ChristopherFritz's Study Log

The author of the code was able to fix the issue.

My manga reading time is about to reduce while I do minor cleanup of auto-generated subtitles (minor text fixes, timing fixes, etc.):

The portions I’ve checked thus far have been 100% accurate, and I didn’t have to spend $3,000 on a GPU. Took about 45 mins to run one episode through whisper.cpp.


My daily manga reading continues to be minimal as I spend time doing random posts for my site (recent postings being on 水の泡になる, 油を売る, and one I had been wondering about for some time now, イヤリング vs ピアス), as well as the occasional subtitle cleanup.

On the subtitle front, I’ve found Whisper to be about 95 to 99% accurate.

It’s interesting to see the difference between transcriptions using the large and medium models, including variation in:

  • timing
  • word recognition
  • where lines are split
  • use of kanji versus hiragana versus katakana
  • spacing and punctuation (or lack thereof)

Here are a few examples. The yellow subtitle on top is from the medium model, and the white subtitle on the bottom is from the large model when running the video through Whisper:

ノミ? のみ? Now I find of wonder if there’s a common preference for one or the other. When fixing up the subtitle for a line like this, I’ll probably end up spending time trying to find an answer for that… (I haven’t yet checked the line for accuracy.)

Listening to the audio, I can’t tell for sure which is correct. Sounds like 「力を持ってる」. But either way, the line sounds like nonsense to me.

The English line is “It’s the dungeon for you, boy”, and the official Japanese subtitle file (which is meant to convey the English script, not act as closed captions for the Japanese script) gives me, 「(ろう)にぶち()んでやる」.

Now that I know (ろう) is a word, 「(ちから)を」 is clearly actually 「地下牢(ちかろう)」. Subtitle file duly updated.

Suddenly, the line actually makes sense: 「さあ、地下牢(ちかろう)()ってるぞ小僧(こぞう)

It’s quite a bit of time and effort to spend on one line out of over 1,000. It’s not like I’ve ever encountered 地下牢 before [1] and I’m sure it doesn’t come up in anything I plan to read[2], but I don’t mind immersing myself in Japanese a bit extra like this.

Errors-to-fix aside, I hope Whisper perfectly trascribed the cave’s lines, because I can’t make out a word it’s saying otherwise.

(Here’s the audio.)

  1. image ↩︎

  2. image ↩︎


It works with medium and large models, and decently fast as well. For 24 min audio => 30 min for medium model, and 50 min for large model.

However, timing is messed up. lol, it is fixed very quickly, like within an hour.

It will probably be tomorrow that I see results of both models, though.

I used WSL2, and created a bash script from Windows to run it. It seems like trouble to make it build natively for Windows.

#!/usr/bin/env bash
$workdir/main -l ja -m $workdir/models/ggml-medium.bin -t 12 -ovtt -osrt $@

With that time, are you manually fixing it using a subtitle editor? That’s what I’m doing with Aegisub.

For the anime Petite Princess Yucie, I now estimate I’m seeing about 95 to 98% accuracy using the large model. (The more subtitles I review, the higher the rate of errors I’m seeing it.)

For the Japanese dub of Disney’s Aladdin, the error rate is a bit higher, so maybe 85 to 90% accuracy? I can’t really say if medium or large has done better for me here.

I really hope to be able to compile a Windows build because there are a few Windows users I know of who are having issues with running the Python version of Whisper. I guess if building it for Windows was easy, someone else would already have put Windows builds online!

1 Like

The repo maintainer fixed the repo himself, so git pull and make.

I had been using Aegisub as well, because that’s the program I knew of from long ago.

Then recently, the reason I reached out to SubtitleEdit, is because this program can edit VTT directly, the direct output of openai/whisper. (While Aegisub seems unable to.)

Aegisub has better UI, probably…

However, whisper.cpp can output SRT directly, so that might not matter.

VTT may be easier to be edited directly by a text editor, though. (Line numbers are optional in this format.)

I didn’t fix the output or manual re-timing this time. Actually, I am at least tempted to, because of too long line lengths.


One “problem” with whisper.cpp, SRT, and Aegisub for me is that whisper.cpp produces SRT line this:

00:00:44.000 --> 00:00:49.000

However, Aegisub expects a comma, not a period.

