The ULTIMATE Beginners Guide to Midjourney in 2024
Summary
TLDRThis in-depth video provides a comprehensive overview of mid-Journey AI art generation capabilities in 2024. It analyzes the key changes in version 6 prompting compared to version 5, explaining how users must now provide more descriptive details to achieve desired results. The video explores various parameters like stylize, chaos, and weird that alter image aesthetics. It demonstrates powerful features like shortcuts, aspect ratio prefixes, image prompting, blending, and describe. The host previews upcoming website improvements and creation mode access for all users. Overall, the video conveys mid-Journey's vast potential through evolving versions and constant updates on a date-based release cycle, predicting capabilities like consistency tools, storytelling, video generation, and 3D creation in the near future.
Takeaways
- Version 6 prompting requires more descriptive details to precisely depict the desired image.
- Parameters like stylize, chaos, and weird enable modifying image aesthetics in different ways.
- Shortcuts and prefixes streamline work by saving preferred settings.
- Image prompting and blending combine images to spark creativity.
- Frequent updates and new versions will unlock more features over time.
- The website will soon provide creation access and social tools for all users.
- Consistency tools for style and characters will better meet audience needs.
- Exciting upcoming capabilities include storytelling, video, and 3D generation.
- A date-based release cycle means constant improvements versus feature-based.
- Mid-Journey has vast creative potential still being unlocked for users.
Q & A
How has prompting changed from version 5 to 6 of mid-Journey?
-Version 6 prompting requires much more descriptive detail from the user compared to version 5 in order to generate the desired image. Users must provide precise specifications instead of broader ideas.
What parameters can alter the aesthetic style of mid-Journey images?
-Parameters like stylize, chaos, and weird enable users to modify the aesthetic style and variation of generated images in different ways based on numeric input values.
How can shortcuts and prefixes improve the mid-Journey user experience?
-Shortcuts allow users to save preferred parameter combinations for efficient reuse across prompts. Aspect ratio prefixes automatically append the desired frame size to images, removing repetitive typing.
What upcoming website and access improvements can users expect?
-The mid-Journey website will soon provide creation mode access for all users beyond the current 10,000 image minimum. Social features will also facilitate community interaction.
What new capability could allow mid-Journey to better meet audience content needs?
-The development team is working on consistency tools that will enable improved style and character consistency over multiple images to better suit user preferences.
How will the shift to a date-based release cycle impact mid-Journey evolution?
-Transitioning from a feature-based to date-based release cycle means users can expect constant updates and improvements on a regular time frame rather than erratic feature-driven releases.
What future creative features are planned for mid-Journey?
-Exciting upcoming capabilities mentioned include narrative storytelling tools, video generation, and 3D image creation, unveiling new creative frontiers.
How much potential remains untapped in the mid-Journey AI?
-Given the rapid evolution of versions thus far, mid-Journey possesses vast creative potential still being incrementally unlocked for users over time through constant updates.
What makes version 6 different from version 5?
-The key differences are more descriptive prompting requirements for desired results, additional parameters that alter image styles, and more powerful merging features like shortcuts, blending, and image prompts.
When can all users access creation modes on the mid-Journey website?
-The updated mid-Journey website will soon enable creation access beyond the current 10,000 image minimum, granting all users creative tools previously restricted.
Outlines
Differences in Prompting from Version 5.2 to Version 6
Explains key differences in how to write effective prompts between mid-Journey version 5.2 and version 6. In version 6, prompts need more detail and specificity. Provides examples contrasting prompts and outputs in the two versions.
Amount of Detail to Include in Prompts
Recommends starting prompts simple and basic, then elaborating by adding descriptive details one at a time. This allows seeing if AI understands and follows the vision. Gives example building up detail about a man's outfit piece by piece.
Adding Text to Images
Shows how to add text to images by putting words inside quotation marks. Notes text generation can be hit or miss currently but is expected to improve soon. Mentions using style raw variation for better text results.
Using Variation Buttons
Covers functionality of Very Subtle and Very Strong variation buttons. Subtle keeps image mostly the same with minor tweaks, while Strong provides almost entirely new interpretations. Also shows how turning on Remix mode allows changing prompt when getting variations.
Stylize Parameter
Explains stylize parameter determines how closely prompt is followed versus AI creativity. Lower values adhere more to prompt, higher values generate more aesthetically pleasing images that may deviate more from prompt. Shows examples across the stylize scale.
Chaos Parameter
Chaos parameter adds variety between images in a grid/set. Shows how increasing chaos progressively diversifies images generated. Recommends chaos 4-8 for some added uniqueness without going too far off prompt.
Saving Preferred Parameter Combinations
Demonstrates how to save preferred parameter combinations as shortcuts that can be easily reused across prompts. Uses acronyms for shortcuts e.g. csw1 = chaos, stylize, weird.
Image Prompting and Blending
Shows how to drag/drop images into prompt box for image prompting. Also covers using blend feature to combine up to 5 images without needing to add text prompt.
More Expected Future Features
Highlights advanced features like InPainting, zooming, panning, style tuning, and consistency tools that are expected to be added in future mid-Journey updates.
Mindmap
Keywords
๐กPrompting
๐กParameters
๐กStylize
๐กChaos
๐กImage Prompting
๐กVariation
๐กRemix
๐กUpscaling
๐กIn-Painting
๐กStyle Transfer
Highlights
Version 6 prompting is very different - you need to add as much detail as you can think of to get the AI to follow your vision.
Rewriting prompts from V5 to V6 is necessary - the old ways of prompting don't work as well anymore.
