Cantonese Progress Report 11/19/2018End Goal: For our next trip to HK (likely early 2019), I would like to be 'conversational' in Cantonese.
Start of current Project: 5/16/2018, Week of year 20/52
Current Week of the year: 47/52, 28 weeks into the project out of ~32.
AnkiCards are Production only: Meaning -> Jyutping
All 4490 Card suspended
InputTotal Content Run Time: 69:00
Last Week: 3:30
OutputWriting: 2650 characters
Last Week: 0
ROL/Shadow/Chorus: 6:00
Last Week: 0
iTalkiTotal Lesson Time: 17:30
Last Week: 0:30
I found the holy grail... for intermediate Cantonese learners at least.
For those of us in that bucket, the biggest issue is the nearly complete lack of audio with accurate matching (electronic) text. For the most part, this content just doesn't exist beyond the small handful of teaching dialogs. Spoken form Cantonese isn't written down very often, and when it is (in some magazines, for example) there is rarely matching audio. Subtitles are usually not written in accurate matching spoken form (almost always SWC), and even when they are, they are usually burned into the video making it an extra slow process to do look-ups as you can't use a popup dictionary or electronically copy/paste them into on online dictionary. The burned in subtitles are especially painful for those who are weak in written Chinese as their ability to see a Chinese character and type it is usually very minimal or non-existent... even hand drawing them is very slow.
While my written Chinese skills and ability to type have now reached a level that the burned in subtitles are less of a problem, it still eats time having to re-type the characters in order to do a quick look up or, even more painful, re-type the whole transcript out so that I can use the Reader app at LanguageTools.io. And accurate spoken form subtitles are still just very hard to find... until now.
This weekend I stumbled on this video by Luke Truman via a post on Reddit:
https://www.youtube.com/watch?v=A_ktCgs ... e=youtu.beThere are three key insights in that video:
1) Putting ", cc" in the search will tell YouTube to only show videos that have a least one soft subtitle track
2) True spoken form Cantonese subtitles use a few characters not found in most Standard Written Chinese
3) When you do a text search in YouTube, it will search in soft subtitles in addition to the Titles and Descriptions of videos
I knew the the first two points, but not the last.
While not a perfect method for a couple of reasons, the upshot is that this makes it possible to somewhat efficiently find videos that have spoken form Cantonese soft subtitles. And I discovered that once you found a channel that has one video with that type of subtitle track, they often have a few more videos with them. (But interestingly, it would almost always only be a small subset of a channel's videos. I don't think I found a single channel that consistently had these subs for all their videos.)
So this weekend, I went to work. I searched and searched and built a big public Playlist with over 300 videos, which can be found here:
https://www.youtube.com/playlist?list=P ... p7vPz-zCegThis is a playlist of all Cantonese videos with a soft subtitle track with spoken form Cantonese. Note that quite a number of these videos have multiple subtitle tracks, so if you start playing them and don't see spoken Cantonese text, check to see if there's another track. If you find a track that doesn't have good subs, let me know and I'll remove it.
And I'd like to add a few words of caution:
If you are easily offended, tread lightly in that playlist as I believe there is plenty of profanity, sexual talk, etc. to be found. I took any video that had, after few second's glance, what looked like accurate subs; so the content can be pretty much anything. There is likely some NSFW content and things I wouldn't show young children.
See here for how to make the most out of soft subtitles in YouTube:
https://forum.language-learners.org/vie ... =17&t=9509I also shared this playlist on the CantoDict forums as I know that I'm not the only person dying to find a big batch of content in this format.
(Note that the above method can be used to find soft subs in other languages as Luke points out in his video.)
This new content comes at a perfect time as the recent discussion on Comprehensible Input (below) got the gears in my head working about my methods of using audio+text:
https://forum.language-learners.org/vie ... =14&t=9513I believe that in the past my audio/video input hasn't been as "comprehensible" as I now think it should be. This happened partially because of the difficulty of making Cantonese audio comprehensible due to the lack of electronic text/accurate subs mentioned above and partially due to my laziness about re-listening to audio. I think there is a bell curve shape on how much time you spend with a bit of content, if you spend too little time you don't get familiar enough with it to really get some relatively 'high comprehension' listening passes with it. If you spend too much time, you're wasting too much time out on the diminishing returns tail of the curve, trying to pick out those last few words that are rare, slurred speech or mis-pronounced. For intermediate learners, I think we need to repeatedly capture the big meat section of the bell curve with each bit of audio: Listen and decode enough to get 75-90% of it clearly on 1-2 passes, then move on.
Since I now have audio content which I can easily make comprehensible due to the matching text; I am going to attempt to use it more consistently in a way where I'm getting those high value passes... i.e. I am going to decode it and re-listen a few times to make sure I'm really hearing it. In the past, I was content with a fairly low % of understanding a large portion of the time, mostly because it was so difficult without matching text to do much more without a HUGE expenditure of time... I'm hoping this new content and refined method will combine and noticeably accelerate my listening comprehension progress. I'm super excited about it!
Onward!