The beginning words are good adjectives, but at the end, maybe switching the worlds into an adverbial form might help. Like equivocate could be equivocal or emend could be emended. Just a thought.
Good video, good advice, good examples, but I'm REALLY upvoting because you added chapters and timestamps. No clickbait, no mystery to get us to watch to the end, just information about prompts.
Fantastic! This is a perfect example of an overlooked positive side to AI-generated work - *improvement in diction* We're improving our vocabulary and how to use it, and seeing the results! Thank you for this. 👍
As McKenna said: Our reality is made of language .. so true and somuch to learn from VisualAi. Its a mirror .. and mine is dull and cloudy, as my english is self taught via yt. Beautiful concepts arose from these words .. thank you for bringing them to my mind, will bring them diffusion. Please keep exploring the linguistic dimensions of VisualAi. So much to learn .. thank you
your videos are truly amazing!! One definitely goes up a few levels in prompting. your efforts in putting these videos together is greatly appreciated.
Love your channel already my first vid , super informative and concise, love how you've put it all together , the vibes the theme, the music, enjoyed this thoroully :D
*i've not used midjourney yet but i have done a lot of interesting things with deep ai, Dall E-2 and nightcafe ai...a total of nearly 10,000 images so far as of this writing in just the few months i've been exploring this aspect of art generation ..my prompts typically employ the same few words at the beginning such as 8K hyper-detailed and hyper-realistic for starters and then wander off down one of many tangents of geeky word salad combining my personal interests...have had good results with hovering jellyfish, Asperitas Cloud formations, crepuscular rays, Richat Structure, crop circles. and strandbeests...plague doctors mask also works pretty well when combined with biomechanical or burtonesque. and dystopian cityscape is typically very nice...have even written short poetry and imputed that as a prompt...interesting results albeit at times very strange or in no way what was expected....your word prompt list will slowly be integrated into my own works...thank you* *and oh yes, subscribed*
wow,Thaeyne, I believe this is a very practical and crazy experiment, some of MJ's words are amazing, they are like some heroes in some games, these heroes have big and small abilities, when some words are in common, the small hero will drown out the big hero, and if we don't test, we don't know what's going on. I guess that's the real reason to subscribe to your youtube channel！
Fantastic videos, thank you. It looks like MJ pattern matches some words well in context and with others, does its best to just match up to the best characters it can. For example in the first video "surreptitious" wasn't so much about the meaning, probably because MJ couldn't put it in context with picture styles. Instead, it did the pattern matching best as it could and used the "repti" part of the word to match with "reptiles!" Very enlightening dive into how it interprets! Thank you!!!
Altough I don't understand anything of what you explain here I entered this prompts in Firefly and the results are stunning. Nearly every prompt I entered so far to create a portrait turned out with humans with missing body parts or people where misshaped. And now with this prompt they're not!
Excellent tutorial and thanks for creating it in such an organized way. Where did you get the seed number? I suppose you do not wish for everyone to use the same seed number, but how do we decide what it should be? Is it random or generated from one of your creations? Thank you.
The seed number is basically a random number that I have been using to make better comparisons. I just chose it and specifically set it to that in all of my prompts. You are free to use this or any other seed number. I have freely shared it on purpose so people would be able to generate the same images and see the results themselves too until the model changes a lot. Usually if you don't set it in Midjourney, the default is a random number out of over 4 billion possibilities.
I'll look into that more. These were just from a randomized list of mine. I prompted a lot more than made to this video. I selected my favorites that seemed to actually influence the look of the images or had a lot of images that I liked. It just happened so that the dark themes kind of influenced the outcomes more :)
V5 seems slower because it's upscaling all images in the background before they show the thumbnails. AI technologies like Blue Willow will become more and more popular over time. Blue Willow recently launched V2 with a brand new model upgrade. Now I got more better-quality images with more aesthetic outputs
Interesting concept. Is there a reason for using some fairly obscure words though, particularly as MJ clearly didn't understand them? Why not instead use a word or 2 from your definitions of these words that MJ could actually understand and provide an appropriate output?
I guess the main reason was to try out something different and see what the outcome is. And also get a sense of the default style of the version 5, because the unrecognized stuff just brings it out more. Plus, I just started out with a few and liked what it did to the images.
When you use --no and words that you don't want in your images. Like --no text. It gives Midjourney slightly more instructions. It is not foolproof, but it does significantly reduce the chance of things you don't want showing up in your images.
