Publications
Automatic speaker clustering from multi-speaker utterances
March 15, 1999
Conference Paper
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. II, 15-19 March 1999, pp. 817-820.
Topic:
R&D area:
Summary
Blind clustering of multi-person utterances by speaker is complicated by the fact that each utterance has at least two talkers. In the case of a two-person conversation, one can simply split each conversation into its respective speaker halves, but this introduces error which ultimately hurts clustering. We propose a clustering algorithm which is capable of associating each conversation with two clusters (and therefore two-speakers) obviating the need for splitting. Results are given for two speaker conversations culled from the Switchboard corpus, and comparisons are made to results obtained on single-speaker utterances. We conclude that although the approach is promising, our technique for computing inter-conversation similarities prior to clustering needs improvement.
Summary
Blind clustering of multi-person utterances by speaker is complicated by the fact that each utterance has at least two talkers. In the case of a two-person conversation, one can simply split each conversation into its respective speaker halves, but this introduces error which ultimately hurts clustering. We propose a clustering...
READ MORE
Corpora for the evaluation of speaker recognition systems
March 15, 1999
Conference Paper
Published in:
ICASSP 1999, Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, 15-19 March 1999.
Topic:
R&D area:
Summary
Using standard speech corpora for development and evaluation has proven to be very valuable in promoting progress in speech and speaker recognition research. In this paper, we present an overview of current publicly available corpora intended for speaker recognition research and evaluation. We outline the corpora's salient features with respect to their suitability for conducting speaker recognition experiments and evaluations. Links to these corpora, and to new corpora, will appear on the web http://www.apl.jhu.edu/Classes/Notes/Campbell/SpkrRec/. We hope to increase the awareness and use of these standard corpora and corresponding evaluation procedures throughout the speaker recognition community.
Summary
Using standard speech corpora for development and evaluation has proven to be very valuable in promoting progress in speech and speaker recognition research. In this paper, we present an overview of current publicly available corpora intended for speaker recognition research and evaluation. We outline the corpora's salient features with respect...
READ MORE
Implications of glottal source for speaker and dialect identification
March 15, 1999
Conference Paper
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. II, 15-19 March 1999, pp. 813-816.
Topic:
R&D area:
Summary
In this paper we explore the importance of speaker specific information carried in the glottal source. We time align utterances of two speakers speaking the same sentence from the TIMIT database of American English. We then extract the glottal flow derivative from each speaker and interchange them. Through time alignment and this glottal flow transformation, we can make a speaker of a northern dialect sound more like his southern counterpart. We also time align the utterances of two speakers of Spanish dialects speaking the same sentence and then perform the glottal waveform transformation. Through these processes a Peruvian speaker is made to sound more Cuban-like. From these experiments we conclude that significant speaker and dialect specific information, such as noise, breathiness or aspiration, and vocalization, is carried in the glottal signal.
Summary
In this paper we explore the importance of speaker specific information carried in the glottal source. We time align utterances of two speakers speaking the same sentence from the TIMIT database of American English. We then extract the glottal flow derivative from each speaker and interchange them. Through time alignment...
READ MORE
'Perfect reconstruction' time-scaling filterbanks
March 15, 1999
Conference Paper
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. III, 15-19 March 1999, pp. 945-948.
Topic:
R&D area:
Summary
A filterbank-based method of time-scale modification is analyzed for elemental signals including clicks, sines, and AM-FM sines. It is shown that with the use of some basic properties of linear systems, as well as FM-to-AM filter transduction, "perfect reconstruction" time-scaling filterbanks can be constructed for these elemental signal classes under certain conditions on the filterbank. Conditions for perfect reconstruction time-scaling are shown analytically for the uniform filterbank case, while empirically for the nonuniform constant-Q (gammatone) case. Extension of perfect reconstruction to multi-component signals is shown to require both filterbank and signal-dependent conditions and indicates the need for a more complete theory of "perfect reconstruction" time-scaling filterbanks.
