@Geoff Nugent cheers Geoff!

@Caitlin Clark thank you!

I use an external upscaler! The upscaler is Upscayl. It's free, open source, and runs on your own computer - no subscription, no watermark, no upload limit. Get it from upscayl.org (Windows, Mac and Linux). Why you need it: Midjourney images come out around 1000 to 1400 pixels on the long edge. That's fine for a screen and nowhere near enough to print. An A2 poster at 300 DPI needs roughly 5000 x 7000 pixels. If you list a Midjourney file as-is, it will print soft and someone will leave you a two star review. How to use it: open Upscayl, drag the image in, pick a model, pick a scale, choose an output folder and hit Upscayl. That's the whole app. I use the High Fidelity model at 4x, which takes my posters up to about 7900 x 11200. That's 480 DPI at A2 and 339 at A1 - more than enough. Five things I wish I'd known: 1. Don't pay for an upscaler. Upscayl does the same job for free. 2. Set your aspect ratio in Midjourney before you generate, not after. Upscaling won't fix a wrong shape, and cropping to fit afterwards eats your composition. For A-series paper the ratio is 0.7069, which is --ar 70:99. 3. Save as JPG, not PNG. A PNG at that size is enormous. 4. Check the result at 100 percent before you list it. Upscalers invent detail. Hands, faces and small text are where it goes wrong, and you won't see it at thumbnail size. 5. It's slow, and it leans on your graphics card. On a laptop with no dedicated GPU a single poster can take several minutes. Queue them up and go and do something else.

I’m lazy and don’t use my own images so I’m unfamiliar but did some digging regardless to see if I can help! --ar 70:99 is a valid ratio and it does work with uploaded images, so it's something small. Three things to check, in order: 1. Make sure it's two plain hyphens, not a dash. This is the one that gets everyone. If you copied my message into a phone or into Word, autocorrect turns -- into a single long dash and Midjourney rejects it. Delete it and retype the two hyphens by hand. 2. If you're in the Editor canvas (the one where you paint over part of the image), the shape comes from the canvas, not from --ar. Drag the canvas to portrait before you generate, or crop your photo to portrait before you upload it. Edit mode automatically matches the shape of the first image you upload, so if the source is already the right shape you don't need --ar at all. 3. In Edit mode your default aspect ratio setting is ignored, so you have to type --ar into every prompt. It won't carry over from your settings. On the look not matching your moodboard: a moodboard nudges, it doesn't dictate. Two things that actually move it. Turn the style weight up with --sw 500 or higher, and add --style raw, which stops Midjourney applying its own house prettiness over the top. And say the medium in words. If you want a photograph, write photograph, 35mm film, specific lens and light. Your camping image has come out as flat vector illustration, which is Midjourney's default when nothing in the prompt tells it otherwise, and annoyingly the moodboard won’t override that on its own.

If you need help brain storming and having more of an idea when it comes to prompting midjourney, chat GPT or Claude are really helpful.

Welcome Stacey!