Audio Generation not authentic

Has anyone else found the audio generation for articles and such to not be very authentic of a native speaker reading it? It seems like a rapid monotone robot, at least for Dutch. I was wondering if there are any other audio generating websites/tools to create an mp3 based off an article or something than anybody uses. For example, when talking with ChatGPT 4, it sounds like an actual native speaker of Dutch with different inflections and pauses etc.

1 Like

I have not found an easy and free way. I like the Microsoft Edge voices. Occasionally I’ll open the “print lesson” view from LingQ, set up Audacity to record the audio “channel” and export as an mp3 that I upload.

I don’t really bother anymore…I might do the same thing (have Edge browser read it for me), but just not record it as that part is a hassle.

1 Like

If you code at all, and if you upload your own content into LingQ, you might try the Chat GPT API. It creates quite realistic English and French narration. It isn’t free, but is pretty inexpensive. If you are getting good sounding Dutch from ChatGPT the API for generating narration is a good bet.

I am really sensitive to robotic sounding voices so now I use the API to narrate texts I want to listen to, and download as an mp3. Then upload the mp3 along with my content into LingQ.

1 Like