Summary
A filterbank-based method of time-scale modification is analyzed for elemental signals including clicks, sines, and AM-FM sines. It is shown that with the use of some basic properties of linear systems, as well as FM-to-AM filter transduction, "perfect reconstruction" time-scaling filterbanks can be constructed for these elemental signal classes under...
READ MORE
A 9PAC system and application programmer's guide
February 16, 1999
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-267
Topic:
R&D area:
Summary
The ASR-9 Processor Augmentation Card (9PAC) is a custom processing card that provides the ASR-9 system with increased beacon and radar processing performance. This paper describes the system and application software that executes on the prototype board, with an emphasis on the interaction between software modules. The application software on the 9PAC determines the position of radar and beacon target reports, replacing software that previously ran on the ASR-9 Array Signal Processor (ASP). The software is organized as a set of cooperating tasks executing under the control of a real-time operating system, PAC/OS, which provides all the services typical of an embedded kernel such as interrupt handling, pre-emptive multitasking, queues, signals, semaphores, mailboxes, and memory management. The deployment of 9PAC will occur in two phases. The Phase I application replaces only the beacon target detector (BTD) and radar/beacon target merge (MRG) functions of the ASP. The Phase I application consists of two executable programs since Phase I uses only two of the C44 processors on the 9PAC. One program, the housekeeping processor, is responsible for all I/O functions and performs the radar/beacon merge operation. The second progam, the beacon processor, is dedicated to processing the raw beacon replies and generating beacon targets which are then returned to the first processor for the merge operation. The Phase II application consists of three executable programs, one for each of the C44 processors on the 9PAC and performs much of the Phase I functionality and adds primary radar processing. The intent of this paper is to provide the 9PAC software support personnel with sufficient information to implement future enhancements without unintentionally compromising some aspect of the overall system.
Summary
The ASR-9 Processor Augmentation Card (9PAC) is a custom processing card that provides the ASR-9 system with increased beacon and radar processing performance. This paper describes the system and application software that executes on the prototype board, with an emphasis on the interaction between software modules. The application software on...
READ MORE
Evaluating intrusion detection systems without attacking your friends: The 1998 DARPA intrusion detection evaluation
February 9, 1999
Conference Paper
Published in:
Third Conf. and Workshop on Intrusion Detection and Response, 9-13 February 1999.
R&D area:
Summary
Intrusion detection systems monitor the use of computers and the network over which they communicate, searching for unauthorized use, anomalous behavior, and attempts to deny users, machines or portions of the network access to services. Potential users of such systems need information that is rarely found in marketing literature, including how well a given system finds intruders and how much work is required to use and maintain that system in a fully functioning network with significant daily traffic. Researchers and developers can specify which prototypical attacks can be found by their systems, but without access to the normal traffic generated by day-to-day work, they can not describe how well their systems detect real attacks while passing background traffic and avoiding false alarms. This information is critical: every declared intrusion requires time to review, regardless of whether it is a correct detection for which a real intrusion occurred, or whether it is merely a false alarm. To meet the needs of researchers, developers and ultimately system administrators we have developed the first objective, repeatable, and realistic measurement of intrusion detection system performance. Network traffic on an Air Force base was measured, characterized and subsequently simulated on an isolated network on which a few computers were used to simulate thousands of different Unix systems and hundreds of different users during periods of normal network traffic. Simulated attackers mapped the network, issued denial of service attacks, illegally gained access to systems, and obtained super-user privileges. Attack types ranged from old, well-known attacks, to new, stealthy attacks. Seven weeks of training data and two weeks of testing data were generated, filling more than 30 CD-ROMs. Methods and results from the 1998 DARPA intrusion detection evaluation will be highlighted, and preliminary plans for the 1999 evaluation will be presented.
Summary
Intrusion detection systems monitor the use of computers and the network over which they communicate, searching for unauthorized use, anomalous behavior, and attempts to deny users, machines or portions of the network access to services. Potential users of such systems need information that is rarely found in marketing literature, including...
READ MORE
Addressing the weather delay problems of the New York City airports with the Integrated Terminal Weather System
January 10, 1999
Conference Paper
Published in:
8th Conf. on Aviation, Range, and Aerospace Meteorology, 10-15 1999.
