Arduino Text to Speech Project?? HELP!!

Question

Hi everyone nice and short I wanted to know how I could get the Arduino to convert text in to speech clearly I have searched everywhere........ not looking shields looking to build from components

Jan Cumps · Answer

I've made such a project here at element14: [Christmas Wreath of Things] Internet of Holiday Lights - Bonus kicker: the IoT talks (a spoken tribute to Science Fair kits)

dougw · Answer

https://www.robotshop.com/letsmakerobots/text-speech-with-speakjet SP0256-AL2 Speech IC RoboVoice Text to Speech IC Magnevation SpeakJet

dougw · Answer

Have you investigated the Google AIY Kit? https://aiyprojects.withgoogle.com/ https://aiyprojects.withgoogle.com/voice/ Here is a project where we installed it on a Raspberry Pi... The Spice of Pi - Blog Doug 9

shabaz · Answer

Hi,

The alexa / google home as I understand, does not do text-to-speech, nor speech-to-text. The service remotely does that, and sends the audio to the device sitting in the home, to play out. This is okay, because audio doesn't take up much throughput.

I just wasn't sure if it was deliberate to do the processing locally, or if you were aware those systems don't do it locally.

If you need an Internet connection for any of the features you're planning to implement, then if it meets your needs, it's also possible to do the conversions remotely on Amazon/Google servers. For more info, there should be many online blogs containing this detail, it is a popular thing.

shabaz · Answer

Hi, If you want local processing, check out things like ReSpeaker. For remote processing, something like Matrix Voice. They are both open source (or at least have a lot of example code) and should give you ideas on how to implement. They don't use Arduino, but by examining them, you may decide if your chosen Arduino device is sufficient, or if you need something else. You may need another processor alongside (possibly running Linux), or you could replace the Arduino entirely with (say) a Pi. Arduino alone could be an uphill struggle.

shabaz · Answer

It's not a server as such, it's a bunch of functionality that resides on many servers in a cloud. Amazon, Google and others have invested millions developing it. You could try to develop such functionality from scratch but it would be very primitive unless you have such resources. Most people I think just use those services (there are APIs to connect to the services). Some are chargeable. To learn the APIs, they would be different for Amazon and Google, so their websites need to be examined for documentation, With Amazon, it's called Amazon Web Services, and Google's is just called Google Cloud, and most likely Google App Engine. The particular service of interest is generally called Natural Language Processing, each of the cloud providers have their own branding for their services (Amazon's is called Comprehend). It doesn't care if the input comes in via speech or text, it's all the same to it. I've not used that one - I've only briefly experimented with the IBM version. If you're interested in that, there are free online video courses.

You can see my brief experimentation here: Pi3 IBM IoT Learn Kit - Review

The example there uses natural language processing based on text input (but it could equally well have been voice input, the concept is the same.

The link to the online course is there too.

Arduino Text to Speech Project?? HELP!!

Top Replies