Feature Request - Rocket Languages Rocket Record

Hi,

I teach a language course. While all the games and interactive learning experiences help my students. 

Is there any possible way you could have one feature similar to Rocket Languages Rocket Recorder? 

Its basically the student listening to the audio first, then recording themselves second to see how well or how different they sound [to a native]. And they are scored 0 -100 on how good, close or bad they are. This will help them in knowing how well they've improved in pronunciation. 

You can find a live example here: https://www.rocketlanguages.com/spanish/rocket-spanish-premium

The content type, Speak the Words Set, is the closest to this but it'll be even better if I could play in a native-speaking audio file rather than the default text as the "solution". 

Thanks,

Shine

Summary: 
Similar feature to Rocket Languages Rocket Record
1
0
Supporter votes Members of the Supporter Network can vote for feature requests. When the supporter network has generated sufficient funding for the top voted feature request it will normally be implemented and released. More about the H5P Supporter Network
BV52's picture

Hi smaoate,

Thank you for contributing your ideas on how to make H5P better! With the H5P supporter program the H5P community can now vote for and fund the top voted H5P features. Also there are developers in the community who every now and then work on a feature they find interesting or useful. You can also help by developing or help (crowd) fund the development of this feature.

-BV

Hey smaoate! I also was quite impressed by the rocket language feature of student voice recording and the match to the native voice. 

Did you find any solution to the problem? I would love to integrate this feature (or sth similar) in my German course.

Do you have right now a feature of recording your student voices and match to the text (or any feedback)?
Would love to hear about it.

Kind regards, Thomas.

otacke's picture

Hi all!

This would actually not be complex to implement, it would just take some time.

  1. There already is a pull request to add optional introductory media to Speak the Words as we have it in "Multiple Choice" and some other content types already. With the latest (not yet published changes) to H5P.Question, Audio will be an option, too. That gives us the audio feature. Could be implemented separately as well, of course.
  2. Speak the Words already receives the written text of what the SpeechRecognition engine understood - you can see that when you answer incorrectly and have a look at the solution. What's left would be to compute the degree of matching between what was expected and what was understood - either based on characters (as on that site) or based on words (as in Dictation).

Cheers,
Oliver

serettig's picture

I've once used the browser's speech-to-text engine for a (school-internal) vocabulary studying site. The problem when using the engine with learners is that it's too tolerant for this context. You can have very bad phonetic pronunciations and it will still recognize the word as correct, as long as the leaner hasn't made a phonemic mistake. So getting a meaningful score that really checks if the learner can imitate the natural sound of a language might be much more difficult than just using the speech-2-text engine.

otacke's picture

@Sebastian That's fully correct. However, since the original posters both liked the solution shown on the other site, I didn't see the need to point that out.