 13807P
`
`-6-
`
`(Other Backward
`Channel Multiplexors)
`
`(Other Home
`Entrtanment
`Processors)
`
`Back Channel
`Multplcxer
`
`1 n formation
`Provider
`
`Backward Flow: Coded Speech Control Data
`
`Forward Flow: Video and Audio Content, Speech
`Enhancement and Speech Coding Parameters
`
`Horne
`Ente rta iimen
`Processor
`
`Hand-Held
`Controller
`
`DVD/VCR
`
`Television
`Display
`
`MOM
`
`Home Compute Server
`
`Figure 1. Block Diagram of System.
`
`Comcast - Exhibit 1011, page 8
`
`Comcast - Exhibit 1011, page 8
`
`