CivArchive
    XavierNocturne - v1.1 | Lumen
    NSFW
    Preview 142385526
    Preview 142379596
    Preview 142377476
    Preview 142379764
    Preview 142380166
    Preview 142386043
    Preview 142388617
    Preview 142389776
    Preview 142387999

    XavierNocturne v1.1 Lumen - Update Notes

    Version Update: Improved Lighting & Refined Anatomy

    Thank you for all the support, downloads, and feedback. This update focuses on two major areas that users requested the most: lighting quality and anatomical consistency.

    What’s New

    • Significantly improved light response and volumetric behavior

    • Better subsurface scattering and rim light handling

    • More stable and natural anatomy (especially hands, torso, and facial structure)

    • Cleaner rendering of fur and skin transitions

    • Improved overall coherence in complex scenes

    Recommended Positive Prompt Base

    text

    score_9, score_8_up, score_7_up, masterpiece, best quality, absurdres, very aesthetic,

    You can freely expand from this base depending on the subject.

    Lighting Control Guide

    Lumen responds very well to directional and atmospheric lighting prompts. Here’s how to guide the light:

    Soft / Natural Light

    text

    soft lighting, diffused light, gentle volumetric lighting

    Dramatic / Cinematic Light

    text

    dramatic lighting, strong rim light, volumetric lighting, cinematic lighting

    Backlight / Rim Light

    text

    backlit, strong rim light, light from behind, glowing edges

    Side Light

    text

    side lighting, light from the left, light from the right, dramatic side light

    Top / Overhead Light

    text

    overhead lighting, light from above, god rays

    Warm vs Cool

    • Warm: warm lighting, golden hour, soft orange light

    • Cool: cool lighting, blue hour, moonlight

    Combining volumetric lighting + rim light usually gives the most cinematic results.

    Recommended Negative Prompt

    text

    score_4, score_5, score_6, low quality, worst quality, bad anatomy, bad hands, missing fingers, extra digits, mutated hands, poorly drawn face, blurry, jpeg artifacts, text, watermark, logo, censored, extra limbs, deformed, ugly

    Recommended Settings

    • Sampler: DPM++ 2M SDE or Euler a

    • Scheduler: Karras , normal , beta

    • Steps: 28–32

    • CFG Scale: 2.8 – 3.5

    • Clip Skip: 2

    • Resolution: 832×1216 / 1024×1024 / 1216×832 (or higher with upscale)

    Recommended Settings for CivitAi generator

    . Sampler: Euler a / DPM++ 2M SDE

    . Steps: 25–35

    . CFG: 3.5–5.5

    . Clip Skip: 2

    Hires.fix Denoise: 0.3– 0.5

    . Hires.fix: 1.3–1.5 + R-ESRGAN 4x+ Anime6B, Remacri

    Free_2U Integration

    Lumen works very well with Free_2U (FreeU v2). It improves structure, sharpness, and overall coherence without needing extra steps.

    Recommended Starting Settings (Balanced)

    Parameter Value Description
    b1 1.2 Backbone scale (Stage 1) – controls overall structure and stability
    b2 1.35 Backbone scale (Stage 2) – affects finer details and contrast
    s1 0.90 Skip scale (Stage 1) – reduces low-frequency noise in early blocks
    s2 0.20 Skip scale (Stage 2) – cleans up high-frequency artifacts

    Alternative Presets

    Mild (safer, less risk of oversharpening):

    • b1: 1.1

    • b2: 1.25

    • s1: 0.95

    • s2: 0.30

    Strong (more detail & punch):

    • b1: 1.3

    • b2: 1.45

    • s1: 0.85

    • s2: 0.15

    Tips

    • Start with the Balanced preset.

    • If you notice halos, crunchy edges, or over-sharpened fur/skin — lower b1 and b2 first.

    • s2 has a strong effect on fine detail cleanliness. Lower values = cleaner but can sometimes lose softness.

    • Free_2U pairs especially well with the improved lighting and anatomy of this version.

    Post-Processing Tips for a Cinematic Look

    For a more film-like and polished result, try this combination after generation:

    1. Image Sharpen - light to medium strength (avoid over-sharpening)

    2. Contrast Fix - subtle increase for better depth

    3. Image Film Grain - low to medium amount for a natural cinematic texture

    This trio helps the image feel less “digital” and more like a still from a movie.

    Thank you again for downloading, testing, and supporting XavierNocturne.
    Your feedback directly shapes the next updates.

    If you create something cool with this version — feel free to share it.
    Enjoy generating.

    XavierNocturne v1

    Indigo/XavierInHell 65/35

    Merge of Indigo Furry Mix (65%) and Xavier (35%).
    This model combines the strong anatomy, body proportions, and lighting of Indigo Furry Mix with the distinctive style and aesthetic of Xavier. It delivers solid structure while preserving a unique visual character, making it well-suited for detailed furry and anthro generations.

