Subs2SRS / Morphman: how to create your own japanese trainer

#1

Hello everyone, I wanted to do a “little” post about some of the experiences I have came across with what I find to be some tremendous tools for learning japanese, those are Subs2SRS decks and the Morphman plugin for Anki.

I did a post some time ago after only a few weeks of using Morphman, here.

Basically I want to show you how Anki could become something like a personal trainer for your japanese studies by using this tools, as you are able to create material adjusted to your own actual knowledge.

This way you can have these two programs throw you:

  • New vocab in a N+1 manner (the previous post shows the basic setup for this)

  • Reading sentences with your known vocab (practice context sentences)

  • Listening practice with your known vocab (real sentences graded to your knowledge)

  • Extra sentences for reinforce vocab you’re having trouble with.

A little about the tools in question (you can skip to the how-to part of the post if you read the previous N+1 post):

Anki: well, I’m hoping you’ve heard about it by now. (if you’re super against using Anki for whatever reason… you might just skip the whole post, as there’s no other more “user friendly” app / website to accomplish the same :man_shrugging:).

Subs2SRS: It’s a separate program that allows you to create Anki decks coming from video of your anime, movie, drama, youtube clips, etc. For this you need such japanese video material to have some form of subtitles associated with it. Japanese subs (are a must) and your native language as well, so you can have have a nice spreadsheet to import into Anki with your video sound (or video clips), a nice image, and the lines from japanese subs / English subs in the same card.

For the Susb2SRS software
Go here http://subs2srs.sourceforge.net/

For a nice tutorial on how to create your own susb2SRS decks go here

For subs2SRS decks for you to try you can go here (though having them from your own shows will prove the most enjoyable experience)

Subs2SRS decks look something like this:

So you get the image for the scene, the audio of the sentence, and the japanese line. Then in the reverse side you’ll get the english translation.

*It’s super important that you have a deck with actual text in japanese (sometimes people have made decks using the subs in image file format, which for our purpose it’s not good at all)

Now about Morphman.
This is a terrific tool, somewhat unknown, and with actually just not enough documentation or tutorials for making an easy use of the many capabilities it offers.

Your can get the plugin here
https://github.com/kaegi/MorphMan (make sure you get the last version, as previous versions require to actually alter the code on the python file to adjust the preferences). The last version has a GUI that it’s pretty easy to use after some clarification.

Morphman uses what are called morphemes, that for our purposes will be treated as they were words in itself. So it identifies words, that could be already located in some actual deck of ours (a vocab deck perhaps, a sentence mining deck, a grammar deck, etc), with that in mind it creates a list of words currently in our knowledge (it uses the maturity status for the actual cards). If your are not using Anki in your current routine you can simply import your list of current known words into the known words file the application uses.

Myself I have my own vocab deck in Anki that includes WK vocab and then some more, besides that, a grammar vocab, that I use to keep track of current grammar concepts that I become aware and make Morphman aware this vocab as well.

If you’re jumping straight from WK to Anki and will like to try this, I would suggest, importing all your known vocab from WK (you can use the handy function from https://wanikaniexplorer.com/ to create a list of your known vocab).
Morphman has a handy function to import a list of vocab (in Anki. Tools>Morphman Database Manager). There you import the CVS file, convert it into a db file and make that one your know.db file that Morphman will identify as your current vocab.


How-To

Ok, all set with the tools you will need. Now I’ll put some pics of my setup, so you get the idea on how you can use it in a similar fashion.

First, ideally have your vocab, grammar, and senteces decks separated, as It will make reviews easier. If you don’t plan on having vocab / grammar decks on Anki then you can just add known vocab to the Morphman database file, and then just have Subs2SRS decks.

In Morphman preferences should be something like this.

You have to choose what type of cards to look into, any specific tagged cards, what field should it serch into and click the modify in the deck you will want to reorder (typically the Subs2SRS sentence cards)

Then if you want to have N+1 sentences (basically learn a new word for every sentence review) your setup should look like this:

With this setup you should be getting Morphman to search into those sentences and finding lines with vocab you know + 1 new word. Will try to get sentences ideally in the 4-5 word count, but just if they are available in your deck. When reviewing you can hit “K” if you already know every word (specially in the begining as there will be particles an contractions that probably wont be in your vocab list) and it will throw you a new one. If the sentence doesn’t strike you as a good one you can suspend the card an it will throw you a new one, if possible with the same word it was testing.

