Ndenyal: Blog

I'm reposting these questions since it appears that very few people saw this post--judging by the lack of replies.

If you're familiar with AI text-to-image sites or programs, I need some help or suggestions. I've found them difficult to use--the basic settings typically produce people with multiple arms and legs arising from all over their bodies, not to mention the six or seven fingers and/or toes that are common. Siamese twins are a frequent result too. And the image which results from the text prompt contains less than 20% of the requested elements. If I ask for a standing person, on 70 generation requests, I will get 50+ instances of sitting, squatting, or prone people.

So, two questions, really. What's a good reliable site(s) and how can I learn to write text prompts that actually work? I see on AI sites some real works of art and that gets me envious. I have no idea how to write a text prompt to generate such images.

Thanks for your suggestions.

I've added over two dozen images throughout the "Freedom" story. Hope that they enhance your reading pleasure.

Here's a question for those of you familiar with AI text-to-image sites or programs. I've found them difficult to use--the basic settings typically produce people with multiple arms and legs arising from all over their bodies, not to mention six or seven fingers and/or toes. Siamese twins are a common result too. And the resulting overall image contains less than 20% of the elements in the text prompt. Ask for a standing person, on multiple generation requests I will get 50+ instances of sitting, squatting, or prone people.

So, two questions, really. What's a good reliable site(s) and how can I learn to write text prompts that actually work?

My own art in these stories is laboriously assembled from multiple single images, themselves composed of anatomical parts drawn from multiple sources and pasted into the final image and graphically edited for the final form. And this needs to be done for each person in an image. Yet I see on AI sites some real works of art and have no idea how to write a prompt to generate such images.

Thanks for suggestions.

I've reposted many chapters of "Freedom to be Free" with a fair amount of custom artwork. I've also added a few images to "Kevin and Denise," "Roger and Cynthia," and "Tom's Troubles." I used various AI sites for base images, but most images are composites of up to twenty individual elements combined and added to the base using GIMP, a graphics program like Photoshop. Hope you think that the images add to the stories.

Many chapters have been supplemented by images, even ones previously posted. Hope you readers enjoy the story.

On a story-score issue, something weird happened around the time I posted the tenth chapter--the score dropped suddenly from about 7.35 to 6.6. I wonder if this is the effect of the new chapter voting method. If a casual reader decides that he doesn't care for a chapter, he can downvote it right in that chapter. And not to mention that trolls have a much easier time with votes. They can give a story a "1" right in the first chapter, or any chapter, without having to go to the last one posted. Anyway, the story score is slowly regaining ground, so many thanks for all your positive votes.

I've been adding images to illustrate scenes in chapters previously posted. See chapters 3 and 6 for the latest ones. Chapter 6 is awaiting moderator release.