A week-long holiday looms here in China. Golden Week is normally a time of travel, but with Covid restrictions still firmly in place, that’s not practical. So what’s a fantasy writer to do? Having Great Walled, Temple of Heavened, Forbidden City’ed and Summer Palaced many times before – I plan to use the hols to visit as many coffee shops as possible. China is deep in its artisanal coffee phase, and while Beijing still lags behind Shanghai, where there are apparently more cafés per capita than anywhere else in the world, it’s no slouch. And in addition to getting caffeinated, I plan to finish (jittery fingers crossed) the final edit of City in the Mist before I begin to hunt for Beta Readers.
As work was winding down, I’d planned to start editing at the beginning of this week, but then MidJourney sucked me in again. I got so lost in text prompts and crazy imagery that I decided to see if I could make a comic version of the book. Not all of it. Hell would freeze before I managed to get that all done on my own. Just a few initial scenes.
Now, I am to Art what Jabba the Hutt is to Weightwatchers – an abject and flagrant failure. After a lifetime of drawing bananas that look more like sausages and sausages that look more, well, let’s not go there, my recent discovery of AI art generators has opened up a whole new world.
Admittedly, it’s hit and miss world. Text prompting is an arcane art up there with talking to angels, and while I try, I’m no John Dee. In addition, MidJourney is still notoriously poor at rendering people, faces, fingers and well, most animals. At least if you expect them to look like actual living creature. My attempt to coax a medieval Arab messenger on a horse out of the bot gave me a slew of riders facing the wrong way on horses made out of straw, with odd, pin-like heads and far too many legs. I suppose to convey motion. But as with most magicks, this one takes time to master and is itself already improving in leaps and bounds.
See what I mean?
But with a little perseverance (and in the comic layout phase, some judicious cropping), it was possible to get more or less what I wanted. You do have to give up on being able to reproduce characters evenly across scenes, though, as the bots reinterpret your text prompts each time. But with a few cheats like basing your character prompts on famous people (can you guess who Leila is based on?), you can get results that are close enough to each other to pass.
I decided to do a graphic retelling of Leila’s first day, her morning by the cave, the arrival of the messenger, her decision to eavesdrop on the governor’s conversation, and her hostile reception at the hammam and at some point, I might even try to add in a few panels on the exchange between Abu Daoud and Ja’afar.
Wait, didn’t I say I was a novelist? Or at least an aspiring novelist. So why did I spend a week playing around with comics? Well for one thing, I’ve always had a secret hankering to be a graphic novelist – just as a writer, not as artist. But I’d been using MidJourney to mock up images of characters and scenes to build up followings on social media sites – I now have a Twitter site, a Facebook page and an Instagram account – so I needed inspiration.
And then when I realised I’d generated enough images to build a narrative, I decided to give it a go. And that was how I fell down the rabbit hole.
I spent two days watching YouTube videos learning to perfect my text prompts for MidJourney and then spent a couple of hours learning how to use it to create comics. One video in particular was helpful, though I took longer more than its enthusiastic host seemed to think I’d need. And I only came up with an 18-page comic, not an entire book.
Now I have the comic, I plan to use the individual pages to illustrate posts and possibly even Insta Reels, but I’ve also compiled an eBook. If enough of you are interested, I’ll make that available somehow, perhaps here or on my Facebook page.
Has it been worth it? As AI means I’m now able to produce something recognisable, as well as turn the pictures into my head into images, however imprecisely, I consider it time well spent. As an author, it’s been a good experience too, forcing me to think what my characters might look like, how they might speak in real life as opposed to on the page of a book, and making me think about pacing and the art of the reveal. I probably wouldn’t do it too often, but it’s fun, (relatively) painless and in future, I could see it being a useful way to break through creative blocks. And for the geek in me, it’s also pretty cool.
Anyway, that’s where my head is at today. I’m off this afternoon on my trusty scooter Esther, in search of a gourmet coffee experience, and to clear my thoughts of AI images, but from tomorrow, it’s back to editing, editing, editing.
Let me know what you think of the pages I’ve posted, and if you’ve had any experience of using AI art in this way, I’d love to hear your thoughts.
Thanks for reading. Enjoy your weekends and catch up with you again next Friday!