�ݺ�ߣ

Editor's Notes

• #2: Thanks to everyone for coming. My name is …..blah blah My educational background is MSc in Information Systems … In this work, we propose a non-visual interface to monitor stock market data , both textual and numerical – and this is what we call an RSS-feed auditory aggregator.
• #3: I’m going to discuss a little on the technologies that are evolved in the system. Firstly, we used one the most commonly used formats for information delivery, the XML and more specifically the RSS. The RSS is widely used to transmit frequently updated web content to feed readers or news aggregators. In the early days of the Internet, there was little need for different websites to communicate with each other and share data. In the new "participatory web", however, sharing data between sites has become an essential capability. To share its data with other sites, a website must be able to generate output in machine-readable formats such as XML (Atom, RSS, etc) and JSON. When a site's data is available in one of these formats, another website can use it to integrate a portion of that site's functionality into itself, linking the two together. When this design pattern is implemented, it ultimately leads to data that is both easier to find and more thoroughly categorized. Secondly, we aim to extend the concept of RSS-based information delivery and aggregation using sonification. Instead of delivering information in textual representation we propose a sonification framework to represent stock data. And lastly, we gave particular emphasis to the information representation concurrency, which is mainly achieved using sound source spatialization techniques and different timbre characteristics.
• #4: RSS stands for Really Simple Syndication . It has quietly become a dominant format for distributing news headlines on the Web. It is a lightweight XML format designed for sharing headlines and other Web content. Think of it as a distributable "What's New" for your site. RSS defines an XML grammar (a set of HTML-like tags) for sharing news. Each RSS text file contains both static information about your site, plus dynamic information about your new stories, all surrounded by matching start and end tags. Each RSS channel can contain up to 15 items and is easily parsed. Say for instance that you want to monitor the latest news of some stocks belonging to your portfolio . Instead of checking the news sites every day for fresh news you now can make use of RSS, and it will automatically fetch the latest related news. Another great thing about RSS feeds is that as your interests change and the sites you follow change you can remove or add subscriptions to your feed. And also RSS is a secure channel that can’t be spammed.
• #5: Text-to-Speech technology, converts normal language text into speech, it synthesises text into speech. Speech synthesis has long been a vital assistive technology tool and its application in this area is significant and widespread. The technology has improved significantly in recent times, and although it does not yet duplicate the quality of recorded human speech, it is still a good option for creating messages from text that cannot be predicted, such as translating web pages for blind users. Sonification is the use of non-speech audio to convey information. Several different techniques for rendering auditory data representations can be categorized as … Many different components can be altered to change the user's perception of the sound. Often, an increase or decrease in some level in this information is indicated by an increase or decrease in pitch, amplitude or tempo, but could also be indicated by varying other less commonly used components like timbre and register. We have suggested the use of text-to-speech technology and non-speech audio cues, called earcons as one way to improve the capacity of the web content transmission through parallelism and as a way of communication for general usage applications such as in-vehicle communication or visually impaired users. Binaural rendering systems is to evoke the illusion of one or more sound sources positioned around the listener using stereo headphones. The positions of the sound sources can preferably be modified in terms of the perceived azimuth, elevation, and distance. Binaural rendering has benefits in the field of research, simulation, and entertainment. Especially in the field of entertainment, the virtual auditory scene should sound very compelling and “real.” In order to achieve such a realistic percept, several aspects have to be taken into account, such as the change in sound source positions with respect to head movement, room acoustic properties such as early reflections and late reverberation, and using system personalization to match the anthropometric properties of the individual user.
• #6: Why stock data? It incorporates multiple parallel and real-time information transmissions, allowing the evaluation of the overall functionality of the proposed system. We can use a wide range of investment options, some of them are stock quotes of a particular market group, portfolio stocks, stock indices and bonds. Earcons are brief musical melodies consisting of a few notes. They are abstract so their meaning must always be learned. Think of all the choices we have to position the audio feeds. What combinations would sound the best? There are so many choices of instruments and combinations of sounds. We focused on orchestration in order to avoid any information losses, due to concurrent RSS-feeds transmissions.
