🙋│image-help

[hardy sentinel]

You need to do it using desktop computer

To change just the color i don't know

[night vault]

use describe by ai, it will generate prompt close to what you have then use phoenix model which can improvise using instructional command like “change hair to blue” or “make a cat character instead”

[real lynx]

Very poor quality of your image

[grizzled swift]

Where can I find the describe by ai?

[night vault]

https://media.discordapp.net/attachments/1170046881183834192/1307803348501532715/Screenshot_20241117-165058-805.png?ex=673cf3c8&is=673ba248&hm=36705dd127e6e0e8d924a085f52e1f1d85592b67e77c3e07dd28045c3a4a4c90&

[hollow breach]

You could use either edge to image or/and depth to image to keep the drawing. And in the prompt you say something like a boy wearing a colour that is not green

[manic root]

Hey Leonardians! Checked out the daily theme yet? Jump into #1108130218260041758 and join the excitement! Remember to cheer on others by voting with the Leonardo emoji . Awesome prizes await! Have a blast!

[marble cape]

Hey so to fix the back color on my image is on real time canvas? First time using it

Don’t know if I’m doing it right

This is the color I was using

Or this color

How do I fix the background quick

[unborn mauve]

Is it possible to change the colour of clothes or outfit on Leonardo ?

[stable carbon]

You want to just change their colour? Yes. Thats possible with the Canvas Editor.
Theres at least two guide videos on it on the community youtube channel.

[stable carbon]

Define 'fix'.
What are you trying to do?

[marble cape]

To add full background to that image

Yeah I don’t know

To my original image on the left

[hardy sentinel]

it will change your image

[hardy sentinel]

I don't know what do you mean with fix ?

[marble cape]

The full color of the background I mean

[hardy sentinel]

this is my result

[marble cape]

Oh that looks nice

[hardy sentinel]

it is not guaranteed to work, you can use inpainting

[marble cape]

Like dark spirits in the pot gives vibes

[hardy sentinel]

yes

[marble cape]

Maybe those question mark symbols

Upside down question marks

Symbols

[hardy sentinel]

I don't understand

[marble cape]

Those upside down question marks on my original image

[hardy sentinel]

the image above is a second iteration from this image 🙂

[marble cape]

You see the full scary ghost in that image

[hardy sentinel]

well, the more iteration I use, it will change the original image, including the pattern on the pot

[marble cape]

Needs some work

[hardy sentinel]

yes

[marble cape]

But I was just wondering for now you could make this full color background in my version of the image

They corner there I was still having hard time

[hardy sentinel]

no, it is not possible, there is a bit random factor

[hardy sentinel]

yes, I have tried the same

[marble cape]

So it can’t be fixed?

[hardy sentinel]

not using realtime canvas, you can use GIMP manually

[marble cape]

Gimp?

What’s that?

[hardy sentinel]

if you use canvas editor for a solid color it is not guaranteed

[hardy sentinel]

an open source software similar to PhotoShop

[marble cape]

Yeah I’m not using canvas editor it’s weird how it does it anyway

Oh Gino

Gimp hmm

I never heard of it before

Hmm

Would that look better without the scary face?

[hardy sentinel]

I don't know

[marble cape]

Hmm

What are your thoughts on it so far?

[hardy sentinel]

art is subjective

[hardy sentinel]

I have a lot of real life problems, I can not think right now. I am busy, see you later

[marble cape]

I know ok I’ll try my best then as always

[feral bough]

Hey @marble cape I've suggested this before, but have you opened a personal mentoring room? Since you are often asking about the same couple of images, it would be better to have a private room where mentors can focus in on helping you and keep track of your progress. Right now this channel is getting a bit cluttered with similar questions by you. Just use /mentoring and it will create a room just for you, where mentors can help you with your projects in a more organized way than currently 🙂

[marble cape]

Ok I understand no one’s there though

Whenever I respond

Not now but I’m
Saying

[feral bough]

Well as I said before, you have to have a bit of patience sometimes. It's just that the mentoring rooms are the preferred methods for things like this because it's more organized

[marble cape]

Ok then your right

Yeah I started around late June or early July on my birthday around

[real lynx]

I've never seen a mentorship from you 👋

[marble cape]

Oh

[real lynx]

Do you know how to activate it?

It is a private room

[marble cape]

Yeah I been having it

[real lynx]

Mmm There is no room with your name

[marble cape]

What you mean?

I’m fixing one of my designs a bit better

[unborn mauve]

Thanks I’ll check it out

[marble cape]

What you mean by no top

Room

[unborn mauve]

Is it possible to do this on leonardo ?

[nimble bay]

Do what? Anthropomorphic images? If so yes

[unborn mauve]

Yes.

[unborn mauve]

This but a jacked physique lol

[nimble bay]

Its pretty easy depending on the animal

[nimble bay]

Different animals 😅

[unborn mauve]

Gotcha

[nimble bay]

Aldebo is quite good for it although so is phoenix now.

[hardy sentinel]

Yes, it is possible

[hardy sentinel]

Cute animal

[real lynx]

ty

[unborn mauve]

Yeah been using albedo, giving mid journey vibes lol

As for Phoenix I wish I had premium lol

[nimble bay]

Aldebo's stuff is better on that front imo

[unborn mauve]

I agree tbh

[marble cape]

working on this

this is my original design

[marble cape]

hey guys

[marble cape]

yeah and i also got to figure the proper prompts just by looking at my images but one thing at a time and of course patience

[real lynx]

And they had no improvement

I generated those images...😐

[marble cape]

Huh?

You whst?

??????

[real lynx]

☝️

[marble cape]

You sure?

I thought Indy’s it first

[real lynx]

...

[marble cape]

I

I could of sworn I came up with yeah no?

No really?

Indent think so look

[manic root]

Hey everyone! Have you checked out today's theme in #1108130218260041758? Jump in, challenge yourself, and don't forget to cheer on others by voting with the Leonardo emoji . There are great prizes up for grabs! Enjoy!

[marble cape]

I don’t think

[real lynx]

It's like elongated

I don't like

[marble cape]

You don’t like that image

Ok

?

[real lynx]

I would try using elements

In negative

[marble cape]

I know it’s not done the herbs

It’s not planted on the ground properly

What was it that right?

It’s the herbs you were thinking right?

Cause this is my original image I did years ago

[real lynx]

I was generating puzzles

[marble cape]

Nice

[real lynx]

And it occurred to me to see what would happen if I named your herbs.

[marble cape]

Like children lol

[worn mountain]

hi ! I want to create multiple illustrations of this character. does anyone know how to achieve consistency throughout all images? maintaining the character and art style as consistent as possible.

[real lynx]

Hi

Do you have premium tools?

[worn mountain]

hey! yes, I have premium tools

[real lynx]

Oki I'm going to create and come back with an idea 👋🏼

[worn mountain]

What do you mean?

[real lynx]

I'm going to try to duplicate your image to show you the best way to reproduce it.

I created this character with this AI?

[worn mountain]

ok..

[worn mountain]

thanks a lot for the help

[real lynx]

I'm back sorry I'm traveling home

Use this tool that allows me to move you and keep your character

With simple images you manipulate the drawing above and below it shows you the result.

I erased his arm so he could say hello

Just take the background color and paint over your original arm and then draw a similar hand where I would like it to be.

AI draws for you, you don't have to do it perfectly

On this side you have a color palette so take approximately the similar background color and erase the hand

You also have to write a prompt below the character

Where I asked for the child to be waving

Below the prompt you have more tools

To modify the character

And the strength that you want to give to the image

[worn mountain]

wow nice! I tried the same tool but was having deficulty trying to erase as well as create heand

will try and share

[real lynx]

And the best thing is that this tool does not consume tokens.

So you can play and have fun for quite a while.

You should also describe the pants in the prompt so that they are not modified.

When I have my character I would like to see it. If you have any other questions, I will be following this channel 👋

[real lynx]

Boy Child waves with one hand raised ,blue shorts, sunglasses, beach in the background

[worn mountain]

I guess I'll have to keep working on it

[real lynx]

I added more info to the prompt, a sunglasses background, and gave the prompt more power.

So that he pays attention to me

[real lynx]

It also helps that you add that action to the prompt

The child covers his mouth with one hand for example

[worn mountain]

yes, I've added the prompt as well

[feral bough]

Didn't we talk about posting things like this in your own mentor channel? It's fine to ask for help but this channel is becoming your private channel, which is not what it's intended for.

[marble cape]

I was just sharing

[feral bough]

Also, if you just say "Working on this" without any context, no one will be able to help you because they won't know what you're asking for

[feral bough]

Okay, that's what #🖼│image-share is for

[real lynx]

/mentoring It is a command that will open a private room for you and the mentors of all communities

[manic root]

Greetings, Leonardians! Have you checked out today’s daily theme in #1108130218260041758 yet? Jump into the action, tackle today's challenge, and make sure to cheer on your peers by voting with the Leonardo emoji . There are significant prize tokens up for grabs! Enjoy!

