Projects
RealTime Interaction
A team at the University of Southampton has been investigating how to improve the real time usability of speech recognition generated text. Real time editing, personal display technologies, and annotation tools are among some of the innovations being researched and developed by the Consortium.
To improve speech recognition caption accuracy, an application called RealTimeEdit was developed to enable corrections of speech recognition errors to be made in real-time. One editor can find and correct errors or the task of finding and correcting errors can be shared between two editors, one using the mouse and the other the keyboard. It is also possible to use multiple editors sequentially to allow a 2nd operator to correct errors that a 1st operator did not have time to correct. In this way a real-time editor can be used in situations where high accuracy captions are required. Investigations have shown that an operator can make up to 15 corrections per minute without training or experience with the system.
While projecting the text onto a large screen in the classroom has been used successfully in Liberated Learning classrooms it is clear that in many situations an individual personalised and customisable display (e.g. font size, formatting, colour, etc) would be preferable or essential. A personalised server and client have been developed to enable users to customise their displays on their own networked computers.
In situations where there is more than one person speaking, using multiple instances of IBM ViaScribe over a network can create captions in multiple windows using the personal display client. To produce a single transcript of the session with speakers identified, the application RealTimeMerge was developed to add the speaker’s name to the text captions and merge the streams from the instances of ViaScribe. The individual and/or merged outputs can be edited in real time.
To support note taking in lectures or meetings, the application RealTimeAnnotate has been developed to enable annotations synchronised with the captions to be made in real time.
For more information about RealTime Interaction, please contact:
Dr. Mike Wald
email: M.Wald@soton.ac.uk
Synote
Mike Wald and his team from Southampton University in the United Kingdom have developed the feature rich multimedia tool Synote. This application allows professors and scholars to upload transcribed lectures and presentations to the web. It takes as input XML files; timestamped text files that correspond to an audio recording. These presentations can also include video and PowerPoint media. This information is synchronized and presented in a single web page that can be easily accessed on multiple browser platforms worldwide.
Having media inputs synchronized allows Synote to automatically adjust the media to correspond to user navigation. As a user navigates the audio recording, the video, PowerPoint and text file will synchronize itself with the audio location. The user can navigate all media types in this way. Having such a wide range of media working coherently provides a rich learning environment for students. The diverse media types also provide many ways to access the information, making this a very practical application for students that have visual or hearing impairments.
Perhaps the most innovative aspect of Synote is the ability to provide media bookmarks, known as Synmarks. Conventional audio and video prove to be particularly hard to use for educational purposes. Navigating such files to locate a particular piece of information is usually trial and error, randomly jumping around the file until you happen across relevant information. Synote eliminates this problem by presenting users with synchronized bookmarks. These allow a user to find relevant information quickly and navigate lengthy media files with ease.
Synote provides a shared learning environment in which peers and mentors can all contribute towards increasingly comprehensive media presentations. As a user it is possible to view Synmarks created by others while adding your own. Comprehensive search features mean that it is easy to locate Synmarks that correspond to a particular user or piece of information. Slide notes can also be used to create synchronized transcripts, with Synmarks being created for each slide. Each slide contains its title and associated text, allowing them to be searched for keywords. It is even possible to perform searches on the transcribed text using search features built into the browser. A fully adjustable web page allows users to customize the presented media to their individual needs.
With the right permissions, users can adjust and enhance the multimedia content presented on the page. A user is able to to input plain text files and manually synchronize it with other media using tools provided within Synote. Users also have the option of using “srt” files, the result of using standard captioning software. Synote also provides a download tool that allows a user to convert narrated PowerPoint files into an XML file, also creating an associated WAV file and slide images. This can then be uploaded to Synote. With such a wide range of ways for users to upload information to Synote, it is easy to see how this technology can be used in a multitude of learning environments. All these features provide a level of flexibility that revolutionizes the way in which multimedia can be used in educational settings, making information even more accessible.
Further innovation and development by the Southampton team continue to push the boundaries in advancing the quality of multimedia in education, allowing what was once cumbersome data to be streamlined within educational institutions.
The Synote homepage can be found here. Once registered, simply go to recordings to view presentations that have been uploaded. The "Synote Guide", provided by Mike Wald, provides a great indroduction to navigating and using Synote.
For more information about Synote, please contact:
Dr. Mike Wald
email: M.Wald@soton.ac.uk
|