Add details one at a time to prompts to see if the AI is following along with your vision.
The higher the stylize value, the more beautiful but less coherent the images become.
The chaos value brings variety into your image grid - each picture starts looking different.
Save preferred parameter combinations as shortcuts to easily reuse them across prompts.
You can now add text to images by putting words in quotation marks.
Image prompting allows you to use existing images as prompts or blend them together.
New features like inpainting, zooming, panning, and style tuning are coming soon from V5.
Midjourney is moving to a date-based release cycle with constant updates every couple months.
Consistency tools for style and characters are in the works.
Storytelling tools will allow new ways to create narratives.
Video generation capabilities are expected in the next couple months.
3D image generation will come sometime after video.
Social features like Discord and workspace organization are slated for the website.
Transcripts
mid Journey has changed a lot in the
past year especially with this newest
update known as version six in this
video I'll give you a breakdown of what
has changed I'll explain all of the
different features of mid journey and
I'll tell you what else you can expect
in the next 12 months there will be time
stamps below the video so feel free to
click around all right first let's talk
about prompting prompting in version 6
is very different in order to start a
prompt you're going to hit forward SL
imagine let me show you an example of a
version five prompt it would look
something like this a blue sports car in
Monaco comma pink Road comma overcast
weather now the first thing you'll
notice is that it didn't really listen
to what I asked for sure we get a blue
car in some of them but no pink Road and
then in images one and two we actually
get a pink car even though we
specifically asked for a blue sports car
this was mid Journey capabilities all
the way up until version 5.2 just a
couple of weeks ago now you could write
this exact same prompt in version six
and it will look something like this we
definitely have a blue sports car now on
a pink road but this isn't exactly the
best way to speak to the AI instead in
version 6 we would write something more
like this wideangle product photo of a
sports car in Monaco the sports car is
bright blue the road is painted pink the
weather is overcast it's not very poetic
but look at the results the details all
get picked up by the bot that's how
powerful mid Journey version 6 is
another result of the same prompt just
look at number one here look how pink
that road is now it might take some time
to get used to prompting like this where
you have to sort of add as much detail
as you can think of any details that you
do not mention are going to be filled in
by mid Journey let me show you one more
example of the difference between
version 5 and version 6 here's a prompt
from version 5.2 a horse running through
the field oil painting Double Exposure
glitch core aggressive inertia and it's
going to create something beautiful like
this version 5 was amazing in its own
right but if you use this prompt in
version six you would get something like
these first of all it doesn't really
look like how it did in version 5 and
these aren't very great I mean I don't
like the look of them so how would you
rewrite this prompt in version six well
in my opinion I would write it similar
to this an oil painting of an aggressive
horse the horse is running through the
field with intense inertia there is a
double exposure effect with a glitchy
nature and I think we get results that
look a lot more like what we wanted in
version 5.2 I'm just trying to show you
that rewriting a prompt is necessary
when transferring your ideas from the
old bot to the new bot and if you're new
here my name is Nolan and it's my goal
to teach AI in a digestible way I want
to make this easy on you and if you
think I'm helping so far please consider
leaving a like on this video so we can
share it with with more people now how
much detail should you add to your
prompts well if you've seen my previous
introductory video on version 6 this
process might seem familiar but we're
going to go through it again here what
I'm suggesting you do is start small
start with a very simple prompt and then
elaborate add as much detail as you can
think of but consider adding these
details one at a time so you can see
whether the AI is following along with
your vision I want to thank a YouTube
comment I got from Alec here I don't
want to mispronounce your name but they
said if you wanted to test the coherence
of a prompt start at the top and work
your way down so here we have the idea
of a full body shot of a man wearing a
Celestial Crown beautiful work so far
mid Journey but what else can we add
well honestly anything you can think of
and I'm going to stick with colors for
now just so you can clearly see the
differences so we have a full body shot
of a man wearing a Celestial crown and
blue sunglasses perfect mid Journey
Nails it now you have to notice that
they're not the same crowns as before
and that's because anytime you add to
your prompt the whole generation changes
this example is just to show you that
mid Journey will listen to each detail
you add and I'm also showing you here
that I like to rate these prompts pretty
matter of factly no poetry is really
necessary I ended the first part with a
period and then I say the man has red
earrings sure enough we have everything
we've asked for so far but we can keep
going the man is wearing a green sweater
mid Journey Nails it different look of a
person though that's for sure I went on
to combine the next two details in one
sentence the man is wearing blue shorts
with a red belt again I'm just using
colors to emphasize all of the details
you can add but this really is Limitless
you could think of anything to put in
your prompts I added that the man is
wearing golden shoes didn't show up in
number one but we have gold shoes in two
three and four the more details you add
the less of a chance you're going to get
a four out of four perfect grid and
that's okay honestly I think even 25% 1
out of four is going to be good enough
for everybody all you need is one
perfect picture I even added afterwards
the man has pink socks on now we have
everything we've asked for Golden Shoes
plus pink socks that is so funny and to
be completely honest I think you could
add details as into what type of
clothing they are maybe pattern socks
maybe you could specify the material on
the sweater I'm telling you there are
kind of Endless Possibilities now and
it's only going to get more powerful as
time goes on so try your best to start
thinking in this new way of prompting I
then added in an extra accessory the man
is standing on an all-white skateboard
in number three and four the proportions
of the body aren't quite right but I
think number one and number two look
okay and we have a white skateboard next
I set my sights on the background and I
wrote it very simply the back background
is a decrepit space station and the
first thing I noticed at least is that
some of the generations turned into a
