Hacker Newsnew | past | comments | ask | show | jobs | submit | alonmln's commentslogin

I enjoyed this a lot! Took me around and hour an a half


Not if the videos are in Hebrew. English and other languages have excellent support and tools that less common languages don't.


Couldn't you feed the English auto-generated subtitles through Google Translate?


How would you get English subtitles from videos in Hebrew?


Anyone who has tried to do significant translation knows that Google Translate is pretty bad. If you just want to translate a word or even a single sentence, it's fine. But if you want to venture into larger pieces of text(like a script for example) then you couldn't do much worse than Google Translate.


There are definitely options available for Hebrew transcription, and Israel has no shortage of domestic computer talent. Human transcription services run about $150 per hour of program material, which is not a huge expense for a business. It's certainly an inconvenience, but a predictable one for the market they're operating in.


Except, OP is not a business; as is read from their post, their content is completely free and they make no money. Then 150 per hour is pretty steep.


Awesome!


I just love PicoTorrent! It's exactly what I want from a Torrent client. Thanks!


Cool, it's impressive how much can it do with a short sample, although this seems like an easy way for end users to deep fake their friends / enemies saying something.


Currently we’re looking at possible solutions, see for example here[1]. If you have suggestions, feel free to chime in!

In the demo we specifically disallowed bulk uploads to hinder such abuses.

[1] https://github.com/coqui-ai/TTS/discussions/1036


I tested it with your comment: https://sndup.net/mghy/ :)

It's also a new possibility to somewhat personalize the text to speech engines. The above example is not really close to my voice.


Maybe the solution is to have a randomly generated paragraph of text to read which expires in short amount of time. So you can't predict it and you don't have enough time to splice together a fake reading from something else.


The problem with any anti abuse measure is someone can create another project which does not have any of this. There are a handful of projects which can do pretty good voice synthesis right now. It would be about as easy as getting a consensus for all photo editing tools to place a watermark on the image to prevent abuse.


I assume it's partially because people zoom / crop images?


So the cost of being able to zoom is data communication and storage consumption?


I must say I'm not a bit surprised. This has became an industry standard:

1. Create a new amazing offer for clients with the promise that it's forever

2. Carefully write the TOS so you can change it in the future

3. Wait some time for users to board your platform with the promise of "forever" in mind

4. Change said offer so reduce costs, and keep the captive audience.


Looks useful! I like the way this library is documented, very well explained.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: