I made a free site with Japanese folk tales you can read from day one (furigana, English, audio) — feedback welcome

Japanese Language Resources
21

You were right to report it, and it caused no headache at all. It was my mistake, not yours.

When I added the dictionary links, I made tapping a word open Jisho — and I did not think about what that took away. You could no longer tap a word to hear it again. That is the main reason the audio is there, so I broke the important thing in order to add the smaller one. I am sorry.

It is fixed now, on all seven stories:

  • Tap a word and it plays again from that word. Tap it as many times as you like. Nothing opens, nothing jumps, and the reading does not run away from you.
  • The dictionary link now appears under the sentence instead — Look up 「大きな」. You only go to Jisho if you choose to tap that.

Tapping the empty part of a line still jumps to the start of that sentence, as before.

Please try it and tell me if it still gets in your way. And please do not hold back next time either — you have now found two things I could not see myself.

23

Thank you so much!!! I am enjoying the first one, したきりすずめ, very much! Also, thank you for making the furigana easy to read. For me (at 70 years old) the furigana is often very hard to read.

24

Thank you so much for saying that — that specific detail really helps. I made the furigana a bit larger and higher-contrast than what you usually see, but I never had anyone actually test it from your angle, so hearing that it works for you is genuinely useful, not just a nice comment.

Enjoy the rest of したきりすずめ, and please keep telling me if anything (good or bad) stands out as you go.

25

I’m very pleased to hear that my comment helped you!

I started wanting to study Japanese five years ago. As I looked into it I found that there were two avenues one could take: 1) Learn how to read (with kana and kanji), and 2) Conversational Japanese. I decided to take the #1 route because I remembered as a child how I would read and re-read books and children’s magazines. I enjoyed doing that a lot and so I wanted to do that with Japanese.

Fortunately I found Tofugu where I learned hiragana and katakana, and WaniKani. But I found that the kanji learning is so steep. I then bought some Japanese children’s books and found that even they are so full of challenging vocabulary and kanji that they are impossible for me to read.

Yours are perfect for me. I don’t need to listen to the reading. I am just able to read the sentences (glancing at the English as needed), and the pictures provide a lot of context for the story. So, thank you very much!

Last night in bed I thought about reading a story to help me go to sleep. But I didn’t want to have to get out my laptop. I looked on your story to see if there is a download button or something and didn’t find one. What I would like to do is to print out the stories so I can read them without the computer.

That said, though, I would really like it if you publish a book(s) with the stories. I would definitely purchase and read them!

26

Thank you — and good news: that is now possible.

Anyone who tips the project gets a link by email to a page with every story as its own PDF — illustrated, with furigana, and now with romaji under each line as well, so you can read it without hunting for the reading. They are made to be printed and read away from a screen, which is exactly what you described.

You can find it here: Support this project | Nihongo Work

The site itself stays free, and always will. The PDFs are just my way of saying thank you to the people who help keep it going.

A properly published book is not something I have decided on yet — but you are the first person to ask for one, and that is how these things start.

27

Fantastic! Thank you! I will happily give you a tip.

I have a concern, though. I hesitate to complain; you are so generous. But here it is: I really really don’t like romaji. I think of it as a crutch for those who will not invest their energy into learning hiragana and katakana. While that is a bit challenging — I know from personal experience — if someone wants to study Japanese it is something they must do. And if the romaji is absent, they must look up the syllable for the character they don’t recognize. Otherwise, why would they?

Tofugu (WK’s parent) offers a free and excellent course for both hiragana and katakana. Here it is for hiragana: Learn Hiragana: Tofugu's Ultimate Guide and here it is for katakana: Learn Katakana: The Ultimate Guide . And they have a fun interactive quiz for them: Practice Hiragana & Katakana

Let me explain further, please. We know that hiragana is an approximation for Japanese. It provides a very helpful way to write Japanese that would otherwise only be written using kanji. But it isn’t perfect. Romaji is an approximation for an approximation. And there is really no need for it except for people who only want to learn how to speak very simple words, maybe for a one-time visit to Japan. Such as こんにちは。

I just gave you a tip on your website and downloaded a PDF. The romaji is in between the lines of the kana and English, making it more difficult for me to ignore it. It’s like watching a movie with subtitles that are so difficult to ignore. I hope you will remove the romaji please.

If you wish to offer romaji please also provide PDFs without it!

Many thanks!

エマ

28

Thank you for the tip — and thank you more for saying this plainly. You wrote that you hesitate to complain: please don’t. Both times you have written to me you have told me something specific, and that is worth more than a compliment.

On romaji, your argument is a good one and I cannot dismiss it. If the reading is always sitting right there, most people will never be pushed to learn the kana. That is true.

