[Feature Request] Text-to-Speech for decks without audio files

Hi, long time memrise user here, studying Japanese. There used to be some great extensions & scripts that worked with the old memrise. Now that the website has been overhauled, I can’t help but miss my most favorite features from those scripts. The most important one was Text-to-Speech.

Text-to-Speech was a google/machine generated TTS script that allowed the user to always have audio for every learning/review session that was missing audio files in the deck. For instance if a deck had words with a native audio file the native file would still be used. However, if the next word being learned/reviewed did not have an audio file, the script automatically includes a machine generated audio sample of the word/phrase being spoken.

[Userscript] Memrise Audio Provider – by cooljingle

Personal Notes

I think being able to have any kind of audio for a review/learning session can 10x learning because being able to hear the word spoken as well as seeing it was always one of the things that greatly streamlined my learning on this site.

Even if it’s a beta option that works sometimes or a paid feature, I would come back to memrise if they found a way to include just this one feature. I think it can be applied to all of the foreign language learning decks. And it really helps speed up the memorization process for any level of learner.

Cooljingle has made a bunch of great userscripts for memrise. By far I think this was the most downloaded and best feature he created for the memrise platform. Aside from their ‘All Multiple Choice’ or ‘All Tapping Tests’ on github. I tried reaching out to them to see if they were updating/creating another, or if the community could take on their work but they have not responded in years. Hopefully they are doing well, and hopefully a dev sees this and considers these ideas.

Since there’s no way to edit someone else’s memrise deck or to really message a user to collab on a deck the only way to include missing audio to one’s favorite language learning decks is to use a third-party resource; e.g. a script. This is a common occurrence once you get to the higher language levels. There tends to be more audio files for the beginner language levels but once you start getting into intermediate and beyond they tend to taper off dramatically.

Thank you for your time,
and to anyone who might read this.

For learning Japanese I like to make courses with audio sentences too, featuring whatever word or grammar point I want to learn, so I would certainly also welcome any new text-to-speech features that get introduced.
After studying Jay1232’s audio sentences course at Japanese Sentences Audio (No Typing) - by Jay1232 - Memrise , I’ve found that this approach works really well, and the vocabulary sticks with me much better than just memorising words and definitions.

For the time being, I use this chrome extension I found to automatically generate audio. It can handle various languages, including Japanese, and basically does what you were asking for.

For Japanese, however, I find that text-to-speech often doesn’t get things right. For example 大分 (daibu) meaning “a lot” gets pronounced by text to speech as the name of the city “Ooita” which is written with the same kanji, so I have to manually correct it a lot. Another example is that text-to-speech can’t tell whether you want to interpret 一足 as “issoku” or “hitoashi”.