I) Midjourney V5 focuses on rapid prototyping of stunning visuals and concept art for graphic designers who need to effectively communicate their visual aesthetic with clients. So in version 5, the quality is absolutely phenomenal. However, cinematic camera control and character consistency didn’t make the priority list. Yet.

For sure is, consistent characters do not come easy. Patience is your best friend and there are several methods. I start with the simplest, and have my wisdom from Tao Prompts otherwise you literally walk 10 miles with your mouse to get to the following:

A) Split Camera-Method (cheapest paid plan on Midjourney)
1	keep you prompts as simple as possible
2	give your character a name (goes for other images-based AIs as well)
3	cinematic full body portrait of Bimbi, a beautiful model, detective style, flowing silver hair, split into multiple different images, shot from multiple angles –s 400s
4	re-run until you like a set fe U4 and click on it
5	make sure remix mode is on in the settings mode
6	Pan once to the left (alternatively you can use the custom zoom function)
7	pan left (or right) and a prompt appears to customize; customize fe eating pizza
8	Alternatively to the tab-button you can use vary to change the prompt, play and repeat until you have enough fotos for your storyboard
9	Save your fotos and cut them in Adobe Photoshop or any other (free app) editor out there
10	For upscaling I use Topaz Gigapixel, but free https://clipdrop.co/ is a good alternative

My next step was Runway and then I headed to Adobe Express et voilà.

B) play with text weights
1	turn on your remix feature (button turns green)
2	give your human like subject a name, keep your prompt simple and prompt fe Full body portrait of an Instagram model, detective style, flowing silver hair –s 400 –ar 16:9
3	re-run until you find the person you like, upscale fe U4 and click variation V4
4	Right click you upscale (U4) and copy image address paste into your prompt box plus paste your original prompt and alter fe.  Full body portrait of an Instagram model, detective style, flowing silver hair, sitting on a park bench –s 400 –ar 16:9 and put a weight to it by fe https…… Full body portrait of an Instagram model, detective style, flowing silver hair, sitting on a park bench::1.5 –s 400 –ar 16:9 [what :: does is, you are adding a weight to your prompt versus your picture (https…), so having her sit on a park bench is 1.5 more times important than the reference image; The default image weight is 1 in version 5, but users can adjust this parameter based on their needs between the range of 0 to 2. Higher image weight values will result in more visually prominent image features in the generated output, while lower values will give more emphasis to the text portion of the prompt; see Prompt Weights etc.]
5	add chaos to the prompt –c 0-100 it changes the variety of the grid you give it https…… Full body portrait of an Instagram model, detective style, flowing silver hair, sitting on a park bench –s 400 –c 4 –ar 16:9
6	use the slider method to enhance your weight further (slide it to the end by repetition) https……  Behind shot of an Instagram model, detective style, flowing silver hair, buying groceries::behind shot::.8 –s 400 –ar 16:9
7	lower the stylize value (Midjourney will follow your prompt more closely) https…….  Behind shot of an Instagram model, detective style, flowing silver hair, buying groceries::behind shot::.8 –s 40 –ar 16:9
8	the no prompt https……  Behind shot of an Instagram model, detective style, flowing silver hair, buying groceries::behind shot::.8 –s 40 –no camera –ar 16:9
9	character sheet https……  Instagram model character sheet, detective style, flowing silver hair, multiple expressions and poses –s 400 –ar 16:9 (works great for cartoons)
10	different scenes https……. Wide angle shot of an Instagram model inspecting a crime scene, detective style, flowing silver hair:: inspecting a crime scene::  –s 400 –ar 16:9
11	generate your background separately and use /blend prompt Prompt: csi crime scene, dramatic lighting –ar 16:9, happy? No, repeat. Right click and copy image address /blend https…… (background) https……(model)  Yes, there is nothing more to add, it works, but I prefer A. I must say this bored me somhow, so I took a little trip to Suno. Suno AI Beta is accessible through the Discord server and is still completely free, how long it will remain that way is not known yet. In any case, it was worth the trip, straightforward, and with 3 prompts, you’re in the game: /chirp, /bark (a Hugging Face Space, see also Github), and one more (😉). In no time you will find out the pros and cons, the cons are mainly the stitching together of your song.

