Sign In
homemodelsimagesvideospostsarticlesbountieschallengeseventsupdatesshop

Crody's Illustrious / NoobAI Image Generation Tips

409

Crody's Avatar

Crody

Sep 6, 2025

(Updated: 22 days ago)

generation guide
illustrioustutorialimageguidegeneration
Crody's Illustrious / NoobAI Image Generation Tips

Hi, this is Crody from Team-C
In this article, I'll explain which setting to use, prompting resources and some example generation settings for Illustrious / NoobAI models

Note: The settings are generally applied for Nova Series

1. Settings

In this section, I'll explain about Samplers (Schedulers), CFG scales, Steps, Image size recommendation and Hires.fix settings

︎

Samplers

Euler A (Euler Ancestral), and that's it for me
Other samplers like DPM++ generally leaves noises a lot so I'd not recommend other than Euler A

︎

CFG Scales

3.0 - 5.0 is the best for most models
If you go more than 6.0 (like Pony or Animagine) the generation gets much vivid or even too vivid
If you use LCM, it can go around 4.0 - 8.0 so balance would be around 6.0 in that case
I use 4.5 a lot and I think 4.5 strikes the balance between controllability and contrast

︎

Steps

20 is the best I think but for some generations (like when you use lighting LoRAs), please follow those models' recommended steps

︎

Image Size

Please keep the rules of either 1024 x 1024 or 1024 x 1536 rules

For the 1024 x 1024 image settings, please read THIS ARTICLE

I generally use 768 x 1344 (4:7) or 832 x 1216 (2:3) for portrait

︎

Hires.fix settings

Denoise factor around 0.4 - 0.5 is the best of following the initial image and detail enhancing

Hires Steps would be 16 if you're using modern systems (like ComfyUI), 40 if you're using Diffusers (like me)

The best scaling factor would be x1.5 more than that (like x2.0 for example) generally creates some errors

︎

2. Prompting guide and Resources

I see lots of people struggles with this section including myself
Eventually I got what I want with below prompting

Note: This way of prompting cannot work on vanilla ComfyUI, vanilla Diffusers and this site's onsite-generator because those system don't have weighting prompt and have some limited token barrier (75 tokens)
To do the prompting, please use advanced method for embedding
For ComfyUI, ADV_CLIP
For Diffusers, sd_embed
For others, I don't know

︎

Prompt format

My prompting comes from my own drawing method to define what to draw which goes something like this:

(Excluded from steps)
All. What should I not include and be careful about?

  1. How much quality do I put to this illustration and which style do I use?

  2. How many subjects are mainly focused and if there are multiple subjects, how do they interact with themselves?
    ︎
    Step 3 and 4 are repeated for each subjects

  3. What character / subject should I draw and what kind of hair, attire and body style the does the subject have?

  4. Apart from interaction with others, what is the character doing?
    ︎

  5. Which angle should the camera looks?
    Where should the camera focuses?
    Where is the camera placed?

  6. What after effects should I use?

So from this, the actual prompting goes something like this:

Note: Separate each segments by using square brackets but on the actual prompt, you mustn't to put them

︎

[1. Quality tags, styles and ratings]
[2. Person count and interaction]
{
[3. Character Name if you need and Body modifier from Top to Bottom including Clothes]
[4. Posing excluding interaction]

} ( Repeat this for each characters, separate each by BREAK)
[5. Camera position and background]
BREAK
[6. Lighting, Depth and After details]
(I add BREAK to make it more visible)
...
[All. Negative Prompts]

For Negative Prompts, there's no appropriate format but I use

modern, recent, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra fingers, extra digits, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), sketch, jpeg artifacts, signature, watermark, username, (censored, bar_censor, mosaic_censor:1.2), simple background, conjoined, bad ai-generated

and if I have some undesired thing additionally, I just add them on first of the Negative Prompt

Useful Pages for Negative Prompts

I'll explain each segments one by one

︎︎

0. Prompt Syntax

Format Tips:

  • DON'T USE CAPS like Best Quality for example unless the tag itself has caps

  • Separate each tags with commas: masterpiece, best quality, ...

  • When spaced tag includes another strong tag like fake rabbit ears and rabbit ears, use underscores instead of spaces like fake_rabbit_ears

  • Symbols included on the right needs backslash before the symbol to work it without confusing the system: ()[]:

  • Use weighting or rearrange the tags instead of repeating the tag itself

︎

1. Quality tags, styles and ratings

Quality tags is the tags like masterpiece, best quality, 4k, high resolution, detailed

From my experience, DON'T use score tags like in Pony models
Score tags results in text that writes score on everything

For me, masterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres

︎

Style tags includes drawn eras, art style, artists and many things

Drawn era tags includes

  • Illustrious era tags like newest, recent, mid, modern, early, old, oldest

  • Danbooru year style tags like 1980s_\(style\), 2000s_\(style\)

For Danbooru year style tags and art styles, please check the links below
https://danbooru.donmai.us/wiki_pages/list_of_style_parodies
https://danbooru.donmai.us/wiki_pages/tag_group%3Avisual_aesthetic
︎

