Audio Formats and speech recognition

You all probably recognize by now, audio formats of wav and MP3. If you are a MAC user, we add to that, the MAC audio format of aiff. The WMA format may be less familiar. What about WMA you might ask. While Microsoft does support many audio formats with various levels of support, it does like its own format, WMA the best. You can learn about this format and sample some of the results of using this format at http://www.microsoft.com/windows/windows... This is Microsoft?s site, so you can be sure that this site only has flattering things to say about this format. You can see someone, LitexMedia, Inc., other than Microsoft?s comparisons at http://www.litexmedia.com/article/compar... What you will find is that indeed WMA is a very fine format, but may not be what you want in all instances. Tucows, the wonderful shareware site, http://www.tucows.com/mmedia/audiocon95.... has many utilities that allow you to convert between different audio formats.

Finding audio sounds How often have you wanted to find a specific sound such as a battle or a dog barking? Which search engine do you use? Do you try http://www.google.com with a wav keyword? Well here is a search engine specifically dedicated to helping you find sound files using key words: http://www.comparisonics.com/FindSounds....

Types of Speech Recognition There are several types of speech recognition. Here are some: ? Speaker-dependent speech recognition implies that an application must be customized or trained to understand a specific speaker?s voice. If you have the latest release of Microsoft Office, you will find that it comes with speech recognition of the type that works best if you train it. ? Constrained-vocabulary speech recognition means that an application recognizes only a fixed, relatively small set of words and accepts them in only a highly constrained context. There is a site, http://studio.tellme.com/ where you can get not only a tutorial on developing constrained-vocabulary applications using voice XML, an elaboration of HTML, but you can also set up an account, use their speech engine, and store your specific application. ? Discrete-speech recognition implies that a speaker must carefully ? enunciate ? each ? word ? separately from the words before and after it, pausing briefly between words. ? Pretty-good-quality speech recognition allows an application to make a few natural mistakes, such as thinking you are saying you want to ?recognize speech? when you are actually saying you want to ?wreck a nice beach.?

Go To Page: 1

Articles in this Topic    Discussions in this Topic