Good evening
I want you to help me in my project
that is Audio Search Engine
Enter keyword want with microvofon and searched the inside of the sounds, and analyze the sound and see whether the word exists or not

Such as: the sun will begin to search files on the network of votes any looking into the audio file is an existing word or not, if they exist shows the file in the Search Results

Like Google and AltaVista, but this depends on the text search

Can anyone help me in this project
What is the language which would deal with them
and how to work?
Thank you

Moderation: Please don't use huge fonts.

Search engines like Google are not going out and searching the internet each time you type in a search word. They read the contents of all of those web sites into their servers and, as they read them, they set up searchable links to them in their database. That way you get very fast searching for any arbitrary terms.

To do the equivalent, and thus make your search engine remotely useful, you would have to scan the Internet for audio files, read and analyze each of them for recognizable words, and create a database based on that. You would not need to store each audio file, just the information in the database. Does this sound practical to you?

my broject is Audio Search Engine
I need to convert input speech to text
and covert the file from speech to text
then compare between them
but I want to help about the language for this project
MATLAB
C#
artificial neural network
I must analys the sound and Sound processing
pleeeeese I must help in this week

Mr.pdq is not like this
I want to search for the keyword within the audio file after treatment

This is called Speech Recognition. It can work very well with a limited vocabulary. For example, if you call your credit card company and the computer asks you to say your account number, or if it asks yes-or-no questions. It's also used for transcribing medical reports. In this case there's a "structure" to the report, and again limited-specialized vocabulary which helpw the computer. (I assume... or hope... the output is validated by a human!) Another thing that's helpful is if the software is "trained" for a particular speaker.

It doesn't work very well with an unlimited vocabulary, or if there is background noise. If you are trying to find lyrics in song files, the background music is going to make "lyric recignition" very difficult! It's difficult for humans, and humans are better at it than computers.

Wikipedia has a list of speech recognition software.

I dont know, but I think you can use whatever language you're comfortable with. The tricky part is going to be the algorithms, and you should be able to apply tha algorithm in any language. You can take a look at the open source programs on the Wikipedia list. And, you should be able to find lots of (very technical) information on speech recognition on the Net an in books.

Incidentally, it usually does not work on a cell phone where network coverage is not very good. Which is extremely annoying...

thank you Mr.DVDdoug & Mr.pawelq

Speech is still very hard for even powerful computers to understand. Each person says the same word slightly different from everyone else. The computer can not compare the sound exactly, it needs to 'know' what the word is. That is Speech Recognition like DVDoug said.

With a song I think it would be impossible with current technology. Singers change the way words sound too much for a computer to understand.

Many people are trying to make this better for a long time. But it is still bad.

It would be better if you allow only a few words to be searched on. Start small.

C# is a very good general purpose language and would be a nice start for a hard project like this. Other languages would work too. Use one that you like and are good with.

I hope you get it working. I would like to see that too!

Goodluck