This setup is ideal for… well learning new words. But it’s not as simple as that. This card for example:

There is the word 人間 , which may be the unknown word. But then Morphman may have marked と and して as 2 words also currently in your known db. But if you are not aware of the grammatical concept of として you might find yourself still struggling to get the meaning of the sentence.

Depending on your grammar studies, you can find this situation more often than not… so I would recommend to put the sentence in jisho.org and identify that grammatical concept. As this could make your reviews somewhat more demanding, limit yourself to 5-10 / a day perhaps.

If you want to review in context sentences your known vocab this if how your setup should look:

The first unchecked box it’s for comprehension cards (sentences just containing mature vocab), The second unchecked box will allow in your sentences unmatured vocab (for anki this is vocab with due time less than a month).

Now this is powerfull stuff too. You’ll get something like this.

My recommended routine for a comprehension card like this one goes like this:

  • (with eyes shut) Listen to the line. Hit “R” to repeat the audio until you grab it (or not)

(optional Shadowing practice)

  • Repeat the sentence an try to match the actual line if you already grasped the words in there

  • Then read the text. Should be all known words. So look for grammatical concepts / particles to ckeck if you get their function there*.

  • Just then hit the anwser button an confirm the meaning (or at least the aproximated translation in the english subs)

These kind of cards are a real pleasure to practice with. 15-30 a day are easy enough, since it’s known vocab. Though you might encounter some unknown grammar in there that could prove an extra effort.

One last feature that could come in handy for reinforcing vocab, specially words you’re having trouble sticking in your mind, is the FocusMorph field.

For this to work you must add a new field into all the type of cards that Morphman is aware of. By default you should name the field “Morphman_FocusMorph”, though in the preferences you can stablish a different name.

This way Morphman will check up all the cards in the Subs2SRS deck and determine for some of them (the ones that have exactly one unknown word) what Morpheme it’s they are currently testing. This will change over time as you learn new words, so you should refesh the database every session (Cmd+M).

So, the situation shoul be something like this:
A new reading of some kanji, a similar difinition you are having trouble to contextualize, etc.

8

In this situation the reading for 度 in this word was the one that I was having trouble and causing me to fail this card a couple of times.

So while reviewing (my vocab deck in this example) I edit the card and put the actual word in the FocusMorph field.

9

Next thing in the review screen hit the “L” key to go to the Browse window and search all the Subs2SRS cards having that particular word as a Focus Morpheme.

In this example, I found 2 cards that contain this word as the only unknown word.
So I pick both and go to the tool menu and select “Learn Now”

11

This will put the cards as your very next cards to be reviewed so you reinforce that particular word right away.


So, I hope you give this a try. I think these are terrific tools to keep your studies on the fun side. Actual immersion it’s just a step ahead giving you’re already using native material. So If you decide to start immersing, you’ll see quickly how you can grab much more content practicing in this manner. :sunglasses:

It’s quite easy to get a large Subs2SRS deck (in the 50,000-100,000 numbers) and it will prove usefull for having a large database of example sentences. Try having ones from different shows, anime, dramas, even movies. As you will find examples for more diverse vocab. I’ve found out that Shounen anime it’s a good source for words related to fighting, war, magic, fantasy :sweat_smile: …If possible have decks of shows you have already or are actually watching, as the retention of the context will be greater.

Anyway, there you have it, all of the ingredients needed are at your disposal to create your own material to improve and reinforce over your actual knowledge. This is a terrific method to work over new vocab, context sentence mining, listening, shadowing … well basically you can do pretty much whatever you can think of.

EDIT:
(*) A very practical way to check grammar on the spot (after checking in jisho.org) would be the DOJG series (maybe made into Anki deck :wink:). If for any reason the DOJG doesn’t do it. This reference book has proven to be my last resource, specially with contractions and spoken terms.