Topic:
R&D area:
R&D group:
Summary
The three major New York City (NYC) air carrier airports (Kennedy, LaGuardia, Newark) currently experience high delays due to adverse terminal weather, both in an absolute sense and relative to other major airport complexes. Significantly expanding the NYC airports (e.g., by adding new runways) to reduce delays is not feasible. One alternative is to provide aviation weather decision support systems to air traffic, airline, and airport operations personnel to help them operate more safely and effectively with the existing runway/taxiway complexes. Under an innovative partnership between the Port Authority of New York and New Jersey and the Federal Aviation Administration (FAA), Massachusetts Institute of Technology, Lincoln Laboratory has installed and is currently operating a functional prototype Integrated Terminal Weather System (ITWS) to conduct research on improving the safety and efficiency of operations at the NYC airports during adverse weather. The New York terminal area provides a stringent test of the ITWS ability to safely reduce delays due to both the meteorology and the operational usage challenges not found at the earlier ITWS test locations of Orlando, Memphis, and Dallas. In this paper, we describe key features of the New York terminal environment and the ITWS prototype, the initial experience in addressing the meteorological and operational usage challenges of the New York terminal area, and describe plans for the coming years.
Summary
The three major New York City (NYC) air carrier airports (Kennedy, LaGuardia, Newark) currently experience high delays due to adverse terminal weather, both in an absolute sense and relative to other major airport complexes. Significantly expanding the NYC airports (e.g., by adding new runways) to reduce delays is not feasible...
READ MORE
Aviation user needs for convective weather forecasts
January 10, 1999
Conference Paper
Published in:
8th Conf. on Aviation, Range, and Aerospace Meteorology (ARAM), 10-15 January 1999.
Topic:
R&D area:
R&D group:
Summary
The prediction of convective weather is very important to aviation, since almost half of the serious delay at major airports in the warm season is caused by thunderstorms. The need for accurate 0-6 hr forecasts for NAS users has been the subject of extensive publications, forums, and advisory committees in the aviation weather community over the last several years (Wolfson, et al; 1997). The Convective Weather Product Development Team (PDT), a core team of scientists and engineers from NCAR, NSSL, and MIT LL, was formed in 1996 as part of the reorganization of the FAA Aviation Weather Research Program. The team is developing convective weather forecast algorithms that produce operationally useful products for both the terminal area and enroute airspace. The products are designed to meet specific users' air traffic planning and safety needs. Before major algorithm development began, PDT members visited terminal and enroute Air Traffic (AT) personnel and airline dispatchers to understand the forecast products that were currently available to them and their needs for a near future product. Also, in order to reach the pilot community, a pilot survey about existing convective weather information and how to improve it, was created and distributed at the OshKosh Fly-In in August of 1997. This needs assessment took advantage of interviewees that had extensively used state-of-the-art weather information products (ITWS) in an operational setting for years. Their requirements, based on personal experiences with operational products during convective weather events, were less stringent than those reported in the recent requirements document pertaining to ARTCC TMUs (Browne, et al; 1999). The results of these investigations were used in the creation of the DFW Terminal Convective Weather Forecast (TCWF) product and the National Convective Weather Forecast (NCWF) products that were demonstrated throughout the summer of 1998 (Hallowell, et al; 1999; Mueller, et al; 1999). These demonstrations also provided additional insight into user needs. In this paper we describe Air Traffic users and their specific responsibilities. We then summarize AT and airline needs based on interviews conducted in 1997 and 1998. Information on pilots' needs for convective weather information is presented at the end.
Summary
The prediction of convective weather is very important to aviation, since almost half of the serious delay at major airports in the warm season is caused by thunderstorms. The need for accurate 0-6 hr forecasts for NAS users has been the subject of extensive publications, forums, and advisory committees in...
READ MORE
Nowcasting requirements for the aircraft vortex spacing system (AVOSS)
January 10, 1999
Conference Paper
Published in:
8th Conf. on Aviation, Range, and Aerospace Meteorology, 10-15 Jan. 1999, pp. 340-344.
