Looking for coding advice

[Dritanis]

Hey, got a quick question for all you guys. I want to start doing independent development on some ideas that I have, basically some simple medical devices. However, I know pretty much nothing about basic coding, and how this could be done. Quick diagram:

input sounds -> check against database of words (imagine a tiny version of that Ford Sync) -> output

If anyone has any ideas or would be willing to throw in their two cents, let me know!

[Hachima]

it depends on the device you are going to use and what API it supports, if any.

[Phelps McManus]

You know nothing about basic coding? This could be an undertaking even for professionals.

If you are lucky, someone has already written some public-source voice recognition software. See if you can find something and determine what language it uses.

I would break it up into three portions.

First, and definitely the most difficult, you need to process the incoming sound and spit out words.

Second, create a dictionary and corresponding function to take a key and spit out a definition. This should be the easiest as there are dozens of ways to accomplish this feat.

Finally, take the definition and output it to whatever device you intend (file, speakers, etc).

If nothing else, pet projects like these are the best way to learn how to program.

[Fog]

Originally Posted by Dritanis

Hey, got a quick question for all you guys. I want to start doing independent development on some ideas that I have, basically some simple medical devices. However, I know pretty much nothing about basic coding, and how this could be done. Quick diagram:

input sounds -> check against database of words (imagine a tiny version of that Ford Sync) -> output

If anyone has any ideas or would be willing to throw in their two cents, let me know!

Speech recognition is a tough problem, but if you have only a small database of possible words to recognize there are some pretty mature solutions available. I haven't personally done anything in this regard myself, but I have a friend who implemented CMU's Sphinx libraries for an application like this and found it satisfactory for the purpose.

I'm sure you're aware that it'll be a lot of work if you don't have prior programming experience.

[Jenison]

1. buy api that already does speech recon
2. take api's output and process it how you want to
3. make a nice UI

KISS - always.

if you haven't seen Dragon Naturally Speaking in action its pretty bad ass.