Two ways to narrate comic panels for blind readers. Which one actually sounds better?
I have been building an app that turns comic book files into audio. Each panel gets a spoken description of the artwork, then the dialogue is voiced per character over sound effects and a score. The voices are synthetic, so, declared up front.
The thing I cannot decide on my own is how the panel DESCRIPTIONS should be read.
Version A reads every description in the same flat, even register, like a reference narrator. Version B reads the exact same words with the delivery matched to what is in the panel, so a quiet underwater shot is read quietly and an explosion is read hard.
I cut one scene twice, back to back, so delivery is the only thing that changes. Same words, same narrator voice, same dialogue, same effects. Spoken markers tell you which version you are in.
https://www.youtube.com/watch?v=Pi3HEUT1ll4
Full transcript/time stamps in the comments, since 21 minutes is a lot to ask.
I'd like your feedback. I can offer both for personal books but not the streaming catalog on my site.
Any other tweak suggestions are welcomed. I am thinking if it should say description or end description or whatever people would like. Also old comics are weird because they had narrator text every few panels so this example is kind of odd compared to modern comics that don't have that part.