We encountered a perplexing issue in our studio concerning the pixel aspect ratio when comparing images, and we’re seeking insights from the community. The challenge arises when dealing with images that possess a pixel aspect ratio (PAR) differing from 1.0, particularly when employing alpha channels and various overlay methods.
Initially, our standard image comparison workflow involved the ‘over’ method. This approach, while seemingly straightforward, led to inaccuracies when comparing images with alpha channels. The alpha channel’s transparency data was inadvertently incorporated into the comparison, skewing the true RGB channel differences.
To circumvent this alpha interference, we transitioned to the ‘replace’ method for image comparisons. This effectively isolated the RGB channels, excluding the alpha mask from the comparison process. For a while, this solved our immediate problem and provided a more accurate visual comparison of the color information.
However, a new challenge emerged when we began working on a project with images that had a pixel aspect ratio different from the standard 1.0 (square pixels). We observed that images were being unexpectedly squished or distorted when using the ‘replace’ method, specifically when adjusting wipe parameters for detailed comparison.
This distortion seemed to be linked to the position of the wipes used to reveal parts of the overlaid images. When wipes were positioned at the image edges, the squishing was less apparent. However, as we moved the wipes inward to focus on specific areas, the distortion became more pronounced on the image with the modified wipes.
Intriguingly, we discovered a temporary workaround. By slightly nudging the second wipe, even by a minimal amount, the images would realign, and the distortion would correct itself, allowing for a proper overlay.
While this manual adjustment provides a temporary fix, it’s far from ideal for a streamlined workflow, especially when frequent comparisons are necessary, and detailed corner inspections are required. Furthermore, precise wipe placement becomes crucial, adding unnecessary complexity to a routine task.
We also experimented with the ‘difference’ operation to further understand the behavior. Similar to the ‘replace’ method, the ‘difference’ operation appeared correct when wipes were at the edges. However, moving the wipes again introduced discrepancies and distorted the comparison.
This behavior suggests a potential issue related to how the software handles image reloading or re-rendering when wipes interact with image boundaries, particularly when pixel aspect ratios are involved. It seems the act of moving the wipes away from the edges triggers a recalculation or reload that incorrectly applies the aspect ratio, leading to the observed distortion.
Has anyone else encountered similar challenges when performing Aspect Ratio Compare operations, especially concerning image overlays and wipe interactions? We are keen to understand the underlying cause of this behavior and explore more robust solutions for accurate image comparison in scenarios involving non-square pixel aspect ratios. Any insights or shared experiences would be greatly appreciated as we work towards a more reliable image comparison workflow.