cartoon and I have found that that
happens with the more details you add so
now what you might want to do is add
some specific terms that reflect a photo
realistic environment I changed the
background part of the prompt to a
decrepit old spooky cinematic space
station and we get these which are all
present in reality but now I want to
show you that there are different ways
to write The Prompt there is no exact
formula I showed you how to do it one
detail by one detail but you can mix up
the way you express those details like
take a look at this first sentence I
turned it into a full body cinematic
still of a man standing on an all-white
skateboard wearing a Celestial crown
blue sunglasses and red earrings the
exact same idea just worded a little
differently you could even try mixing up
the order of the details in this case
the background full body cinematic still
of a man standing in a decrepit old
spooky space station he is wearing the
crown the glasses and the earrings he is
wearing the green sweater the blue
shorts the red belt he has pink socks
underneath golden shoes and he is
standing on a white skateboard and look
at number two here like that that's what
we described that's how powerful mid
journey is and these were just examples
of showing you a lot of detail plus
showing you that you can write it in a
few different ways again there's no
right answer be as descriptive as you
can or mid journey is going to fill in
the blanks for you I have a few more
examples of different prompting
techniques but first let me show you how
to add text to your images in mid
Journey version 6 you can add text to
your images by putting words in
quotation marks so here we have an old
timey computer screen that says hello
world don't worry about C4 for now we're
going to get to that in the parameter
section what I want you to take away
from this is the fact that you can
express certain words inside of your
Generations now number one worked really
well number three kind of didn't work at
all text is a bit Hit or Miss at the
moment first it's way easier to
accomplish with short words however I do
want to point out there is an expected
update to the text generation coming
very soon probably by the time you're
watching this which means that any
limitations you see now might not be a
problem in the future just keep that in
mind the main point is that you can now
add text to your pictures by placing
words inside of quotation marks in your
prompt one little extra trick that you
might want to try with regards to text
and honestly any prompt is by using Dash
Dash style raw this is a variation of
the mid Journey bot that will have a
more literal interpretation of your
prompt style raw is recommended in two
situations first is when you've written
a longer prompt you know what you want
and you want mid journey to listen to
you style raw has a good chance of
getting you there the second time you
might might want to use style raw is
when the base mid Journey bot isn't
working it's just not listening to you
there is no harm in trying style raw and
in this case when generating text style
raw might give you better results it's
sort of the first thing I experiment
with when I'm trying to figure out a
promp here's a cool example of text a
neon sign on a dark black wall
background it didn't include everything
I wanted I wanted that little heart
symbol at the end and it just couldn't
fit it in although mid Journey worked
out pretty well here and it's not quite
front-facing but maybe I just didn't
describe that in the correct way and I
wanted to show you just how cool this
generation is that the J extends below
turning into the V for the version 6
like that's that's art right there that
is so cool if you want more tips on
generating text I suggest you check out
one of my videos I think it should be
maybe right here but make sure you come
back to this video after we have a lot
more to go over and with regards to text
the more complicated your prompt is the
less likely it's going to to work for
now again these limitations might be
erased in the future a beautiful woman
with mid Journey tattooed across her
forehead kind of V6 is tattooed on her
left cheek kind of and a heart symbol is
tattooed on her right cheek now with
version 6 you can specify where things
appear in the picture however maybe I'm
getting confused I was under the
impression that when we say left we're
getting it from our perspective so her
left cheek would technically be her
right cheek you know if she was facing
us you know what I mean but in this I
asked for V6 on her left cheek and it
interpreted that as her real left cheek
I don't know if that's a glitch I'm
pretty sure it's all from our
perspective so maybe this complicated
nature of the prompt just made mid
Journey bug out a little just wanted to
point that out and now the last thing I
want to mention to you with regards to
prompting is what to do when the bot
just isn't quite getting what you're
asking for here's a prom from version
5.2 we're going to take that sports car
in Monaco same idea as before but at the
end we want some double exposure we want
some glitchiness and we're going to get
results like this some pretty cool
looking images the details like the
color of the car they're not really
important here I'm more worried about
the overall aesthetic and if we Tred to
generate that same effect in version six
by using words like there is a glitchy
Double Exposure effect on the photo it
doesn't quite turn out at least how I
had in my mind so what do we do we're
going to use multi- prompting multi-
prompting is quite powerful in mid
Journey version 6 maybe a little too
powerful right now it involves a process
of splitting up your prompt by using two
colons and then adding in different
weighted values so the first part would
look something like this there is a
double exposure effect with glitchy
elements on the photo colon colon and
then we're going to use this sort of
call back effect where we're going to
repeat what we wanted similar to the
cowbell method or maybe the slider
method as it used to be known in version
5 similar we're basically going to
repeat what we want and add a small
weight to it in this case it's the words
glitch core Double Exposure 0.5 as a
weight when you don't specify a weight
like for the first part of our prompt
the default value is one if we would
have made this value 1.5 or anything
above one glitch core Double Exposure
would be more important to the bot and
we don't want that we just want a little
bit of an accent a little bit of salt a
little bit of spicee
and we're starting to get some results
like these which maybe is a little
closer like here in number two we're
starting to get some sort of effect so
you could bump up that value a little
glitch core Double Exposure 0.8 and like
look at number three look at number four
that is so cool and that's the power of
multi- prompting when you're not getting
what you want you can add a little bit
more of it to the end with just a little
bit extra weight it's really powerful
but it works in these specific
situations now what happens when you
find an image you really like what
options do you have well let me show you