For artist tags, check this link below
https://danbooru.donmai.us/artists

︎
Rating tags are given to maintain sexuality
This includes

  • general

  • sensitive

  • questionable

  • explicit

︎

2. Person Count and interaction

Person count includes tags like 1girl, solo
For this, please look at the links below

https://danbooru.donmai.us/wiki_pages/tag_group%3Agroups
https://danbooru.donmai.us/wiki_pages/tag_group%3Afamily_relationships

︎
If there are multiple characters and each one of them interacts with others, prompt it after person counts like 2girls, back-to-back

This can be found on the link below

https://danbooru.donmai.us/wiki_pages/tag_group%3Aposture#dtext-two

︎

3. Character and Body Modifier

This includes character name, hair style, body parts and clothing
Characters can be found on below page
https://danbooru.donmai.us/tags?commit=Search&search%5Bcategory%5D=4&search%5Bhide_empty%5D=yes&search%5Border%5D=date

︎

I generally prompt from top to bottom which goes like this:
Note: links below for each one of them leads to useful tag groups for the segment

  1. Skin, Head, Hair color and hair style including hair length
    https://danbooru.donmai.us/wiki_pages/tag_group%3Askin_color
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aears_tags
    https://danbooru.donmai.us/wiki_pages/tag_group%3Abody_parts#dtext-head
    https://danbooru.donmai.us/wiki_pages/tag_group%3Ahair_color
    https://danbooru.donmai.us/wiki_pages/tag_group%3Ahair_styles
    ︎

  2. Headwear and Hair attires
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aheadwear
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aeyewear
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aattire#dtext-headwear
    https://danbooru.donmai.us/wiki_pages/tag_group%3Ahair#dtext-objects
    ︎

  3. Eye color and description of eyes
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aeyes_tags
    ︎

  4. Expressions
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aface_tags
    ︎

  5. Neck and neck attire
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aneck_and_neckwear
    ︎

  6. Upper torso and those attires
    https://danbooru.donmai.us/wiki_pages/tag_group%3Ashoulders
    https://danbooru.donmai.us/wiki_pages/tag_group%3Abody_parts#dtext-upper
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aattire#dtext-shirts
    ︎

  7. Lower torso and those attires
    https://danbooru.donmai.us/wiki_pages/tag_group%3Abody_parts#dtext-lower
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aattire#dtext-pants
    ︎

  8. Legs, feet and those attires
    https://danbooru.donmai.us/wiki_pages/tag_group%3Abody_parts#dtext-appendages
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aattire#dtext-legs
    https://danbooru.donmai.us/wiki_pages/tag_group%3Aattire#dtext-shoes
    ︎

Overall attires

︎

4. Posing excluding interactions

This includes every pose structures from top to bottom as well

︎
5. Camera position and background

Camera positions includes not only where the camera is but also camera's angle, view and composition of the camera itself

I generally use something like dutch angle, wide shot, fisheye and this allows to create warping and rotating the cameras greatly

You can find them on these pages below

︎

For backgrounds, you can find on this page
https://danbooru.donmai.us/wiki_pages/tag_group%3Abackgrounds
︎
Also, extra things can be found on other tag groups as well so check them out as well
https://danbooru.donmai.us/wiki_pages/tag_groups

︎

6. Lighting, Depth and After details

Lighting includes light source, light style and shadows which is described in this page
https://danbooru.donmai.us/wiki_pages/tag_group%3Alighting

︎

Depths of view like depth of field can be found on the same page as Image Composition
https://danbooru.donmai.us/wiki_pages/tag_group%3Aimage_composition#dtext-depth

︎

After details would be where you want to put extra details like detailed eyes for example
You just need to do detailed + what you want for this part

3. Example Images and Settings

System: Diffusers + sd_embed custom notebook

00000_428649103.png

Generated using Nova Anime XL IL v11.0

Initial image size: 768 x 1344
Steps: 20
CFG Scale: 4.5
Sampler: Euler A
Hires Factor: x1.5
Hires Steps: 16
Hires denoise: 0.4
Upscaler: Nearest-exact
Hires Sampler: Euler A

Prompt:

masterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery, 1girl, solo, cute, pink hair, long hair, choppy bangs, long sidelocks, nebulae cosmic purple eyes, rimlit eyes, facing to the side, looking at viewer, downturned eyes, light smile, red annular solar eclipse halo, red choker, detailed purple serafuku, big red neckerchief, fingers, glowing stars in hand, arched back, from below, dutch angle, portrait, upper body, head tilt, colorful, rim light, backlit, (colorful light particles:1.2), cosmic sky, aurora, chaos, perfect night, fantasy background, BREAK, detailed background, blurry foreground, bokeh, depth of field, volumetric lighting


Segmented Prompt:

[masterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery]
[1girl, solo]
[cute, pink hair, long hair, choppy bangs, long sidelocks, nebulae cosmic purple eyes, rimlit eyes, facing to the side, looking at viewer, downturned eyes, light smile, red annular solar eclipse halo, red choker, detailed purple serafuku, big red neckerchief]
[fingers, glowing stars in hand, arched back]
[from below, dutch angle, portrait, upper body, head tilt, colorful, rim light, backlit, (colorful light particles:1.2), cosmic sky, aurora, chaos, perfect night, fantasy background]
BREAK
[detailed background, blurry foreground, bokeh, depth of field, volumetric lighting]