The reason it is there at all is that this site is not only read in English. A large share of the people using it are Vietnamese, Nepali, Indonesian and Thai, and many are at day one, unable to read a single character. For them romaji is not a crutch — it is the only way in at all. So I have not removed it.

But you were right that it should not be forced on you, and you gave me the answer yourself. There are now two PDFs for every story, with romaji and without. The no-romaji version is on your storybook page as of today.

While I was in there I also put up a first attempt at reading the stories on the screen — one sentence at a time, with the recording attached and every word tappable for its meaning. Romaji is off by default there, with a switch for anyone who wants it. It went up today and is still rough; I would rather put it in your hands early and keep fixing it than polish it alone for a month. If anything about it is awkward, please tell me.

Thank you. That page is better today because you told me it was in your way.

— Sazanami

29

A few things changed this week, all of them because people here said something.

・Someone told me the romaji was getting in the way of learning kana. That was a fair complaint, so every story now has a second PDF with no romaji at all. The original stays for people who cannot read kana yet.
・There is now a way to read the stories on screen, one sentence at a time, with the recording attached and every word tappable for its meaning.
・Every line now has romaji as an option rather than a fixture — off by default, with a switch.

Still free, still no account. And still the same request: if something is wrong or awkward, tell me. Three separate people in this thread have now changed the site, and that is worth more to me than a download count.

30

Good work, and thank you for putting so much effort into this. Would it be possible to use as much kanji as in normal writing with furigana instead of mostly hiragana? That way learners of all levels can benefit from reading these stories. Yes, the reading speed seems way too slow. I used 2x speed on YouTube and it sounds about right. The sentences are short enough that I don’t think you have to read too slowly, even for beginners.

31

Thank you for asking this. I’m not experienced or fluent enough to know that more kanji would be more “normal Japanese writing” but it did occur to me that there is a lot of hiragana in the stories, so much that it looked odd to me. I agree with you: although I may get more than I can digest at this stage, having the text be more “normal” (i.e., with more kanji and furigana) makes sense to me.

32

@nihonzuki @EmmaJGray

Both of you are right, and I went and measured it rather than argue.

On the speed. The stories run at 2.1 to 3.5 characters per second. Natural Japanese narration is 6 to 8; a news broadcast is about 7. So the slowest one — Momotaro, which is the first story most people open — is running at under a third of normal speed. Your 2x works out to 5.6, which lands exactly at the bottom of the natural range. You were not being harsh. You were being accurate.

I should say plainly how that happened, because it was not a considered decision. I looked at a channel that reads Japanese folk tales very successfully, and I copied its pacing. It took me until this morning to notice that channel is made for people falling asleep. Slow is correct there. It is not correct for someone actively trying to catch words, which is what this is for. I copied the surface of something that worked without checking whether it was doing the same job.

On the kanji. 11.9% of the text is kanji. Ordinary Japanese writing runs 25 to 35%. So it sits at about a third of normal density — and it is not even consistent: one story is at 6%, another at 23%. There was no rule, it just drifted that way. Since there is furigana on every kanji, raising it costs a beginner nothing, and it gives back the word boundaries that all-hiragana quietly takes away. That is a real loss and I had not weighed it.

Both fixes mean regenerating the audio and re-rendering the videos, so this is not a same-day change. But it is the right one, and I would rather do it now at eight stories than later at thirty.

Thank you — genuinely. This is the second time this thread has changed the material, and neither change started with me.

33

さざなみさん、

Wow, I am very excited about your stories and getting to know you a little. On your website I read about your career and some of the job interviewing tips that you provide there. You are a remarkable person.

I completely understand that making these changes is a larger amount of work for you.

I’m also excited by the prospects of their being more stories, maybe even thirty or more. I downloaded the PDFs without romaji and am in bed now reading them on my laptop!

ありがとうございます、

エマ

34

Emma — thank you. “In bed with the laptop” is precisely the thing those
PDFs were made for, so that landed well. I will take the compliment, but
in fairness the site is nine days old. Most of what looks deliberate is
just recent.

Now, the kanji is done. It went out this morning.

The text now runs at 23.8% kanji, up from 11.9%. Ordinary Japanese
writing sits at 25 to 35%, so it is in range rather than at a third of
it. Furigana is on every kanji, as before. The PDFs on the support page
have been rebuilt too — Emma, the ones you downloaded yesterday are the
old version, so it is worth grabbing them again.

One thing worth reporting, since it is the sort of risk this change
creates: converting to kanji is not safe by default. 「つかれて」 came
out as 着かれて instead of 疲れて, and 「ないて」 as 無いて instead of
泣いて — the analyser guessed the wrong word, and in plain hiragana
there was nothing to guess wrong. I caught fifteen readings that no
longer matched the audio and fixed them, but it is a good reminder that
adding kanji is a real edit, not a cosmetic one.

