Say something with the IMP

hello i need my imp to say some words (sounds…)

i saw this
http://www.hobbytronics.co.uk/emic2-text-to-speech?keyword=sound
Do i really need something link this?
Is there any more options?

I need it to say simple things like numbers and letters or play some audio files from the internet.

Regards
António

I just started working on a walkie talkie project with the imp… so far just using the built in ADC & DAC on the imp for audio (8 kHz).

Instead of the approach you mentioned – you could also consider generating the audio at server side (some sort of text to speech service), and just sending the audio data down to the imp.

We will soon have the Lala audio reference design available at MakeDeck. It has the kind of functionality that you are looking for.

We built an overly simple TTS example for Maker Faire this year.

You can find the code here.

This example uses a SparkFun Mono Audio Amp Breakout.

NOTE: The sample code will crash if you try to send more than a couple seconds of audio data to the imp. The proper way to do this would be to chunk the data, and stream it down to the imp - or use a SPI flash to store the data.

Also worth mentioning, there’s an article in the Dev Center about working with Audio:

http://electricimp.com/docs/resources/sampler_ffd/

@beardedinventor

the readspeaker.com cloud service is excellent and easy to implement.

but it has a cost per word of 0.01 usd

@axisdev
that seems the best solution for me but i need to first create the sound data and store it then feed it to the imp.
Is there a way to just convert some mp3 to raw data that the imp can understand?

I’ve actually considered making something like this as one of the MakeDeck cloud services. Could you give more specifics on exactly what you would like to do?

I am starting an app to manage the locations of laboratory samples.

The room for the sample deposit will have a wireless barcode scanner.
On the receiving dongle i will adapt an imp to connect it to my app. avoiding a regular desktop pc. i think i can do this…

all places to deposit the samples will be marked with barcodes
one for check in command, one for check out command ,and other with the code of the place itself.
This way the user only needs the scanner and not a tablet with a page for data input.
To deposit a sample he scans the input code, the sample code and the place to deposit the sample code.
To remove its similar.

So the user is blind.He does not know if the system is processing the data or need to repeat the input.

For this i would like to implement sound warning. Audio messages would be nice.

This is roughly my idea for this project.

In the back end i will start an app that receives data from the imp and can tell the user if all ok or not.

There are other details of course but not relevant for this