we'll take a look at this prompt Chrono
Trigger character concept pixel art
scene of a galactic Knight in sparkling
armor I was a big fan of image number
two I'm going to hit the U2 underneath
the generation this stands for upscale
which means it's going to send us a
singled out version of this picture boom
right here and now what options do we
have well for version six at the moment
we're given four options underneath the
picture first I'll explain the variation
buttons we have very subtle and very
strong now if you were to click one of
these variation buttons mid journey is
going to send you four more images based
on this picture very subtle we'll turn
into these which is like 89% the same
image a night gold armor red cloth
nighttime background two columns on the
right side some sort of a ruined castle
and that's a very highle inter
interpretation of the image but the
image itself is different you're getting
four different subtle variations on the
image you chose and you're going to want
to choose this when you're mostly happy
with a picture but you want to see what
small tweaks could be made to it this is
also good for creating like cool
slideshows on social media where it's
just one small change after the other
but back in our image if you click very
strong this is what you'll get now these
are quite different okay maybe like 85%
the same Maybe 80% we have the same idea
of a nighttime background some castle
ruins golden armor red cloth but the
poses are a little different like in
number three he's looking a completely
different weight so this is known as a
strong variation because most of the
time it's not going to resemble the
original picture it's going to be quite
different as different you can get while
keeping the general ingredients the same
does that make sense this option is
really powerful for fixing small
imperfections in a picture you're
basically giving the bot a redo and it
can often create something you like a
lot more than the original but we are
not done with the variation buttons If
you go into your settings by typing in
forward SL settings into the prompt box
on Discord you hit enter first of all I
want to point out that you can select
the subtle or strong variation to be
your default from here in this case I
have high variation mode selected which
means strong variation and now if we go
back to the original grid instead of up
scaling number two if I hit V2 variation
for image 2 I'm going to get a strong
variation of that picture if you go back
into your settings and you put low
variation mode on and then you hit V2 a
subtle variation will be generated every
time if you're curious I suggest
sticking with strong variation but as
always you can upscale the image first
and then choose which type you want from
there just had to let you know now I
don't want to confuse you but there is
an extra thing we can do with with the
variation buttons back in your settings
if you click this button right here
remix this means remix mode will be
turned on and this means every time you
hit a variation button a new prompt box
will appear and you have a chance to
regenerate that image with a change
prompt let me show you how it works so
variation buttons will be blue when
remix is turned off but when remix is
turned on and we hit very subtle The
Prompt box will appear and we can change
this prompt however we like you can
erase the whole thing but I don't
necessarily recommend that in my opinion
you should just add to the prompt so
we'll say a galactic Knight in sparkling
armor standing in a field of flowers
you'll hit submit and we'll get
something like these a subtle remix of
that image it's not the same as
variations remix definitely changes the
picture we are not getting the exact
same night in these images but we're
able to change things and you can do
this for time of day you could do this
for eye color honestly any detail you
want to add and the bigger details you
do change can often result in some of
the coolest images you'll ever see
there's just really no Rhyme or Reason
you're going to have to experiment on
your own but I want to point out that
the results will differ if you use very
strong we'll do the same thing Galactic
Knight standing in a field of flowers
you hit submit and we get these maybe
similar looking kns but quite different
results from the original like I said
you can take remix in a bunch of
different different directions this was
just to show you that you can access the
remix feature first through your
settings and then through the variation
button each variation button has two
jobs one without remix and one with blue
is without green is with I hope I'm
making that clear now the other option
you have beneath one of your upscaled
images is two more upscales again this
may change in the future you might see a
2X upscale or a 4X upscale but for now
at the launch of version 6 we have
upscale subtle and up scale creative let
me show you both clicking upscale subtle
will turn this image into this image
just a higher resolution version of our
picture you'll want to choose upscale
subtle when you're really happy with
what you already have however if there's
any wonky parts of your photo you might
consider hitting upscale creative and
that will turn into this it's going to
change your image I mean obviously it's
the same but you can clearly tell that
they're different it made changes to it
and again there's really no rhyme a
reason for what it's going to do it's
basically just going to reinterpret your
prompt again and it's going to make some
corrections now because you can't
control these Corrections I'm not sure
that a creative upscaler will always be
your first choice but just keep in mind
it's there if you want to experiment
this will raise the resolution of that
image by twofold or maybe even more now
if you've learned something so far I
have something special to tell you about
for the last 6 months I've been working
on a mid Journey Master course where
I'll teach you everything I know in a
more structured setting I'm not sure
when it will be completed and available
to you but if you're interested in an
exclusive discount you should consider
signing up for my newsletter I'll leave
a link to that in the description below
okay now we get to move on to the fun
stuff parameters what are they well
there are things you can add to your
prompt that will change the outcome
things that will move you away from the
default look of mid journey and the
first parameter we're going to focus on
is the stylize value I use this a ton
the stylize value basically deter
determines how beautiful your image is
going to be and I know that might sound
simple but it's a little more
complicated than that the styliz value
is a tossup between how much the bot
listens to your prompt and how much
creative freedom it has the values go on
a scale from 0 to 1,000 with the default
being 100 if you don't mention a
stylized value it's going to be 100 if
you place a value lower than 100 mid
Journey will follow your prompt more
closely and if you raise the value above
100 you're going to get more beautiful
images that might come at the cost of
some coherence or prompt accuracy let's
