Sound file storing in Database & accessing these files

Question

Hi Friends
i have working on the project of Sound Recognization system
the requirement is that sound should be passed through mike input device & that sound should be stored in to database along with the user & password
for recognization & identification of user
But problem in storing the sound files in the database so how can i store the files into the database so please guide me.

Answer 1

brad103 wrote:
In my particular application, I process the byte array from the TargetDataLine. Then I save this to a db using new ByteArrayInputStream(soundData), soundData.length); where soundData is my byte[].
Am I correct that this will save the data in the relevant format (eg PCM_SIGNED, 8k, 16bit, etc), but the header information is not included? That's correct. Essentially, in JavaSound, that data has an associated "Format" object that was part of the line the data came from, would need to be given to any line the data was going to...and that format object is essentially a container for all of the header data...
And if so, then if I control the saving and retrieval then I can omit that step anyway as I know the format?You can omit the AudioSystem.write step and store the byte[] directly into the database...as long as you're capable of, as above, recreating the Format object that was originally associated with the data.
And my next question (digressing slightly, sorry umeshS), is that when using the tritonus gsm library and converting from above format to gsm, should I be using AudioSystem.write to write to an outputstream prior to calling AudioSystem.getAudioInputStream(targetEncoding, originalAIS) to convert encodings?AudioSystem.write adds the file header to the stream, and that's all. If you're using a "stream", you have to have the Format object in order to create an AudioInputStream, so there's no need to add the file header (and the file header will screw things up if you're not using the File or URL versions of getAudioInputStream)...
So, as a giant summary...
AudioSystem.write adds a file header, which will include format information. This should only be used if you're planning on using AudioSystem.getAudioInputStream(File f) to read the file the next time.
Otherwise, just save the data and the Format object associated with the data, so you can read it all back in later by building an AudioInputStream around your favorite IO stream.
And in retrospect, after explaining all of that...umeshS, the problem with the way you're doing it is that when you copy the file into the database, you're copying the file header. And then, because you're not using the "File" version of AudioSystem.getAudioInputStream, the system isn't handling the file header being there correctly.
You need to copy things into the database by creating an audioinputstream of your file, and then saving out the raw byte[] data. You can just call "read" on the ais a bunch of times and throw the results into your stream. Then, you can play it by pulling the data out of the database via a stream, and playing that stream by throwing it back into an AudioInputStream.