How to generate the most REALISTIC images possible with Midjourney v6
HTML-код
- Опубликовано: 7 окт 2024
- Midjourney just released it's latest model (v6) and this is how to set it up so that you can use it to.
Let's look at what's new, including, more accurate prompt following, longer prompts, and minor text drawing ability.
You're amazing in the way you explain, and your presentation skills are mind-blowing.
Thank you, so nice of you to say!
Thanks for reviewing the new midjourney v6 update. This stuff is extremely valuable to me as a beginner.
Glad to hear it's helping you 😉
Excellent introduction - step by step "how to" for a beginner. Thank you!
So glad it was helpful 😊
I'm in love.....more please. He is so good at this.
I appreciate you so much 🙏
Thanks, I needed that MJ beginner's refresher.
Very informative. Love your style! Can’t wait to try it!
Let me know how it goes!
Not used MJ yet, found it a bit overwhelming - however you’ve now given me the inspiration - love your videos - best I’ve seen ❤️
Yeah, give it a shot! It's actually super easy and intuitive to use.
Glad it inspired you ❤️
Dude, you've got skill in communicating clearly. Well done. Great vids. 👍
So nice of you to say that!
enjoyed this vid. Would like to see more of you explaining each feature like this. Thanks
Awesome - I put it in the video pipeline!
All features please. You explain amazing
Okay, I'll definitely make a video about it soon 😊
Appreciate all your videos!
Thank you so much!
great video man! keep it on! huge fan of your content
So glad to have you watching!
@@WesGPT If you can make a video of how to use images as prompts and achieve good results that would be great! 🙏🏻
@@germanmairen I kind of do this in the AI Model video, have you seen that one?
It's here: ruclips.net/video/p-1YLcdOPIc/видео.html
Great video, and amazing channel bro.
Appreciate it!
Cover all the following features, thanks for explaining.
Sure thing 😊
Please do more Midjourney and the photorealistic fine tuning settings. Cheers
This video is in the pipeline!
Use the words: "associated press photo" or "unsplash" in ur promts nd u get realistic images.🎉
Oh I like those keywords! Makes perfect sense - both of those imply HD/High Quality images
Great content. Please run over all the features in the future😊
I definitely will!
great videos thanks for being awesome!
No worries and I think you're awesome too!
Great video over all
Appreciate it 🙏
The v6 model is way better but because of discord I was considering cancelling my subscription and moving over to paid Leonardo but now I’m tore…. the photorealistic images are mind blowing. Thanks for another great knowledge share. Happy New Year Wes. See you next year!😁
Happy New Year, buddy!
So soon (hopefully within the next few months), we should be able to create Midjourney images on their official website and remove the Discord part entirely. So maybe revisit in the future once that's available?
@@WesGPT I'll stick with mj for now and use the awesome workflow you shared for instagram marketing. It's perfect for me. I'll let you know when I make my first $200. :) Thanks Wes. It's going to be an awesome year.
amazing review❤❤
Glad you enjoyed ❤️
00:26 🆕 Midjourney's latest model, version 6, offers more accurate prompt following and longer prompts compared to the previous version 5.2.
01:38 🔄 The new 'Remix Mode' allows users to transform a photo-realistic image into an illustration, enhancing the creative possibilities.
02:06 📝 Version 6 introduces minor text drawing capabilities, a significant improvement over previous versions, where generating text was challenging.
02:31 🔍 Improved upscaling feature in version 6 enables enhancing small images to larger sizes while maintaining high resolution.
03:09 🎯 Midjourney V6 demands a different approach to prompting, favoring short, straightforward prompts for better results, a shift from the lengthy prompts used in older versions.
03:38 📈 Continuous improvements in speed, image quality, coherence, prompt following, and text accuracy are expected in Midjourney V6 over the coming weeks.
04:05 🌐 Midjourney operates through Discord, and setting up involves connecting to a Discord profile for API and bot integration.
04:44 💳 Subscription plans for Midjourney vary, with options for monthly or yearly payments and different tiers offering more image generations and faster GPU access.
05:40 🤖 Midjourney can be added to personal Discord servers for private image generation, with commands initiated by the forward slash (/) key.
07:17 ⚙ Users must change settings in Midjourney to switch to the latest version 6 for image generation.
08:11 🤝 Combining Midjourney with GPT-4 or ChatGPT for prompt generation can enhance the creative process for image generation.