go through a bunch of examples I'm going
to show you this prompt in all the
different values a cartoon bear dressed
like a firefighter the bear is wearing a
red firefighter uniform there is a small
yellow duck in the right corner the duck
is wearing a blue police uniform no
mention of the stylized value so the
default is 100 and this is what we get
some cool looking pictures I love these
now you can trigger the stylized value
by adding dh- s and then the number in
your prompt like right here I have it
highlighted and we moved it all the way
down to D- s0 and what's the first thing
you notice that these aren't very
goodlooking the default is 100 for a
reason it's the perfect blend of
listening to you and creating beautiful
pictures a lower value style Z is going
to get you some ugly images it's going
to follow your prompt more closely but
you're going to lose some of the
aesthetic beauty that mid Journey can
bring to your prompt you're really only
going to want to lower the stylized
value when you find that mid Journey
just isn't listening to you at higher
values I hope that's making sense same
prompt at stylize 25 I think these are
still on the ugly side of things but
it's around s50 where I think you'll be
comfortable you can still get some
pretty cool images this way and then we
have s75 just to show you what it looks
like as you approach the default once
again I'll show you the default here
S100 but now take a look at what happens
as we raise the value here it is at 200
and you can already notice some straying
from your promt in terms of the duck no
longer having a police uniform on in two
of these pictures 300 is very
interesting number one and number three
is what I'm starting to expect from the
higher values a much more opinionated
mid Journey like I asked for cartoon and
it's giving me these sort of 3D graphic
look but number two and number four are
just beautiful cartoons so it's around
this value where you're going to start
to see the biggest differences here we
have S400 and it's honestly starting to
look like some sort of real life movie
like these are beautiful you can't deny
that 500 we definitely lost most of the
police uniform like look at this Bear
right here that's so cool these are just
gorgeous 600 yes mid journey is ignoring
most of our prompt except for the main
character the firefighting bear and it's
placing it in these exact same scenarios
what I mean is moving away from the
cartoon effect into this more 3d viby
effect the higher the stylize value the
more you see what midj Journey thinks
you want 700 look at number two that wet
fur that's so cool 800 again a very
similar look in all four of them 900
more of the same I love number four here
and 1,000 look there's a police duck in
the corner I was not expecting that I
think number four is really beautiful I
think they're all really beautiful now
am I suggesting you use a higher
stylized value honestly no unless you
don't know what you want and even if you
do know what you want maybe a higher
value is just fun to experiment with to
see what mid journey is thinking on its
own but if you're looking for prompt
accuracy you're going to want to stick
to the default or consider lowering it
even more all right now we're moving on
to my favorite parameter the chaos value
you can trigger it by adding D- C to
your prompt and then a value between 0
and 100 the default is zero keep that in
mind and we'll use this ProMed as a
template vintage Polaroid photo of a
small alien dressed like Optimus Prime
very cool mid Journey it listened to us
it listened to what I wanted I like
these images and one thing you'll notice
off the start with no chaos value is
that they all look kind of similar and
what the chaos parameter does is it
brings variety into your grid meaning
each of these pictures is going to start
looking different than one another and
what are the costs of raising the chaos
value honestly chaos a little bit of
Chaos chaos is going to get you some
really cool pictures a lot of chaos and
you have no idea what to expect let me
show you same prompt at chaos 4 and
honestly it's around this digit that
you're going to want to stick to it just
adds a little bit of spice to the
generations while still maintaining what
you wrote in your prompt I'd say even
going up to chaos 8 might be okay in
certain situations it's fun to at least
experiment with but you'll see in number
two we're already moving away from what
we asked for I don't really think this
is an alien dress like Optimus Prime but
it is a small being at least maybe and
at chaos 12 12 out of 100 is honestly
too much these are amazing images I
think they look so cool but you'll see
that we've just strayed so far from The
Prompt that is this really worth it I
don't know to me it is I like seeing
what mid Journey can create but I I
don't recommend using this very often
you might be more frustrated than normal
chaos 16 is pretty hilarious like look
at this little guy down here I'd say
it'd be hard to get these images on
their own and the cool thing about chaos
is that you can just add chaos to any
promp and get some really varied
Generations like these 20 love number
three that's so funny we'll jump ahead a
bit here to 25 what can I say except uh
except laugh honestly you're barely
going to get what you asked for but
they're still pretty cool to look at
chaos 30 we get a vintage Polaroid but
no real sign of Optimus Prime although I
do like number one that's that's a cool
Transformer picture 35 we straight up
just get Optimus here and then I have no
idea what's going on in these 40 number
two is cool look at this reptile alien
thing like that is hilarious 45 love
them and just take a look at how
different they are from each other
that's the key point of the chaos value
chaos 50 you'll still see Optimus Prime
come out in some sort of way in these
generations and then you'll get
something like number four which is just
just somebody in a field like okay like
I said I don't recommend anything over
like chaos 8 probably at the most 55
number four here is awesome 60 we got
two Optimus that's good I guess 65 only
one slight Optimus picture and the rest
are nonsense 70 is kind of cool I like
number one 75 one clear example of a
transformer and more nonsense chaos 80
notice Optimus is number to and then
look how strange the other ones are like
what does number four have to do with
anything that is how much variety you
can get inside of your grid I'm not
suggesting you try this this is throwing
darts with a blindfold on completely 85
still two pictures of Optimus that's
kind of cool chaos 90 just making me
laugh 95 I like these a lot especially
these stacked Stones here in a field and
Chaos 100 somehow Optimus shows up twice
one is in a hotel room those are just
some examples of what a chaos value you
can do to one prompt and feel free to
use this at your own risk it is one of
my favorite parameters like I said
though and the next parameter I want to
show you is called the weird parameter
you can trigger this by using d-w inside
of your prompt and then a numeric value
between 0 and 3,000 zero is the default
by the way and what does the weird
parameter do well it just makes your
images look more weird it's honestly the