FAQ

Glossary: MJ Midjourney Token – a small part of the prompt that contains a few words /describe – a command allows you to upload an image and generate four possible prompts based on that image /Shorten – a command analyses your prompt, highlights some of your prompt’s most influential words, and suggests unnecessary words you could remove

Research Mode: Make a cup of tea and start hunting for ideas. f.e. You could scrape ten thousands of messages from your MJ Discord channel, upload the JSON to the ChatGPT Code Interpreter and extract thousands of unique tokens. Sort them by popularity and choose your top.

What is hot (please forgive me Babyboomers, we know that we are hot, we do not need special words for that, but as of Gen Y they are very special):   Holographic in the style of kawaii aesthetic (I always have a good laugh at this one, as anyone who lived in Japan knows it is a street word, you hear it easily 100 times a day, kawaii or sugoi (awesome). Your are probably done with your 100 on a short commute from Shibuya to Roppongi.), colorful garden  *  in a style of surrealism  *  light silver and light gold  *  neon light pastel accent colors  *  in the style of circular abstraction  *  atmospheric lighting  *  candid moments captured  *  in the style of conceptual installations  *  emotional and dramatic scenes   *  in the style of surreal engine  *  carpetpunk    troubadour style  *  celebration of rural life * liquid foil gradient  *  euphoria  *  in the style of minimalistic monochromatic landscape  *  surreal fashion photography  *  post-internet aesthetics  *  in the style of hyperrealstic photography  *  in the style of organic material  *  algeapunk  *  transfixing marine scenes    roman art and architecture  *  hyperrealistic animal portraits   *  adventurecore  *  unicorncore    hyperbolic expression  *  kodak   * muted tones  *  ektachrome  *  forestpunk  *  Cambodian art   *  mirror rooms   *   dark pink and aquamarine  *  macro photography  *  fractal geometry  * national geographic photos   *   California impressionism  *  dreamy.   The endings with -core and -punk are special nowadays. Read the Austrian Falter today and a journalist was lamenting that some new-rich told him he’s moneypunk. As a baby boomer I know exactly how Punks looks like, because they were visible on the street on a daily base. They disappeared just like the Mods did. PS I also miss the Krishnas, such a lovely sight in orange.

What would be the tools you need for a nature documentary AI clip narrated by David Attenborough: Midjourney, Runway, Trinity, ChatGPT, Eleven Labs, CapCut (also available for PC)

Make any picture sing: Revive

C) face swap

D)

II) Dall-E

III) Stable Diffusion

IV) Meanwhile over 100 image generators are on the market

Make €

Clear Voice	Revolutionizes freelance writingno bidding, showcase portfolio, set ratespartnered with Fiverr for more opportunitiesearn $ 200 to $ 1,000 per assignmentquick turnaround, rewarding career
GigWalk	task include shelf fotos, menu audits, event assessmentsearn $100 to $ 300 based on task complexity
Snapwire	transforms Instagram photos into cashshare captivating images on Snapwire to monetizesell photos to global brands, earn top dollarbuild a portfolio, earn commissions on purchases
Scribie	offers a home-based income through transcriptionconvert audio and video into valuable written contentnumerous companies pay well for transcription servicesbeginners can earn up to $800/month; skille typist up to $3,200
Field Agent	Earn by completing uncomplicated tasks for global brandsUse the user-friendly app for capturing data and opinionsContribute to product enhancement and earn simultaneouslyDownload the app, complete tasks, and earn effortlessly
Trymata	Engage in diverse tasks, share insights on website and appEarn 10$ per completed test, taking 20-30 minutes eachTake as many tests as desired dailyMove beyond survey sites and embrace more fulfilling earning
Fiverr	Turn your talents into income on FiverrOffer logo design, video editing, NFT artwork, and moreMonetize your creativity and skills with FiverrFiverr lets you earn from virtually any talent