Topic:
R&D area:
R&D group:
Summary
Aircraft wake vortices are counter-rotating tubes of air that are generated from aircraft as a consequence of the lift on the aircraft. The safety concern of wake vortices, particularly when lighter aircraft are following heavy planes, has caused the Federal Aviation Administration (FAA) to enact minimum separation requirements during the arrival phase of flight. These separation standards are imposed at the arrival threshold during Instrument Flight Rules (IFR) and are a significant constraint on arrival capacity at the largest U.S. airports. Any movement toward increasing air traffic efficiency, such as concepts toward free-flight, must address increasing runway capacity if they are to be fully effective. Decades of past wake vortex measurements clearly show that current wake vortex separations are overconservative in many weather conditions, and that adapting the separations to the current weather state could safely reduce these separations...This paper describes the known meteorological influences on vortex behavior and gives an overview of AVOSS. Airport climatology is studied to discuss the prevalence of conditions that are conducive to capacity increases with AVOSS technology. Finally, additional constraints on AVOSS nowcasts are discussed.
Summary
Aircraft wake vortices are counter-rotating tubes of air that are generated from aircraft as a consequence of the lift on the aircraft. The safety concern of wake vortices, particularly when lighter aircraft are following heavy planes, has caused the Federal Aviation Administration (FAA) to enact minimum separation requirements during the...
READ MORE
Optimizing the ITWS algorithm designed to remove anomalous propagation ground clutter from the ASR-9 precipitation product
January 10, 1999
Conference Paper
Published in:
8th Conf. on Aviation, Range and Aerospace Meteorology, ARAM, 10-15 January 1999.
Topic:
R&D area:
R&D group:
Summary
A key product within the Integrated Terminal Weather System (ITWS) Initial Operating Capability (IOC) product suite removes anomalous propagation (AP) ground clutter from the ASR-9 precipitation product. This has been identified as a critical component of ITWS due to the frequent occurrence of AP when storms or outflows move over an ASR-9. Editing is accomplished by comparing the raw ASR-9 weather data to composite maps generated by the Next Generation Weather Radar (NEXRAD) and the Terminal Doppler Weather Radar (TDWR). An editing template, containing regions of AP, is created based on the ASR-9 data collected at the middle of the composite volume scan to minimize the difference in update rates. The template is used to edit the ASR-9 scan immediately after the composite map and all subsequent scans until a new composite map is received. This algorithm has been shown to perform quite well, especially if the weather and AP returns are not co-located. During the 1994 Demonstration and Validation Operational Test and Evaluation in Memphis (MEM) and Orlando (MCO), the probability of editing AP (PEAP) in the absence of weather was 0.97 for level 2 and greater returns (Klingle-Wilson, 1995). The probability of editing weather (PEW) for those cases with weather only was quite low, i.e., 0.01. In order to minimize the removal of weather returns in those cases where the AP and weather are located in close proximity, the editing thresholds are quite conservative. This is reflected by the 1994 results which show a PEAP of 0.81 and a PEW of 0.02 for this class of event. Besides the conservative thresholds, another area of concern is the fact that the AP regions can expand or increase in intensity after the AP editing template is created. This rapid variation frequently occurs with convectively generated AP and can cause the performance of the algorithm to decrease with time until a new template is created. In this study, we will examine the algorithm failure mechanisms in detail to identify possible site-adaptable parameter changes that can be used to improve the performance for the mixed weather/AP events. This is especially germane since the parameter set was not re-evaluated after the TDWR composite map was incorporated in 1995. In the critical region over the airport during hazardous weather conditions, this radar updates more frequently than the NEXRAD. Since the parameters were designed to account for the NEXRAD volume update rate, they are probably too conservative for the current algorithm (which uses both composite maps).
Summary
A key product within the Integrated Terminal Weather System (ITWS) Initial Operating Capability (IOC) product suite removes anomalous propagation (AP) ground clutter from the ASR-9 precipitation product. This has been identified as a critical component of ITWS due to the frequent occurrence of AP when storms or outflows move over...
READ MORE