perfect way to describe what happens I
will note that it sticks to your prompt
a lot more closely than the stylized
value and definitely more than chaos but
I don't know how often you'll want to
use the rear parameter let me just show
you the examples we're going to start
with the base prompt a cyberpunk dolphin
made of mechanical Parts swimming in the
fluorescent Caribbean Coral I think
these turned out great but what happens
when we add dash dash weird to The
Prompt at just w50 we're already getting
these and again I don't know how you
would describe them other than more
weird weirder you know weird 100 I like
number three that's sort of reminds me
of a disco ball 200 you can still get
some cool pictures 400 look at number
two that's hilarious and I like number
one a lot but these are weird right 600
H I don't know why you'd want to be
getting this high and keep in mind it
goes all the way up to
3,800 honestly the different parts of
the picture are starting to like blend
together in my mind 1,000 and maybe one
out of four is appropriate I like number
one and then the rest are just hard to
tell what's going on we'll jump ahead a
little more here's 1300 I like number
three 1,600 reminds me a lot of Lego
like literally this is oh man these look
so funny 1900 and these are starting to
get really difficult to look at the
color palette just awful is a good word
weird is a good word we got lucky with
2100 and I think these honestly look
more normal than the prev previous one
2400 and as you can see there's really
no way to know what you're going to get
with the weird parameter 2700 I like
number three a lot that light coming off
of the tail that's that's really cool
and then weird 3000 the highest value I
don't even mind these that much but I
can't in good conscious suggest using
the weird parameter inside of your
prompts although you could see that it
still sticks to your prompt fairly
closely but the results are just not
ideal they're not beautiful in any way
they are weird however I do think it's
fun to combine the parameters I
previously showed you that being a
styliz value a chaos value and a weird
value there is no perfect formula you
might try something like this chaos 18
stylize 700 weird 450 and I used the
prompt futuristic sports car on the moon
I think number one is amazing I like
number two a lot that's a fun little
combo of parameters to try or you might
might try portrait of an Instagram model
old Hollywood style surreal nature with
the combo of chaos 6 stylize 80 and
weird 35 some low values just to add an
extra element of spice to the pictures I
think these turned out great but now
what do you do when you find a recipe of
parameters you like you'll want to save
it as a shortcut and here's how you do
that in the prompt box on Discord you're
going to hit forward slash prefer and
then we're looking for for prefer option
set when you click on an option box
you'll see a list of shortcuts you've
already created we don't want to deal
with that for now right now in the
option box we want to create the name
for our shortcut what do we want to type
into our prompt that will then expand
into the parameters we're looking for I
don't know about you or how your memory
Works keep in mind you can name this
shortcut anything you like for me I like
to sort of stick to acronyms so if I'm
doing a chaos value a stylized value and
a weird value maybe I'll just call this
like
csw1 so you type that in here and the
next important thing to do is click on
the plus one more that appears beside it
this is going to open up another option
above called value we're going to click
on value and then down here back in the
prompt box we're going to input what we
want the shortcut to evolve into this
being the combination of parameters we
just found now when I hit enter it's
going to set the shortcut of
csw1 to expand into the values we set if
I were to type in and prompt for a cute
dog the way you trigger the shortcut is
by using dash dash so we'll say a cute
dog Dash Dash
csw1 when I hit enter you'll see that it
expanded into our values so while you're
experimenting and you find something
that you really like you save it as a
shortcut so you can easily experiment
with it across all of your prompts and
look how cute these dogs are oh that's
amazing and have you noticed that I
always get my Generations in a wide
aspect ratio 16x9 but I've never typed
that into the prompts well that's
because I have it saved as a suffix and
here's how you can do that yourself back
in the prompt box you're going to hit
forward slash prefer and you can type
out prefer suffix or you can find it
here in the menu you're going to see a
new value pop up above click on it and
then down here in the prompt box you can
write whatever you like in my opinion I
think this is best to use for aspect
ratios so I know that I use 16 by9 a lot
I can put that right here and then I
never have to type it into my prompts
mid Journey will add it at the end of
each prompt itself you could add words
you commonly use you could add styliz
and Chaos values again I recommend just
sticking to an aspect ratio and the
reason why I like it for an aspect ratio
is because it's easily overridden in the
sense that I could type a cute dog
aspect ratio 2x3 and now even though
16x9 is at the end of all my prompts
because I type 2x3 it will appear first
and take priority and that's how we end
up with cute dogs like these I hope
you're not sick of mid Journey just yet
there are a few more things you can do
inside of version 6 that are extremely
powerful one of them that I want to talk
about is image prompting the easiest way
to do this on Discord is to click this
little plus button beside the prompt box
and upload a file you can choose the
files off of your computer I chose these
to experiment with hit enter and Discord
will throw them into a little grid right
here for image prompting we're going to
want to take this picture grab the link
and place it into the prompt so mid
Journey knows what we are referring to
one way you can do this is by clicking
on the picture once it will expand now
you can rightclick copy image address go
into the prompt box and hit paste contrl
+v command V alternatively you can just
click and drag the picture into the
prompt box that's probably way easier
and what you'll want to do most often
now we can hit space and we can even
drag another picture down here and then
the beauty of image prompting is that we
can add some words I have no idea what's
going to happen when I type in the CN
tower a Toronto skyscraper I have no
idea what's going to happen we'll hit
enter and then maybe we can take this
picture of the car and say a dragon
flying through the sky here's that
combination with the CN tower some
pretty cool pictures I really did not
know what to expect and then here's the
dragon flying through the sky from that
picture of the car like that is
hilarious now I'm not going to go very
in-depth into image prompting here there
is a lot to know about it a lot of
different things you can try and if you
really want to learn more maybe that's
when you should check out my course
again sign up for my newsletter in the
description below if you want to be the