[marble cape]

Yeah I know the thing I got to figure is how to properly make the herb plants be planted on the grass

Are there other things thst need to change on my image cami that you mentioned?

[mild pelican]

Hello there, I just bought the premium for leonardo ai in the hope to receive consistent images fitting to the reference image i've provided ( checked the tutorial on the landing page for creating consistent characters ) , but whatever I do.. it just doesn't seem to work and i get completely different results than expected. Does anybody have any idea on how to achieve that or might have information on what I’m missing? I understood that the creation can be enhanced by adding training data, but I would have thought ( since its a fairly simple image ), that Leonardo can easily handle it and that there must be a flaw in the prompt or configs Ive provide. 

Any idea on solving that would be really appreciated!

[night vault]

character reference is just for human face (and anthropomorphic face work sometimes). it could not work with for whole character like that. (yeah the feature is a bit misleading something like facial consistency could be better)

[mild pelican]

Oh okey, thank you for the info! back to figma i guess haha

[thin tinsel]

there are others ways to work with Leonardo on your character. Just because character reference isn't for frogs it doesn't mean you have no other options.
Why don't you signup for mentoring?
just type /mentoring

[worn sapphire]

mentoring

[feral bough]

Please type /mentoring (so with a slash) and you'll be good

[molten pulsar]

I'm new to the group here so I'm sorry if this question has been asked or gets asked often but... Does anyone know how to get Weapons to generate accurately? I run a History site and create content of the same and it is difficult at times when I can't seem to get a Bow and Arrow or even a sword to look correct. Any tips?

[real lynx]

Hi

I tried this tool

[torn sigil]

Heyy, I wonder is there any way for me to change texts in images. I want to change the text on the headband of the book in this image

[real lynx]

With this tool you can modify your drawing in real time and see the changes while keeping your original images.

With simple strokes and a prompt I was able to get the child to raise his arm to greet me.

[torn sigil]

[thin tinsel]

have you tried image to image or any of the image reference options

[molten pulsar]

Yes, and this can work sometimes but it's not always easy to find reference images other than just the weapon itself, which I can't use as a Character unfortunately lol 🙂

[thin tinsel]

wdym

[torn sigil]

help mee~êe

[molten pulsar]

well for example, a macuahuitl which is an Aztec Obisdian bladed sword that sort of looks like a baseball bat with obsidian edging along the sides... I can get an image of the weapon itself, but to find an example of it being used in combat, especially one that resembles an Aztec Warrior is difficult. I can do Image to Image or I can do content reference but I'd need a good starting image (at least) first and they are just not readily available.

[thin tinsel]

so you're saying you always want someone holding it

[torn sigil]

am I a joke to you guys?

[molten pulsar]

Always is not quite accurate but I'd say 90% of the time, yes

[mild pelican]

hey thank you i try it out later!

[mild pelican]

thank you ill give it a look in a sec!

[mild pelican]

hey i would assume ( from what other people pointed out ) that you can do that in the realtime-canvas. Probably just a highlight of the text

and as a prompt whatever you want to have there

#🙋│image-help message

[torn sigil]

Thank you so muchh

I'll try and comeback soon ^^

[mild pelican]

I'm new too haha so dont take my word for that

[torn sigil]

[thin tinsel]

Is part of the problem around accuracy is that the AI does not know those specific weapons?

[molten pulsar]

No, I was using that particular weapon to help you understand my difficulties in finding certain reference images. It does just as poorly with swords, bows and arrows, etc. Again, finding a person holding a bow and arrow is not difficult. Finding a horde or Mongols firing 1000's of arrows is... or even 1 on horseback. I hope I'm making the distinction better? 🙂

[thin tinsel]

No I am not sure what the main challlenge is. But maybe there are several challenges

I can tell you now tho, you're not going to get a horde or Mongols firing 1000's of arrows made with Leonardo.

• You want the weapon to be accurate.
• And you want the person with the weapon to be historically accurate
• And you want the weapon to be held correctly. (Yes bow and arrow, and holding a sword are known-issues).

[molten pulsar]

Examples from today

[thin tinsel]

I see a hand issue with image 2

[molten pulsar]

I wouldnt say that, it's just very hard lol

yes, and shields held backwards or with 2 hands...

[thin tinsel]

I dont see any issues with these two images

[molten pulsar]

lol... I was showing you those images to say that it's not impossible to get a hord of mongols, it's just hard to do. 🙂 Same with producing weapons... It can happen, it just costs me an arm and a leg in points and basically seems like i need to get lucky, rather than how my prompt(s) are written. I was hoping that there was some sort of solution I was unaware of for this.

[thin tinsel]

Well do you want to fix existing images, or imrove results of future images

For example the guy holding the sheild with two arms; we could get rid of his left arm.

[molten pulsar]

Honestly, I'd rather improve results of future images so that it happens less moving forward.

[thin tinsel]

So then you need to think about how to utilize image reference

[thin tinsel]

And you can use realtime canvas to make silhoutte images as a base or to control square images that can then be expanded out in Canvas editor

or use a sketch.

[molten pulsar]

thank you! I'm not a very good Drawrer unfortunately but I do want to get better at all of this so I will have to try that.

[thin tinsel]

Yes, you can do this using Canvas Editor.
First remove the text by drawing over it with the background colour
Then use the type tool to add the text you want
Mask over the whole lot.
Generate with inpaint on somewhere between 0.2 - 0.3

[thin tinsel]

Have a go with realtime canvas first then

[real lynx]

Hi, do you still need help?

[torn sigil]

yup

[real lynx]

How can I help you?

[torn sigil]

you see there is a nonsense word in the headband of the book in the image, I want to change it~

[real lynx]

I would delete it with canvas editor

[torn sigil]

like the whole image it is that bad? 😭

[molten pulsar]

That's not bad! 🙂 Thank you! Is there a way to slide the sheild over more so his hand is completely hidden?

[real lynx]

No, I didn't like the image, it's missing a cat on the other side and it's the battle of the garbage dump.

Haikyuu style

[torn sigil]

I was kidding but I can't copy, english is not my first language sowy..

[real lynx]

What language do you speak?

[torn sigil]

Turkish ;-;

I am english student tho dw I can understand if you do not really go hard on me

[real lynx]

Nasılsın

[torn sigil]

awhh

fenasın!

[real lynx]

Ana dilim İspanyolca

[torn sigil]

I will be learning spanish next year

[real lynx]

Satranç tahtanızdaki zavallı siyah şah aynı anda iki imparatorlukla savaşıyordu.

İki beyaz kral var ve etrafı sarılmış

[torn sigil]

Neyseki siyah şah süper!

Avarage turkish playing chess lets say

1v17

[manic root]

Hey Leonardians! Did you get a chance to check out today's daily theme in #1108130218260041758? Jump into the action, tackle the challenge, and make sure to cheer on your peers by casting your votes with the Leonardo emoji . There are awesome prizes up for grabs! Enjoy!

[torn sigil]

historical facts...

well I am aware of the distortions on the board but I can still fix those. Still tho I am quite unexperienced and I wanted to know how to change the text in bottom line of the board

[real lynx]

Harfleri değiştirmeyi deneyebilir veya istediğiniz kelimelere benzer bir görsel elde edebiliriz.

Phoenix ön ayarı, metin kullanmak için en iyisidir; yapay zekanın bize görüntüsünü açıklayan bir istem vermesini sağlayabiliriz. Ancak bir görüntü alamayacağız, elimizde benzer bir görüntü olacak.

Diğer yol ise satranç tahtasını değiştirmemek ve aynı görüntüyü korumak için tuvaldir.

[torn sigil]

Well can you speak english if it does not offend you, as you can see Turkish colloquial and formal versions are quite complicated which makes hard to speak and write generally so as you might expect the meanings of the sentences are shifting, its would be easier for me to understand in english..

[torn sigil]

I am mostly intrested on canvas

[real lynx]

There are many tutorials on Youtube

If I can speak English or Turkish it is thanks to two translators I use.

One translates the screen and the other translates what I write.

I'm going to use your image to see what I can modify.

What would you like it to say on your chess board or do you just want to erase the letters?

[torn sigil]

thats a good thing but being aware of things such as canva which is a name translated to "tuval" which means canva in Turkish but you are meaning the application with canva so its should not be translated, you might like to predict those things maybe, just a piece of advice from Türkiye!

[torn sigil]

Taylan Gültürk I want it to say

[thin tinsel]

yes, you're going to draw the sheild further over his hand

[thin tinsel]

I gave you the steps here:
https://discordapp.com/channels/989166677390426132/1170046881183834192/1309249199593754755

[torn sigil]

Welp this font is so far than original but thanks for your efforts its still looks nice!

[real lynx]

What I did was delete the old letters to keep the original image in canvas editor then I went to create the new plate in Phoenix

[real lynx]

Once I chose the plate I erased the background to put it in the chess image

[torn sigil]

I see

[torn sigil]

this one looks totally amazing

[real lynx]

This is a description of your original image, that's why it looks different, I just modified the text with the prompt

[torn sigil]

how exactly do we modify the text with the prompt?

