nwatkins
Member
From: usa
Registered: 2007-08-26
Posts: 45
So there seems to a demand for audio/video transcripts to help reinforce what we are learning through SRS methods. The transcripts, though, are a little difficult to come by. And the ones that are available are usually only accessible if you already read Japanese.
So I put together a website called J.Dialog (http://jdialog.appspot.com). This site, I hope, will be come a collaborative environment for creating transcripts. However right now it is very crude, and need of everyone's advice.
You can login with any Google/Gmail account.
The basics are: post a video/audio, add/edit/collaborate on transcript, get transcript verified.
What is available right now: the basic layouts, a sample video with some partial transcript written.
What I am looking for: What type of protocol can I implement that everyone thinks would allow us as a community to edit/create the transcripts together. I think the verified users checking our work is a great idea.
I'm super busy so I want to get your feedback before proceeding in order to ideally prevent a lot of work that gets thrown away.
Thanks again!
NOah
Dragg
Member
From: Sacramento, California
Registered: 2007-09-21
Posts: 369
If it is just transcripts without translation, then you shouldn't need native speakers. I think a lot of us could probably perfectly transcribe short audio segments given enough time. Even if the transcribers screw up from time, I feel pretty confident at catching mistakes. Therefore, going through multiple layers of verifiers would probably do the trick.
However, I don't think the site will be very successful unless you provide more structured rules for content. For example, are you suggesting that people upload full movies? Most users will be too overwhelmed. My guess is that you should probably designate a maximum length of somewhere between 5 and 15 minutes. Also, what are your copyright concerns? Do you have restrictions of what you don't want uploaded? For example, it might be more simple and safer if you limit transcripts to things like Japanese podcasts that are probably less likely to get you sued since they are freely available online through services like ITunes.
Last edited by Dragg (2008 December 22, 8:25 pm)
Dragg
Member
From: Sacramento, California
Registered: 2007-09-21
Posts: 369
@Tobberoth
For this, there might need to be a grading system in which a fluent or near-fluent person gives users a rating based on accuracy of past transcriptions. A few mistakes are probably still inevitable, but I for one, am not quite so picky as to disregard the usefulness of the whole project because of it. But, you are right; I'm sure some people here would.
The transcripts should remain unlocked, and therefore symbolically a "use at your own risk" document, until one of the highest rated users reviews it.
The more I think about it, the more I think that even smaller chunks might be better, as in two minutes in length, that could be later joined together.
Last edited by Dragg (2008 December 22, 9:06 pm)
nwatkins
Member
From: usa
Registered: 2007-08-26
Posts: 45
I think at this point it isn't necessary to worry too much about who the verified users are, if in fact that is an integral part of the way the website may work. What I mean by this is, assuming verified users exist, the next step is to design the functionality of the website. Finding users may or may not be difficult.
In terms of copyright issues, well...I'm not too worried about it. Podcasts are a great idea as you guys mentioned, and short clips from tv shows or movies (yes, 5-10 minute maximum), given our terms, would probably fall under fair use. If not, its no big deal to take them down if there are complaints, but I can worry about those issues.
The third issue, where to get the content, is in the works. One, I think users should be able to upload videos if they want. Otherwise, a mechanism for linking to youtube videos would be desirable. And third, I just finished RTK, so I will be invariably wanting content as well.
But from my perspective, the most important thing is: I hope some people here will take on these issues if the project seems feasible. Finding answers to the questions raised is our first concern, and even devising a set of stages for development with the first being a limited set of users for the purpose of creating initial content.
The point being, I don't have large quantities of time to be a part of the debate, but I do have the know-how to make development suggestions become reality very quickly.
Thoughts?
Last edited by nwatkins (2008 December 22, 9:05 pm)
Amset
Member
Registered: 2008-09-07
Posts: 94
It's always good to have more websites to get transcripts from... 頑張って
I think the transcripts, if they are typed out from just the audio and not written material, should probably just be put in by native speakers, but if you have subtitles in japanese, anyone could type that up... it'd also just be useful to have a site which hosts a lot of transcripts, so you don't have to search around for them so much, on dramanote or whatever.
Is this relevant? From House of Leaves (Mark Danielewski) -
A monk joins an abbey ready to dedicate his life to copying ancient books by hand. After the first day though, he reports to the head priest. He’s concerned that all the monks have been copying from copies made from still more copies.
"If someone makes a mistake," he points out. "It would be impossible to detect. Even worse the error would continue to be made."
A bit startled, the priest decides that he better check their latest effort against the original which is kept in a vault beneath the abbey. A place only he has access to.
Well two days, then three days pass without the priest resurfacing. Finally the new monk decides to see if the old guy’s alright. When he gets down there though, he discovers the priest hunched over both a newly copied book and the ancient original text. He is sobbing and by the look of things has been sobbing for a long time.
"Father?" the monk whispers.
"Oh Lord Jesus," the priest wails. "The word is 'celebrate'!"
