Show HN: Audible for the web, Switch seamlessly between audio and text | Hacker News Reader

Show HN: Audible for the web, Switch seamlessly between audio and text

6 pointsavantion9 comments

I’ve wanted to listen to my favorite articles or web novels and switch seamlessly between the audio and text so I can read, and if I want to cook or whatever else, I can continue off the audio. I could not find it anywhere, so I created one: http://bit.ly/2qeHLev. Click on the audio version, scrub the player and scroll the content to test it out. All of it is done programatically, so it works with anyone's voice. Also, it works on the mobile web so feel free to check it out on your phone.

What do you guys think about my side project? I'd love to hear your thoughts

View on HN

I don't see any audio controls, or audio download progress, in either Safari, Chrome, or Firefox. Looks like a syntax error, there's some HTML in a Javascript file.

Conceptually, since I can't try it, it reminds me a bit of hyperaudio. Have you seen http://hyperaud.io?

I also did a hypertranscript similar to this here: http://vitor.io/uxr101 Click on any word to scrub to it, or click in whitespace to toggle play/pause.

whoops, thank you for that input. Just fixed it.... I tried both of those links. It's pretty cool. It is very similar to what I've created except I took UI cues from Audible's mobile app. ...Man, just goes to show how big the world is. I thought I was a bit original on that. Thank you for sharing.

How did u manage the sync between the audio and words? ... In my little demo, I've used IBM Watson's Speech to text API to analyze the audio. I used those results - it had timestamps, to the recognized words - and used some algorithms to match it with the text. Naturally, it did not catch like 20% of the words but was able to massage it enough that it looked smooth.

I had a verbatim transcript already, and I used CMU Sphinx to do the alignment between that and the recording.

With raw audio only, I'd probably use a transcription service that provides timings, or software like Trint to do machine learning first, and then clean up by hand, like you did.

The software I wrote to handle the playback is here: https://github.com/vitorio/hyperaudio-lite

I have a similar system implemented with TTS. It highlights text as it reads it. Do you depend on human generated audio files?

I'd love to see it. Do you have a link? ... Yes, I depend on the human generated audio file over TTS. The synthesized voice is not so great though it's getting better every year. I wanted something similar to Audible's WisperSync (http://www.audible.com/mt/wfs) which uses real voice narrators.

It's just a bookmarklet for Chrome on OSX, using the Alex voice from the operating system. I put inside it a ton of JS, to find text and read it back, while changing the color of words. You can trigger it with clicking on a word. The script would also restyle text so it is easier to read.

https://gist.github.com/Visarga/e6597edcb7ec6993521829ef5c17...

The code is uglified for bookmark use, but the original is not something I would be proud to share (it's almost a mess).

By the way, I have special treatment for YC, reddit and arXiv. Only the posts are converted into speech, skipping other texts.

that's The Thing 5+

I'm not really sure what you mean by 'The Thing 5+'

Reply on news.ycombinator.com