[real lynx]

With this tool he described his image to me

Side view. Image is a digital artwork featuring a detailed and ornate chess set placed on a rectangular platform. The chess pieces are intricately designed, with a mix of traditional and fantastical elements, including a dragon head. A large raven, Cuervo grande, perches on the edge of the platform, outside of the board. The platform is engraved with the text "Taylan Gültürk" in an elegant serif font. Surrounding the chess set are decorative roses, both in full bloom and wilted, adding a touch of gothic romance. The background is a smoky, ethereal scene, enhancing the mystical atmosphere. The overall color palette is a blend of dark and muted tones, with highlights on the chess pieces and roses, creating a dramatic and captivating visual effect.

In the description there were the old letters, I just modified them according to the instructions you gave me.

Anything you want to appear in text has to have a small description and then the text goes in quotation marks. For example, minimalist gold letters that say "Hello."

Graffiti style letters

Neon

[torn sigil]

Thats brilliant! but I can not see the "describe with AI" option in the web, I am on macbook is there any differences?

[real lynx]

Those weren't neon, those are graffiti

[torn sigil]

I see

[torn sigil]

I meant I can't see such a option on pc

[real lynx]

Canvas is only used on PC But I generated these images in Phoenix from my cell phone

[thin tinsel]

It's next to the generate button

✨ Behind the sparkles

[torn sigil]

I can't see sparkles

Yep I got the prompt too, ehe

[real lynx]

Turn off the old legacy

In our AI we had a very big update and you are on the old one if you turn it off the new tools appear

[real lynx]

☝️

[torn sigil]

ooh.

welp

everything became so smooth suddenly

[real lynx]

I will not be available for 2 hours

[torn sigil]

lol thanks for all your helps and nice to meet you!

[real lynx]

Nice to meet you 👋

[torn sigil]

Awh, thanks so much tho I hope to see you around again!

[real lynx]

@tribal sequoia hi

[tribal sequoia]

Hello

This was what it first did, I hit edit with ai and just prompted to make dress longer and the background gothic

That's the next generation

I eventually got it to this but it cut off the tops of the first generations and I wasted a bunch of coins

[real lynx]

I really like this negative element

[tribal sequoia]

Those are really cool. I never thought of bringing in the negative element. I mean with paintings it's habit but with digital creations it's a whole new movement to get used to

[manic root]

Hey everyone! Have you checked out today's daily theme in #1108130218260041758 yet? Jump into the action, participate, and make sure to cheer on your peers by voting with the Leonardo emoji . There are awesome prizes up for grabs! Enjoy!

[real lynx]

Your prompt

I liked it a lot

Here I am using the anime preset

If you want to create text you have to use the Phoenix preset

Whatever you want in letters has to be in quotes and a small description of the text

For example, graffiti-style letters that say "Good morning"

[tribal sequoia]

Thats the prompt for the last one I shared

Sexy vampire couple embracing each other in a romantic pose, with the man on the left and woman on the right, their faces inches apart, gazing into each other's eyes. The man has light skin, dark blonde short-hair, and glowing, blue eyes. The woman has long black hair with deep red ombre highlights, light skin, and sparkling purple eyes. They stand in a dimly lit, ominous hallway of a dark, gothic castle with purple-tinged smoke swirling in the background, casting an eerie, sensual atmosphere. The man wears a sleek black suit with deep red and purple semi-transparent accents, while the woman wears an elegant, deep purple long-sleeved gown with a long slit up the side, revealing her tattoo, her leg raised and pressed against the man's body.

[real lynx]

You can also use a /mentoring command that opens an exclusive room for you where all the mentors in the community will be helping you.

[unborn mauve]

Is anime much better or semi realistic ? Because semi realistic making my picture look like painting idk why

[real lynx]

Anime semirealista

[keen hatch]

@real lynx i have a question

i want to make a character in my photo cry, but i cant because it says that its against TOS, is there a way to bypass?

[pseudo schooner]

do you get content filter warning or something else?

[pseudo schooner]

check for words that could tigger the nsfw filter. If that does not help i am sure cami or other mentore is glad to help you. Just give it little time so they can respond

[real lynx]

In your personal photos or in an image? generated by ai

[keen hatch]

wdym?

[real lynx]

AI generated image

my modified photo

A photo of a woman crying uncontrollably. Tears are pouring out of her eyes, falling like water, and streaming down her cheeks, dripping onto her black dress. Her dark hair frames her sorrowful face. She sits slumped in a chair in a room with a patterned floor, surrounded by the quietness of sadness. The room has a door and a window, and the soft, melancholic lighting highlights the depth of her sorrow.

this prompt use

[real lynx]

A photorealistic close-up portrait of a blonde woman crying intensely, tears streaming uncontrollably down her cheeks, her face filled with raw emotion and sadness. The background is softly blurred, creating a bokeh effect that emphasizes her distress. Shot in natural lighting, with a slight vignette effect around the edges, captured using a Canon 85mm f/1.4 lens to add depth to her expressions. Fine details are visible in the tears and skin texture, capturing the sorrow etched in her face. High-resolution, with a moody and somber atmosphere.

7

[manic root]

Hey Leonardians! Have you joined today’s theme challenge in #1108130218260041758 yet? Jump in, try out the challenge, and encourage your peers by voting with the Leonardo emoji . There are some great prizes up for grabs! Enjoy!

[keen hatch]

iokk

[unborn mauve]

These are so realistic, but compared to mine, they feel like oil painting lol

[vital crescent]

using AI. This looks like to be the final version of my prompt.

The digital illustration is a chilling reimagining of Pikachu, the beloved Pokémon, transformed into a creature of nightmare fuel. Inspired by the haunting aesthetic of Junji Ito, this grotesque rendition blends hyperrealistic textures with horror elements, creating a disturbing and unsettling image. The once-familiar yellow rodent is now a monstrous figure, its fur rendered with uncanny detail, each hair catching the light in a mesmerizing yet unsettling way. Its skin, with a lifelike sheen, reveals pores and blood vessels, blurring the line between digital and organic. The creature’s enlarged, glassy, lifeless eyes stare directly at the viewer, their dilated pupils promising madness. The artist’s meticulous attention to detail—from intricate manga-style linework to hyperrealistic texturing—creates an uncanny realism. Each wrinkle and fold in its flesh pushes the boundaries of the uncanny valley, evoking dread. By referencing Junji Ito, the artist transforms a symbol of joy into a nightmare, unsettling even the most familiar forms.

[vital crescent]

Going by what others have told me so far, using a prompt this long and detailed the available art style options won’t have much of an effect

[wise canyon]

The prompt and the Model do all the heavy lifting, think of the style presets as a type of image guidance,
so a very long detailed prompt may reduce their affect, particularly with Phoenix as it is very prompt adherent.

[marble cape]

Is anyone on mentoring?

[open echo]

@hushed pebble here’s a quick exemple I made using Content Reference, AlbedoBaseXL and the prompt « Ice texture, see through, shimmering, shining, splendid »

[hushed pebble]

well its good but not like the video right ?

[open echo]

Of course it’s not like the video, not the same website, not the same prompt, not the same model, not the same base image

A LOT is different, so the results will be different too 😄

Variations:

[hardy sentinel]

beautiful images

[prisma terrace]

anyone help with some advice. when i generate an image with text then try to remove the background it also removes the text

[real lynx]

You could make the image and the text separate, the text, you remove the background and put them on the image you want the text with canvas editor

[hardy sentinel]

please post your images as well, so people could understand your question better

[hardy sentinel]

is the text color the same or similar to the background ?

[prisma terrace]

@real lynxthanks Cami, i'm lazy and hoped there would be a easy solution 🤣

[real lynx]

[prisma terrace]

this is the sort of image i'm trying to work with

[hardy sentinel]

cute dog image, ok, the text is gone too ? could you add it manually using GIMP or PhotoShop ?

[prisma terrace]

i never thought of using gimp or affinity. when i try remove background it removes everything other than the dog

[hardy sentinel]

the problem is your background is not one color

[hardy sentinel]

well, I never use the feature 🙂

[prisma terrace]

yeah, i thought about that but even when i use black background on no background it always adds something

[real lynx]

Is this used?

[prisma terrace]

that's the one i used first Cami

[real lynx]

I go back to my first idea so first the letters and then I remove the background from the letters and join the two images

[hardy sentinel]

ok

[prisma terrace]

yeah, it sounds like a great idea - thanks. then overlay the text with the image using something like affinity designer?

[hardy sentinel]

yes

[prisma terrace]

is it true that any text generated is copyright free?

[real lynx]

Now I am going to join these two images with canvas editor, it is another tool from leonardo.ai

[prisma terrace]

i'll try that option too

[real lynx]

you use phoenix it's free

[manic root]

Hey Leonardians! Have you checked out today's daily theme yet in #1108130218260041758? Jump in, take part in the challenge, and make sure to cheer on your peers by voting with the Leonardo emoji . There are some great prize tokens up for grabs! Enjoy!

