This article will give you a general framework for your workflow and explain some of the reasoning behind it however, you may find that a different workflow is best for your use case. Please feel free to share this knowledge with other users on our forums.
In general, you should use your AI products as early in your workflow as possible. This is because our AI models have been trained on millions of data sets consisting of original image files so you will get the best results from original image files. Typically, you won't see significant adverse effects if small changes have been made so don't worry if you have some important pre processing in your workflow that needs to happen before you use our products.
Gigapixel AI Last?
In our AI products, the framework should be DeNoise AI -> Sharpen AI -> Gigapixel AI.
The reason Sharpen AI needs to come after DeNoise AI is to reduce the risk of sharpening any noise and creating artifacts.
We recommend using Gigapixel AI last because it increases your image and file size which means it takes longer to process in our other AI products. This doesn't necessarily mean it has to come last however it is the most commonly used workflow and if you're editing large batches can end up saving you quite a bit of time.
One notable exception to the Gigapixel AI last approach is when you have a very small image. An example would be a very old image or if you've suffered a data loss and only the previews were recovered.
In this case it can be beneficial to upscale by a small amount in Gigapixel AI first such as x1.5 or x2, then use DeNoise AI -> Sharpen AI, then use Gigapixel AI again to get to the size you want.