00:00:44,000 --> 00:00:49,000

It’s an easy fix, though. I just open the SRT in a text edit and for Find/Replace to change all . to , and then Aegisub can load it just fine.

Then from Aegisub I save it as Advanced SubStation Alpha format.


I originally misunderstood what you meant about fixing in one hour, so my reply was probably confusing. I understand now!

It’s interesting how some things I run through, the timing is very good. And other things I run through, it comes out off by a lot. Well, it’s still better than nothing!

1 Like

whisper.cpp got a wrong format, it’s easier the change the code in main.cpp (and more robust than simple text replace too).

Probably in Line 265 of /main.cpp of function output_srt, change to

fout << to_timestamp(t0).replace(8, 1, ",") << " --> " << to_timestamp(t1).replace(8, 1, ",") << "\n";

Also, Aegisub is good for re-timing, when keyboard shortcuts and audio spectrogram are get used to.


It’s so easy, I didn’t even have to do anything. I didn’t even have to post it as an issue!


Saturday morning: “My copy of the new Pokemon games won’t arrive until tonight, so I’ll start coding a new project.”

Saturday evening: “Pokemon games arrived. Time to play them.”

Saturday night: “I really want to get back to that coding… I’ll continue Pokemon Sunday.”

A couple of hours past midnight: “This is a good place to stop with the code. Time for sleep.”

Sunday morning: “Back to where I left off with the code.”

Sunday evening (almost): “I’ve made really good progress with this code.”

As much as I like what I was able to get together with my Google Sheets document that builds my book club threads for me, it has always had major issues that cannot be worked around. And those issues created enough friction that I stopped using it after completing the club for the first volume of Shadows House.

Recently, I put together a web page that lets me populate some data, and then it saves it to a file for me to put on my website.

Saturday I pondered, “Why not use some of the same concepts for my book club builder?”

Things get technical from here. Now’s the time to click away to something else to read. Don’t say I didn’t warn you.

I could:

  • store the data as a JSON file
  • load it into a web page
  • click a button to copy the markdown text for posting a new thread or weekly post
  • if needed, modify values on the web page and save them

Sure, I could be reading manga or playing Pokemon, but when there’s a coding project that will better streamline things for me…

Thus my coding began.

First, I took one of the book clubs I’m running and put its information into a JSON file:

My long-term plan is to not have to touch JSON directly, and instead, do everything on the web page.

I’ve never worked much with Javascript (preferring to make web pages that work well without), and I was completely unfamiliar with methods to load a file from a local computer onto a web page.

The options look to be:

  1. prompt for a file to open
  2. drag and drop

I prefer the latter, as it’s faster for me.

And I was able to find a helpful post online showing how to put a red square on a web page, and allow dragging a file onto it.

But that red square was a bit painful on the eyes, so I generated an AI artwork on Takagi to drag a file onto:


Dragging the JSON file over to Takagi loads data from it.

Series-level data:

Volume-level data:

(I only have volume three in the JSON file, but with more volumes, I can select the volume to show on the left.)


From here, I can:

  1. make changes and then save them
  2. copy a volume thread’s markdown
  3. copy a week’s thread/comment’s markdown


There’s still a lot of work to be done, but I’m liking my progress so far.

I do want to get back to coding (after coding almost non-stop for the past five and a half hours), but I’ve got some Pokemon waiting for me.


Recent milestones for my latest project:

  1. Code is available on GitHub.

Once it reaches a point where others may benefit from this project, I plan to host a copy online as well. (This will remove the need to download a copy, as that’s a barrier to entry.)

  1. Basic code for generating a vocabulary spreadsheet is in place.

The weekly template will likely save me five minutes per week on book club duties.

The volume template will save a fair deal of time on top of that.

But the real time saved will be in generating a vocabulary spreadsheet within a few clicks. That’s including separate tabs for each chapter.

For now, I’m looking forward to seeing how tomorrow’s book club posts go.

From there, for the first time ever, I’m almost looking forward to doing new volume threads, with new vocabulary sheets.

It’s only “almost” because I haven’t yet implemented conditional formatting to the spreadsheet-generation macro.

Well, it’s not like I have a new volume thread coming up anytime soon. Just in eight days. I know what I’ll be focusing on next for coding.

