undefined | Better HN

0 pointsDamnInteresting5y ago0 comments

I operate a long-form nonfiction website, and back in 2012 we wanted to introduce a podcast episode version for each of our articles. I started tinkering with making recordings, but I suffered from the cringe of hearing my own voice, which was amplified by hearing every mistake and retake.

I finally got past this problem by rolling my own audio recording software. I paste in the script, and the software highlights one sentence at a time. I tap a key to start recording that sentence. If I make a mistake or I'm not happy with how it sounded, one key press discards the audio and restarts the recording. When satisfied, one key press moves it to the next sentence.

When I'm done, I can download all of the recordings into one merged WAV, and I never have to hear any bad takes (it also trims 0.1 seconds from both ends of each sentence's audio so the key press noise is removed).

It's a little clunky, being home-grown and all, but it works great. Recording stuff is pretty easy now. I keep thinking about making it available to the public, but it's not yet polished enough for that.

0 comments

9 comments · 4 top-level

dlivingston5y ago· 3 in thread

That’s so cool. Any chance you’d be okay sharing the script?

DamnInterestingOP5y ago

This is a little embarrassing, but the tool is actually browser-based, powered by Javascript and PHP. I know that's not the ideal foundation, but those were the skills I had at the time. I do have a copy of it online that I used to demo it for a friend.

Being browser-based is not really a problem for me, because I run it on my local dev server and everything is instant. But over the web it has to upload each wav file to the server, which can be slow depending on bandwidth. It does however indicate when uploads are in progress so you can avoid interrupting it.

One of the ways the tool is unsophisticated is that it's designed for a single user. So anyone with the login can see/hear others' recordings, overwrite them, etc.

Anyway, if you're still interested in seeing it, just email me at: alan at [my HN username] dot com. I can send you the URL, username, and password.

genderismadeup5y ago

A fiend of mine works at an Audio Reading service and they have wanted to improve their volunteer recoding workflow for ages. This sounds like it could have potential for them, and js / php might even be better than alternatives in their case.

rovr1385y ago

Honestly, I like the idea. Sounds great and easy to deploy!

toast05y ago· 2 in thread

> it also trims 0.1 seconds from both ends of each sentence's audio so the key press noise is removed

If you wanted to, you could probably use something to wait for sound level to drop, and then cut until the silence ended. Although a fixed cut probably works fine, too.

OJFord5y ago

Or trigger on key up? Though I suppose maybe there's a slight sound still when the cap hits it's maximum travel if it releases the switch before that?

toast05y ago

I would think key up is going to be noisy, plus maybe you need a bit to compose yourself before you start speaking.

rcurry5y ago

This is a fantastic idea. I’ve found that nothing is more annoying than to get through like five minutes of a take when recording a video and then having to redo it because I slipped up. I would love to be able to write a script, be prompted a few sentences at a time and then have it stitched together seamlessly. You need to refine and market this. If you could do this with video you’d kill it.

serjester5y ago

Agreed this is really neat! A couple months ago I had to make a product demo and it took way too long because I kept having to re-record bits.

Really hope you opensource it.

j / k navigate · click thread line to collapse