Negative Prompt:

modern, recent, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra fingers, extra digits, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), sketch, jpeg artifacts, signature, watermark, username, (censored, bar_censor, mosaic_censor:1.2), simple background, conjoined, bad ai-generated

Thumbnail / Cover Image would be exact same setting except initial size is 1488x696

︎

00002_3282896328.png

Generated using Nova Furry XL Illustrious v11.0, use it as non-sorted prompt example

Initial image size: 768 x 1344
Steps: 20
CFG Scale: 4.5
Sampler: Euler A
Hires Factor: x1.5
Hires Steps: 16
Hires denoise: 0.4
Upscaler: Nearest-exact
Hires Sampler: Euler A

Prompt:

masterpiece, best quality, amazing quality, very aesthetic, 4k, high resolution, ultra-detailed, absurdres, newest, scenery, 1girl, furry, anthro, phoenix fox girl, orange fur, long hair, multicolored glowing hair, gradient hair, flame hair, glowing orange hair, floating hair, hair between eyes, gradient orange yellow eyes, very large eyes, uneven eyes, medium breasts, facing to the side, looking at viewer, upturned eyes, intimidating, smirk, smug, grin, crazy smile, long snout, sharp teeth, black headband with fire pattern, black collar with white spikes, cropped leather flaming red jacket over collared impossible white shirt, underbust belt, cinematic lighting, fluffy, portrait, upper body, close-up, foreshortening, dutch angle, (detailed fire forest background, fiery background:1.2), (blue fire:1.2) AND red fire, dark ambient, glow, BREAK, detailed eyes, depth of field, detailed fluffy fur, volumetric lighting


Segmented Prompt:

[masterpiece, best quality, amazing quality, very aesthetic, 4k, high resolution, ultra-detailed, absurdres, newest, scenery]
[1girl, furry, anthro]
[phoenix fox girl, orange fur, long hair, multicolored glowing hair, gradient hair, flame hair, glowing orange hair, floating hair, hair between eyes, gradient orange yellow eyes, very large eyes, uneven eyes, medium breasts, facing to the side, looking at viewer, upturned eyes, intimidating, smirk, smug, grin, crazy smile, long snout, sharp teeth, black headband with fire pattern, black collar with white spikes, cropped leather flaming red jacket over collared impossible white shirt, underbust belt]
[cinematic lighting, fluffy, portrait, upper body, close-up, foreshortening, dutch angle, (detailed fire forest background, fiery background:1.2), (blue fire:1.2) AND red fire, dark ambient, glow] BREAK
[detailed eyes, depth of field, detailed fluffy fur, volumetric lighting]


Negative Prompt:

particles, 3d, rendered, human, multiple tails, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, very displeasing, (worst quality, bad quality:1.2), bad anatomy, sketch, jpeg artifacts, signature, watermark, username, simple background, conjoined, bad ai-generated

Fixed prompt version

00002_3282896328 (1).png

Other settings are same as before

Prompt:

masterpiece, best quality, amazing quality, very aesthetic, 4k, high resolution, ultra-detailed, absurdres, newest, scenery, 1girl, furry, anthro, solo, phoenix fox girl, orange fur, long hair, multicolored glowing hair, gradient hair, flame hair, glowing orange hair, floating hair, hair between eyes, black headband with fire pattern, gradient orange yellow eyes, very large eyes, uneven eyes, facing to the side, looking at viewer, upturned eyes, intimidating, smirk, smug, grin, crazy smile, long snout, sharp teeth, black collar with white spikes, medium breasts, cropped leather flaming red jacket over collared impossible white shirt, underbust belt, portrait, upper body, close-up, foreshortening, dutch angle, (detailed fire forest background, fiery background:1.2), (blue fire:1.2) AND red fire, dark ambient, glow, BREAK, detailed eyes, detailed fluffy fur, depth of field, volumetric lighting

Segmented Prompt:

[masterpiece, best quality, amazing quality, very aesthetic, 4k, high resolution, ultra-detailed, absurdres, newest, scenery]
[1girl, furry, anthro, solo]
[phoenix fox girl, orange fur, long hair, multicolored glowing hair, gradient hair, flame hair, glowing orange hair, floating hair, hair between eyes, black headband with fire pattern, gradient orange yellow eyes, very large eyes, uneven eyes, facing to the side, looking at viewer, upturned eyes, intimidating, smirk, smug, grin, crazy smile, long snout, sharp teeth, black collar with white spikes, medium breasts, cropped leather flaming red jacket over collared impossible white shirt, underbust belt]
[portrait, upper body, close-up, foreshortening]
[dutch angle, (detailed fire forest background, fiery background:1.2), (blue fire:1.2) AND red fire, dark ambient, glow]
BREAK
[detailed eyes, detailed fluffy fur, depth of field, volumetric lighting]

You can see that clothing is better, fur marks are gone and some details are controlled

Other Useful Articles

Thanks for reading

409

Comments