Recent Japanese study progress: 0%

But I have been getting a little reading done. (I’m falling behind on Orange again.)


I feel like the more I implement (formatting and conditional formatting when auto-generating vocabulary sheets), the more I need to implement (settings and user interface to select which formattings to apply).


Ah waw, I had seen the sheet before, cool to see how it’s done!


I’ve been fairly lax on my reading lately. Between working on coding and playing Pokémon (in English), reading progress has been fairly bad. I’ll probably consider joining the next season’s daily reading challenge to really get back into it, beyond minimal daily reading.

On the plus side, slacking off a bit at the end of 2022 means it’ll be that much easier to beat my “most volumes read in one year” in 2023.

For 2022, I’m currently sitting at 76 volumes complete. Considering 2021 totaled 64 volumes read, it’s a good increase!

My goal for December is to finish up every volume I’m in the middle of, except for book club volumes (unless they’re due to complete in early January, in which case I may finish them off at the end of December).

This leaves me with the following to focus on right now:

On the coding side of things, I’ve made good progress at getting implemented what I need to get implemented, and I’ve continued to be as bad at user interface as I always am.


Main interface and Series values:

Volume values:

Chapter values:

Week values:

(Still need to fix whatever’s causing the widths to be crazy on this one.)

Template values:

One of the most important things for this, if others are to consider giving it a try, will be having good default templates. That’s also last on my to-do list to work out before reaching version 1.0.

Vocabulary values:


Weekend plans: Read lots.

Weekend reality: Multiple computer system issues occurring one after another.

Fixes were delayed for a few hours Saturday morning due to my Internet connection being down. That did at least give me some Nintendo Switch playing time, once I solved the Switch crashing by finally swapping in a new SD card.

Unfortunately, as part of the issue, my browser extensions got a bit messed up, so my Migaku extension isn’t working and will probably need to be wiped out and reinstalled from scratch. Luckily, I succeeded in extracting all my reading progress, so I didn’t have to figure out where I left off on things I haven’t read in a while.

The “good” side of that is I decided to put together a web page where I can easily access the Mokuro files for everything I’m reading. That means I can do dictionary look-ups straight from the manga text without having to type the word into a dictionary. (This is good considering I’m currently unable to type in Japanese. I’ll get that fixed later.)

To reach my reading goal for the remainder of 2022, it looks like I have to read at least 47 pages per day, not counting book club reading. I think fixing my other broken systems may have to wait a bit, as I have a lot of reading to do!

Time to go do some more reading to see how much I can knock down that daily goal.


ambitious!!! I wish I had that kind of time… good luck!!!


If I cut back on sleep and skip a few meals, I can make it =D


I finally learned how to do local storage in the browser. This means I no longer need to work with external files for the book club manager tool I’m working on:


With a few (unrelated) bug fixes, and getting some default templates set up, this can reach a point where others will be able to make use of it, should they wish to do so.


For the intersection of people who:

  1. Use Mokuro for reading their digital manga purchases.
  2. Are comfortable installing Python software from a GitHub repository.
  3. Read this study log.

(All zero of you.)

…I have put online the source code to my “Mokuro Bookshelf” page.

What this does it take everything one is reading via Mokuro and puts it onto a single web page:

There are a few things still pending, such as a UI for removing a book from the page, as well as putting “finished” books in their own section.

This version of the page doesn’t include my personal countdown for pages to read each day through the end of the year (which is sitting at 47 pages per day).

Since this uses my fork of Mokuro which adds extra Javascript, this bookshelf page only works with Mokuro HTML files generated from my fork. It ignores any generated from the vanilla Mokuro due to it lacking necessary fields.


Looking Back on 2022 and Forward to 2023


I kind of dropped out of doing SRS. It’s just not a thing for me at the moment.

I figure as soon as Migaku’s new release with their own SRS application (eventually) comes out, I’ll get back to learning new kanji and vocabulary (based on my frequency lists).


This past year, I wanted to do more intensive reading, which I planned on using 名探偵コナン for. That went well for a little bit until I fell into extensive reading mode…

I think I need to pick something I haven’t read in English for intensive reading. The problem is knowing what to pick. I can easily look up how much known and unknown vocabulary a series has, but I cannot do the same for grammar.