[prisma terrace]

@real lynx@hardy sentinel thanks for your help and advice, i'll try those suggestions

[real lynx]

Come back if you need more help or if you have another question we love to help

[hardy sentinel]

good luck

[prisma terrace]

i'm always finding problems with what i generate 😂

[open echo]

I would love to give you a definitive answer, but It depends on the country/state laws regarding AI images where it will be used/sold 😅

But as for Leonardo, they don’t keep any copyrights over what you generate. So you can use it without limits from Leonardo itself.
Keep in mind that anything marked as « public » can be freely remixed by other users.

[hardy sentinel]

beautiful work

[prisma terrace]

that's what i thought and i did see a youtube video which said that all text generated draws data from thousands of fonts sp whatever is generated is copyright free

[prisma terrace]

so that was using canvas and overlaying the text?

[open echo]

As my favourite law YouTuber would say… « it’s complicated » 🤣

[prisma terrace]

can't disagree with you on that one 😁

[real lynx]

The ideal is to do it all at once and let the AI do all the work.

[real lynx]

Only with prompt But if demanding

You have to make them separately :3

[real lynx]

pokemon letters that say "Gengar" graffiti style stiker sin fondo .Cinematic photography of a surreal and ghostly gengar from the Pokémon Horror Edition, reminiscent of the vintage aesthetic of the Pokémon Red and Blue Edition, set against a dark, eerie background with a mesmerizing 8K resolution. Gengar intricately detailed, translucent body glows with an otherworldly luminescence, enveloped in a dark, toxic aura that appears to seep into the surroundings. Its glaring eyes, an unsettling shade of glowing red, pierce through the darkness, radiating an aura of malevolence. The atmosphere is heavy with a sense of foreboding, as if gengar is lurking in the shadows, waiting to strike. The overall mood is ominous, with subtle, muted colors and heavy shading that accentuates the haunting, ghostly quality of the Pokémon.

[prisma terrace]

@real lynxthanks Cami, i'll try all the options and see how it goes

[real lynx]

Whatever you want in text goes between quotes "hello for example"

[prisma terrace]

what happens if you try remove background with that image?

[real lynx]

I'm going to try

[prisma terrace]

@real lynxthat seems to have worked, i'll try that and see what the results are, thanks again Cami

[thin tinsel]

there is also the transparency toggle so you don't have to remove background after

[prisma terrace]

i've overlooked that option

[prisma terrace]

@thin tinselnice image

[thin tinsel]

thanks.

[hardy sentinel]

awesome

[nimble bay]

Where's the gherkin. 🙂

[sonic swift]

I like this

[open echo]

Hey there @ripe prairie

Which preset and style are you using?
Different settings, apart from the prompt, can bring « studio » lighting, so that’s something to investigate 🤔

[ripe prairie]

I am trying cinematic kino. I have been advised to use depth to image, i have tried image to image and content reference...what settings should be used?

[real lynx]

Hello, what would you like to create?

[manic root]

Hey Leonardians! Have you checked out today's daily theme yet in #1108130218260041758? Jump into the action, tackle the challenge, and don’t forget to cheer on your peers by giving their entries a ! Great prizes are up for grabs! Enjoy!

[glass minnow]

I need to create a prompt that will let me generate a girl from a reference in a restaurant with an empty plate in front of her, but the AI gives me a full plate even tho I specified EMPTY. I also tried "nothing on it", but the AI canceled it because it associated the phrase with the child. Anyone who could help?

[ripe prairie]

I have this thing and basically I want proper anatomy on top to than project onto the mesh...

[real lynx]

oki

[keen hatch]

i think @real lynx knows

[real lynx]

@glass minnow @ripe prairie

I invite you both to activate a private /mentoring room, use this command

[stable carbon]

can you share the prompt you used, so we can help get you where you wanna get..?

[glass minnow]

sure "a full and fat girl in a restaurant with an EMPTY plate with nothing on it in front of her"

i went with just empty first, but it still gave me a full plate

and then i got this

[stable carbon]

what exactly do you expect by using the word 'girl'?
are you looking for a little kid, or a young woman..?

[glass minnow]

yea, little kid i could say

[stable carbon]

try using Realtime Generation. it doesnt cost tokens and it'll help you find the right words that you can then use over in the actual image generation.
i tried a few things myself and i see that the AI prefers to put stuff on the plate, yea.. but with the right wording it's less frequent. "young child, she is at a table with an empty plate" seems to work often.

[glass minnow]

alright. thank you 🙏

[real lynx]

Willow helped create the image, my elf also wants to eat :3

[open echo]

I wouldn’t use Kino as its specifically trained to have cinematic lights.

Leo Lightning gave ok results but still some highlights.
It’s a model trained for photorealism with little to no training of « fantasy » or « studio » aspects.

Also if you have a version without the grid in the background, it will help to have a blank background. (Else the grid is integrated in the generation)

I tried to add « shadows » and « highlights » in the negative prompt but I guess the training is too strongly biased towards realistic lights.

Do you have a version of the model with a flat color? I feel maybe the shadows on the base image itself might interact with the generation.

[ripe prairie]

Here is a small screencap that should work. No shadows, consistent grey tone

What type shoud I use if not kino? Phoenix doesn't work great either...
I just want to give this thing skin... that can't be too hard...is it?

[open echo]

Having never worked with 3D models, I can’t vouch for if it’s should be easy or hard 😅

Nonetheless, like I said, the preset « Lightning » should give you better results than Kino.

Results are based on training. The system will always attempt to return what has the highest odds of being correct. So if there is a billion pictures of bodies with highlight, but a thousand with flat lighting, the results will be skewed towards dynamic lightning.

Sometimes, it’s just to know what keyword in the prompt will trigger the results. Exemple if that is « turnaround character sheets » when people want multiple angles within the same generation.

So for now I’ve tried « Front facing Human body, skin,front lighting, blank background » with the negative prompt « Side light, highlights, shadows »

[marble cape]

is anyone on mentoring?

[shut locust]

Hey people, I’m creating content for YouTube using Leonardo and I realised the free plan struggles to give detailed outputs, for example, “a close up of a dentist performing surgery on a patients crooked teeth showing gum disease and other bacteria” and there would be multiple sets of teeth and disfigurement. Is this because it’s a free plan, a bad prompt, or the wrong settings? I use Leonardo diffusion and phoenix currently. Can someone give me some insight? Thanks

[hardy sentinel]

Good morning

[unborn mauve]

Is it possible to generate omni man in leonardo ?

[unborn mauve]

Albedo is good, try different models for best result

[ripe junco]

Is there a place here on your discord to discuss technical issues?

[stable carbon]

if i remember correctly, the AIs knowledge was trained in 2021. So if its something that was released after that, the AI will not know it.
the more recent things unfortunately are not easy to achieve.

[unborn mauve]

Dang gotcha

[stable carbon]

you can still get pretty close to anything you want if your prompt is specific enough 🙂

[hardy sentinel]

we are mostly users here, please click this symbol at the bottom right of your leonardo website

[real lynx]

You never activated your mentoring channel

[unborn mauve]

Yeah but it’s not giving the accurate omni man lol

[manic root]

Hey everyone! Checked out today's daily theme in #1108130218260041758 yet? Jump into the action, tackle the challenge, and make sure to cheer on your peers by casting your votes with the Leonardo emoji . Awesome prizes are up for grabs! Enjoy!

[real lynx]

I have never tried it, I will try it and I will show you the results in premium and free

[hardy sentinel]

Hi, you can use content reference

[marble cape]

But it already shows mentoring on it

[unborn mauve]

How.?

Is it available in legacy mode ?

[hardy sentinel]

it is available in both user interface, but I am not sure if it is available for free accounts

[marble cape]

I’m just having a hard time on something

[hardy sentinel]

first you need to download the reference image

[marble cape]

Really quick

[hardy sentinel]

Hi, I am having hard time too

[marble cape]

Oh

[hardy sentinel]

i wish you a great day

[unborn mauve]

Oh it’s not free ?

[real lynx]

comand /mentoring

[marble cape]

Type it?

[real lynx]

Yes sir /mentoring

[unborn mauve]

Is it style reference right ?

[hardy sentinel]

content reference

[stable carbon]

do you have paid membership or no?
if yes, there's an image guidance type called content reference. or character reference, which may both be useful in re-making Omniman.
free users only have limited image guidance types, like style reference and a few older ones.

[stable carbon]

you have your own private mentoring thread, correct..?
maybe its only Cami who is unable to see it.. 🤷‍♂️

i have a little time, i'll give you one chance. if you can explain exactly what you're trying to do in one post here, i'll see if i'm able to help.

[unborn mauve]

I don’t have membership, but yeah I think I’ll just use chatgpt to describe what my characters looks like

It’s getting there slowly I’ll just keep generating lol

[stable carbon]

yea ofc. you can also just change small things in an almost perfect image with the canvas editor. like adding the moustache aftzerwords. or fixing the colours on the uniform, if everything else comes out nicely. 😉
dont try and get every detail right in the image, try to get most of it, and then do the rest yourself with canvas 🙂

[unborn mauve]

Yeah I don’t know how canvas works though.