first to know about it there are still a
couple more things that I can share with
you here though now if you want to just
combine a couple of pictures without
adding extra words you can hit forward
SL blend that will give you the option
to blend up to five images at first
you'll just see one and two but you can
click down here at plus four more and
that will create a popup up here where
you can choose images three four and
five as well as change the dimensions of
the blend let's do image three and then
now we can pick any pictures from our
computer I'll choose that Spartan we'll
choose that scary guy and then we'll add
some color there what do you think this
is going to create blending is a lot
like image prompting it's basically the
same thing except you don't have the
chance to add words to The Prompt that's
really the key difference blend goes a
lot quicker image prompting might have
more options available to you oh my God
these are so cool that is basically a
blend of all of those pictures I think I
like number one the most and the last
thing I want to show you in version six
for now is the describe feature you're
going to hit forward slash describe this
will give you the chance to upload an
image we'll try this Spartan again and
when you hit enter mid journey is going
to give you four prompt descriptions
based on that picture now this is an old
version of describe this was made back
in version 5 but we should be expecting
an update to this feature sometime soon
so SL describe is going to get a lot
more powerful moving forward and that's
why I think you should know about it now
you can imagine all of them you can
click on any of the prompts that you
like in fact you can even click on them
and tweak The Prompt yourself a little
side note an extra special tip that I'll
share with you here if you really want
to recreate a picture you would run it
through the describe function you would
then click on the picture copy image
address and then we're going to paste
that inside of the described prompt that
will add an extra reference image to the
description super powerful stuff oh like
these are so beautiful I basically
showed you everything that's available
in version 6 at this moment but there
are a lot more features we can expect
coming soon in order to show you these
I'm going to take you back to version
5.2 here we have the prompt moon goddess
of the ninth realm we'll choose one of
these to upscale and then we'll see a
lot more options underneath that image
first first of all like I mentioned
earlier we might see these upscales in
the future that being a 2X and a 4X but
I really want to point to all of these
buttons down here why don't we start
with vary region it is another variation
button but this is specifically in
painting I'm sure you've heard that term
before if we click on this a new box
will appear where we can either take the
rectangle selection or the lasso tool
and we can change individual parts of
this image we can basically ask for an
entire L new face we could change her
accessories we could try to add
something in this Blank Space over here
imp painting is a super powerful feature
it's just not available yet at this time
of recording but it's on the table
they're working on it we'll see that in
version 6 soon here's the generation of
her that changed her face you see how
powerful it could be other than in
painting we also have the chance to zoom
out we can zoom out at a 1.5 scale or a
2X scale now those are pretty simil ilar
I'm not going to lie you're not going to
see that big of a difference but this
works amazing when mid Journey doesn't
quite get your framing right especially
in this case maybe it cut off her hair
maybe you weren't happy with that and
you want to see more of the image here's
what the 1.5 would look like and here's
what the 2x would look like but we
aren't done there you can also use
custom zoom and the prompt box will
appear so the zoom value isn't going to
change very much I don't think you can
raise it too high I don't think you can
raise it past two to be honest with you
but the main point of this is that you
can change your prompt so that when you
zoom out you can change what gets added
around the image pretty cool pretty
powerful in its own right but if this
ever comes to version 6 here's why you
would want to use it we're going to want
to change the zoom from two to one and
then we can change the aspect ratio to
anything we want let's try 2x3 and what
that's going to do is stay focused on
this image but it's going to change the
aspect ratio like this which means we
might see more on the bottom and the top
it's a different way of zooming out and
it's the only way to change your aspect
ratio change the frame of the
composition and save any image you've
ever made and then we come to this
option make square and this is just
maybe ideal for Instagram easily make
any image into a square aspect ratio you
can do that from Custom Zoom for sure
but here's just a quick button for you
there you go what do you think but we
are not done in 5.2 and in the future
you should have these options as well
these arrows down here trigger the
panning feature if I click to the left
it's going to take this image and expand
the frame to the left and create
something new we can do that to the
right we can do that up and we can do
that down now I do want to note that if
remix is turned on again you can find
that in your settings or you could type
in prefer remix into the prompt box hit
enter and that will toggle the remix
feature now if you were to hit pan in a
direction a prompt box appears and you
can change the prompt we could say the
legs of an octopus now this doesn't work
very well in version 5.2 but you should
know that this feature is available and
will probably be way more powerful when
it comes to version 6 here are the
results of panning left we get a larger
view of space here it is panning to the
right and we get some more of whatever
this background circular device is
panning up and that's kind of cool I
guess I like where her hair is going
here and panning down we're going to get
to see more of the character wo and then
panning down with the remix feature
enabled the legs of an octopus we're
going to get some tentacle like stuff
going on here that is pretty cool now
there is one more really cool feature
that used to be available in mid journey
and you should be excited about seeing
this come to fruition in version 6 and
I'm talking about style tuning if you
type in forward SL tune into the prompt
box you're going to be given the chance
to write a new prompt in this tuning box
you could type in literally anything you
want here let's just try something like
a tropical Sunset it's going to create a
test for you a tuning test where mid
journey is going to take that prompt in
a bunch of different directions and you
get to choose which style best suits
what you were looking for you can choose
between 16 32 64 and 128 directions I
recommend just sticking with like 16 to
start until you get used to what you're
about to do we'll stick with the default
mode hit submit are you sure yes this
can get quite expensive just want to
point that out I have a whole video on