ChatGPT tries to find the relation but comes with hedge explaining "both can be used metaphorically", hm... I'm not convinced. Equivocate --> hedge --> horse?
@@thaeyne Thank you for your informative video and prompt response. I have learned a lot from you! Just to clarify, it's just a random parameter to fix the AI from generate too different image, and we don't need to upload a specific image to do that right? I apologize for my lack of knowledge, as I have only recently joined MJ as a member and am a year behind everyone else. Thank you again!
@@lulugoh3935 No worries, you'll catch up :) Yes, it is just a random parameter. It is a number between 0 and something around 4 billion that is assigned randomly to an image each time you do /imagine. With the --seed (number), you can just specify the number yourself instead of Midjourney selecting a random number for you. I have just happened to choose the number 777. And I usually use the same number each time so I know what it is and I don't have to look it up for each image. It has nothing to do with uploading an image. And you don't have to apologize for not knowing this stuff. And normally you don't have to bother with the seed number at all, you can leave it out from your prompt.
Just a random number I use, it can be anything from over 4 billion numbers and it is usually selected randomly unless it is specified. This way I know the differences in the images come from something else like the aspect ratio or the prompt text.
i get that the bot thought unfathomable was water because fathom is a unit of measurement for the depth of water, so the bot is interpreting the sound of the word rather than its actual meaning
Abhorrent resulting in metal band covers and promo pics makes sense. a metal band with that name is entirely plausible. and thanks to the channel, now i got to browse a thesaurus for more prompt input.
Ha, Malicious Architecture is about harmful software architecture used by hackers. Hence IT architecture diagrams of aggressive harmful software systems :D
What I've done is use wordpress for the first time in my life, select a template and then try all kinds of things and plugins and try to automate posts and wiki pages as much as possible.
How can I make couples look different from each other? For example, a Korean man and a Caucasian woman? Because when I try to do it, the two end up being the same ethnicity.
Describe the woman in full details using commas to separate the description. For example ethnicity, complexion, hair texture, body figure… when you’re done with your description of her end it with a period. Next add the activity you want her doing and place a period at the end. Now you want to start next sentence with the word “With her” and from there list the ethnicity of the male, his complexion, and so on. End his description with a period as well. I hope this helps I get opposite race couples all the time. I say it’s the husband or boyfriend. Let me know how this works for you.
@@oluwadare12 I tried and in 3 of 4 he was ginger like her. But the real problem is that both were "ugly" and I said that they were beautiful like movie star. I need really beautiful people for book cover
@@ArethaVGuedes That was so funny, I feel you though. You have to sometimes put in your prompts like a formula. Try out my prompt below. It’s a known photographer, type of shot, and desired film. Then your subject instructions. You can play with it after to tweak the look. I hope this helps, and congrats on the book☺️👏🏾👏🏾 ALBERT WATSON. DOCUMENTARY SHOT, FUJIFILM PRO 400H. BEAUTIFUL WHITE young woman, FAIR complexion, long curly hair texture, FULL - Figure hourglass body shape. Laughing with her young handsome ASIAN Husband, PORCELAIN skin COMPLEXION. Vivacious. --v 5
You forgetting who is writting this app, enginers mostly from India and other countries, so why in the world they would include worlds they never heard off??
I am fluent in English and French. Most of these words are rarely used in everyday English while you hear them often in everyday French. Conclusion: learn French, you’ll improve your prompting skills.
Very nice
vivacious | Supple | Querulous | Eloquent | Jubilant | Abhorrent | Invincible | Diabolical | Unfanthomable | Insidious | Malicous | Idyllic | Effulgent | Iridescent | Immutable | Emend | Conjecture | Vitriolic | Equivocate | Ameliorate | Abyss | Dirge | Caricature | Dogged | Buffonery
Thanks
The beginning words are good adjectives, but at the end, maybe switching the worlds into an adverbial form might help. Like equivocate could be equivocal or emend could be emended. Just a thought.
Good video, good advice, good examples, but I'm REALLY upvoting because you added chapters and timestamps. No clickbait, no mystery to get us to watch to the end, just information about prompts.
