With the popularity of the methods described on the AJATT website and so forth, a lot of people are collecting sentences to study from. There's also various ways to obtain sentences (from wwwjdic, kodonline etc.) However, there's always the worry of where the sentence came from, who wrote it, what context it's appropriate in etc. There's also the problem that when you try and say the sentence, you'll get the intonation all wrong.
So why not collect transcribed sound files instead? It would be far easier to tell if the speaker is native Japanese or not, it would naturally focus more heavily on spoken Japanese rather than written Japanese, you could learn the pronunciation and intonation at the same time...
The bad point seems to be the space required to hold a large collection of sound files. However the biggest strength seems to be that foreigners could contribute with confidence. Get your favourite TV program or film or radio program and use Audacity or something to sample ten sound clips. Everyone does that and we start to build a sizeable collection...
I'm thinking about proposing it to the folks at wwwjdic. The Tanaka corpus has improved a lot but you'll never know if the sentence was originally Japanese or originally English. With contributed sound files, there's no doubt.
What do other people think?
So why not collect transcribed sound files instead? It would be far easier to tell if the speaker is native Japanese or not, it would naturally focus more heavily on spoken Japanese rather than written Japanese, you could learn the pronunciation and intonation at the same time...
The bad point seems to be the space required to hold a large collection of sound files. However the biggest strength seems to be that foreigners could contribute with confidence. Get your favourite TV program or film or radio program and use Audacity or something to sample ten sound clips. Everyone does that and we start to build a sizeable collection...
I'm thinking about proposing it to the folks at wwwjdic. The Tanaka corpus has improved a lot but you'll never know if the sentence was originally Japanese or originally English. With contributed sound files, there's no doubt.
What do other people think?