In total, in 2022, I completed 91 manga volumes and one light novel. This total includes both some manga with shorter volumes (140 pages) and some manga with longer volumes (360 pages).



I picked this series based entirely on my knowing most of the vocabulary in it. It’s been a fairly easy read, and I’m enjoying it. The mangaka does a decent job of pacing the reveals, as well as timing them for the end of each volume. I was afraid going into it that the whole “ghost friend” thing might be a bit of a gimmick, so I’m glad to see much of volume two fleshed it out (so to speak).

The only issue I have with the series so far is that male characters have these *super-wide* necks when seen from the side...

2022 feat: Two volumes.
2023 goal: Plan to read the final volume.



Having long ago watched the anime subtitled in English and having read much of the manga in English has definitely impacted my speed on getting through this series in Japanese.

I think another contributing factor is simply reading so many new things that I don’t spend as much time with reading things I’m familiar with (with some exceptions).

2022 feat: One volume (but it’s an omnibus, so it’s kind of like two volumes).
2023 goal: Plan to read two volumes. Actually, that brings me to the end of the series, doesn’t it?



It’s nice to be reading this one after having it on my “to consider reading” list for a while. I don’t recall where I first heard about it, whether it was a mention on the forums or a recommendation from Kobo.

When it comes to the reading experience, orange is much like アオハライド and 俺物語!! for me in that there are somewhat longer chapters that I tend to read in one go.

I keep going back and forth between “caught up with the book club” and “behind the book club”

2022 feat: Five volumes.
2023 goal: Complete the series, which I’m on track to do with the book club. Two more volumes!




I found this one difficult due to the high number of unknown vocabulary words. Even though Mokuro + Migaku makes lookups quick and easy, if I’m looking up too many words then I forget my lookups from the first half of a sentence before I finish lookups in the second half.

I feel like I wouldn’t mind continuing the series eventually, but I think even if I read the kind of manga I like for the next ten years, I still won’t gain the vocabulary needed for this one!

2022 feat: One volume.

SUPERMAN vs飯 スーパーマンのひとり飯


I’m not into comics about people eating and enjoying food. But the comedic element of a connoisseur cast of Superman made it enjoyable. I don’t plan to continue the series, but it was a fun read.

2022 feat: One volume.



I like the characters and events well enough in this series, but for some reason, I often can’t push myself to read more. I think it’s partly because sometimes I get a bit stuck on remembering who’s who, and distinguishing some of the characters. I have this series on hiatus for now.

2022 feat: Two volumes.




This series is currently my main “something easy to read between reading more difficult things” series as well as my “oops, I didn’t read anything yet today, and I need to get to bed, so let’s get in something real quick” series.

2022 feat: Nine volumes.
2023 goal: Try not to run out of material?




Like its spinoff above, this series is one of my easy reads, only at a weekly pace as I’m reading it alongside the book club.

2022 feat: Six volumes.
2023 goal: Continue reading alongside the book club.




I somewhat enjoy this series, but not enough to feel compelled to keep going, so this one’s on hiatus for now.

2022 feat: Two volumes.




This series does a good job of portraying stories centered around various characters, with artwork that is always lovely to look at, and oh yeah some rather unfortunate outfit designs.

One day, we'll be able to run a whole manga series through AI and have it re-outfit the characters.

(This isn’t quite what she wears in the series.)

2022 feat: Six volumes.
2023 goal: Run out of material and resort to waiting for new volumes. I’m down to one more volume, and I don’t anticipate another coming out until the latter half of the year.

ご注文はうさぎですか? Complete Blend



This one’s been slow-going as I intensively read the first half at the start of my manga-reading journey back in 2018, and the second half I extensively read over time in 2019 and 2020, on top of having seen the first season of the anime (covers same material) subtitled in English back in 2018.

2022 feat: One volume.
2023 goal: Casually work my way through Complete Blend volume 2. I’d love to make more progress, but I’m reading the higher-resolution Complete Blend releases, so if I go too fast I’ll turn out of material.




If this series had furigana, I probably wouldn’t keep falling behind, as I am enjoying it.

I think the main barrier has been my using Migaku Reader for reading manga, as I can’t do instant vocabulary look-up that way. Anything without furigana, I’ve been reading through Mokuro. Now that Migaku Reader randomly started having issues for me, I’ve switched to using Mokuro for all my manga reading for now. This should allow me to stay caught up. At least, once I catch up again.

