|
 |
Audio in Anki: Worth it? | ||
Tags: Anki | ||
Share with:
| ||
 Language Learning Forum : Learning Techniques, Methods & Strategies |  | |
| kujichagulia Senior Member Japan Joined 4642 days ago 1031 posts - 1571 votes  Speaks: English* Studies: Japanese, Portuguese | Message 1 of 31 06 January 2014 at 7:20am | IP Logged | ||
|
My focus in 2014 is to improve my listening skills, especially where Japanese is concerned. With that in mind, I recently started adding audio to my Anki cards. Most of the audio comes from text-to-speech sources, such as Google Translate or, in my opinion, the even-better Oddcast TTS service. The rest come from textbook audio, podcasts, or wherever else I can get audio.
This is how I set up my cards. (1) RECOGNITION CARDS These cards have a sentence on the front with one unknown word, and the L2 definition or the English translation on the back (and sometimes a translation of the entire sentence, if needed). I just stick the audio on the back, so that it plays when I show the answer. Not really needed, but it helps me to learn aurally as well as visually. (2) "LISTEN" CARDS I put "Listen.." on the front of a card, and when the card comes up, the audio plays. I determine if I understand what was said, then I show the answer, which is the sentence that was spoken, unknown word and definition/translation. I've been testing this for 2 days now, and it seems to be great. However... it makes creating Anki cards more troublesome and time-consuming. With the thousands of mp3s, syncing to AnkiDroid takes longer (although I could manually sync, so that is not a huge problem - just inconvenient). Moreover, there are many times when I won't be able to review Anki cards with headphones on. I mainly review cards during my lunch breaks at work, and wearing headphones would seem a bit awkward, especially if students or other teachers want to chat with me in the office. I could do it at home or on the train, but I prefer using that time to get new input and exposure to native materials, rather than reviewing Anki cards. Those are problems that I can figure out on my own. But the big question I have for those of you who use audio in Anki is this: Is audio in Anki worth it long-term? Does it improve your listening skills so much that it is worth the extra hassle of making such cards? I feel like the answers will mainly be something like "Try it for yourself and see." I understand that. But no need to invent the wheel, right? If any of you have experience with this and feel that it is worth it or not worth it, I would like to hear from you. I should point out that this is not the only listening I do. I review audio from my textbook and podcasts, and I have Japanese TV on at home whenever I can, and of course I try talking to natives at work and around town. 1 person has voted this message useful | |||
| Aik Newbie Japan Joined 4324 days ago 5 posts - 6 votes Speaks: English* Studies: Japanese | Message 2 of 31 06 January 2014 at 8:17am | IP Logged | ||
|
Very much so, I think.
All of my cards are made using subs2srs, so the audio is basically free. I find words stick really well with it - sometimes if I can't remember a word, I can if I try and remember it in the voice of the character who says it. I think it's largely responsible for anything that's good about my listening and accent. So, it's good for me. I'm not sure about text-to-speech stuff though. I have the audio both on the back and front of the cards, and a pass is if I can repeat the sentence back and understand it (while looking at the text - no furigana). 1 person has voted this message useful | |||
| vermillon Triglot Senior Member United Kingdom Joined 4473 days ago 602 posts - 1042 votes Speaks: French*, EnglishC2, Mandarin Studies: Japanese, German | Message 3 of 31 06 January 2014 at 10:12am | IP Logged | ||
|
Having tested it for over a year, it definitely helps A LOT, if you use sentences. In no time, you will remember entire sentences and their prosody (which is great), your accent will get better and you will basically have a lot of sentences ready to be re-used in your memory.
And I'm not very good at listening skills in normal time. However, what I use is native speaker recorded sentences (textbooks). Not sure how text-to-speech would fare, given that it's usually not great are prosody, and that the general rhythm of the sentence is very important. And I also have a lot of times when I can't use earphones: as long as your card doesn't rely only on audio, this is fine: you should still try to speak the sentence out loud or at least subvocalize it in your head. 1 person has voted this message useful | |||
| kujichagulia Senior Member Japan Joined 4642 days ago 1031 posts - 1571 votes Speaks: English* Studies: Japanese, Portuguese | Message 4 of 31 06 January 2014 at 10:49am | IP Logged | ||
|
OK, so it seems that this works great with real, human, native voices, but we are not so sure about text-to-speech.
