Denoting chat messages in audio books

[BobbyBrandt]

There have been several threads discussing different ways to format text when denoting chat or text messages in a story, but what ideas do you all have on how to translate that formatting into speech?

Mary texted: This is a message
Bill texted: Message received

It seems like too much of that would be chaotic for a listener. Let's share some ideas.

 

[AwkwardMD]

Did you mean to put this thread in the Audio Creators Booth subforum?

 

[BobbyBrandt]

No. I thought that since the other threads on formatting chat and text messages were all placed in this forum, I would get the best ideas from putting it here.

I believe that would be especially true for those who have used text to speech as an editing tool and have heard how their formatting for messages sounded when listening to it rather than viewing it.

 

[Voyeurkenneth]

@BobbyBrandt I’ll give my two cents. Indeed, TTS for my text messaging exchanges is… not smooth. As in, the way I’ve formatted it on the page is by a first name initial then colon.

X: …

Y: …

X: …


For editing, it is what it is. But it obviously doesn’t translate well into audio book format.

It’d seem logical to just treat the exchange like regular dialogue? And then have some sentence that denotes the text exchange is over.

 

[AwkwardMD]

In my uses of TTS as an editing tool, so that I can hear my own writing, I already know what the words should look like on the page. What I am creating is a story on a page, a reading experience, and not an auditory experience.

If you want to create an auditory experience, then the people you should be asking are voice actors. Pay artists, and don't use AI to replace them.

 

[Bramblethorn]

I read stories to others regularly. Although I haven't had to deal with text messages, I often have to read passages with untagged speech, or which mixes speech and unspoken thoughts. On the printed page, punctuation tells me when one speaker stops and another starts, or when we switch from dialogue to thoughts or narration, but my audience can't see that punctuation.

As far as I can tell, it's never been much of an issue. Things like pauses and changes in intonation seem to be enough to mark those transitions, when combined with whatever tags and other cues do exist in the text. When I switch between speakers, I'll naturally pause a little longer than I would within a continuous block of speech; when I switch from speech to thoughts, I'll drop my voice a little. I'm no expert at reading but AFAICT it's enough for readers to understand what's happening. (Though I should probably tag in @Omenainen who's heard me read.)

If I were narrating text messages, having cued readers that a text exchange is approaching, I'd switch to a slightly artificial-sounding voice - perhaps flatten it towards monotone and more even timing, since text strips out those volume and timing cues.

 

[Omenainen]

I read stories to others regularly. Although I haven't had to deal with text messages, I often have to read passages with untagged speech, or which mixes speech and unspoken thoughts. On the printed page, punctuation tells me when one speaker stops and another starts, or when we switch from dialogue to thoughts or narration, but my audience can't see that punctuation.

As far as I can tell, it's never been much of an issue. Things like pauses and changes in intonation seem to be enough to mark those transitions, when combined with whatever tags and other cues do exist in the text. When I switch between speakers, I'll naturally pause a little longer than I would within a continuous block of speech; when I switch from speech to thoughts, I'll drop my voice a little. I'm no expert at reading but AFAICT it's enough for readers to understand what's happening. (Though I should probably tag in @Omenainen who's heard me read.)

If I were narrating text messages, having cued readers that a text exchange is approaching, I'd switch to a slightly artificial-sounding voice - perhaps flatten it towards monotone and more even timing, since text strips out those volume and timing cues.

…in other words, automated text to speech thingy wingies don’t perfectly (or even very well) replace human readers? So surprising