    Compatibility

    • Fully works in ComfyUI

    • Fully works in Automatic1111 / Forge / reForge

    Sampler:

    • euler / euler_a (best)

    • dpmpp_2m

    • dpmpp_sde

    Scheduler:

    • normal

    • karras

    • beta (best)

    CFG Scale: 3.5 – 6
    Steps: 25 – 35

    Recommended prompt:

    Positive:
    masterpiece, best quality, very aesthetic, detailed eyes, detailed face, intricate details, beautiful lighting, dramatic lighting, volumetric lighting, sharp focus, highly detailed,

    Negative:
    lowres, worst quality, bad quality, bad anatomy, bad hands, extra fingers, missing fingers, blurry, low quality, text, watermark, ugly, poorly drawn face, poorly drawn hands,

    Strengths

    • Good anatomy and body structure

    • Quality lighting and depth

    • Distinctive style from Xavier

    • Handles detailed fur, clothing, and complex scenes well

    Notes

    • Best results at moderate CFG

    • Compatible with most SDXL LoRAs

    • Optional FreeU_V2 or ADetailer can further improve details

    • Works reliably in both ComfyUI and A1111-based interfaces

    Description

    XavierNocturne v1.1 Lumen - Update Notes

    Version Update: Improved Lighting & Refined Anatomy

    Thank you for all the support, downloads, and feedback. This update focuses on two major areas that users requested the most: lighting quality and anatomical consistency.

    What’s New

    • Significantly improved light response and volumetric behavior

    • Better subsurface scattering and rim light handling

    • More stable and natural anatomy (especially hands, torso, and facial structure)

    • Cleaner rendering of fur and skin transitions

    • Improved overall coherence in complex scenes

    Recommended Positive Prompt Base

    text

    score_9, score_8_up, score_7_up, masterpiece, best quality, absurdres, very aesthetic,

    You can freely expand from this base depending on the subject.

    Lighting Control Guide

    Lumen responds very well to directional and atmospheric lighting prompts. Here’s how to guide the light:

    Soft / Natural Light

    text

    soft lighting, diffused light, gentle volumetric lighting

    Dramatic / Cinematic Light

    text

    dramatic lighting, strong rim light, volumetric lighting, cinematic lighting

    Backlight / Rim Light

    text

    backlit, strong rim light, light from behind, glowing edges

    Side Light

    text

    side lighting, light from the left, light from the right, dramatic side light

    Top / Overhead Light

    text

    overhead lighting, light from above, god rays

    Warm vs Cool

    • Warm: warm lighting, golden hour, soft orange light

    • Cool: cool lighting, blue hour, moonlight

    Combining volumetric lighting + rim light usually gives the most cinematic results.

    Recommended Negative Prompt

    text

    score_4, score_5, score_6, low quality, worst quality, bad anatomy, bad hands, missing fingers, extra digits, mutated hands, poorly drawn face, blurry, jpeg artifacts, text, watermark, logo, censored, extra limbs, deformed, ugly

    Recommended Settings

    • Sampler: DPM++ 2M SDE or Euler a

    • Scheduler: Karras , normal , beta

    • Steps: 28–32

    • CFG Scale: 2.8 – 3.5

    • Clip Skip: 2

    • Resolution: 832×1216 / 1024×1024 / 1216×832 (or higher with upscale)

    Free_2U Integration

    Lumen works very well with Free_2U (FreeU v2). It improves structure, sharpness, and overall coherence without needing extra steps.

    Recommended Starting Settings (Balanced)

    ParameterValueDescriptionb11.2Backbone scale (Stage 1) – controls overall structure and stabilityb21.35Backbone scale (Stage 2) – affects finer details and contrasts10.90Skip scale (Stage 1) – reduces low-frequency noise in early blockss20.20Skip scale (Stage 2) – cleans up high-frequency artifacts

    Alternative Presets

    Mild (safer, less risk of oversharpening):

    • b1: 1.1

    • b2: 1.25

    • s1: 0.95

    • s2: 0.30

    Strong (more detail & punch):

    • b1: 1.3

    • b2: 1.45

    • s1: 0.85

    • s2: 0.15

    Tips

    • Start with the Balanced preset.

    • If you notice halos, crunchy edges, or over-sharpened fur/skin — lower b1 and b2 first.

    • s2 has a strong effect on fine detail cleanliness. Lower values = cleaner but can sometimes lose softness.

    • Free_2U pairs especially well with the improved lighting and anatomy of this version.

    Post-Processing Tips for a Cinematic Look

    For a more film-like and polished result, try this combination after generation:

    1. Image Sharpen - light to medium strength (avoid over-sharpening)

    2. Contrast Fix - subtle increase for better depth

    3. Image Film Grain - low to medium amount for a natural cinematic texture

    This trio helps the image feel less “digital” and more like a still from a movie.

    Thank you again for downloading, testing, and supporting XavierNocturne.
    Your feedback directly shapes the next updates.

    If you create something cool with this version — feel free to share it.
    Enjoy generating.

    FAQ

    Checkpoint
    Illustrious

    Details

    Downloads
    158
    Platform
    CivitAI
    Platform Status
    Available
    Created
    9/10/2026
    Updated
    9/15/2026
    Deleted
    -

    Files

    xaviernocturne_v11Lumen.safetensors

    Mirrors