I have native audio for a few sentences I put into Anki, but most sentences come from sources without audio, which is why I wanted to use TTS. I suppose for such sentences, I can submit them to RhinoSpike and have a native speaker record them. But again, that takes time, and it makes me wonder if it is worth it. 1 person has voted this message useful | |||
| Bao Diglot Senior Member Germany tinyurl.com/pe4kqe5 Joined 5561 days ago 2256 posts - 4046 votes Speaks: German*, English Studies: French, Spanish, Japanese, Mandarin | Message 5 of 31 06 January 2014 at 11:21am | IP Logged | ||
|
I just tried google translator TTS and, even though it is comprehensible it sounds pretty bad. Not the kind of material I'd use for training listening comprehension. Of course, the prosody is somewhat off, and to counteract that it's pretty slow. I am under the impression that native speakers often speed up or slur on set phrases or things like verb forms the native listener knows to expect, and that's one of the most difficult parts for me. That and Sinojapanese words combined with particles.
I tried audio cards on my mobile device before and then had to realize it cut down the time I reviewed because there were a lot of situations when I didn't want to use headphones or it would have been really weird to or take too long, and I never managed to filter my deck so that the audio cards weren't included in my mobile deck. (And then I stopped reviewing altogether, once again.) So I decided to go back to getting as much diverse input as I can stomach and randomly transcribe whatever I find interesting. (When transcribing, I have to decide whether I am hearing a certain word/a combination of certain sounds, and that makes it stick, once I've gotten a transcript/correction and have the difficult spots in red ink in front of me when listening to the audio once again.) But if I ever decide to work with audio cards again, I'll use sub2srs and shared decks to get a big big deck, clean it out for easy peasy stuff, and then suspend everything and unsuspend cards only as examples for words I've just encountered in the real world. 1 person has voted this message useful | |||
| kujichagulia Senior Member Japan Joined 4642 days ago 1031 posts - 1571 votes Speaks: English* Studies: Japanese, Portuguese | Message 6 of 31 06 January 2014 at 11:46am | IP Logged | ||
Yeah, I made audio cards with the expectation that TTS was going to be bad. (The Oddcast site linked above has a better Japanese TTS, in my opinion.) But I was thinking that it is better to have bad audio than no audio. The idea was to aurally get the words in my head, then figure out prosody later. But perhaps I was mistaken. Anyway, I'm this close to nuking my audio cards and just going back to what I was doing before. Or if I do do it, just use subs2srs. Edited by kujichagulia on 06 January 2014 at 11:48am 1 person has voted this message useful | |||
| betelgeuzah Diglot Groupie Finland Joined 4196 days ago 51 posts - 82 votes Speaks: Finnish*, English Studies: Japanese, Italian | Message 7 of 31 06 January 2014 at 11:47am | IP Logged | ||
|
Do you guys think this will be useful even when the language in question is almost 1=1 with your native language pronounciation and spelling wise?
I've pondered this a lot since Japanese and Finnish have very few differences (more voiced consonants and pitch accent in Japanese as the main ones). So when I see a new word I can guess how it's pronounced and I feel like I'm not getting anything useful out of the audio. It's almost the same with Italian basically. As long as I include the stress of the word then I can replicate the word easily. 1 person has voted this message useful | |||
 emk Diglot Moderator United States Joined 5327 days ago 2615 posts - 8806 votes Speaks: English*, FrenchB2 Studies: Spanish, Ancient Egyptian Personal Language Map | Message 8 of 31 06 January 2014 at 12:45pm | IP Logged | ||
Subs2srs is definitely the key piece here. Just find a movie with subs, rip it, run the whole thing through subs2srs, and presto! you've got ~900 cards. Delete at least 80% of these cards on first review, and you're in business. And yeah, it's a remarkably good intensive listening exercise, and that dialog will get burned into your brain. Text-to-speech sounds like a really dubious approach to me. Anki audio cards will get pounded into your head to a remarkable degree after a month, and I can't imagine wanting to do that with an artificial voice. As a general rule, I never SRS anything that isn't genuine native content, because the SRS system will magnify any errors considerably. 6 persons have voted this message useful |
|  | 
Printable version
|
| Copyright 2024 FX Micheloud - All rights reserved No part of this website may be copied by any means without my written authorization. |
