Request for opinion: less worst my crappy English or text to speak?

I am producing a blended-learning course on machine learning and Julia, and I would like to ask you if you think that for the average student would be less challenging following the slides with the audio recording of my crappy English or the text to speak voice…

Here is an example slide with the text to speak:

And here the same slide with my crappy English:

Note that the second part of the slide is more representative of the rest of the course, as tipically I don’t read the slide content…

Thank you :slight_smile:

1 Like

They are both easy to understand for me (native).

Non-native English speaker (and listener) here. I understand both, but I would definitely prefer the second one. Although computer-generated voices have come a long way, the first version sounds very monotonous to me. So while both might be understandable, I could imagine that the second recording with your own voice would actually be more engaging for the students!


Crappy English is easier for me. Your English is very easy to understand!

Second the crappy english, very understandable.
text-to-speak i do not like.
There is a single thing you should improve: your intonation of “as” is quite funny and can be misunderstood badly :wink:

teacher: “…such as visual perception…”
student: “did he said “arse”?”

1 Like

Another option would be a better AI voice, or even cloning your own voice. See Lyrebird

1 Like

The AI voice is not bad at all, but it feels a little like your’e in your car with the GPS on. I would suggest you use your own voice, because it will be good practice for you to work on your pronunciation. Another thing I would suggest is to simplify the language and go for a proper spoken style (say based on BBC English), e.g. not use words like “hence”, but instead say “Now we”, avoid things like “gonna/wanna” and avoid ambiguous sounds like “they’re” (easily mistaken for “their”). But if you’re out of time to make your own voice-over, the AI is fine. :grinning:

I tried your link, but it is now “” and they have a Win/MacOS only client, and obviously, I am on Linux :wink:
Is it really worth it? Is it better than Google voices ?

Thanks everyone. At the end so I went for my own voice (subtitles are available).

This is the “kick-off” meeting (approx 3h)… I’ll record the bunch of the other slides in the next weeks: SPMLJ 00 KOM 1: Course introduction [Intr to Sc Programming and ML with Julia] - YouTube

By the way, do you know a reliable permanent public installation of Moodle ? I could host it in my university instance, but I want the wider public to be able to access the exercises…