The speed is not fixed yet, and I want to be straight about why. It
needs the audio re-recorded rather than re-timed, and instead of
regenerating it I am going to have the stories read by an actual person.
That may take a little time to arrange, but I am not going to let it
sit.

Thank you both. Two days, two things fixed that I could not see myself.

35

Thank you for updating the kanji — I think. I’m comprehending now how you’re able to make all this (except for the recordings) so quickly. As you wrote earlier, you’re building it using AI so making the changes and performing the analyses is almost instant. I use ChatGPT for many things and helping me edit my ばめん writings is like this.

I’m concerned, though, that AI-introduced errors (slop) such as 着かれて instead of 疲れて would be impossible for me to catch and would generate lots of confusion. Maybe I should have been more careful what I wished for especially because now you’ll need to read it all and manually make corrections.

Earlier in this thread someone wrote to you about how challenging it is for us to read when words lack spaces between them. You responded, “Not being able to see where one word ends is the thing that makes early Japanese so much harder than it needs to be.” And then, you wrote about how the audio is linked in a way that discriminates between the words. I don’t use or see a need for the audio so that doesn’t help me.

The problem I had last night was exactly what the other commenter was describing. Trying to parse the words in the kana. I resolved to print out a couple of stories and pencil in a mark that adds that separation.

I had this same problem with ChatGPT: when I started using it all of the words were squished together. I told it not to do that, to have a space between each word, and now it does so for me. Maybe your AI can do that too.

36

Both of the things you raised are now live.

On the risk you flagged. You were right that it existed. Converting to kanji did introduce errors: 「つかれて」became 着かれて instead of 疲れて, 「ないて」became 無いて instead of 泣いて. In plain hiragana neither was possible — the analyser only gets a chance to guess wrong once you give it kanji to choose from.

But please do not feel you asked for something costly. I did read the whole thing, before and after, side by side. I am a native speaker and checking text is close to what I do for a living, so it goes quickly — and eight stories is nothing. Doing this at thirty stories would have been the painful version. You asked at the cheapest possible moment.

I also ran a check rather than trusting my eyes alone: for all 256 lines, generate the reading of the new text and compare it to the reading of the old. Any line where the two differ is a line where something moved. Fifteen came back — 宝物 read as ほうもつ instead of たからもの, 私 as わたくし, 都 as と. It takes about a second, and I will run it every time from now on.

And it cut both ways. One of those fifteen was a mistake that had been sitting there before I touched anything: 守る人 was being read まもるにん, which is simply wrong. Adding kanji is what made it visible. So the change did not only carry risk — it also cleaned something up.

On the spaces. You should not have to print anything and pencil it in. The site already knew where the word boundaries were — that is how tapping a word plays that word. It just was not showing them. Now it can.

There is a third PDF on your storybook page, “word spacing”. The on-screen reader has a switch for it too, off by default.

昔々、 おじいさんと おばあさんが いました。

The kanji helps with the same problem from the other side — 桃太郎 and 二人 break up the line visually in a way that ももたろう and ふたり did not.

Three changes in three days, and all three are yours: romaji made optional, the kanji brought up to normal, and now the spacing. Almost everyone who visits says nothing. The few who say something usually say it is nice. You have told me three times exactly what was wrong, and all three times you were right and I had not seen it.

You wrote earlier that you should have been more careful what you wished for. Please do the opposite.

I would rather say the rest in Japanese. It is written the way you asked for today — with spaces between the words — and at the level of the stories, so I think you can read it.

エマさん、 ほんとうに ありがとうございます。

みっかで みっつ、 このサイトは かわりました。
ぜんぶ エマさんが おしえて くれた ことです。

エマさんは、 はじめて この サイトを たすけて くれた ひとです。
それは わすれません。

これからも、 おしえて ください。

— さざなみ

37

さざなみさん！

ありがとうございます。I’m delighted and honored that added a note in Japanese to me.

Because of my psychology, I guess, it can be hard for me to ask for things. It feels risky, like with friends and when I was working in my career to ask of colleagues and employees. Sigh. I was a senior manager of semiconductor business management and often didn’t have direct reports. I was a very good individual contributor! Of course, all this has its roots in my childhood… :slight_smile:

Yesterday morning I wrote to you on your ko-fi.com site with some of my asks. I guess you haven’t seen it yet. Another ask is this, please:

  • Provide PDFs without the furigana, and have the sentences repeated on the last page(s) with the furigana. In this way, I can try to read the text with the kanji and when I get stuck I can easily (if I have a printed copy) flip to the last page and look up what I didn’t recognize. This will encourage me to really learn and read the kanji as well as the English translation.