Translated into Chinese language by one of my friend:
0:00 Intro 介紹
2:00 Supple 柔軟的
3:00 Vivacious 活潑的
4:25 Querulous 抱怨的
5:46 Eloquent 雄辯的
6:40 Jubilant 歡欣鼓舞的
7:33 Abhorrent 令人憎惡的
8:25 Invincible 無敵的
9:21 Diabolical 魔鬼的
10:30 Unfathomable 深不可測的
11:23 Insidious 陰險的
12:09 Malicious 惡毒的
13:01 Idyllic 田園風光的
13:46 Effulgent 燦爛的
14:32 Iridescent 彩虹色的
15:29 Immutable 不變的
16:15 Emend 修改
16:53 Conjecture 推測
17:36 Vitriolic 尖刻的
18:32 Equivocate 支吾其詞
19:26 Ameliorate 改善
20:12 Abyss 深淵
21:16 Dirge 輓歌
22:12 Caricature 諷刺漫畫
23:14 Dogged 堅決的
24:01 Buffoonery 滑稽的行為
Fantastic video! I can't wait to improve my prompts and discover new possibilities with Midjourney v5 by using these word samples.
how a simple concept for a youtube video can be sooo helpful for people interested in one topic. bravo!
Fantastic! This is a perfect example of an overlooked positive side to AI-generated work - *improvement in diction*
We're improving our vocabulary and how to use it, and seeing the results! Thank you for this. 👍
Everything about your videos are great. Script, process, structure, technical detail, opinions, range, format, voiceover, sound design and mix.
One word change the result... This is amazing... Thxs for sharing!
I love that you used the negative prompts this time! Nice job
As McKenna said: Our reality is made of language .. so true and somuch to learn from VisualAi. Its a mirror .. and mine is dull and cloudy, as my english is self taught via yt. Beautiful concepts arose from these words .. thank you for bringing them to my mind, will bring them diffusion. Please keep exploring the linguistic dimensions of VisualAi. So much to learn .. thank you
Love the format of your videos
Well done!
Midjourney understands iridescent well because it's a not an emotion per se and less abstract. But it did well for all, really.
I've randomly tossed in "nightmare" and achieved interesting results. Thanks for sharing. I may try some of these.
Quel bonheur de voir une vidéo concrète sans blabla inutiles ! Merci 🙏
Very useful. Helps me understand the power of precise language.
your videos are truly amazing!! One definitely goes up a few levels in prompting. your efforts in putting these videos together is greatly appreciated.
As a newbie with Midjourney, this video and your website are amazing resources! Great work and thank you!😊❤
I learned more on this video than in the last course I bought on Udemy. So, thank you!
what course so I know not to buy it jajaja
You deserve 100x in clicks and likes. That good.
This is amazing! Thanks for sharing.
You did really great! No reason to criticize.
OMG 😳. Immediate subscribe. Best thing I’ve ever seen on the topic. And I’ve seen a lot of them.
thanks for this AWESOME content, this is the best prompt video Ive seen yet
I get why unfathomable does water. Fathoms...... this is a really interesting video. I need to experiment with these. Thanks
for this vid
great channel Thaeyne and interesting research in an ever evolving world 🌻
REALLY enjoyed this, very helpful, thank you!!!
Love your channel already my first vid , super informative and concise, love how you've put it all together , the vibes the theme, the music, enjoyed this thoroully :D
I'm completely blown away by the stunning artwork showcased in this video! Each piece is so incredibly beautiful and thought-provoking.
Fabulous, thank you very much!
*i've not used midjourney yet but i have done a lot of interesting things with deep ai, Dall E-2 and nightcafe ai...a total of nearly 10,000 images so far as of this writing in just the few months i've been exploring this aspect of art generation ..my prompts typically employ the same few words at the beginning such as 8K hyper-detailed and hyper-realistic for starters and then wander off down one of many tangents of geeky word salad combining my personal interests...have had good results with hovering jellyfish, Asperitas Cloud formations, crepuscular rays, Richat Structure, crop circles. and strandbeests...plague doctors mask also works pretty well when combined with biomechanical or burtonesque. and dystopian cityscape is typically very nice...have even written short poetry and imputed that as a prompt...interesting results albeit at times very strange or in no way what was expected....your word prompt list will slowly be integrated into my own works...thank you*
*and oh yes, subscribed*
wow,Thaeyne, I believe this is a very practical and crazy experiment, some of MJ's words are amazing, they are like some heroes in some games, these heroes have big and small abilities, when some words are in common, the small hero will drown out the big hero, and if we don't test, we don't know what's going on. I guess that's the real reason to subscribe to your youtube channel！
I was looking so bad for a guide like this. A big thank you dear artist 🙏🏻❤. It helps a lot
This is quality video, big thanks! ❤
wow! your wiki is has so much information. i know what I'm doing for the next 12 hours
Glad it was helpful!