[stable carbon]

there's some guide videos on our community youtube channels. one sec

[unborn mauve]

I watched it but still didn’t understood it

[stable carbon]

here's my basics guide video for canvas.
https://www.youtube.com/watch?v=YFECJegzc-c

[unborn mauve]

Thanks I’ll tune in

[stable carbon]

sure thing. i'm not big on canvas myself. i made that video for other people who struggle with canvas, like myself 😂

[unborn mauve]

Gotcha, I’ll see how it works lol it’s just I have heard it takes a lot of coins and you need it for image generation right

[unborn mauve]

This is what happened when I try to remove the emblem

[marble cape]

Ok well you know I’ve been trying To generate these herbs exactly the same to be planted on the grass but it keeps giving me different one as you see

[stable carbon]

when you open canvas, there's a blue/purple square. when you generate with canvas, it will only generate inside that square.
this image explains what canvas does with what is inside that square

basically. if you want to remove a logo and make it pure white, just paint white over it and make sure the generation frame is over it. then use prompt white spandex or something.

[stable carbon]

i already need to ask for more info because i dont understand what you want...

you want the layout of 5x6 "herbs" in your image, but you want them all to look exactly identical..?
how is one supposed to look?

[unborn mauve]

thanks and what if i want to add a different logo in that same image ?

sorry different emblem

[marble cape]

No just to make these same herbs be properly planted on the grass

There off is what I’m saying

[unborn mauve]

i think i understand how to remove emblem but what if i would like to add a different logo instead of it, can you help with that?

[golden abyss]

Hey guys, what are the best models and setup to start creating PIXEL ART?

[thin tinsel]

There is no difference between what he free plan will produce and what the paid plan will produce. It is the same product.
There are just a few extra features behind the pay wall.
If you want help with your prompt you could share it here, along with which model and preset you were using

[feral bough]

@steady tree ... What Willow said

[thin tinsel]

It depends how pixellated you want to go. But basically all the models in the new UI will do it with the right prompt.

[steady tree]

Is there any method to generate prompt from an image which can generate the same with original?

[thin tinsel]

Describe with AI @steady tree

[steady tree]

where is this?

[thin tinsel]

Do you know where the generate button is?

[steady tree]

Please guide me.

[thin tinsel]

Go to http://leonardo.ai

2. Log in

3. Go to Image Creation

4. Look for the Generate button on the right-hand side

[steady tree]

thanks !!!

[thin tinsel]

5. Next the Generate button you will see the Sparkles ✨ .
6. Press the Sparkles
7. Select Describe with AI

8. Upload your reference image

[steady tree]

but can we recover the original image with the prompt?

[thin tinsel]

I don't know what you mean.

[steady tree]

If I upload an image, the app will generate a prompt, right?
Then, if I generate an image with that prompt, the generated image will be the same with the original one?

[thin tinsel]

You already have the image, why do you need another version of the exact same thing?

[steady tree]

I want to insert other face to that image. But I can't find AI model to deals with 2 images. So I tried to extract prompt from an image and then edit another photo with that prompt. My idea may sound like stupid. Do you have another suggestion?

[thin tinsel]

You have an image you just want to change the person's face?

[steady tree]

Yes. Could you see the website?
loverse.ai
The result is awesome.

[thin tinsel]

No I'm not going to look at that website.

[steady tree]

why?

[thin tinsel]

If you have a paid plan you can use Character Reference to get the face.

[steady tree]

I am not going to use the website. I want to develop my own ai model with high quality than that website.

[thin tinsel]

https://youtu.be/wdkdvjWDxPQ?si=mkCTEu5PFqs4Agmn

[thin tinsel]

THen what do you need Leonardo for?

[steady tree]

I mean the loverse.ai. The Leonardo is useful.

@thin tinsel Do you have experience with development?

[thin tinsel]

Do you want another time out?

[steady tree]

If you don't have it, there is no point in this discussion. I need help. I don't want to waist my time here.

[thin tinsel]

It's not just your time you wasted here

[steady tree]

Nice talking.

[manic root]

Hey Leonardians! Checked out today's theme in #1108130218260041758 yet? Jump in, share your creations, and don’t forget to cheer on others by voting with the Leonardo emoji . There are awesome prizes up for grabs! Enjoy!

[thin tinsel]

https://tenor.com/view/what-do-you-want-scar-ugh-the-lion-king-eyeroll-gif-14590952

[marble cape]

Yeah

[unborn mauve]

How do you blend a different picture into the original one using canvas ?

[feral bough]

Hey, maybe you posted this in the wrong channel?

[marble cape]

No I mentioned in mentoring

[feral bough]

Well just so you know, messages from that channel aren't visible for anyone other than mentors

[thin tinsel]

You can't. It doesn't have any sort of transparency.
If you say what your aim is, then I could suggest other ways of acheiving it

[marble cape]

Command/mentoring

[feral bough]

You dont need to type command in front of it, only /mentoring

[unborn mauve]

I’d like to replace an emblem with a different one

[thin tinsel]

you want to put your logo on there?
https://discordapp.com/channels/989166677390426132/1170046881183834192/1310610702976811038

[unborn mauve]

Let me explain

Let’s say I want to remove superman logo and add omni man logo in its place

But it looks weird when I do it

Like it doesn’t blend in

[thin tinsel]

Did you teach yourself how to use Canvas Editor?

[unborn mauve]

Yes

Did the mask and all that but still it won’t work

[marble cape]

Need some help

[thin tinsel]

The only person who has the patience to help you is Cami, and she is not here at the moment. So I suggest you wait for her.

[thin tinsel]

You need to have inpaint on really low.

[unborn mauve]

Wait a minute

[unborn mauve]

How low are we talking about here ?

[thin tinsel]

Experiment.
Start on about 0.2 and go up or down depending on how much it changed.

[unborn mauve]

Gotch

And do I have to mask or sketch on the symbol ?

I mean the emblem

The one I’m going to add in replacement

[thin tinsel]

Mask over the emblem and over a bit further outside of it

[unborn mauve]

Like the outside or the inside ?

[thin tinsel]

If what you're adding is not the right angle and not the same colour tone or style, it's still going to look out of place

[thin tinsel]

outside

[unborn mauve]

Alright I think I’ll have to find what blends in then

Another question, @thin tinsel can I like sketch or something on my picture to make an emblem or something myself ?

Like self create

Because if I don’t have a picture that blends in so I’ll need a technique to make it myself

[thin tinsel]

Yes you can draw the logo into position and then you could put the inpaint up 0.4 - 0.5 to transform it more.

[unborn mauve]

Alright I’ll try that out

[thin tinsel]

https://youtu.be/adiCB8Esm6U?si=12cozMevrfCg0Sua&t=471

[unborn mauve]

close enough ?

forgot to paint the insides red lol

[thin tinsel]

looks great

[unborn mauve]

yeah just some editing left to do

i am gonna do a redo to make it better

[thin tinsel]

does he normally have a large package?

[unborn mauve]

All superheroes do 💀

[thin tinsel]

Ah that's perfect then. 😊

[unborn mauve]

But honestly this is just me experimenting today, I’ve learnt a lot canvas today and finally can put different emblems and change the outfits without the need of generating 100 images lol, appreciate the help really.

[thin tinsel]

that's great news.
Hope you had fun along the way

[marble cape]

Ok I’ll wait

[marble cape]

well someone said there was going to do this for me

[marble cape]

yeah if you can help me man that will be great

[marble cape]

im just dealing with certain things

[marble cape]

Forgive me whatever

[onyx flicker]

I keep getting a "forbidden" error and am unable to render anything. Anyone know why?

[leaden jacinth]

🙏

[leaden jacinth]

Currently they are working to find the reason.

[hollow breach]

Which model are you using?

[leaden jacinth]

The site is having issues.

[leaden jacinth]

Tap in when ur ready.

My current situation is over 14 minutes of buffering.

[night vault]

delete it and it should unblock new submission

[leaden jacinth]

Will do as I've logged out for a while and just went back...it's up to 23 mins lol.

[leaden jacinth]

I cannot delete anything.

By now I would've expected a failed generation.

[night vault]

oof, i think you need to contact support. or if you have ios/andriod app maybe it can be delete from that instead

[leaden jacinth]

Ok will do. thx.

[lyric hedge]

Hi everyone, I was wondering what would be a good preset to use for glitch art ? I saw a prompt and not achieving the expected results from it. "webcore-infused digital collage, merging elements from early internet websites, such as rotating GIFs, sparkling cursors, and Comic Sans text'"

[night vault]

do you need readable text? if not some elements combination can be good at that

[lyric hedge]

@night vault No, I would not need readable text. Would that be illustrative albedo ? I posted the image I am trying to recreate.

[night vault]

albedo, animexl works. try these elements combination

-0.2 abstract line art
-1.1 3d
0.2 glitch art

prompt add (chromatic abberation:1.3)

[lyric hedge]

Thanks. I will have to work on it since I wanted to make sure that the laptops come through more like the original art.

