Please help.
I canāt wait to try it all out though I realize there will be lots of things to work out.
One of the reasons Iāve continued to invest (i.e., subscribe!) Is hoping youād develop a workable flashcard solution! Iāve seen a steady and dedicated development process so will continue to support 
1 Like
Hi !
I would like to say I enjoy the app very much learning japanese and thai, and find it so simple and powerful.
This new version is full of good ideas! And on another aspect, as J_M3 mentionned earlier, I also worry a little as I developped a color coding on my own as well. I use it for a while now and find it quite powerful, so I am afraid this will be much harder to use it efficiently, as I can definitly not see clearly the tag color with only the thin underlining.
Therefore, if it is possible, may I suggest to make a much thicker underlining. Anyway, people not using it can discard it in the options if they donāt use, so no noise for them.
Another possibility which came up to my mind and get rid of the underlining thing, would be to be able to switch between the new āgreen/orangeā mode and the old mode, letās call it ācustomized modeā :
- the ānew modeā for users willing to have an all-in-one āautomaticā experience with the embedded flashcards, etc.
- the ācustomized modeā for users that like to keep the control and use colors to their liking and learning specificities.
BTW, maybe it is what you meant with the option āShow tags in textā, or does this option only show tags as underline?
I donāt know if it will be of any help, just in case, I am sharing my color system principle, which basically has a convergent philosophy with the new version, but with slightly more ānuancesā in colors which I find very useful in my personal use. Maybe some ideas can be picked up here for future updateā¦
- GREEN: trivial words (for examples words like āI, you, it, whatā, ābe, have, getā , Names, ā¦)
- RED: Well-known words, that I want to understand perfectly but also that I want to be able to use naturally without any hesitation or mistake (i.e., āActive phaseā words)
- YELLOW: Recently learned words, which I have to work on (ultimately, I want them to become RED one day). The yellow is very vivid, and these words are easy to spot.
- BLUE: Words which are not so common, that I may not need to use myself but that I need at least to understand, for example, words from the news, technical words, very bad words, very formal words, etcā¦ (i.e āPassive phaseā words).
Also note that this Coloring principle is coupled with my ANKI deck philosophy:
- RED, YELLOW words are revised with either ātheme / versionā sentences Notes, or very detailed Vocab Notes (with as much as possible audio extracts from my inputs).
- BLUE words are in a separate deck, with just simple note extracted from dictionnary without any audio and quite ārelaxedā deck parameters (for example the note maximum interval is only 600 days for the BLUE words, as it is 200 days for the notes with RED & YELLOW words).
This way,
- Remaining WHITE words (which I can spot easily!): they are words I need to investigate (it is a new word to me, or if not, something āfell through the netā and I may need to color it / add it in my ANKI decksā¦),.
- Remaining PURPLE words: I donāt need to investigate on it as being above my level (unless for some reasons it is of special interest to me).
Here printscreens showing you how it looks:
Also, when some new word or unclear idiom needs further investigation in one sentence, I click on the Star so as to make the investigation later, after I finish the episode. So, the āsaved sentencesā Tab is only used as an āInboxā which I need to clear.
The other interest of this colour coding is when using the tab āWordsā (cf. printscreen) for one movie or episode:
- I can assess at once its degree of difficulty, seeing each percentage
- I can work easily on the new words (i.e. WHITE words), for example say: āon this episode, I want to check and investigate on all level 3 & level 4 words which are still in WHITEā.
- I can make a dedicated seance of revision (especially of the YELLOW and BLUE words in my case) based on the episode.
BTW does this āWordsā tab will remains more or less as it is now with the new update?
Thanks and Best Regards.
2 Likes
Super excited about this update! Love the new flashcards feature, and PhrasePump looks very powerful as well. As a fellow dev Iām impressed with how much youāve been able to build with such a small team. Sorry to see thereās a minority of vocal customers complaining about things on this forum ā just know that most of us are very happy with language reactor, thereās nothing else available right now that comes even close.
5 Likes
Really loving the news features! So glad I subscribed to the Pro version and this new update is a reaffirmation that I will continue to subscribe to such an awesome language learning product. Thanks so much!
2 Likes
Hi Asane. 
Regarding 4th and 5th colours. The new system has a couple of nuances. The orange colour is for all words that you should be generally paying attention to (osmosis zone). Words which you are actively learning in the flashcards (~Yellow in your arrangement) additionally have a solid outline. You can also mark any word (even if itās already orange) as āMarked to Learnā, which, prioritises it in the flashcard queue over other suggestions, and it shows with a dashed outline (until it enters the active flashcard pool and shows with a solid outline)ā¦ it would be fine to use this for words that you know better than other orange words (something like your red colour).
Currently the flashcard algorithm adds words to the study pool by first taking all the āMarked to Learnā (queued) words first, then falls back to suggestions (most frequent words that you donāt know). We plan to add a mode shortly (could do it today) that selects appropriately between the two categoriesā¦ taking āMarked to Learnā words first, as long as they are not too infrequent (still in the orange colour range). This would mean you could mark a word as āMarked to Learnā when you start to recognise it in context, but it wouldnāt be jammed straight to the front of the flashcard queue, but would show as soon as it can justified to learn intensively.