I noticed this last night. My eyes automatically go to the furigana and ignore the kanji. I’m thus not learning the kanji with the reading.

That said, I also noticed that having the English translation below the Japanese is perfectly fine. It’s easy for me to focus on the Japanese and only jump to the English when I need it. So, in summary, I am currently thinking that an ideal PDF for me would have:

  • The Japanese text with the kanji and kana (of course), with spaces between words. Maybe some people will want to have a version with the same Japanese text, but also with the furigana. I may appreciate this too because I am still an “intermediate beginner” so I may practice reading that version too.
  • No furigana; have the sentences on the last page with the furigana for reference.

I’m not 100% sure that this is the last comment!

エマ

38

Emma,

This is a wonderful summary, and I’m happy to say you don’t have to choose between the two — I built both, plus one more:

  • Level 1 — furigana + romaji, for anyone not yet comfortable with kana
  • Level 2 — furigana + kanji, spaces between words, no romaji (this is your “intermediate beginner” version)
  • Level 3 — no furigana at all. Every sentence is numbered, and the full text with all the readings is printed again at the very back, so you can flip there when you’re stuck instead of having the answer sitting right above the word

All three now also open with a short illustrated page explaining how to read that particular version, since the differences between them aren’t obvious at a glance.

Your storybook link is the same one as before — it’ll always show the newest versions, no need to ask for a new link.

Thank you again, truly, for pushing on this. Level 3 exists because of you.

さざなみ

39

さざなみさん、

Thank you so much. I’m so appreciative. Another day, another comment. Actually, a couple:

  1. Just now I looked at a Level 3 example, したきりすずめ. The sentence list at the end is spread over 5 pages. Did you do this intentionally? (If so, please explain so I may understand.) I would prefer to have its length be on at most two pages.
  2. The filenames. For example, you have “shitakiri-suzume-2-furigana” and “shitakiri-suzume-3-kanji-practice”. I suggest changing them to: “shitakiri-suzume-2” and shitakiri-suzume-3". Why? Because when they are downloaded they will sort better in directories. Please also make the Level 1 one like this: “shitakiri-suzume-1”. I don’t see the need to add the “-furigana” etc. because you have done such a nice job on your website to explain the levels, and on the first page of the PDFs you explain it again.
  3. For the story, いっすんぼうし, the title is: “Issun-boshi, the One-Inch Boy”. I am wondering why you use “One-Inch” in the title. After all, Japan use the metric system, right? I agree though that the “Issun-boshi, the One-Centimeter Boy” sounds strange to my ear, so I don’t know what would be a better title.
  4. Within your PDF files I am unable to select any of the text except from the title page. Maybe this is a switch in your PDF file generator? It would be nice to be able to select the text and copy it to, for example, Google Translate or ChatGPT. I would then be able to check for translation or other discrepancies. I noticed this a couple of days ago and didn’t mention it because I thought maybe you have this to prevent copying. But, we have the files themselves so that doesn’t make sense to me.

Thank you,

エマ

40

エマさんへ

All four, in order:

  1. Fixed — the reading list was using a fixed number of lines per page instead of filling the page. Every story’s appendix is now 2 pages, evenly split.
  2. Done — filenames are now just {story}-{level}.pdf (e.g. shitakiri-suzume-2.pdf). The folder names still spell out the level for browsing.
  3. “一寸法師” (Issun-boshi) isn’t a metric inconsistency — 一寸 (issun) is part of the character’s traditional name, the way “Tom Thumb” doesn’t get converted to centimeters in English. It’s a fixed proper noun, not a unit being used literally.
  4. This one took more digging — you were right that it wasn’t intentional. The Japanese text is drawn as an image (so the furigana lines up exactly where I want it), and images can’t be selected. I’ve added an invisible text layer underneath every image, so now you can select and copy the Japanese, romaji, and English on every page — nothing changes visually. I tested it by extracting the text back out of the PDF to confirm it copies correctly.

All of this is live now, so any file you download from here on will already have it.

さざなみ

41

さざなみさんへ

Wow, どうもう ありがとうございました。ほんとに。

About #3, what you wrote makes perfect sense to me, thank you.

I just sent one of the files out to be printed in color. It is expensive, more than $20! So a printed book would be much better economically I’m sure. But of course, such books wouldn’t include the audio with them, but maybe a link could be provided to access the audio online.

I look forward to hearing more from you in the coming days,

エマ

P.S. Unfortunately, I think I now understand some of the difficulties you’re having with enabling a user to select text. I just tried it on a Level 2 and Level 3 file (The Tongue-Cut Sparrow) and found that the kanji (with or without furigana) isn’t selected. So that prevents me from using it as I described, copying and pasting the text somewhere.