45 Likes

Are those who are completing in 7 days really learning everything?
Is my Anki Layout good? Worrying over layout too much
新 Monthly Progress Challenge
How not to get bored reading Genki?
What have been your most effective self-study methods?
Should I just stick with Wanikani?
Anyone doing / did AJATT?
新 Monthly Progress Challenge
Anyone doing / did AJATT?
Questions about learning
"Exam" every ten levels
Anyone else facing burnout?
Recommended books/materials for intermediate Japanese
Time spent doing flashcards?
[SCRIPT] Add Context To Testing Material
Shadowing with Glossika
An argument for AJATT
新 Monthly Progress Challenge
What have been your most effective self-study methods?
Any way to speed things up...?
新 Monthly Progress Challenge
The best way to study vocab lessons?
Sentences in context
Spirited Away Anki Deck (Sen to Chihiro no Kamikakushi)
If you had two months to learn, what would you focus on?
Going monolingual
SRS intervals and WK long time value
How Do I Practice Listening To Japanese?
Learn Japanese with your Favourite Anime
Monolingual cards. Anki experts, help!
Do you still use other Apps beside WK?
JLPT or themed vocabulary
A good, all-round self-study routine?
What are the best supplementary resources for WaniKani? What about after WaniKani?
What resources you use to study daily?
Studying without textbooks
Making learning Japanese fun and enjoyable
Struggling with listening
Listening Practice to Japanese
#2

Fantastic! I was intrigued about Subs2SRS/Morphman since you mentioned in another thread you made, but I found it very complicated because I’m a noob at anki and its plugins, but this helps immensely.

Thanks a lot for taking the time to explain it all! I really appreciate it I’m very excited to give it a try!

1 Like

Intended use with sentences in SRS
#3

Glad you find it useful. I think setting the whole routine can be demmanding at first, but after some reviews and clearing the first cards, it’s a setup that can work flawlessly and keeping your learning quite efficiently for some time.

Creating Subs2SRS decks is somethat easier now than somewhile ago. You can start using the pre-ade decks, the link in the post it’s great and recently people are contributing more and more with new decks. Probably soon you’ll be digging other content and finding resources like Netflix and J-Drama fansubs sites being a well of content for getting as much raw material to transform into japanese study as you might want (specially for getting more diverse meterial).

0 Likes

#4

This seems pretty useful, I definitely think i+1 and native sentences are the way to go.

Having trouble getting morphman to see my existing anki decks or the downloaded vocab list. Do you have any guides floating around more on this initial set up?

0 Likes

#5

For it to identify your current decks (under Morphman Preferences>Note Filter) make sure to select the actual card type that those cards have. put attention to the field it’s checking (there can be no spaces in the field’s name). You should see the K number in the Anki main screen when properly set.
1

As for adding with the database I haven’t done it myself as I already had an Anki vocab deck when I started with this and I’ve kept all my vocab there, but this video explains that:

It’s from an older version, but the data managger function it’s pretty much the same in the current version.

0 Likes

#6

Yeah I had to change some names and stuff, seems to be working now! :slight_smile:
Just need to add another new set of subs2srs decks as I already know most of the one I’m using now xD

Thanks again.

1 Like

#7

The extra perk in this method is that makes anything you can turn into decks into useful content.
Some obscure movies I wasn’t very eager to watch happened to have Jsubs and ended up being new decks. One of this movies (Nichiren) has proven to be a terrific source for religious and monk’s related sentences :smile: , as well as an introduction to humble speech.

0 Likes

#8

Nice summary! And this should be helpful because as someone who also set it up I can say that the information that is already available online could be better.

I can only agree with @Ncastaneda: this makes such a difference for my Anki studies. I think it is really worth the trouble of setting it up. I also had to set up a virtual machine because subs2srs only runs on Windows but still… totally worth it!

1 Like

#9

Thanks. My routine has come a long way since you told me about this Morphman thing :+1:

I’m sure there’re still more novels ways to use this plugin. Will do another post if I come up with anything new.

1 Like

#10

Oooh, it’s you! I didn’t realize that since you had changed your avatar.