[night vault]

try add more weight to laptop like (laptop:1.3) decrease chromatic aberration

[night vault]

if glitch is too messy, move negative element closer to 0 to reduce it influences

[manic root]

Hey Leonardians! Did you jump into today’s theme yet? Check it out in #1108130218260041758! Get involved, challenge yourself, and don’t forget to cheer on others by voting with the Leonardo emoji . There's big prizes at stake! Enjoy!

[leaden jacinth]

thx, my issue was resolved.

[night vault]

https://tenor.com/view/awesome-minions-excited-gah-i-cant-even-gif-313403926715866522

[lyric hedge]

Thank you

[leaden jacinth]

🆒

[leaden jacinth]

Good to know whenever I'm experimenting with elements.

[night vault]

those 2 negative elements will help create complex scene. give it a try.

[hardy sentinel]

awesome logo

[hardy sentinel]

Albedo Base XL for realism or western illustration, Leonardo Anime XL for Anime or Western cartoon look

[hardy sentinel]

Hi, do you have anything in your prompt that could indicate a younger person ?

[hardy sentinel]

or any indication of violence, blood, military ?

[hardy sentinel]

Looks good

[unborn mauve]

It’s almost close but I worked all night to make it accurate enough to omni man but I’m learning

[hardy sentinel]

great result, it is time and token intensive

Good morning, how about this?

[hardy sentinel]

have you tried to clear your cache? it is working for me right now

[hollow breach]

Great

[onyx flicker]

No, the prompt was, "An icon of a phoenix." It's working now.

[hardy sentinel]

awesome

[hardy sentinel]

it is probably a server issue then 🤔

[unborn mauve]

i didn't knew the wonders leonardo upscaler could do lol but this is the final result @thin tinsel

[thin tinsel]

wow, looking goood! 🔥

[unborn mauve]

yeah it's legit omni man lol

just gotta face swap now

dayummm

[thin tinsel]

nice one!

[unborn mauve]

i think leonardo upscaler saved me a lot of time, or i would have been at this all over again lol

[lilac minnow]

who have this Channel

[hardy sentinel]

Awesome result

[unborn mauve]

Yeah I’m also amazed lol

[marble cape]

That looks ok it looks like it’s almost look like in 3D. I like the mist is it circling around? but where is the 3rd candle at?

Also it would look better if it was in the center

[hardy sentinel]

yes, I add only the mist

[leaden jacinth]

@hardy sentinel Yes all is well 😊

[marble cape]

Is it circling around? Also where is the last candle?

[hardy sentinel]

yes, I need to put the candles at center

[hardy sentinel]

not generated somehow

[leaden jacinth]

@hollow breach

[marble cape]

Can you make another candle?

[hardy sentinel]

no I did not wrote circling around, because your image is already depict the position you have wished

[marble cape]

Oh

[hardy sentinel]

perfect, that was maybe a server problem ?

[hardy sentinel]

sure, what kind of effect do you like ?

[marble cape]

So you can add another circle with the mist?

[hardy sentinel]

not really, because your image aspect ratio is not 1:1

[marble cape]

Like the most circle around the candle maybe just an idea or just leave it like that

Is there a way to add another candle though?

Maybe I can copy and paste it

[leaden jacinth]

Yes there was a brief moment when nothing was generated 😕

[hardy sentinel]

I need to draw using inpainting and after several iterations I get this image

[hardy sentinel]

yes, I managed to add several candles, you can manually draw using inpainting

[marble cape]

The other looked better that you did 🤩

That’s looks really nice that you did

[hardy sentinel]

thank you, glad that it is more similar to your original idea

[hardy sentinel]

I am using realtime canvas

[marble cape]

Maybe I can copy another candle somehow add it or

[hardy sentinel]

only using instant refine after that, so it is free

[hardy sentinel]

oh ok, you want to add a candle to the image above #🙋│image-help message

[marble cape]

Yeah the one you did

With the mist

[hardy sentinel]

ok

[marble cape]

Thanks can’t I want to do it myself to learn but I got another project

Different project I’m working on

[hardy sentinel]

this is the second iteration

third iteration of three candles

[hardy sentinel]

I wish you success for your projects

[marble cape]

I liked this better

[hardy sentinel]

ok, it would be difficult to add another candle

[marble cape]

Could you make the mist switch around to?

Swerve

I ment

[hardy sentinel]

there is no space left 😄

[hardy sentinel]

yes, I am out of idea right now how to do it 🤔

[marble cape]

I’m going to trying to see if I can copy the same candle and somehow add it unless you were going to do it?

[lyric hedge]

Thanks 👍 I will try that out.

[hardy sentinel]

you are welcome

[hardy sentinel]

it is not exactly swerve around it, i add the amount of candle

or this one

I need to go, wish you a good day, see you later

[marble cape]

Yeah I’m trying to add another candle with this one

[manic root]

Hey Leonardians! Checked out today's theme in #1108130218260041758 yet? Hop in, join the challenge, and use the to cheer on entries you love! Fantastic prizes are up for grabs! Enjoy!

[thin tinsel]

Stop trying to solicit for business which is against the terms of agreement for this discord server.
If you would like to join in as a participating member of our community, you are more than welcome.
If you are only here for personal gain, then you will not be around for very long.

[thin tinsel]

There is an element called Glitch art.
(edit: Sorry I later saw that Zorgi had already recommended to try it). 🙂

[thin tinsel]

You can use your original image as content reference or style reference

Here is an example with Glitch Art set to 1.00

[thin tinsel]

Hi 👋 glad it's working OK now. I think I've had that before, it just refused do make any images, and there didn't seem to be any particular reason.

[thin tinsel]

https://tenor.com/view/cat-answer-bop-question-gif-26201781

[thin tinsel]

https://tenor.com/view/no-not-the-lion-king-gif-12689717

[marble cape]

Hey

☝️

[heavy dome]

Hi all,
I'm looking for your expert help, for a not so common question.
We are creating a character creator. For this we need pixar style assets like facial hair or hair styles.
I'm lookin to create something like this:
Plus one extra where there is a stubble beard.

How can I create this? multiple promts is also fine, but in the end I need to have the same face with different facial hair features to ask our customers to choose what is closes to what they want to create.

We need to create the same for hairstyles, Female, Male, Boy and girl

[thin tinsel]

So you already have the faces you want?

[heavy dome]

nope it should just be pixar art style faces.. Don't realy care about the actual face as long as it's more or less consistent

even without face would be okay, something like this. This would be okay for hair style. We would need the same for facial hair.
There is a lot of stuff to be found on the interwebs, but it's hard to get consistent style when it comes to boys, girls, male, female. And for hair and facial hair

[thin tinsel]

You said

"in the end I need to have the same face"

[heavy dome]

Sorry for not being clear, I guess I mean every picture on the row should be more or less the same style/face or figure

[pure geyser]

Hey @thin tinsel which preset and elements should I go for - to achieve this type of image

[thin tinsel]

Hi Striky, 👋
Let me just finish helping Youla, and then I can have a think about your image. 🙂

[pure geyser]

Fair enough

[thin tinsel]

I feel like you'll need to do each hairstyle seperately and then arrange them on a page like that once you have the variations.

[thin tinsel]

Looks like an acrylic painting to me. 🤔
Let me test a couple of things and come back to you

[pure geyser]

👍

[thin tinsel]

I think I'd probably go with Concept Art for the model and then Stylised Illustration for the preset.
Do you have a good prompt for it?

[pure geyser]

Nope, probably use chatgpt or describe with ai

[thin tinsel]

Yeah go with Describe with AI to get your prompt and make sure 'Acrylic Painting is mentioned. And 'Vibrant' colours.

Also, 'hightly saturated' is always a good one.

[pure geyser]

https://tenor.com/view/thinking-think-about-it-let-me-think-jerry-mouse-gif-1879525588246727560

[heavy dome]

That would be good! the hairstyles should be more different but that's something I can promt. How dit dyou get this result?

[thin tinsel]

I used your image of the silhoutte heads, and used Describe with AI to get the prompt.
Then I used Leonardo Phoenix Model, with the 3D render preset.
You can then edit the prompt to mention the facial hair and stubble

and also 'pixar-style

[heavy dome]

Wooow! Super nice!

i'll give it a shot

[thin tinsel]

Is that more what you had in mind?

[heavy dome]

jep, but when we are talking beards the hair should remain the same, when we are talking hair style the facial hair should remain the same. But I'll give it a shot

what did you promt chat gpt to get a description suitable for leonardo to use?

[thin tinsel]

You might need to sort that out by editing afterwards

[thin tinsel]

No ChatGPT.
Leonardo's Describe with AI.
As I suggested to Striky Above.

Do you know where to find it?

[heavy dome]

nope, looking for it now 🙂

[thin tinsel]

Next to the Generate Button look for the Sparkles ✨
It's in there

[lyric hedge]

Can you use glitch art elements in the new mode or does it have to be in legacy ?

