First off I just wanted to say awesome job with LipSurf! I use it primarily for WaniKani and it’s great help when I don’t want to bring my arms out from under the kotatsu
I have a quick bit of feedback, but I know you’re working every day to improve already! When inputting answers in Japanese on WaniKani, it’d be great if the Japanese speech recognition would only use hiragana to avoid situations such as this:
Sometimes it will run through several words with the same pronunciation before arriving at your intended answer, and others (like here) it doesn’t accept it at all. I’m not sure how easy it is to separate the input formats on a certain site from the general application, but it would be a huge improvement.
Thanks again for the awesome app! Looking forward to seeing how it develops in the future
The live text displays with Kanji because it’s guessing the meaning of what you’re saying given the context (not just the phonology) But don’t fear, all the input is normalized to hiragana before checking against Japanese commands (in this case WaniKani answers). Kanji in the live text is useful for other commands, like googling things in Japanese.
Yes, we work with the kanji and go backwards into hiragana, and sometimes the kanji have multiple readings that we need to try.
The specific example you gave me is very interesting! I dug into the issue and found that 迷惑メール is actually incorrect in the open source dictionary we’re using. I’m going to get a higher quality dictionary, and make sure this example in particular is fixed, and upload the new dictionary ASAP!
Thanks very much for bringing this to my attention!
I have a temporary fix for this in v1.6.6 (along with some other unrelated fixes).
The Japanese language Kanji->Kana functionality of LipSurf is being overhauled, and I was hoping to get that done before posting this message here, but that’s still a work in progress. In the meantime, I put a temporary fix in for you.
Thanks for your quick reply and help! There are a couple others that do the same thing as well, but enough to the point that it’d be silly to list them all individually. Glad to hear you guys are already aware and working on a solution, too. I know the Japanese recognition is still in Beta as is, but want you guys to know that it really is great software, and i appreciate your work!
Thanks for the encouraging words. Yes, we’re working on a solution that looks promising enough to fix what you described and even take Japanese out of beta.
