Question

im using an open source program called Festival that generates text to speech, and in ubuntu i call its method text2wave that converts text into a wave file. I am looking into converting short paragraphs to wave files, but the problem is that each wave file ends up being approximately 1.2MB in size. The wave file is recorded at 16khz, and while recording at 8khz halves the size of the wave file (sacrificing sound quality), the wave file is just simply too large. These paragraphs have to be served many times to the clients and our server can not support that much bandwidth.

Does anyone have any idea how to bypass this problem? Is it possible to compress the file on the server, and decompress it in my flash script (on the client side) and play the wave? How much size would i be saving? Thanks for any help.

Was it helpful?

Solution

Flash supports the mp3 format. You can encode your wave file on the server side with LAME, and let Flash do the decoding on the client side. I'm assuming you can use arbitrary CGIs on your server. How much size you'll save this way depends on your encoding settings. you should experiment.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top