Audio to Text Components

[David Champion 48]

Is there a VCL library available to convert human voice audio into text?

It has to work without access to the internet, the English language preferably.

[programmerdelphi2k 237]

https://blogs.embarcadero.com/this-google-api-easily-adds-powerful-speech-recognition-to-apps/

https://github.com/halilhanbadem/delphi-google-speech-to-text

 

now, if you want just "read the text", you can use the "ISpeechVoice" from MS, just importing it in your IDE, nothing more:

• Component -> Import component... type library sapi

 

Edited December 1, 2022 by programmerdelphi2k

[David Champion 48]

@programmerdelphi2k This doesn't achieve what I'm trying to do which is to transliterate without any cloud services.

 

The client's environment prohibits the use of the internet; there is only a local network.

 

[programmerdelphi2k 237]

ok! if I know some I let you know!

[David Champion 48]

I have found the Windows.Media.SpeechRecognition namespace in Microsoft WinRT.

That may be a way forward without the Speech Recognition needing to connect to Azure.

 

[PeteG 0]

Hi David,

Long time later but did you ever get anywhere with this? I've found https://github.com/ggerganov/whisper.cpp which is a C++ library, could work at getting that doing something with a fair bit of work.

 

Pete

[David Champion 48]

Thanks for the recommendation. The feature that I was suggesting as part of an on going project was not thought to be worthwhile.

So, no, it was canned.

 

Edited May 2 by David Champion

[Rollo62 536]

You can have a look here, from Grijjy, its quite old, but worked well for me under iOS and Android, so I assume Windows is OK too.

[David Champion 48]

@Rollo62 It was more the other way round; limited Speech Recognition.

Logging to text at various intervals what people are saying, so that positions in Audio log can be sparsely described.

Also, the application cannot connect to the internet.

 

 

Edited May 2 by David Champion

[Rollo62 536]

Oh yes, of course.
I had skimmed the title too quickly, normally terms like TTS, TextToSpeech, SAPI, SpeechToText trigger me in the right direction.

 

Maybe this will be helpful
"https://learn.microsoft.com/de-de/windows/apps/develop/speech"
and an older article with Rx1.4.2 sources by Brian Long
"http://www.blong.com/Conferences/DCon2002/Speech/SAPI51/SAPI51.htm"