I donāt say this is a complete analogue to your current arrangement, Iām just trying to show what is possible in the new system. Have a play with it, interested to hear any further thoughts. The new word colouring system, for better or worse, does not prioritise the question āshould I understand these wordsā, but āshould I pay attention to these words.ā
Weāre trying to balance offering a ready-to-use workflow for new users, while offering enough flexibility to not interfere too much with anyones existing systems/worklow. Some questions are quite difficult.
2 Likes
Yeah the vocabulary highlighting doesnāt work for Korean but maybe I donāt understand it fully
Noted. Are you using the Language Reactor site with the extension installed? Autopause is problematic for TurtleTV in that case, but itās hard to fix. Ah, like, a pause for 2s then continue? We considered it a couple of times, also requested was to pause at the beginning of subs. Also infinite loop feature. The difficulty is in making a UI that isnāt overwhelming, and that playback control code is pretty scary, I prefer not to touch it when itās working. 
Iāll keep it in mind when I get time to look at that code again.
2 Likes
Hmm, didnāt remove anything on purposeā¦ Netflix made a change ago a while ago that meant our hack to āshow all possible subtitlesā (often 20+ langauges) stopped working, but as far as I know itās working again now. Who knowsā¦ maybe they got complaints and took notice. If Chinese subs were available but pinyin wasnāt working, that sounds like a problem on our end. Can you post some links to Netflix movies you had problems with? Weāll check them. Also, make sure your using our extension (installed from the link on languagereactor.com), not a copy cat one.
1 Like
Korean, itās a little tricky, we didnāt have time to look at the issue in detail.
Infinite repeat, we had in for a while on the āsā key, they changed the functionaity. Iāll keep it in mind.
Entire caption colourā¦ thereās probably less of a need with the new automatic colouring system, I havenāt considered it actually.
Yeah, I can add an extra mode for the blurring (translation first, then source).
1 Like
Did they drop all the Russian-language content immeadiatly? I wasnāt paying attention. The first channel in this list has a ton of soviet-era cinema: Language Reactor
For this word, the dictionary should show like this, if you have your translation language set to English:
The settings (translation language) are not shared between dev.languagereactor.com and languagereactor.com, I donāt think. Can you send a link to the video you are watching? Sometimes the word is not identifed correctly by our NLP code. We can notes to words.
Weāre going to improve TurtleTube videos soon, Iāve been working on it for a while actually.
Click on OHA, yes, needs fixing.
No clozeā¦ did you try PhrasePump feature? I think itās what you want.
We use nice TTS (imo) for the dictionary and example sentences, for ātext modeā we fall back to browser TTS which isā¦ not nice in Chrome, much better in MS Edge browser. We experimented with making our own voices, actually recorded a dataset last year for English to release that was better than LJSpeech. Weāll try make improvements.
Pause behaviour needs some attention, you are correct.
RE: āstreamlined exports that could tie new words to the sourceā, how are you exporting words? With HTML export, or to Anki?
We have about 8-12 days of work left, Iād say. I might try sneak in a little vacation for a few days first though.
Kirianās response is correct, Iāll just add, the new features are based are based around a list of words, ordered by frequency, which you can see in the UI. We made the lists by sampling from hundreds of thousands of Youtube and movie subtiles, text from a web crawler, and the sentences from tatoeba. Itās biased towards spoken language, rather than books and news etc.
If you donāt have stuff queued to learn, flashcards will suggest words for you, starting at the most frequent word you donāt know, and work itās way up.
2 Likes
This is a good linkā¦ Iāll see if I can fix this now.
EDIT: Ben Windingās answer has a potential issue with HTTPS certifcatesā¦ hmm. Tricky. Iām thinking to change the auth library in any case soon.
Thanks Dave for your response and explanations. I expiremented with this on the demo server and it seems to work very well. Iām looking forward to these features being integrated in the Netflix and Youtube extensions. Do you have a planned release date for that?
Thanks for your reponse. I expiremented with this on the demo server and it seems to work very well. Iām looking forward to these features being integrated in the Netflix and Youtube extensions.
Dear David Wilkinson and the team,
First of all, thank you you guys for the amazing work. This extension is by far the best tool I have ever used to learn language on youtube and I have been using its premium features for 3 months
I am learning Japanese, and I hope that 2 things can be improved in Youtube:
- Vocabulary graded by level: Is there any thing we can make this tab optional (i.e. disable it so that there are only 2 columns: Subtitle and Saved vocabulary?) Since I am already studying at an advanced level, in a video with a lot of words, it takes very long to load the video and cause some lagging while I never bother to look into the Vocabulary graded by level column because I learn by reading example, not reading individual vocabulary
- Subtitle in-video: Since in Japanese videos there are a lot of burnt-in subtitles which are big; so when we turn on LLWN this subtitle is obscured by the burnt-in subtitles, making it very hard to see and cause eyestrain. Is there any ways to make the LLWN subtitle smaller or change the position of LLWN subtitle in the video? I think it would be really appreciated by the Japanese/Chinese learner community!
Thank you, have a great week. Look forward to hearing you guys soon!
Language Reacter is helping me a lot in studying foreign languages. Looks like you made a successful update, congratulations !!!. and Iām eagerly awaiting release iOS and Android based program to be released. 