[thin tinsel]

The NEW UI models will let you use the Elements, the only exception is Phoenix.

Use Albedo or Vision

[thin tinsel]

Right! It's Leonardo DIFFUSIION that will do Acrylic Painting style and is only available in Legacy UI, which is why I couldn't work out why nothing would do a painted style!

[marble cape]

Hey Steven

[hardy sentinel]

hello

[marble cape]

Did you managed to add another candle to the image?

[lyric hedge]

After trying this, the elements are taking away from the design path I wanted. It generates a woman in glitch rather then the abstract collage aesthetic I was looking for. Thanks for letting me know where to find them.

[real lynx]

Can this help you?

[timid obsidian]

can anyone help me generate a prompt that will give me a image of the number 1 with legs arms and a face a cartoon figure ??

[real lynx]

Hi

We send you images in general

[marble cape]

Well thanks for trying

[marble cape]

I really like this that you did

[marble cape]

Sorry cami I know it doesn’t mean anything

From saying sorry

[manic root]

Hey Leonardians, have you checked out today's daily theme in #1108130218260041758? Jump into the challenge and make sure to cheer on your peers by voting with the Leonardo emoji . Great prizes await! Enjoy!

[real lynx]

It means a lot to me, don't worry.

[marble cape]

Ok

Um

Trying to add another candle with that style

[night vault]

canvas IS the generative fill. same thing.
you can use brush to mark where you want to inpaint rather than where you want to freeze.

[marble cape]

hey guys how do i turn my herbs upright?

[karmic pulsar]

Yooo pls help

Can someone send me a custom prompt I want to generate images with this exact dog, I tried everything making ai prompt in leonardo, with chatgpt, nothing can be close to this dog, thanksss

[leaden jacinth]

Have you tried AI Describe?

Perhaps Image to Image?

[feral bough]

Is this the same dog in all the images? AI is based on randomness and it cannot recreate the exact individual dog. But if you know the dog's race you can probably get quite close by describing the size, color, pose, etc.

[karmic pulsar]

Yes

[karmic pulsar]

Already tried

[feral bough]

Well there's really not much more you can do. AI doesn't know specific dogs, maybe only super famous ones. So you'll never get the exact dog since AI is based on randomness. If you have a paid version you could use tools like Content Reference to get closer, but it'll never be exact

[leaden jacinth]

The Leonardo AI Described all four images and nothing was remotely close 😭?

A Pom in a Hoodie Nya! A super cute fluffy pompom dog, all snuggled up in a cozy white hoodie.
Its fur is a beautiful, rich reddish-brown color, and its eyes are looking right at you with such a sweet expression.
The hoodie fits perfectly, and the little paws are peeking out from the sleeves.
It looks so comfy and adorable lying on a soft light beige couch.
Meow! So fluffy and precious, nyaa!

[leaden jacinth]

Try this one

[glass minnow]

i need help with prompts for multiple images with "lifelike vision, 3d render", with a girl images as the guidance:

1. i need sonic the hedgehog in a cave in minecraft with a zombie standing behind him (for some reason i can't get that to be generated). sonic does not look like a minecraft skin
2. i need sonic the hedgehog and the girl in the reference in a minecraft forest scared. they do not look like minecraft skins.
3. need basically the same image as no. 2, but they are screaming their lungs out

[marble cape]

i still need help with this

i maneged to do a few things

maye this is better than but i need to add anotehr candle and i dont know how to

and i forgot on how to put it in the center?

[glass minnow]

@leaden jacinth can you help me?

[leaden jacinth]

HI. I am not sure about this one but you can ask for a mentor by typing /mentoring for 1:1 guidance.

[glass minnow]

alr thanks

[leaden jacinth]

Question: Is it that Leo is not rendering Sonic correctly or is your prompt iteration incorrect?

[glass minnow]

sometimes it renders sonic as a minecraft skin, which i don't want, but it also renders day instead of night and no zombie. that is for the first. for the others, they either get rendered as skins, or it's making a combination of the two. and i can almst never make them have scared or panicked faces

[leaden jacinth]

Pls share the prompt. Let's try to dissect the issue.

[glass minnow]

this is one of the prompts: "there is a girl together with sonic the hedgehog in a forest in minecraft at night. they are frightened and scared and they are screaming their lungs out"; this is another: "sonic the hedgehog is in a forest in minecraft at night. he is not shaped like a skin. behind him stands a minecraft zombie"

[leaden jacinth]

What preset were you using to generate the image?

[glass minnow]

the preset style is 3d render

[leaden jacinth]

Sometimes you do have to try several different presets and styles if you have the credits due to trial and error of AI. BRB

[glass minnow]

the thing is that i have to use these settings. i need a very specific thing

[leaden jacinth]

I understand. I used a few keywords to test Sonic. This is what I got You may just need to build your prompt more. Sonic the Hedgehog, girl, Minecraft, nightcore, frightened, screaming, eerie

https://cdn.leonardo.ai/users/46484726-47e0-47c2-8d2b-6ecb58d855f2/generations/4103b397-2629-479c-8b16-84db3bde06ae/Leonardo_Anime_XL_Sonic_the_Hedgehog_girl_Minecraft_nightcore_2.jpg

Anime XL Realism

https://cdn.leonardo.ai/users/46484726-47e0-47c2-8d2b-6ecb58d855f2/generations/4103b397-2629-479c-8b16-84db3bde06ae/Leonardo_Anime_XL_Sonic_the_Hedgehog_girl_Minecraft_nightcore_1.jpg

lol

[manic root]

Hey Leonardians! Have you checked out today's theme in #1108130218260041758? Jump into the excitement, tackle the challenge, and use the to vote for your favorite entries. Incredible prizes are up for grabs! Enjoy!

[thin tinsel]

SONIC THE HEDGEHOG is a trademark of SEGA CORPORATION.
Do you have a licensing agreement from Sega to use their intellectual property?

[manic charm]

Someone have a trick to make the pg not look for the screen?

[thin tinsel]

don't know what you mean. whaat is pg?

[manic charm]

oh sorry, character

pg is in portuguese, i forget

[thin tinsel]

Do you mean the thing where the character always looks directly at you?

[manic charm]

yes

exactly

[thin tinsel]

Are You using a Character Reference image of someone looking at you?

[manic charm]

the head is turn for the screen, but the eyes arent looking for the screen

[thin tinsel]

The head is facing the screen but the eyes are looking away.
What would you like to have happen?

[manic charm]

A man blonde, dirty with dust, archaeologist medieval clothes. Holding an artefact, looking for the artefact (highest quality face), (highest quality clothes), (best shadow), intricate details, cinematic, detailed, dynamic shadows, art draw style, realistic eyes, correct pupils. Ruins background.

this the prompt, I'm would like the character is looking for the artefact

[thin tinsel]

Look "at" not "for"

[manic charm]

oh, okay, i will try this

my english is poor, I trying to improve haha

[thin tinsel]

You can prompt in Portuguese rather than English if you prefer

[manic charm]

I have the feeling that english prompts generate better images, its unrelated?

[thin tinsel]

It is still a difficult thing to achieve.
You might use Content Reference if you have an image to use as the reference image.

Or you can use phoenix, because phoenix pays attention to your prompt

[manic charm]

worked

[thin tinsel]

Oh! Perfect!
I went with "Inspecting it"

[desert herald]

Hopefully this is an old topic as I see it was discussed going back many months. Is there a good strategy for keeping text from appearing in so many Phoenix module image generations?

[marble cape]

Hello all

[real lynx]

I would have to see your prompt to tell you what you are doing wrong.

[real lynx]

Good night

[marble cape]

If someone can help me add this

Another candle

[real lynx]

I would tell the AI to describe the image and at the prompt it would add one more candle

[marble cape]

On canvas editor

?

And how do I center the image I forgot?

[marble cape]

That shining thing?

[desert herald]

Thanks, prompt example in this case is:

A visceral exploration of humanity’s defiance of natural laws, digging into the sacrifices and consequences of breaking free from the forces that bind us. Evocative, abstract, mysterious, very creative and highly artistic

[desert herald]

The image generated in this case is this

[real lynx]

What a good image

I'll see if I get good results and I'll let you know.

[desert herald]

Thanks! FWIW, I was using No for prompt adjust, Dynamic preset and high contrast. Of course if I generate 4-5 times it will make something without text.

[frail spire]

Hey guys is there a way to have Leonardo stop generating texts in gibberrish and in english instead?

[real lynx]

16 images and they had no letters with this setting

[desert herald]

Ahh, I have mostly used the "Quality" mode, when I switched to "Ultra" the text wasn't there in 10 generations. Ultra is more expensive but in the end it will actually save credits because I generally have been by default queuing up a bunch of gens in advance to get some without text.

Hopefully this setting does continue to turn out to give more predictable text apperance results. Thank you very much!

[desert herald]

Now if only I could get the text to go away with the "Graphic Design Vector" style. haha. Regardless of Ultra and Quality, text 95% of the time is appearing when I use that style.