tuning maybe I can link it up here if
you want to check that out but come back
to this video cuz we are not done yet
once the test is done it's going to send
you a link and now this is going to open
you're going to ask to choose between
Pairs keep in mind you don't have to
make a selection on every pair if that's
sort of the fun with it but let's say
you like this style here and then maybe
this style down here oh but you like the
minimalism scene here and scrolling down
you really like this plain color palette
here you're going to see that a code
gets created at the bottom and this code
is dynamically updated anytime I make a
selection it gets changed now you can
copy the code here take it back in a
Discord make sure version 5.2 is set and
then you're going to see this amazing
outcome get created just using a style
code style tuning is so powerful it's
one of my favorite features ever
unfortunately like I mentioned it's not
available in version 6 at the moment and
when they do make it available it's
probably going to look and act a little
differently than what you just saw but I
wanted to show you what you can expect
in the future keep an eye out for style
tuning it's the best I do also want to
mention that if you're interested in
anime Aesthetics you should be checking
out nii journey n ji this is a sort of
subsidiary of mid journey in a sense
basically it's a model that has been
focused trained on anime you can create
similar Aesthetics in mid Journey but
nii journei is all anime we should be
expecting version six of nii journei
sometime soon in January but for now
I'll just give you a comparison between
mid Journey version 6 and M5 here's the
prompt Gundam Mecca in version 6 and
this is what we get kind of cool but if
you were to put Dash Das ni5 in the
prompt you're going to get these this is
the default nii look and what's really
cool about what this team accomplished
is that they added a bunch more styles
that you could choose from what I mean
to say is that if you added Dash Dash
style expressive with the same prompt
this is the type of look you would get
you could also try style Scenic for
these more expansive compositions lots
of background detail and lastly we have
style cute which is clearly focused on
creating the cutest images possible now
I'm not saying all of those Styles will
be available in version 6 I don't know
if that's the case but this kind of idea
is something that we can look forward to
in the future the next thing I want to
show you is mid Journey's new website it
went through a complete redesign in the
last couple of months and it is quite
impressive you'll notice right here at
the top imagine is coming soon we're not
there yet but the rest of the mid
Journey homepage will look like this the
explore page where you can browse the
most popular and recent Creations from
the entire community in fact you can
even search for
prompts here's every prompt involving
Batman this goes on forever you see how
fast that was that search is like
instantaneous if you go over here on the
left to my images you'll be presented
with your entire gallery in order to
download you have a few different
options you can select all next to a
specific day and download everything
that you created on that day or you can
simply drag and select photos that way
alternatively you can hold down shift
and deselect or select new pictures now
I'm pretty sure they will be adding an
archive feature soon where you can look
through what you've created on any day
ever however until then if you want to
search through your prompts you'll click
up here and let's try looking for the
word Neon do you see how quickly that
populated every time I've ever entered
the word neon in a prompt it shows up
that quickly and you can just keep
scrolling and scrolling and scrolling
and it's all there just for you super
fast this new website is amazing another
feature they have is rate images over
here on the left if you click that
you'll be asked to choose which image
you like more left or right one or two
rating these images helps tell the bot
what we find most beautiful and we can
sort of correct the course of the entire
model now if you're one of the top 2,000
people to rate images in a single day
you'll be given one free hour of GPU
time for fast Generations it's a really
good deal it helps mid Journey out and
you get a nice reward out of it now I
know I said creation will be coming soon
but I can give you a sneak peek of what
that looks like if you've made more than
10,000 images you have access to the
alpha mid Journey website and here is
where creation is enabled let's prompt
for an underwater rainbow all of your
settings will be over here we can make
it in portrait landscape we can adjust
the slider anywhere in between you
remember stylization weird and Chaos
known as variety here we can bump that
up and down you can use the standard
model or raw mode change which version
you're on and whether you're generating
in private or not we hit enter and it
will create our images for us over here
these look okay maybe not the best
prompts but whatever you get the point
image prompting is actually really easy
to do from the website all of your image
prompts will appear here and then you
can drag and drop as you see fit let's
just blend those two together oh and
it's still has my styliz value the weird
and the chaos oops let's see what it
turns into here we go these are amazing
those are gorgeous now I want to talk
about what we can expect in the future
as far as updates to the website is
concerned we can expect creation to be
enabled for everyone pretty soon maybe
by February the website will also be
updated with social features similar to
Discord a place to communicate create
with each other and receive help we'll
also have organizational features like
folders Collections and personal
workspaces then as far as overall tools
that we can expect from mid Journey
they're working on things like
consistency tools that's for both style
and character consistency they know
their audience wants it and they're
working on it we can also look forward
to some storytelling tools it's hard to
say how that will actually be visualized
but the team sounds excited about that
Prospect furthermore there will be some
sort of video generation coming in the
next couple of months and 3D generation
probably sometime after that right now
version 6 was just launched but we
should expect version 7 sooner than
later I mean within a couple of months
and a version 8 after that and I think
it's likely we see a version 9 before
the end of 2024 I'm confident in saying
that because of mid Journey's new
strategy and that is a date-based
release cycle rather than a feature
based release cycle we are going to see
constant updates more often there's
everything you need to know about mid
journey in 2024 I hope you found this
video useful leave a like on it so we
can share it with more people subscribe
so you don't miss all of the tutorials I
have planned if you want access to some
exclusive prompts you can support me
over on patreon I'm Nolan I hope you're
doing well take care and I'll see you
next time peace
Browse More Related Video
5.0 / 5 (0 votes)