HN user

holdenc137

81 karma
Posts1
Comments23
View on HN

Currently working on http://untamed.earth

Scratching an itch, the intention is that its a map, centered on the users that shows all (configurable) things of interest near by. Think of Atlas-Obscura but much more local - eg AO doesn't list every prehistoric burial mound on the planet, but I want to know where they are ;)

Listened to a few podcasts, read a few - then tried to come up with some combinations that (I hoped) were funny :)

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

Totally agree. I think we're probably only a year or so off TTS that can put some proper intonation into a sentence - hopefully then they'll be indistiguishable from live speech.

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.

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 really good - I'd have just used one of the shelf)

The generation of the 'script' uses a kind of L-System - like production rules. There's a big file along the lines of:

[the podcast] = [intro] [main body] [outro]

[main body] = [main prediction] [lucky colours] [alibi] etc

// these rules finally break down to text, eg

[main prediction start] = "A bite from a wild animal" or "A chance encounter"§

So the script has lots of combinations and is semi random - but it should always make sense.

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 tell it was stitched together phrases.

I like your thinking :) I thought horoscopes (Because they are kind of repetitive) was a good first fit - because the scripts could be generated from stock fragments 'A chance encounter' etc...

I think when the little glitches are ironed out of 'real' TTS - we'll be awash with generated content.