Fantastic videos, thank you. It looks like MJ pattern matches some words well in context and with others, does its best to just match up to the best characters it can. For example in the first video "surreptitious" wasn't so much about the meaning, probably because MJ couldn't put it in context with picture styles. Instead, it did the pattern matching best as it could and used the "repti" part of the word to match with "reptiles!" Very enlightening dive into how it interprets! Thank you!!!
Specified word: Ameliorate:
In 20:10, the software prompt processed the women’s Portraits and converted the EXACT faces into Fantasy portraits.
thank you so much for your work. You are a good person. Art will be with you
Sophisticated eloquence seems to work.
This. I am in love with this content. Keep up the great work!
Thank You, great work!
Another great video, your voiceovers are almost hypnotic, what software do you use ? Thanks
I use Murf for the voiceovers, you can find the link in the description to it.
Hi Thaeyne, I missed the part of using ar 9: 16 -seed 777 - thank you, your video and everything in it, is fabulous
Do you have the prompts for your amazing Intro? ❤
Thanks for the programer!
You are amazing, thank you for sharing. So much work!
Altough I don't understand anything of what you explain here I entered this prompts in Firefly and the results are stunning. Nearly every prompt I entered so far to create a portrait turned out with humans with missing body parts or people where misshaped. And now with this prompt they're not!
Smart content , thank you so much !❤
you help me so much! thank you
Excellent tutorial and thanks for creating it in such an organized way. Where did you get the seed number? I suppose you do not wish for everyone to use the same seed number, but how do we decide what it should be? Is it random or generated from one of your creations? Thank you.
The seed number is basically a random number that I have been using to make better comparisons. I just chose it and specifically set it to that in all of my prompts. You are free to use this or any other seed number. I have freely shared it on purpose so people would be able to generate the same images and see the results themselves too until the model changes a lot. Usually if you don't set it in Midjourney, the default is a random number out of over 4 billion possibilities.
@@thaeyne I appreciate your response. Thank you.
Hello tharyne,
can you teach me step by step for. make like that?

How about a some more positive words. I little too many dark sidedness. Ugh! With that said, great job! Looking forward to what’s next!
I'll look into that more. These were just from a randomized list of mine. I prompted a lot more than made to this video. I selected my favorites that seemed to actually influence the look of the images or had a lot of images that I liked. It just happened so that the dark themes kind of influenced the outcomes more :)
we are now seeing how much dark and unkind imagery midjourney produces - that represents the mental state of man, sad...
V5 seems slower because it's upscaling all images in the background before they show the thumbnails. AI technologies like Blue Willow will become more and more popular over time. Blue Willow recently launched V2 with a brand new model upgrade. Now I got more better-quality images with more aesthetic outputs
It's great to hear that you're enjoying your experience with Blue Willow
Thank you ♥️
The link to the first video could be on the description
Amazing ❤❤
Excellent!
very useful!
You are the best !
Interesting concept. Is there a reason for using some fairly obscure words though, particularly as MJ clearly didn't understand them? Why not instead use a word or 2 from your definitions of these words that MJ could actually understand and provide an appropriate output?
I guess the main reason was to try out something different and see what the outcome is. And also get a sense of the default style of the version 5, because the unrecognized stuff just brings it out more. Plus, I just started out with a few and liked what it did to the images.
Amazing, thank you :)
These are great!
Very inspirational 👍
Nice! Useful.
Thank you.
Killer video!
F…….amazing video , thank you
Vitriol is an archaic word for poison, like a poisonous potion.
I would like to ask what are negative prompts
When you use --no and words that you don't want in your images. Like --no text. It gives Midjourney slightly more instructions. It is not foolproof, but it does significantly reduce the chance of things you don't want showing up in your images.
Laughed when invincible showed Man City jersey lol
I would have liked to see each version produced without the comma as well
ChatGPT tries to find the relation but comes with hedge explaining "both can be used metaphorically", hm... I'm not convinced. Equivocate --> hedge --> horse?
thanks
May I know now how does the --seed777 looks like? Or it just a random image? tQ.