2022 feat: Three volumes.
2023 goal: Catch up with the book club and don’t fall behind?




I dropped this series after finishing the first volume. I only finished the first volume because I didn’t want to leave a volume unfinished. That’s about all there is for me to say on this one.

2022 feat: One volume.




Back when I first started this series (before it was nominated for a book club), I almost read the first volume in a single sitting. At one point, it became a bit of a tradition for me to read each new volume the day it comes out. I’m looking forward to the next volume’s release while enjoying slowly revisiting the material in the offshoot book club.

2022 feat: Five volumes.
2023 goal: Maintain status quo.



This made for a nice easy read, but I do admit it didn’t hold my interest very well. I expect I would have enjoyed it more if it was one of my first manga read in Japanese, and as such I’m interested to see if it gets picked and enjoyed by new readers in the ABBC sometime.

2022 feat: Two volumes.




This is from the same mangaka as キラキラ100%, but so far I’m enjoying this one more. Things took an interesting turn right at the last page of the first volume, and if I wasn’t focused on finishing up every volume I’m in the middle of for year-end I would have started volume two already.

2022 feat: One volume.
2023 goal: Read the last three volumes.

ななか 6⁄17


The male lead character in this series is a bit difficult to understand with his “tough guy” way of talking, but I’m toughing it out. Still, that’s making me less likely to read it. I enjoyed the anime long ago, so hopefully I’ll start to really get into reading the manga over the next volume or two.

2022 feat: One volume.
2023 goal: I don’t have any specific goals in mind for this series, so perhaps just keep going forward reading it?



This one I managed to spread out over a long time, as it didn’t capture my interest much. I’m undecided about whether to give the second volume a try. If the first volume ever gets picked for the BBC and gets an offshoot club, I’d read along.

2022 feat: One volume.



I enjoyed the anime for this series enough back when I saw it that I decided to make reading the manga (which lacks furigana) a goal of mine. I think maybe it was intended at the end of 2019 to be a 2022 goal, as I didn’t consider how difficult it would be for me to learn certain kanji, and thus hit a lack of progress in WaniKani. That said, one volume in, and it’s going better than I expected!

2022 feat: One volumes.
2023 goal: I don’t have any specific goals, aside from continued reading.




I like that this series is a fairly easy read, but the lack of furigana gets in the way when I happen upon an unknown kanji. Now that I’m using Mokuro more, I expect to make good progress in reading this series going forward.

2022 feat: One volume.
2023 goal: Perhaps complete two volumes?



This series fits squarely into “I wanted to like it going in, but it just isn’t catching my attention.” I’m undecided whether I’ll continue reading it.

2022 feat: One volume.




This series went by so fast… I feel like I can see myself reading it again in a few years, but if I have other things I’m focused on reading then maybe not?

2022 feat: Seven volumes.



One volume per year. I’m looking forward to reading more in a few months.

2022 feat: One volume.
2023 goal: Learn the magic spell that allows mangaka to increase their productivity.




Waaaay below my planned reading for this series. This volume had too many characters speaking with a dialect, role language, etc. for me to feel up to reading it each day or week. There’s just one more volume in the series with the current cast, then hopefully the next cast will be a little easier for me to follow.

2022 feat: One volume.
2023 goal: Finish up the Kanto series (one more volume) and start the Johto series. No specific volume count goals at this time, but I hope to at least reach the volumes with legible text!




I started reading this one because I figured the difficulty would be similar to 名探偵コナン, but where I’ve read half of that series in English, this one I haven’t.

The difficulty has been higher, but that’s partly due to the overly casual way some characters speak.

Storywise, まじっく快斗 feels like fantasy and science fiction compared to 名探偵コナン feeling “real-world” (aside from the whole poison-turns-teenager-into-child thing). It’s a bit jarring considering the two series co-exist with Kaito Kid crossing over into Conan’s stories from time to time.

2022 feat: One volume.
2023 goal: Aim to finish the series, which is four more volumes.




I enjoyed the series, even if some chapters were a bit slow.

The final story arc felt a bit rushed/sloppy to me, but it wasn’t bad, and it led up to a nice finish.

