# Arcana FilmMaking Process

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2F0FYhZKBLU6qTd3TD4dqB%2Fimage.png?alt=media&#x26;token=cc073b19-2e40-4d1c-9b72-3f1f177e9b40" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2FKDtMjNOj818SdXo8UKU0%2Fimage.png?alt=media&#x26;token=8b24009d-565a-4086-9c1a-097381bb33d6" alt=""><figcaption></figcaption></figure>

## OVERVIEW:

1. Train Models for your World and your Characters
2. Use X-Ray to turn Storyboards into Image Assets
3. Use Transform to turn image assets into your world's aesthetic
4. Inpaint your assets with your character specific models for character consistency
5. Send images to Motion Tab

{% embed url="<https://youtu.be/3mRZ1pRvEhc>" %}
From Storyboard to X-Ray Generated Images to Final Video
{% endembed %}

## STEP 1 - Model Training

Go select CREATE YOUR OWN MODEL:

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2FrgoYO07APGqzDcs25lVA%2Fimage.png?alt=media&#x26;token=9df357e2-18e9-41b5-a870-202d79f4d90a" alt=""><figcaption></figcaption></figure>

For a MOVIE you need at least TWO Models. One trained on your world and Aesthetic, and another trained on a specific actor you want to cast. Without the latter you cannot achieve character consistency, and without the former you cannot create visual consistency across your shots.&#x20;

## STEP 2 - Using X-Ray to turn Storyboards into Image Assets

Upload your storyboard into X-Ray (soft, sharp or ultrasharp), prompt what you want to see, don't be precious about the output image just make sure it's accurate enough so you can transform it in step 3.

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2FrGmDRYwH5BKdC0uEeQA5%2Fimage.png?alt=media&#x26;token=8a59a573-b10e-4e0e-9cf4-e9ca37f0e53e" alt=""><figcaption></figcaption></figure>

## STEP 3 - Using Transform to turn image assets into your world

Play with transformation strength between 45%-75%. Make Sure you select your WORLD MODEL that you custom trained. Don't be afraid to iterate on the image with smaller transformation strength until you are happy.&#x20;

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2FVF0P1vHejy73KRgnowqN%2Fimage.png?alt=media&#x26;token=9a5abffe-534e-45f7-bb32-1a2ce6133691" alt=""><figcaption></figcaption></figure>

## STEP 4 - INPAINTING FOR CHARACTER CONSISTENCY

SELECT your CHARACTER MODEL and begin inpainting the faces. Usually 65-81% transformation strength gets you home. Iterate until happy.&#x20;

<figure><img src="https://2624966838-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKcFb4ihnfBnfd0stkhMX%2Fuploads%2FVN3usrqyV920XTBUK2Ho%2Fimage.png?alt=media&#x26;token=6695c7cd-e5ac-4f29-8e05-17999c790cb4" alt=""><figcaption></figcaption></figure>

## STEP 5 -SEND IMAGE TO MOTION TAB

1. Drag your image to motion, or select the motion button to send it to Motion (duh lol)
2. Select the Motion Video Model You Want (We love kling)
3. Select the Creativity Level (we tend to stick to default 0.5 but increase if facing prompt adherence issues)
4. Prompt what you want to see (e.g. CAMER PUSH IN, Character walking, etc...)
5. For best practices do at least 4 generations and cherry pick your favorites

## VIDEO TUTORIAL

{% embed url="<https://youtu.be/xsuhZSmQhI4>" %}

## GALLERY OF RESULTS

{% embed url="<https://www.instagram.com/arcana_ai/reel/DGTR5RsutiY/>" %}
Click to play
{% endembed %}

{% embed url="<https://www.instagram.com/arcana_ai/reel/DGRnRfBucvf/>" %}

{% embed url="<https://www.instagram.com/arcana_ai/p/DGRo_pANvP4/>" %}

{% embed url="<https://www.instagram.com/arcana_ai/p/DGaB2nMOTmg/>" %}