[stray crater]

Hi, I launched a job, then got an "Error: database query error", the job wasn't displaying time, I ctrl+f5, timer started, but no output, queued another job, same situation, first job is running for multiple minutes now, what's happening ? :/

(and now when I cltr+F5 I have directly that error message and previous job aren't displayed, nor the ones that should be running :/)

nvm, just noticed the top banner mentionning issue, gl hf 😉

[dreamy birch]

Hello, I'm trying to generate an image of a woman with butterfly wings on her back, but she's not facing the camera, sometimes she's generated from the side or with her back to the camera.
prompt: Hello, I'm trying to generate an image of a woman with butterfly wings on her back, but she's not facing the camera, sometimes she's generated from the side or with her back to the camera.

[ebon sandal]

Hi , How could I make a square car wheel in Leonardo? Have I to draw the square wheel well beforehand with the vizcom brush? with prompts it is possible ?

[manic root]

Hey Leonardians! Checked out today’s daily theme in #1108130218260041758 yet? 🌟 Jump into the challenge and don’t forget to cheer on your peers by voting with the Leonardo emoji . Great prizes await! Enjoy!

[thin tinsel]

This is a known issue. I would recommend you use phoenix as it's more responsive and then use **describe with AI **to note what changes you want.

[candid perch]

Hello, wich preset must I use to have same picture and quality please

[thin tinsel]

Why are we talking about Vizcom. Do you need help with Leonardo?

[candid perch]

I try différents things but I don t have images with all details

[thin tinsel]

The quality of the images is not about preset.
These images are high quality because they are high resolution.
Therefore use the biggest size option to make the images (once you are happy with your prompt)
You can also use the Universal upscaler to increase the resolution.
Or do you mean something else?

[candid perch]

@thin tinsel i talk about détails : on my picture in all my prompts i don t have many elements

Look

You see the différence ?

[thin tinsel]

You cannot show me an image and expect me to see whatever you see.
Do you mean the image resolution or not?

[candid perch]

Yes and view

[thin tinsel]

view means what?

camera angle?

[thin tinsel]

This is not photo style it is like a drawing

[thin tinsel]

Using this one with EDGE TO IMAGE

[sonic swift]

Describe her eyes or smile 🙂 That usually helps me

[ebon sandal]

Hi WillowB, sorry for the mistake. I want to make the image with Leonardo. The complete image with Leonardo.

[thin tinsel]

please confirm if you mean that you'd like to have square wheels in new images, or whesther you want to edit your existing image.

[marble cape]

Steven?

[marble cape]

Man this is hard

[marble cape]

I just don’t know how to add another candle

im trying to as you can see but its hard

[feral bough]

Sometimes with AI, things are just not possible. If you've tried multiple things and it hasn't worked, then it might be time to move on

[marble cape]

Oh man I was looking forward to adding another candle cause my original design looked like this though:

[thin tinsel]

Have you even tried Cami's advice?

[marble cape]

For adding another candle?

[thin tinsel]

Yes for adding another candle.
I was just looking for a yes or no on whether you'd followed Cami's advice

[marble cape]

It’s there describe right?

[thin tinsel]

https://discordapp.com/channels/989166677390426132/1170046881183834192/1311536136056737812

[marble cape]

Describe with ai?

[thin tinsel]

Read what Cami has said.
You will not get any spoon feeding from me.

[marble cape]

Why

[thin tinsel]

Why read it, or why am I not going to spoon-feed you?

[marble cape]

Spoon feed me?

But I am the mother of ♋️

[thin tinsel]

Because you've had a huge amount of spoon-feeding from a lot of different people, who have come to resent having to continue spoon feeding you.
You are not a newbie anymore you can start feeding yourself. OR show that you understand how to feed yourself , by showing that you are making the effort

[marble cape]

Yes mom ♋️

[thin tinsel]

So let's try again.
Have you TRIED what cami suggested to you. Describe with AI?
Yes or No?

[marble cape]

Not yet but I will describe with Ai

With that image

Right now I will I was busy with thanksgiving

[thin tinsel]

Yes that's a great idea.
And you can use Steven's image as Style ref so it looks a similar style if you like

[thin tinsel]

Ah you're in America are you?

[marble cape]

Yeah um u.s..

New Jersey…

[thin tinsel]

Nice

[marble cape]

Not really

It’s not nice

[thin tinsel]

Sorry to hear that.
Everywhere has nice and not nice parts to it

[marble cape]

Well it’s not that I hate it I’m just not crazy about how things are

[thin tinsel]

Ok, well you focus on getting your candles done for now

[marble cape]

Oh but um I used up my tokens when I tried to..

I forgo to say that

Wait but won’t describe with ai will change the whole image

[thin tinsel]

No, it's looking at the image and doing it's best to say exactly what it looked like

Would be easier if we knew which model etc but not essential

[marble cape]

Well whether I use describe with ai since that is 3D I used phoenix

[thin tinsel]

Ok, So when you have your tokens you can try Describe with AI, with Phoenix, and just add number of candles to three. Job done

[marble cape]

Wait what’s the prompt to write to add another candle? Just type another candle right?

[thin tinsel]

No.
You want THREE CANDLES. So make sure the prompt has THREE CANDLES written in it

[marble cape]

Oh ok

Also going to generate this

It’s one of those pendulum clocks with black triangle ferocious teeth

[manic root]

Hey everyone! Checked out today's daily theme in #1108130218260041758 yet? Jump into the action, participate, and make sure to use the to vote for your favorite entries. Great prizes are up for grabs! Enjoy and good luck!

[verbal ledge]

Hey guys i need help with creating something similar but a dog holding a phone, can you anyone help?

[ancient halo]

something like this is what you mean

[verbal ledge]

yea but i plan to animate so i need it either under water or falling from the skye

[ancient halo]

what do you mean by that

[verbal ledge]

yo those are cool can you share what prompt you used

[ancient halo]

for the 1st or the 2nd

[verbal ledge]

1st

[ancient halo]

set 1

?

[verbal ledge]

this one

[verbal ledge]

tysm

[ancient halo]

is it that good otherwise I will continue

[verbal ledge]

no it's perfect thanks

[ancient halo]

you're welcome and good luck 👍

[marble cape]

Hello guys happy thanksgiving

[dreamy birch]

two came out perfect and two didn't

[dreamy birch]

Thanks for the tip

[marble cape]

ok i have 150 tokens again

[marble cape]

wait it genereates the prompt for me so i should remove all the prompt it gaved me and just put 3 candles then. also which should i use for it phoenix right?

which should i use to get the same?

yeah i should of told you that whenever i click describe image they generate the prompt for me.

[real lynx]

Phoenix spends a lot of tokens I would use albedo

[marble cape]

ok

but the prompt should be just 3 candles with white mist than right? nothing else to add?

[real lynx]

Read the description of the prompt where there are no more than two candles, modify it and ask for 3

[marble cape]

ok well it gaved me this:

whats the stregnth should be on?

[dreamy birch]

with male character he generated more front

[marble cape]

it gaved me this too\

[marble cape]

i still stuck though

[marble cape]

I really like this one though

[remote root]

what do you write in promt to get all character in the picture? i have 9:16 ratio and write: (((full body view with high heels))) but still get half of her😞

[manic root]

Hey Leonardians! Have you checked out today's daily theme in #1108130218260041758? Jump into the excitement, tackle the challenge, and don’t forget to cheer on your peers by voting with the Leonardo emoji . Awesome prizes are up for grabs! Enjoy!

[real lynx]

Hi

[real lynx]

Maybe this combination of elements will help you.

The prompt is simple, you would have to modify it to the features of your blonde wife wearing blue clothes.

Prompt

An alcohol ink illustration of a fiery-haired woman with a modern, avant-garde outfit. She is surrounded by a swirl of warm colors.

[leaden jacinth]

Hi. Try using "rose petals on the ground" or something that's near her shoes like "diamonds near her classy strap on heels 👠 " see if this works...

[remote root]

why negativ number on the elements??

[leaden jacinth]

Full Body:

**(as in an image of a person standing upright, capturing their entire figure from head to toe). - this view is good for making sure that you get ALL of the person (or object) in the image, not just part of them. It can also be used for animals and objects.

[stable carbon]

as VS just mentioned above, specify something about the ground or the feet in your prompt. thats a lil "exploit" if you will, to get full body shots. even something dumb like "standing on ground" or on white floor. it works pretty well in my experience.

you're already using a vertical format, rather than horizontal. that would have been my other suggestion 🙂

[remote root]

thanks❤️

[remote root]

thanks❤️

[stable carbon]

if it still struggles, feel free to share your prompt here and we can help optimise it. 🙂

[leaden jacinth]

No problem 😊

[remote root]

it worked🥰

[hardy sentinel]

Hi Marina, do you mean the whole body ?

[remote root]

yes with head and boots😊

[hardy sentinel]

awesome, then mention head and boots 🙂

[remote root]

i got tips here to write (as in an image of a person standing upright, capturing their entire figure from head to toe).. it gives me 1 of 4 generated pictures with it and 3 half body😊