• #7: The proposed sonification-enabled RSS-feed aggregator is subscribed to N different RSS-feeds. Depending on their type – textual or numerical, the parsed information is organized into M information streams (where M≤N, a value that exclusively depends on the type of the received data). Currently, this information categorization is performed manually by the user during the initial subscription to a specific feed. The information is transmitted to the Earcon Design Engine, which is responsible for producing the appropriate earcons in real-time, taking into account the type of each information category. The binaural processing module is used to spatialize the incoming audio messages. What we achieved is to provide a robust and efficient (in terms of acoustic perception) mean of concurrency during sonification. Finally, the derived spatialized earcon signals are forwarded to the Auditory Display Synthesis module, which mixes the corresponding binaural signals and reprodes the complete auditory display.
• #8: Let’s assume that our portfolio consists of two stocks , the Exxon Mobil Corporation (XOM) and the Marathon Oil Corporation (MRO), both belonging to the oil companies group. We want to monitor both the daily stock data values and the latest news about these companies. Snapshots of the numerical data feeds are shown on the left, while a snapshot of the latest news feeds about these companies are on the right. We chose to monitor the % Change of a stock quote and we will see how to do this in the next slides.
• #9: Let’s see how to setup the speech parameters. We have selected 2 informational data feeds that we want to concurrently transmit and reproduce them in the auditory display. This is a screenshot of the speech parameterization form that we built. Firstly, we set the narrator parameters, like select the voice, the volume and the voice speed. Then we set each source localization parameters, that is, the azimuth and the vertical position. (We can personalize much more the resulting spatial audio clip by assigning our head diameter). We can also set the absolute start time of this audio clip. In the text to speak box the text that is set, was parsed and trimmed from the selected web feed.
• #10: Our next step, is to setup the sonification parameters for the other 2 numerical data feeds. We assigned a different timbre (musical instrument) for each stock quote, and we used the dynamics to refer to the volatility of each stock quote. As the stock price rise, we increase the volume and as it falls we decrease the volume. We choose a different pitch or register , depending on our instrument selection. In this work we have used a single pitch on the earcon construction. Finally, we have to find a way to map the % stock value change to musical parameters that the end user can easily perceive. For this , we have used different note values and note numbers in the measure, according to a numerical scale that we built. ( Timbre is the quality of a musical note or sound or tone that distinguishes different types of sound production, such as voices and musical instruments. Register is the relative "height" or range of a note, set of pitches or pitch classes, melody, part, instrument or group of instruments Dynamics usually refers to the volume of a sound or note Tempo is the speed or pace of a given piece (beats per minute). It establishes the musical meter )
• #11: This is the earcon parameterization screen where we can see an example of the information that we described in the previous slide.
• #12: On this slide we are going to set up the binaural parameters of the 4 stocks that we have used in our example. We tried to choose instruments from different families (woodwinds, brass, strings), basson which belongs to the woodwinds instruments and piano which belongs either to the percussion or strings instruments (there is some debate here… ) or is used for solo performances. Similar news and data are better to be positioned in the same direction (proximity), so that objects close to each other are grouped together. Also, the use of male and female voices alternatively results in a better perception. On the right we can see the visual diagram of the aforementioned example.
• #13: We must understand what an instrument can and cannot do. Ranges (Middle C). Piccolo is a bad choice for the note D4. It can play well in higher register. The same is true for contrabassoon. It can play well in lower register. Besides they belong to the same family. The selection of the same 2 female voices is a bad choice , and their horizontal position is leads to further confusion to the user.
• #15: Users were given a short training period time and then were presented with sounds and they had to indicate how the system was set up. Results showed that even with small amounts of training , users could get good perception rates.
• #16: Let me leave you with these closing thoughts
• #17: That completes my presentation, thank you for listening. I'd be glad to try and answer any questions.