It is not possible to only prompt --seed 777 by itself, it has to be accompanied with some text or an image.
@@thaeyne Thank you for your informative video and prompt response. I have learned a lot from you! Just to clarify, it's just a random parameter to fix the AI from generate too different image, and we don't need to upload a specific image to do that right? I apologize for my lack of knowledge, as I have only recently joined MJ as a member and am a year behind everyone else. Thank you again!
@@lulugoh3935 No worries, you'll catch up :) Yes, it is just a random parameter. It is a number between 0 and something around 4 billion that is assigned randomly to an image each time you do /imagine. With the --seed (number), you can just specify the number yourself instead of Midjourney selecting a random number for you. I have just happened to choose the number 777. And I usually use the same number each time so I know what it is and I don't have to look it up for each image. It has nothing to do with uploading an image. And you don't have to apologize for not knowing this stuff. And normally you don't have to bother with the seed number at all, you can leave it out from your prompt.
🧡💛❤💛🧡
Fantastic! I know parameters, but why seed is 777?
Just a random number I use, it can be anything from over 4 billion numbers and it is usually selected randomly unless it is specified. This way I know the differences in the images come from something else like the aspect ratio or the prompt text.
i get that the bot thought unfathomable was water because fathom is a unit of measurement for the depth of water, so the bot is interpreting the sound of the word rather than its actual meaning
Abhorrent resulting in metal band covers and promo pics makes sense. a metal band with that name is entirely plausible.
and thanks to the channel, now i got to browse a thesaurus for more prompt input.
Ha, Malicious Architecture is about harmful software architecture used by hackers. Hence IT architecture diagrams of aggressive harmful software systems :D
Invincible is the name of a comic, so training data would lean heavily on its imagry.
great
Hi, Thaeyne, how to make an website like yours?
What I've done is use wordpress for the first time in my life, select a template and then try all kinds of things and plugins and try to automate posts and wiki pages as much as possible.
@@thaeyne wow that's truly inspiring. 👍
Insidious grannies are referencing a movie. :)
I didn't know that, but that does make perfect sense.
🌎🩵💛🩵🙏😇🙏✨️Amazing!!!!
Thanks for the video ..your channel monetized?
Not yet
when the word invincible tries to recreate the character invincible. lol
How can I make couples look different from each other? For example, a Korean man and a Caucasian woman? Because when I try to do it, the two end up being the same ethnicity.
Describe the woman in full details using commas to separate the description. For example ethnicity, complexion, hair texture, body figure… when you’re done with your description of her end it with a period. Next add the activity you want her doing and place a period at the end. Now you want to start next sentence with the word “With her” and from there list the ethnicity of the male, his complexion, and so on. End his description with a period as well. I hope this helps I get opposite race couples all the time. I say it’s the husband or boyfriend. Let me know how this works for you.
@@oluwadare12 thanks! I Will try
@@ArethaVGuedes You’re welcome!
@@oluwadare12 I tried and in 3 of 4 he was ginger like her. But the real problem is that both were "ugly" and I said that they were beautiful like movie star. I need really beautiful people for book cover
@@ArethaVGuedes That was so funny, I feel you though. You have to sometimes put in your prompts like a formula. Try out my prompt below. It’s a known photographer, type of shot, and desired film. Then your subject instructions. You can play with it after to tweak the look. I hope this helps, and congrats on the book☺️👏🏾👏🏾
ALBERT WATSON. DOCUMENTARY SHOT, FUJIFILM PRO 400H. BEAUTIFUL WHITE young woman, FAIR complexion, long curly hair texture, FULL - Figure hourglass body shape. Laughing with her young handsome ASIAN Husband, PORCELAIN skin COMPLEXION. Vivacious. --v 5
Very useful, thank you!
You forgetting who is writting this app, enginers mostly from India and other countries, so why in the world they would include worlds they never heard off??
Why not put the words on the video instead of just saying it? I have to stop and reverse back so to copy the "words". Tedious!! Anyway good video.
Are these arts all created by AI??!
Yes, all images were created with Midjourney.
@@thaeyne Unbelievable!
Does anyone else get the impression that she likes fantasy?
I am fluent in English and French. Most of these words are rarely used in everyday English while you hear them often in everyday French. Conclusion: learn French, you’ll improve your prompting skills.
I hate computer generated voices.
Sorry about that, unfortunately my own voice would sound worse.
thanks