Looking for: 1. audio with transcripts and 2. articles with translations
September 24, 2019 8:33 AM   Subscribe

I'm looking for two kinds of web content: 1. audio with transcripts and 2. articles with translations. Any languages are fine.

1. I'm looking for audio - typically podcasts, but other audio is fine - with reasonably verbatim transcripts. (I suppose video is fine, too, but audio is better.)

I'm looking for content that's easily accessible, so FREE ONLY - no paid content.

A few examples:

* France Culture's Le Journal des Idees has transcripts

* The Engines of our Ingenuity podcast has transcripts

2. I'm also looking for pages that have fairly direct translations (that is, not just summaries or paraphrases).

A few examples:

* Hukumusume has stories with Japanese and English (AND audio!)

* Many (all?) Canadian government websites have information in both French and English

I am not looking for machine-generated transcripts or translations unless they are of exceptionally high quality.

So: any and all pointers to FREE audio (or video) that has transcripts, and articles/web pages with translations in any languages!

Thank you!
posted by kristi to Computers & Internet (13 answers total) 5 users marked this as a favorite
 
Best answer: Most NPR content has audio files and fairly close transcripts. E.g. here's one from today on UK parliament, but every day there are lots of stories posted across many topics.
posted by SaltySalticid at 8:44 AM on September 24, 2019


Best answer: Radio France has le journal en francais facile that provides a transcript and audio broadcast.

Learn French by Podcast does podcast lessons that are often about current events but the transcripts are $1.
posted by shoesietart at 8:51 AM on September 24, 2019


Best answer: The Smart Bitches, Trashy Books podcast is transcribed.
posted by Lexica at 8:54 AM on September 24, 2019


Best answer: The blogger/economist Tyler Cowen has Conversations With Tyler where he holds conversations with several public intellectuals in varied fields. It features audio podcasts alongside verbatim transcripts of the conversations. Free to all.
posted by caveatz at 9:04 AM on September 24, 2019


Best answer: 99 percent invisible has transcripts that are more or less verbatim. This is an SEO-ish sort of page but it lists a number of accessible podcasts which come with full transcripts.
posted by jessamyn at 9:31 AM on September 24, 2019


Best answer: The excellent "Scriptnotes" screenwriting podcast has the most recent 20 episodes available for free, and they have human-written verbatim transcripts.

There's a delay between the audio episode dropping, and the transcript being provided, e.g. a couple of weeks. The page for each episode will indicate this.

For instance, at the bottom of the page for episode 416, "Fantasy Worldbuilding", which dropped on Sept. 3, it says:

You can download the episode here.

UPDATE 9-12-19: The transcript of this episode can be found here.


And includes links.
posted by theatro at 9:31 AM on September 24, 2019


Best answer: For 1:
The Lingthuasiam podcast has transcriptions, with a little delay.

For 2:
Some of the EU websites are translated into multiple languages (the dropdown has 23, though not all pages are available).
posted by scorbet at 9:42 AM on September 24, 2019


...oh I guess that Scriptnotes transcript delay is more like a bit over a week, not a couple of weeks. What is math.
posted by theatro at 9:45 AM on September 24, 2019


Best answer: Ologies - a science interview podcast by Ali Ward has transcripts - click on the individual episodes for the link to the transcript, again with some delay for the newest one.
posted by metahawk at 9:53 AM on September 24, 2019


Best answer: Maximum Fun lists their podcasts with transcripts.
posted by mmascolino at 10:31 AM on September 24, 2019


Best answer: This American Life has good transcripts.
posted by gideonfrog at 10:55 AM on September 24, 2019


Many (all?) Canadian government websites have information in both French and English

I see you've got the learning tag in your header. If you're planning to base machine learning on parallel corpora of English and French from the Canadian government (our Hansard is - or was - even available as aligned texts!) you'll find something very disappointing. Your model will learn English - Canadian French, which is not remotely applicable to octagonal/world French.
posted by scruss at 2:05 PM on September 24, 2019


Response by poster: These are all excellent answers! I was unaware of the NPR stuff - that will be very helpful.

Special thanks to scorbet for the EU websites link - that's fantastic.

scurss - I'm actually interested in human learning rather than machine learning, so the Canadian French thing is not a problem, but I greatly appreciate the caveat.

Many thanks to everyone - this is all tremendously helpful!
posted by kristi at 12:10 PM on September 28, 2019


« Older It feels like there are fleas on me, there are not...   |   Where to buy carnauba wax in Chicago? Newer »
This thread is closed to new comments.