-
公开(公告)号：GB2332541A
公开(公告)日：1999-06-23
申请号：GB9726835
申请日：1997-12-20
Applicant: IBM
Inventor： COBBETT MICHAEL
Abstract: A system and method for boot failure recovery in a digital computer as a result of an attempt to load a procedure, in which a flag is associated, 12, with the procedure following a loading attempt which results in failure. The flag is read, 10, on a further attempt to load the same procedure, to abort the further attempt and initiate an orderly exit thereby avoiding the potential for entry into an endless loop.
-
公开(公告)号：GB2305830A
公开(公告)日：1997-04-16
申请号：GB9520023
申请日：1995-09-30
Applicant: IBM
Inventor： BOWATER RONALD JOHN , COBBETT MICHAEL , STATON MERVYN ANTHONY
Abstract: The invention relates to a voice processing system capable of varying the speed of output of digitised audio data stored therein. The digitised audio data is stored using blocks of LPC coefficients. Each block is sufficient to allow twenty milliseconds of speech to be generated therefrom. Periodically, or selectably, the utilisation of particular blocks is repeated resulting in a decrease in the speed of output of the speech synthesised therefrom. Alternatively, selectably blocks of LPC coefficients are omitted from use thereby producing a corresponding increase in speech output. Figure 5 is a flow-chart illustrating the invention.
-
Patent Agency Ranking
公开(公告)号：GB2348304B
公开(公告)日：2003-07-09
申请号：GB9906644
申请日：1999-03-24
Applicant: IBM
Inventor： COBBETT MICHAEL , DUFFIN PAUL ANDREW , WONG YUK-LUN
Abstract: The invention relates to a method for executing, in a computer system 5, a device driver 10 which is used by a number of applications running on that system. A first 12 and second version 11 of the device driver are provided, with the first version being optimised and the second version containing debug code for providing trace information. Upon discovering a problem with the device driver, it is possible to selectively switch to the debug version of the code without any of the applications already using the device driver being aware of the switch. Having gathered the necessary trace data, it is then possible to transparently switch back to the optimised version. Conventionally, device drivers form part of the operating system within a computer and so are difficult to remove and re-load without taking the whole system down and re-booting. The invention however affords continuous service to any application already using the device driver and does not require a time-wasting reboot which may in any event remove the error condition.
-
公开(公告)号：GB2325122B
公开(公告)日：2002-03-27
申请号：GB9709183
申请日：1997-05-06
Applicant: IBM
Inventor： BOWATER RONALD JOHN , COBBETT MICHAEL , JORDAN ROBERT MICHAEL , RITCHIE ANDREW
-
公开(公告)号：GB2348304A
公开(公告)日：2000-09-27
申请号：GB9906644
申请日：1999-03-24
Applicant: IBM
Inventor： COBBETT MICHAEL , DUFFIN PAUL ANDREW , WONG YUK-LUN
Abstract: In a method for executing, in a computer system (5, Figure 1) a device driver 10 which is used by applications running on the system, first 12 and second 11, versions of the device driver code are provided. The first version is optimised, and the second version contains debug code for providing trace information, whereby upon discovering a problem with the device driver, it is possible to selectively switch to the debug version of the code without any of the applications already using the device driver being aware of the switch. Having gathered the necessary trace data, it is then possible to transparently switch back to the optimised version. Switching between the first and second versions of the software component occurs in response to a system call, which sets a boolean flag of which the value determines which version of the software component is executed. Conventionally, device drivers form part of the operating system within a computer and so are difficult to remove and re-load without taking the whole system down and re-booting. This method, however, affords continuous service to any application already using the device driver and does not require a time-wasting reboot which may, in any event, remove the error condition.
-
公开(公告)号：GB2308781A
公开(公告)日：1997-07-02
申请号：GB9526663
申请日：1995-12-29
Applicant: IBM
Inventor： BOWATER RONALD JOHN , COBBETT MICHAEL , PORTER LAWRENCE L , STATON MERVYN ANTHONY
Abstract: Using the Internet World Wide Web (WWW) network 320, a WWW Client 310 can communicate with a WWW Server 330 to request the reconfiguration of or to generate software which controls a voice processing application. A voice response system client communicates with a voice response system server to alter the configuration of the voice response system or control the execution of software on said voice response system which enables a voice application program to be generated. The output of the voice response system ordinarily destined for display on a visual display unit of a local terminal is directed to the voice response system server. The voice response system server forwards the data to a voice response system client. The voice response system client generates data in a first format useable by said WWW client terminal from data in a second format received from said voice response system and generates data in said second format useable by said voice response system from data in said first format received from WWW client terminal, that is, the voice response system client dynamically generates HTML data from the data generated by the voice response system for transmission to and subsequent display at the WWW client terminal and visa versa.
-
公开(公告)号：GB2305288A
公开(公告)日：1997-04-02
申请号：GB9518949
申请日：1995-09-15
Applicant: IBM
Abstract: A speech recognition system for discrete words uses a Hidden Markov Model (HMM), which is nominally adapted to recognise N different isolated words, but which is trained to recognise M different words, where M>N. This is achieved by providing M sets of audio recordings, each set comprising multiple recordings of a respective one of said M words being spoken. Only N different labels are assigned to the M sets of audio recordings, so that at least one of the N labels has two or more sets of audio recordings assigned thereto. These two or more sets of audio recordings correspond to phonetically dissimilar words. The HMM is then trained by inputting each set of audio recordings and its assigned label. The HMM can effectively compensate for the phonetic variations between the different words assigned the same label, thereby avoiding the need to utilise a larger model (ie to use M labels).
-
-
-
-
-
-
Search Results
17
records
(took 0.017 seconds)
