IllustratorsLeak
jlabjapanese
jlabjapanese

patreon


December update

It's the last day of 2023 and high time for an update! The last weeks were packed with job and family related things, so I did not post anything. However, I also continued with the project: In addition to creating a few new cards, I was mostly working on two new types of content: First, the vocab deck you voted for: I extracted all the vocabulary from the main deck and tried different ways to get good audio. Second, I thought that it is a good idea to provide you with the captions I use for video mining. I used AI/whisper to generate them and while they are far from perfect, the quality is very good - much better than the subs on YouTube. Both the deck and the captions will be finished soon, I will provide more details in a dedicated post each.

Off topic: A chore I did was to switch IDEs: For jlab, I had used pycharm for the anki addon and visual studio 2017 for the mining tools in the past. I switched to vs code and visual studio 2022 now to be able to use github copilot in the project. I've been using this at work a lot and found it very useful. This is just a technical detail, but since Teppei mentioned a couple of times that many of his students have a background in computer science in his podcast, I thought this is maybe interesting ;).

I wish you all the best for 2024, a lot of success with your Japanese and I will get back with a content update soon. Thank you for all of your support!

Comments

Thank you for the hint on HyperTTS. TTS was my method of choice, I tried it on google cloud and a few others during the last days. But I ended up hiring a freelancer to get the pronunciations right. I'm currently waiting for the audio file...

Joe

Thanks for all your work and support. I hae had good luck using the HyperTTS addon to produce audio for morphman cards, where I have audio for the sentence but not for the morph. HyperTTS lets one round-robin a bunch of voices that you like, so you can avoid the "same voice, over and over problem" you can get with tts.

stephen fuqua


More Creators