[Userscript] WaniKani Pitch Info

Awesome, thanks!

All the ones you grabbed are only pitch form 2?

Is there a reason you’re including the pitch pattern in the definition instead of letting it be generated?

Ah, no I mean I also have them all as pitch: [0] or [1] or [2] or [n]

"お酒": [0],
"明治": [1],

OJAD uses a pattern in order to draw their diagrams so those were the initial results, but I’ve now used my pattern generator to reverse engineer the pitch numbers like weblios [n].

I just pasted the patterns since it’s easier to visualize them to confirm (and I pasted early before converting :stuck_out_tongue:)

Ah ok sweet, well for sure then submit a pull request if you’d like and I’ll check it out and hopefully get it pushed out to main!

Merged the changes, update should be going out to people now, thanks :slight_smile:

Hey, just wanted to say thank you for making this, it’s helping me a lot with getting the hang of pitch accent :slight_smile:

1 Like

@Invertex When you get a chance, can you add the pitch information for the new vocab?

I’m pretty sure it isn’t necessary to manually add pitch info for new vocab - it gets pulled automatically from an external site.

If the pitch info is not showing, it usually just means that the external site doesn’t recognize the word (this happens sometimes, especially if the word is just a suffix or a combination of words).

Yeah, but if they add it to the script it doesn’t have to do a web search every time.

I added the missing pitches for the existing list of vocab, and added the new vocab. There are about 300 items that don’t have pitch info in the dictionary or can’t found in the dictionary at all.

3 Likes

I was wondering, how reliant is wanikani + this script combo ?

I’m new to pitch accent and I’m going crazy. Most of the time, what I see and what I hear match quite well but some times, even after listening to it ten times the two clearly don’t match for me.

Plausible scenarios:

  • Both wanikani audio and weblio pitch accent always match, i.e my hearing sucks (this is the most plausible scenario)
  • Wanikani audio and weblio pitch accent might differ but both are equally right.
  • Wanikani audio is always right, i.e weblio might suggest unused pattern
  • Weblio is always right, i.e. Wanikani speaker mispronounced/use some dialect specific accent.

Which is it ?

1 Like

There are other pitch accent resources. For instance, have you tried checking the Online Japanese Accent Dictionary to confirm things?

As an aside, the Suzuki-kun tool on that site is cool. It lets you put whole sentences in and see how accent and intonation flow in more realistic scenarios.

As for choosing one of your 4 options… kind of hard to do for things generally. I’d lean toward mishearing and poor recording/audio (clipped, volume issues, etc) before assuming they were wrong or regionalized. But it would depend on specific examples.

1 Like

Thanks, I was looking for a ressource to cross check. I’ll try to come back with a good example.

If you really want to check the WK recordings, you can use a recording program to generate a spectrogram. Here are instructions to do so with the free program Audacity.

It may look like overkill at first, but I used this method when I was first starting out with pitch accent and it helped a lot. I played the vocab from my speakers while recording, then I recorded myself and compared the two spectrograms. This method probably helped the most with the early difficulty of hearing the wrong pitch.

3 Likes

I really like this script, but for some reason it disables autoplaying of audio, at least during lessons.
Any idea on how to fix this?

Edit: It seems that autoplay audio only doesn’t work during the lesson, but the mini quiz at the end and normal reviews have autoplay audio on, so it really isn’t that big of a deal, i can live with that :stuck_out_tongue:

I’m glad I found this, thanks!

@Invertex Lately the J hotkey for playing audio sometimes doesn’t work. I’m not sure that it is caused by this script. But this script occasionally messes with the audio in odd ways, such as duplicating the audio (sometimes making it twice as loud and other times playing twice while overlapping). Those other issues happen when I try to play the audio from tabs other than Reading and gets resolved when you go to the Reading tab. But this issue doesn’t seem to go away by going to that tab.

Has anyone else encountered a problem with this script not working on item pages? (not lessons or reviews, the vocabulary page itself)

Show example

Error on 方, should be pitch pattern 2, not 1.

Edit: Why do I only see the charts during the reviews and not word pages?

Yes, I also have this problem.

It must have detected the reading of 頭高 方(ほう)rather than 尾高 方(かた)Can’t exactly see the problem without the script working on the item page though.

Do hope it can get some attention and be fixed soon, though!

I don’t use WaniKani on pc as much. I mainly do everything straight from my phone because of the ease of reviews and access for study. Is there anything like this for mobile devices?