09:21 🌊 An example demonstrates Midjourney's ability to generate complex scenes, such as a mansion on a beach with sailboats, showing its versatility.
10:16 🔁 The 'variation' feature in Midjourney allows users to generate alternate versions of an image if the original generation is close to the desired outcome.
11:09 🖼 Shorter prompts in Midjourney V6 still yield high-quality, aesthetically pleasing images, indicating improved AI understanding.
12:17 ✍ The video tests Midjourney's text drawing ability with a specific prompt, demonstrating its capability but also showing room for improvement.
13:12 🎭 An example of generating a photo-realistic image of a woman in a theater illustrates the advanced capabilities of Midjourney V6.
14:05 🎨 The video suggests exploring other features like aspect ratio, chaos, weird tile, and stylized options in future episodes.
Thank you for the chapters 🙏
How can I generate higher resolution and which command do I have to use, and thank you very much for your video so cool u are the man
Try the UPSCALE button after you have generated the image
@@WesGPT Thank you very much
Man 6 months ago I took a prompt certification course which is mostly useless now
I know eh. Things are changing so fast.
Great video, please do more. In the Basic Plan, you get roughly 200 generations per month. Is that 200 sets of 4 images? Or does each of the 4 images count as a "generation" meaning you get about 50 sets of 4 images?
200 sets of 4 images, since 1 generation makes 4 images.
@lukazashovski answered it for me. Will definitely make more content around this!
I wonder how I can make a the person that I created wear clothes that I have a photo of
So I'm getting asked this a lot and all I know of is Adobe Firefly - but haven't experimented with it yet. I'll see if I can make a video about it!
Thanks very much for the video! I did a very nice image of a woman with wet hair in a scuba suit in V6, but when I tried face swap, it came out TERRIBLE! It looked like I cut her face out of a magazine and pasted it over the other face.
Hmm, how high is your stylize number?
@@WesGPT No idea, I've just been running it normal. I have found that certain shots don't work, but others do so I'm sticking to those that do. New posts since yesterday are all V6 and I think they look good.
they'll look back at this video in the year 3000 and laugh
At how much better AI image generators are...
Have you heard any updates on when Midjourney will totally separate from the Discord platform?
The only thing I've heard is that SOME accounts have already gotten access to generate from the main website and that it's slowly rolling out in a beta.
Thanks!@@WesGPT
How to use Midjourney v6 (for the algo wes - do a cta at end of yr videos about the title to rank higher in YT algo😉)
That's a good tip! I'm going to start mentioning the end cards to see how that does too.
Short simple straight to the point prompts work better with every AI model in my opinion. All of the wordy extraneous conversational prompts are a waste of the AI's time and totally unnecessary. I'm sure it discards 50% of the mind floss people throw at it.
I'm starting to believe this too. However, and it's a big however, if I want to get a certain image style, I do still have to use those wordy prompts with keywords. Even on this v6 model (I've been running some experiments all day today).
@@WesGPT I actually agree with you . But I am thinking about precise, information dense, carefully crafted prompts. We all recognize the skill great writers have in turning a phrase to illicit rich understanding of their intent, I can't help but feel that this has some value in this application. I suppose an interesting experiment would be to take passages from extraordinary writers describing memorable parts of their books and using them as a prompt and seeing whet we get when going for artistic results.
I was annoyed lol w v6 cause u need raw at but for realism photo cinematic they rock description off precious version and can't in paint etc and coherence yes it's great toll to go gpt ask for chat to help but text doesn't always work and or numbers and or make own gpt prompting super sensitive its wild
I experimented with it a bit today and if you bump the stylize up to 1000, you'll get some crazy photorealistic images. But it doesn't listen to your prompt that well...
Hi Wes, are your services available for hire?
You can contact me through email at heywesfrank@gmail.com
Really not happy with V6. In particular on photorealistic. Faces look to illustrated
Awe, sorry to hear that! I'm loving it so far 😊
👋
This content is amazing man. I've noticed something very specific with your nightclub photography video and want to inquire about doing business with you in a specific market. Do you have a discord/email I can reach you at?
Thanks for the kind words! You can contact me at heywesfrank@gmail.com
Boom V6, always makes me laugh when I heard this stuff will take 3 years or till 2030 to get good...compare V1 to V6 and its mind bending and in such a short time!
I know! Can you imagine v7 or v8 in 6 months?
Continue 👏
🔥