I’m super happy to hear that you really started using it and that unlike myself you took the time to write down more details so that more people can get into it. :slight_smile::+1:

1 Like

#11

I just realized while posting about this elsewhere that I might not have mentioned this yet: I have started using a little tool for my subs2srs cards you might also find helpful. I use this Anki plugin to automatically generate a list of translations for items on that card: https://ankiweb.net/shared/info/1015321168

That way, I can just go through the cards without having to look anything up and it is much more convenient than in the beginning. I know that there aren’t supposed to be that many new vocabulary items per page but still… I like it much better this way. I’d rather get through 50 cards quickly than through 10 and get tired because I keep looking stuff up. The translations are obviously not perfect but they are usually a good enough starting point.

The only problem that this plugin had out of the box was that it was creating too much noise but there is a suggested fi for this on github and the author is, afaik working on integrating it in an update. Let me know if you are interested in using the plugin and have a problem with the duplicates.

2 Likes

#12

Thanks!! Seems like a great addition. Though looking at that example looks like there’re quite a few extra definitions.
For now I find rather enjoyable to go through my “lessons” while doing +1 sentences and put new lines through jisho.org (actually thanks to some previous advice of yours I’ve managed take apart reviews / lessons :+1: )

39

I’m keeping those at 15 per day, and about 2-3 days a week only I do +1 sentences (so to keep the new vocab stream flowing :nerd_face: ). Rest of the week it’s mainly comprehension cards.

I’m still keeping a vocab deck, but I imagine that will disapear eventually. I think I’ll give the plugin a try by then when only working sentences.

Would you happen to know if there’s a possibility to do same thing but for JP deffinitions??

0 Likes

#13

Yup, it is definitely far from perfect. Just wait until you have a card where there is for example “そう” in the content. But in 90% of cases I find what I’m looking for and it has made the process much more enjoyable and faster for me and that’s my main goal with Anki: make it fun so that I never hesitate to do it. And my subs2srs cards are by far my favorite part of all my Anki decks.

I was wondering the same thing today. I’m not yet at a state where I feel confident enough for J > J but I will post an update here when I have looked into it and found something that works.

If you want to have a look yourself in the meantime, you could give this one a try: https://ankiweb.net/shared/info/1967553085
If it doesn’t work and we can’t find an alternative I might at some point look into modifying the original plugin by switching it to a different dictionary. But so many things to do, so little time…

1 Like

#14

Thanks for the morph man recommendation. I was actually in need of a way to sensibly sort my own listening practise cards I (somewhat more manually) create from tv shows and subtitles!

0 Likes

#15

I’m not quite there myself. I’m working on improving my reading, so I guess that will dictate when I will be able to go J-J.

I think this one works over words (from the comments, I haven’t ried yet). Will try it on my vocab deck for now.

:star_struck::star_struck: … wouldn’t mind waiting to whatever improvement you can come up with :+1:

0 Likes

#16

:smiley: If you don’t hear from me then remind me when you feel like you are ready to switch to J-J. Maybe I’ll need the extra push to consider switching too and investing the time in finding a J-J solution…

My main open question about it at the moment is what source to use. For J>E there is the JMDict. I’m hoping there is something like it for J>J but I haven’t found it yet.

1 Like

#17

There is already a plugin that fetches definitions from sanseido. Maybe you could see if there’s something to lift from there?

0 Likes

#18

If possible, I wanted to avoid querying data live. That sounds like a painful thing to do for hundreds of subs2srs cards with sentences and not just single words. That’s at least how I like to use the J>E plugin: create a lot of cards, generate all the translations and then I don’t have to worry about it for months and just slowly work through the cards one by one.

0 Likes

#19

I was just going to go through all the trouble of making a Lucky Star deck, but I checked that site and someone just made one a week ago, how lucky :wink:

2 Likes

#20

:astonished::astonished:
that blog has grown very much indeed. Lots of Subs2SRS decks now.
I still find decks of the shows I’m watching to be the most memorable for reviews… but having some hassle-free extra thousand cards as sentence bank it’s great too :sunglasses:

1 Like