Feature Requests for Gigapixel AI
This document outlines feature requests for Gigapixel AI, aimed at improving its functionality and user experience.
1. Resize to Longest Side
-
Description: Implement an option to resize images based on their longest side while preserving aspect ratio.
-
Motivation: This streamlines the upscaling workflow and provides greater control over output image dimensions, enabling easier megapixel estimation. This feature is already present in Photo AI and would be beneficial in Gigapixel AI, which is primarily focused on upscaling.
2. Grain Control
-
Description: Introduce a feature to manage grain levels during upscaling.
-
Motivation: This allows fine-tuning detail levels for a more natural or artistic aesthetic, particularly valuable when upscaling older or low-resolution images. Similar features exist in Sharpen AI, Topaz Video Enhance AI, and various competitor software.
3. Second Pass Upscaling and Detail Recovery
-
Description: Implement a two-part enhancement:
-
Second Pass Upscaling: An option for a second upscaling pass dedicated to recovering fine details, further enhancing sharpness and fidelity in complex textures and patterns.
-
Recover Original Details: An option to blend the upscaled output with the original image, minimizing artifacts and simplifying the process of achieving high-quality results.
-
4. Adobe After Effects Integration
-
Description: Develop a dedicated plugin or wrapper for seamless integration with Adobe After Effects.
-
Motivation: Allows direct incorporation of Gigapixel AI’s upscaling capabilities into video editing workflows, unlocking new creative possibilities. The existence of previous (albeit limited) attempts at similar integrations suggests its feasibility.
5. Enhanced Batch Processing
-
Description: Improve batch processing by allowing users to define the maximum number of simultaneous processes.
-
Motivation: Significantly reduces processing time for large image batches.
6. Updated Models and 1x Downscaling
-
Description:
-
High-Compression Model: Introduce a new or updated AI model specifically trained for upscaling highly compressed images to address compression artifacts effectively.
-
Face Enhancement Models: Update the existing face enhancement models for improved accuracy and performance.
-
1x Downscaling Fix: Address the issue with the 1x downscale option producing pixelated results. Consider reverting to previous 1x models or ensure the 2x to 1x downscaling process functions correctly.
-
Closing
These enhancements would significantly improve Gigapixel AI’s usability, efficiency, and overall user experience. Thank you for your time and consideration of these requests.
