Show HN: A fully automated podcast – actually 12 podcasts
21–30 of 35 posts
Re: Show HN: A fully automated podcast – actually 12 podcasts
#22This makes me dread that soon many other podcasts will be automated like this, and it'll be orders of magnitude more difficult to find good content than it already is.
Also, I promise to only churn out inane content for LOLs.
Re: Show HN: A fully automated podcast – actually 12 podcasts
#23Earlier quoted context omitted.
I can imagine. Imagine a podcast generation system. You'd simply have to describe the personalities of the host(s), the topic, the general vibe of the theme music, length, hardcode any sponsors, and a GPT-4 powered MLops service could produce something liked by a list of demographics 8 times out of 10 in 5 minutes. There really is no stopping this train. In 2030 90% of internet content adopting the guise of being fro…
In a parallel to email spam-bots, we get personal agents which (by your choice) filter out the generated content and make sure you only get the real deal. The fight is real.
Re: Show HN: A fully automated podcast – actually 12 podcasts
#24What vocal synthesis program did you use? Sounds 100% real at parts.
Basically it is real. Because the possible scripts that can be generated are known - fragments of speech (eg 3,4,5 word phrases) were recorded (so the intonation is free). Would be great to do it with an off-the-shelf TTS engine but I don't think there quite there yet. I know my recording skills and microphone technique is rubbish - but if I knew what I was doing on that front - I think you'd be really hard pushed to…
Re: Show HN: A fully automated podcast – actually 12 podcasts
#25Can you elaborate more on the TTS? Did you prerecord fragments (how many did you actually do?) and you just stich them together? So there is a may.mp3, 22.mp3 and your scripts just puts them together?
For dates etc - you got it. I think from memory it would be 'Wednesday' + 'the 18th' + 'of' + 'may...' + '20' + '22'
For the narrative speech it would be more words in a file. There are plenty of files (EDIT: just checked 350ish files that cover all the variations of script that can be generated at the moment)
In general the TTS - part of the project is the 'art of the almost possible' (if TTS engines sounded really good - I'd have just used one of the shelf)
Re: Show HN: A fully automated podcast – actually 12 podcasts
#26Earlier quoted context omitted.
Basically it is real. Because the possible scripts that can be generated are known - fragments of speech (eg 3,4,5 word phrases) were recorded (so the intonation is free). Would be great to do it with an off-the-shelf TTS engine but I don't think there quite there yet. I know my recording skills and microphone technique is rubbish - but if I knew what I was doing on that front - I think you'd be really hard pushed to…
The potential is 100x more with vocal synthesis imo. No need to make programmatic mad-libs style formats. Complete freedom, even though the quality isn't optimal.
I've tried to listen to books with today's TTS and it soon becomes really grating (To my ears at least). It only needs the tinyest slip every few sentences and you can't listen any more.
Re: Show HN: A fully automated podcast – actually 12 podcasts
#27Can you elaborate more on the TTS? Did you prerecord fragments (how many did you actually do?) and you just stich them together? So there is a may.mp3, 22.mp3 and your scripts just puts them together?
Sure. For dates etc - you got it. I think from memory it would be 'Wednesday' + 'the 18th' + 'of' + 'may...' + '20' + '22' For the narrative speech it would be more words in a file. There are plenty of files (EDIT: just checked 350ish files that cover all the variations of script that can be generated at the moment) In general the TTS - part of the project is the 'art of the almost possible' (if TTS engines sounded r…
Re: Show HN: A fully automated podcast – actually 12 podcasts
#28Re: Show HN: A fully automated podcast – actually 12 podcasts
#29You have a text error in the description. A daily horoscope podcast for Aquariums. Also the episode for May 22nd narrates the date as April 22nd. But I love this concept. Is this an AI speech engine or pre recorded snippets? Where did you get the text snippets from? Have you thought of incorporating GPT-3 into your horoscopes ala co-star?
It's prerecorded snippets that came out of my mouth ;)
Re: Show HN: A fully automated podcast – actually 12 podcasts
#30Earlier quoted context omitted.
Sure. For dates etc - you got it. I think from memory it would be 'Wednesday' + 'the 18th' + 'of' + 'may...' + '20' + '22' For the narrative speech it would be more words in a file. There are plenty of files (EDIT: just checked 350ish files that cover all the variations of script that can be generated at the moment) In general the TTS - part of the project is the 'art of the almost possible' (if TTS engines sounded r…
How did you come up with the initial list of phrases? Did you do some kind of analysis of other horoscopes?
Here's all of the current 'starts' for the main prediction:
(Note they're all pretty non commital - so anything could come next)
A bite from a wild animal
A financial matter coming to a head
Completion of a long delayed task
A seemingly generous gesture
A sudden realisation
An agreement with a headstrong peer
An unavoidable slowdown
Being pulled between two emotional options
A sudden eruption of feelings
Investigating a proverbial - light in the woods
Involvement with a purely privte project
Making peace with the past
The chance of a big win
Todays socialising