3 ms·

I have something pretty similar that I built some time ago for personal use: listen to articles while I am working on things and can't dedicate my attention to

I have something pretty similar that I built some time ago for personal use: listen to articles while I am working on things and can't dedicate my attention to anything else. I've been wondering if I should open source it(shove it all into a docker container that people can use locally). I would have loved to be able to publish it as a service but two things are stopping me: 1 is cost to run this and 2 is potential licensing issues I might face with it. But open source... And it runs pretty smoothly on a CPU(albeit a 14-core xeon and an 11-th gen i7).

yunusabd 5y ago
Same here, we built a scraper to (more or less) intelligently extract text from websites, I think adding TTS to it would be trivial. I'm guessing the hardware requirements are for the TTS engine? Could the Web Speech API be used to generate the audio locally to bring down costs? Even though you'd lose support for IE and FF.
- axegon_ 5y ago
  Exactly what I did, simple web interface I run inside rambox along with all chats and email clients and whatnot, paste a url and it scrapes it and generates the audio to be played in browser(rambox in this case). OK, I do keep a copy of the audio and the text and some other metadata but that's completely optional. As far as tts, while the cloud speech is unmatched, it's still proprietary and not free. However there are plenty of really good open source solutions which work extremely well(sample from the one I'm using: https://storage.googleapis.com/adocs_g/example.wav https://storage.googleapis.com/adocs_g/example.wav). I use it all the time and in terms of resource usage it's completely unnoticeable. But again, I'm the sole user and on pretty powerful hardware so...
  - yunusabd 5y ago
    That sample does sound pretty convincing. Just out of curiosity, that spec'd out machine is sitting idle most of the time, only to generate some TTS occasionally?
    - axegon_ 5y ago
      No, nothing else as far as tts is concerned. It is my workstation however so it's doing a million other things simultaneously. Full disclosure, I have no idea how far the tts can be pushed on it but it is single threaded so there is room for more should there be a need for it.
  - call-me-al 5y ago
    that sample sounds very convincing! if you don't mind me asking, what TTS library did you use to generate it?

yunusabd 5y ago

Same here, we built a scraper to (more or less) intelligently extract text from websites, I think adding TTS to it would be trivial. I'm guessing the hardware requirements are for the TTS engine? Could the Web Speech API be used to generate the audio locally to bring down costs? Even though you'd lose support for IE and FF.

axegon_ 5y ago
Exactly what I did, simple web interface I run inside rambox along with all chats and email clients and whatnot, paste a url and it scrapes it and generates the audio to be played in browser(rambox in this case). OK, I do keep a copy of the audio and the text and some other metadata but that's completely optional. As far as tts, while the cloud speech is unmatched, it's still proprietary and not free. However there are plenty of really good open source solutions which work extremely well(sample from the one I'm using: https://storage.googleapis.com/adocs_g/example.wav https://storage.googleapis.com/adocs_g/example.wav). I use it all the time and in terms of resource usage it's completely unnoticeable. But again, I'm the sole user and on pretty powerful hardware so...
- yunusabd 5y ago
  That sample does sound pretty convincing. Just out of curiosity, that spec'd out machine is sitting idle most of the time, only to generate some TTS occasionally?
  - axegon_ 5y ago
    No, nothing else as far as tts is concerned. It is my workstation however so it's doing a million other things simultaneously. Full disclosure, I have no idea how far the tts can be pushed on it but it is single threaded so there is room for more should there be a need for it.
- call-me-al 5y ago
  that sample sounds very convincing! if you don't mind me asking, what TTS library did you use to generate it?