Text prompts can be used to describe movement but can be challenging to convey all of the visual detail. Dreamina AI does not just use a text prompt, but also provides visual prompts to help with image-to-video generation. It has an Omni reference feature for multiple references (where applicable depending on selected workflow). Another option provided by Dreamina Seedance 2.5 is the ability to interpret movement, spatial relationships, and character positioning through reference-to-video (R2V). These tools are used to help set a creative direction but do not ensure that it will be reproduced exactly. The selection of appropriate references, prompt writing, and analysis of results are still crucial.
What Is Omni Reference in Dreamina AI?
Dreamina AI’s Omni reference feature allows you to convey your visual intent via several reference images. Images of characters can be used to explain appearance, and images of the environment can be used to explain atmosphere, composition, and lighting. Product references can draw attention to unique shapes, materials, and design elements. These references, together with the starting image and written prompt, offer further context for the video generation. When credits and features allow, this may be a useful technique to use in a AI video generator workflow. However, the options and behaviours of the reference are different for each interface and selected model. Supplied images are used to help and guide generation and not to ensure that the generation is the same image.
Choose References According to Their Creative Purpose
Choose references to convey the information required in each image. Facial features, hairstyle, clothes, and even distinguishing accessories should be evident in character identity references. Environment can be used to create lighting, atmosphere, backgrounds, and spatial arrangements. Product images should be clearly seen and demonstrate key product geometry, materials, textures, and branding. Style references are used to convey color treatments and artistic direction, not identity requirements. Prevent the use of images that are in conflict with each other or present incompatible designs or appearances. Clear references help Dreamina AI in interpreting the creative brief and minimize unneeded ambiguity.
Steps to Try Omni Reference to Guide Every Image-to-Video Result Precisely
Step 1: Upload your images and write a guiding prompt
- Sign up or log in to Dreamina, then upload all the reference images you need.
- Open the “Create” tab from the left menu and then select the “AI video” option.
- Write a detailed text prompt that explains exactly how the video should look and feel. State video elements, theme, and concept so the AI understands your vision from the start.
Step 2: Apply Omni reference, choose settings, and generate
- Next, choose a model from the available options at the bottom of the dialogue box.
- Models to test include Dreamina Seedance 2.5, Dreamina Seedance 2.0, and Dreamina Seedance 2.0 Fast. Dreamina Seedance 2.0 Mini, Dreamina Seedance 1.0, and many more versions appear in that list.
- Click on “Omni reference” and guide the AI with multiple reference images for accurate results. Each added image then steers the AI toward the exact video result that you picture.
- Adjust the custom aspect ratio and resolution in the “Aspect ratio” and “Resolution” tabs first.
- Finally, click the “Generate” tab to begin creating your reference-guided video from all your images.
Step 3: Review the result and export your final video
- If the video misses your vision, you can “Edit prompt” or “Regenerate” to refine it.
- Open the “Edit video” option to modify the generated video exactly as you need it.
- Pick “Edit segment” to edit only part of the video while the rest stays intact.
- Lengthen the clip with “Extend video” if you need a bit more time on screen. Raise the clarity with “Creative upscale” to make every detail from your references stand out.
- Dreamina gives you complete freedom to edit your video until it fully matches your references.
- Finally, click the “Download” tab sitting at the very top right of the main interface. This action saves the video file straight onto your local device for any later use.
Six Principles for Using Omni Reference More Effectively
- Reference clarity: Choose images that show key details clearly, without hidden or confusing details.
- Purposeful selection: Give each reference a specific function rather than include irrelevant visual data.
- Instructional consistency: Provide a correlation between written prompts and references, including a description of any intentional changes in pose, setting, or appearance.
- Identity preservation: Explain defining features of character/objects and look at them as they move.
- Spatial awareness: Give clear information about the position of the subject and the relationship of foreground and background elements and other objects in the scene.
- Iterative evaluation: Check each result for comparison with your goal; add or remove references or prompts as needed when certain information is lost.
Understand How R2V Reference-Based Generation Supports Creative Direction
The R2V technology is available in Dreamina Seedance 2.5 for visual material and other inputs to generate videos based on their content. These can vary from images, video, green screen footage, and white model references, depending on the workflow. These inputs may convey character location, character relations, and movement. Omni Reference is a particular aspect of the interface, while R2V refers to the general approach of reference to video. Both require checking the footage that has been produced for mistakes. Even if the references and instructions are consistent, there may be variations when it comes to interpretation.
Use Omni Reference for Different Image-to-Video Projects
Dreamina AI offers diverse creative uses with its visual generation powered by references. Portraits can be used to help convey appearance and direction of a scene in character animation. Materials, shapes, and unique design features can be accentuated in product presentations. References of atmosphere, lighting, and composition are beneficial in a cinematic environment. Character design can be linked with appropriate setting references through illustrated storytelling. Promotional graphics may be a mix of the existing artwork and additional direction for style. Different occurrences of a multi-character scene can be referenced separately, and positioning can be explained. The results are dependent on the quality of the reference, model behavior, and complexity of the movement.
Troubleshoot Reference Conflicts and Unexpected Results
If there are any sudden changes of characters, make descriptions of defining characteristics more forceful and remove contradictory information. If composition changes, explain what you want the subject to be in and eliminate any confusing background/camera directions. If it is a motion other than appearance requirements, explain the motion separately. If it is a visual artifact, examine confusing information in the source and find the affected parts. Make localized edits with support if needed or remake the clip if extensive edits are required. Every change should be for a specific issue, not for over-complications.
Evaluate the Final Video Against the Original Creative Brief
See if the subject, setting, and style of art are identifiable throughout the clip. Look for inconsistencies in movement for unstable details, unexpected changes, and spatial relationships. Check camera behaviour – does it help the intended action? Check against the purpose given for each reference image. Select fast revisions, regeneration, or editing at local spots depending on the remaining problem. Check the video you export before you post it or use it in a larger production.
Conclusion
Dreamina AI includes source images, written prompts, and reference-guided generation to assist with more intentional video creation. It has some Seedance features that can be used to convey character identity, scene set-up, and movement. But references can’t ensure precise reproduction and cannot fix all discrepancies in the picture. The image selection, instructions, and iterative evaluation are clearer, and this offers a better foundation. Use Omni reference as creative guidance, and then check each of the results against the original artistic goal.
