views:

110

answers:

5

Does anyone know of any well-supported/proven methods for converting a Microsoft Word document to an MP3 or WAV format such that hearing-impaired folks could "listen" to documents that I have stored in my web-based document management system?

I already have the interface built such that someone can use the telephone to get the list of documents available, with the dates and titles "read" to them over the phone, but now I would like the ability to let someone actually listen to the contents of Word files stored in the system.

Ideally a .NET library or utility that would let me convert the DOC -> MP3 after each upload would be best, but one that "read" the file on demand would be okay too.

+1  A: 

You could try out Panopreter it has a free and professional version. I have used it before with pretty good success.

I'm not sure if the free version can save directly to mp3 but you could always use another program to record the output to an mp3.

gruntled
+1  A: 

I'd possibly use mintext command line version to first convert the .doc file into text and then store this text in your database (upon upload perhaps?)

Then use any Text to speech library to convert and save this file, with perhaps a reference Id that links this file in the database.

Darknight
+2  A: 

Getting the text out of Word is the easy part. There are some tts libraries available but the goods ones tend to be expensive. Acapela (http://www.acapela-group.com) is probably the best I've found and the API supports output to sound files. I'm not sure of MP3 is supported, but converting sound formats is trivial.

Best of luck.

EricBudd
+4  A: 

You could use the System.Speech.Recognition for configuring your TTS engine and the System.Speech.Synthesis namespace for recording/saving (saves to .WAV). In System.Speech.AudioFormat you can configure the output parameters of the WAV file. For Word, you can just use COM Interop to grab the file, open it and get it's text in any kind of chunk you need and feed that into the managed SAPI. I'd probably seperate things into Paragraphs/Runs, Tables, References and Other.

If you're looking for a far more robust technology, one of the more well-known TTS engines is NaturallySpeaking's Server SDK - they've recently added an "Audio Streaming" feature that allows for on-the-fly TTS, so there would be no need for pre-recording to a permanent audio file (but if you want, it can also pre-record). You can also check out their AudioMining SDK, which allows for full indexing and searchability of spoken text within audio files. Their SDKs are all COM though, not .NET.

Otaku
+2  A: 

If your Word document is stored in the new OpenXML format introduced with Office 2007 you might want to have a look at the Daisy toolchain.

There is a free converter available on SourceForge to create a Daisy file from your Word document. The Daisy format is understood by screen readers and there are also converters available to convert a Daisy file into an audio file.

For details see:

0xA3