emer1ca wrote:You can count me in as well.
I'm using the Assimil method along with the MP3 files as my main learning material. I'm curious to hear what you're offering.
Sorry to have taken so long to get back to you – I let a few things get on top of me and my little coding project slipped. Here’s a basic summary of what I’m doing:
I know several people on HTLAL used to talk about how they would split up the Assimil audio files in Audacity to allow them to listen to just the dialogues or just the exercises. I always liked the idea, but it always seemed too much work (Audacity lets you automate repetitive work through scripting, but it’s not straightforward) and the fact that you can’t legally share your work with others means lots of people doing the same thing and lots of time wasted.
What I’ve been knocking together is a simple web-browser-based app that lets you quickly mark out where the different parts of the files are. You will be able to work at the level of individual lessons, but also build up the full course as a single file.
Because the file will contain no data taken directly from the audio files, it shouldn’t breach copyright to distribute it, as it will be useless without the original audio.
I’ve got the logging of times working fine, and it’ll be very simple to get the page letting you play (for example) any dialogue by number, a random dialogue from a range, etc etc.
The next step will be to write a simple program that will take the data collected by the web-page and use it to split the audio into segments, so that you can put just the dialogues on your MP3 player, or turn the L2→L1 translation exercise prompts into cards to import into Anki (or similar). (These audio files will be for your personal use, so legal in most places. It is not my intention that you share these.)
I am trying to create a sensible, logical file format for this so that anyone who develops language apps can use the data generated easily.
The reason I want to do this is that I’ve seen loads of great ideas for apps and algorithms, but they always end up dead and forgotten about because the creator doesn’t have access to enough content, and an algorithm isn’t much use without content to run it on. I want innovators to be able to take material that customers have already bought and paid for and help them get the most value out of it.
I’m starting with Assimil, but in the longer term I intend to make it generally applicable. The reasons for doing Assimil first are simple:
1) there aren’t many single brands as well-regarded on this forum
2) the audio files are almost entirely content (no filler) so you get the most material at the lowest time commitment
3) the audio files have a completely regular structure. Every recording starts with the lesson number, then the name of the lesson, then the dialogue (twice), then “exercise: translate” followed by 6 prompts in the target language
4) the target-language-only audio style means that there are no regional variations, so the timings are the same regardless of which base language you are learning from and therefore more people will benefit from it.
I’ll be releasing the software as open source, but I’m considering setting up a website to host the files (timings only, of course – no audio will be shared) and enable people to collaborate proactively in processing entire courses quickly. I’m going to order an Assimil superpack in Russian so that I can look into how timings on the CDs match with those on the MP3s, but in the meantime you’d be looking at either CD or MP3.
I have a couple of relatively pressing matters to attend to in the next few days, but I should be working on the software again very shortly, so I should be able to provide a prototype in a week or two.
Again, very sorry for going quiet on you, and hopefully things should start moving shortly.
Regards,
Niall (Cainntear)