(warning: links to furry, incest, yaoi, who knows what)
The easiest thing (which I usually do) is to simply <<
hypnohub.net/post/show/13111|expand the image to the left or right>> to get some space for text. Something about 500 pixels is usually good. First, make sure there's no locked "Background" layer (at least in PS 6 - yes my software is that old) because it will automatically fill new canvas space with the background color, and you really don't want to mess with the original image's layer like that, as it makes things like changing the background of the new area harder. Then use Edit -> Change canvas size (something like that) and add about 500 pixels, making sure it's expanding from the left or right, not the center. Adjust as desired and put a layer behind the image that's a solid color or gradient appropriate for however you want it to contrast with the text and contrast or blend with the edge of the image. By segregating the text it tricks the brain into not caring that it doesn't really match the rest of the image. If the original image has a [[simple_background]] then you can often <<
hypnohub.net/post/show/19956|cheat it out>> to give yourself some more room even when you're following the features of the image.
When putting text over anything that's not a single solid color, using the "stroke" layer style at 2 or 3 px, in a color that contrasts strongly (by luminosity) with your text color, usually white or black, is
immensely helpful to the readability of the text. I consider it mandatory if the text needs to be clearly readable (though in this fetish it often doesn't). Often that's not enough because there's too much visual noise behind the text; in that case you can <<
hypnohub.net/post/show/12186|mute it with a solid color>> (useful if the image is already too wide to expand but you don't want to crop side features out completely) or block it out completely with a box. If adding any other features like <<
hypnohub.net/post/show/19368|speech bubbles>>, borders are much more optional, but can make them pop, and might improve readability if you're packing text close to the edges.
Make sure to worry about:
* Readability (flowery scripts and small font sizes tend to be harder to read, color must contrast with the background)
* Font (I prefer human voices to be in <<
en.wikipedia.org/wiki/Sans-serif|sans serif>> because that more resembles handwriting, machine/robot stuff should be in <<
hypnohub.net/post/show/10437|angular fonts>>, etc.)
* Spacing (for paragraphs of exposition I often prefer justified-left, while for speech I prefer left-aligned, and on text as sparse as a manip it helps not to have a sentence's first word be all the way over on the right side before a new line)
* Stylistic clarity (Font <<
hypnohub.net/post/show/19973/|face>>, <<
hypnohub.net/post/show/12861|color>>, and <<
hypnohub.net/post/show/20101|size>>, <<
hypnohub.net/post/show/20122|blank space>>, <<
hypnohub.net/post/show/10357|alignment/positioning>>, and <<
hypnohub.net/post/show/20054|stylistic flairs>> can all give clues to the reader about who is speaking when about what in which tone of voice and at what volume)
Finally, this applies to all manips in general, but is definitely true for text:
never, ever upload a manip on the same day you did the bulk of the work on it. Your quality will improve massively if you sleep and then look at it again with completely fresh eyes. This is the #1 way to catch typos, sections of text that don't convey what you want as well as you thought it did, and when text is not as easy to read as you thought it was when you had just typed it and basically already had it memorized when you were evaluating its readability.