2022 feat: Two volumes.




I feel like I should have gotten in just one more volume to average one per month.

The number would have been even higher except I took a few breaks from Conan to avoid it overtaking all my reading time. I look forward to running out of material to read in about a decade.

2022 feat: 11 volumes.
2023 goal: Perhaps aim for an average of one volume per month.



I’m glad I stuck it out until the end of the book, but for now, it looks like I’m focusing on manga.

2022 feat: One volume.



A new volume is out in January!

I really enjoy this series, but considering it doesn’t start to really get good until maybe volume 3 or 4, and at its best it’s still only average, it’s hard to recommend it.

2022 feat: Six volumes.
2023 goal: Don’t read more than one chapter in one sitting.



The characters and story haven’t compelled me to continue, so this one’s on hiatus.

2022 feat: Two volumes.



I forced myself to complete a whole volume, but I’m not interested in continuing the series.

2022 feat: One volume.



I was surprised to find the series ended after only five volumes.

Although they move to different story beats, anyone who likes からかい上手の高木さん would probably enjoy this series.

2022 feat: Two volumes.



After forcing myself to push through to the end, this series qualifies as “I wanted to like it more than I did”. The artist did a good job with it.

2022 feat: Two volumes.




After taking a long time to get through volume one, I figured my improved reading ability would make volume two more enjoyable. And yet I slowly worked my way through it over much of the year. I’ve put this series on hiatus as well, as I just didn’t feel compelled to continue.

2022 feat: One volume.

New Series for 2023

For 2023, I’m adding to my reading list:

Colori Colore Creare: Picked at random. It’s from the mangaka of ARIA. I don’t know if it’s a one-shot volume, or if it’s still ongoing.

ストロボ・エッジ: From the mangaka of アオハル. I’ve been meaning to read more of their work. This series came out before アオハル, which I figure is the better side to start on in case the latter works are much better than the earlier works. If I like the series, will I read through all 10 volumes in one year?

古見さんは、コミュ症です。: I figured I’d read this one regardless of the Beginner Book Club poll results. Since it won, it looks like I’ll be reading it with the club at club pace. This is a fairly long-running series, so we’ll be reading it for years to come!

夢みる太陽: Random pick from orange’s mangaka. I may start this while we’re on break between orange volumes, as there seems to be a crossover story with this one around the end of orange. This series is completed with 10 volumes, so if I like it maybe I can finish it within a year.

春色アストロノート: This one comes included with orange, so I’ll likely read it after finishing orange.

日々蝶々: I have no idea what to expect from this one. After I finish up 6月のラブレター, this is next on my list of “I know most of the words and kanji, so it should be an easy read” manga.

2023 Book Covers

Here’s everything I’m looking to read going into 2023:


I decided on a spur-of-the-moment whim to start listening practice.

Some time back, I was watching some anime for practice, but that was mostly reading along with the subtitles.

The plan for this time is to watch with scripts for the following:

  1. Pause at the end of each line of dialogue.
  2. Press key to re-play line of dialogue.
  3. Press key to show subtitle and re-play line of dialogue.

I wonder if I can get this to tie in with Migaku somehow so it skips over dialogue with words I don’t know, so I can focus on lines where I know all the words.


I’ve started two small projects this year, both recent.

As I work on these projects, I often find that learning how to implement something on one leads to improvements on the other.

Wanikani Book Club Manager

This recent project is a tool for managing book clubs, including reducing posting each week to a few clicks.


Whenever I think I’m getting close to being able to bump the version number up to 1.0, I notice more things that absolutely need to be implemented before I could expect anyone else to consider using this. Not to mention to make it easier for me to use with future book clubs, as my current templates include some series-specific data in them.

Mokuro Bookshelf

This recent project is a single-page document that auto-tracks my reading process when using Mokuro.


This replaces Migakur Reader for me for reading manga, so I’m looking forward to seeing how I improve it (slowly) over time.

Website Post Builder

This one’s specific to my Japanese-learning website.


Since upgrading my site to use templates, and having improved my post builder tool further, posting content to my site’s become a lot easier. I don’t know how much I’ll post on it in 2023, but between this and some other tools I’ve written to streamline the process, all friction has been removed from the process. Now it’s just a matter of laziness on my part.