Selective Recover through Photoshop Automate

The Preview facility in GigaPixel 8 for restore allows amall areas to be previewed as to how they may look. If I create a small selection in Photoshop as a new layer via a selection tool and then try to restore that layer, the software does not detect the “empty” pixels, but restores the entire image. Surely the ability to select small areas for restoration using this method simply mimics the preview facility.

If it worked as it should then users can restore parts of an image that are important and not huge areas of sky or other landscapy elements where restored detail is not an issue.

Ian

Hi Ian.

You’re correct

If Topaz somehow are able to implement this feature great however, in the meantime here’s a work around.

You already have this ability to Create a smarter Restore from Photoshop layers.

If you turn your new Layer into a Smart Object

Then double click on the Smart Object to open it and send this Layer to Gigapixel via File Automate

Next apply your Enhancements and send back to Photoshop and save the Smart Object

Only the contents of the Smart Object will be restored

You now have the choice of leaving the Layer as a Smart Object or Rasterize the Layer

Hope this helps

Wow, a possible workaround, many, many thanks!

Might persist a bit longer!

Ian

I have noticed that the image brought back into photoshop is cropped and in a different position as a result, is this me or the GigaPixel/Photoshop software. Doesnt matter whether its rasterised or not, same effect happens.

** Fixed ** That was me, had to match resolution of image to the DPI of Gigapixel!

1 Like

My woes continue, whilst from a distance the restoration creates more definition, the process introduces too many artifacts back into the image, very pixelated and too noticeable to use! Not sure whether this is the model, or not.

Very, Very Disappointing…

Hi Ian.

If you’re still asking about the Generative AI Models Recover & Redefine then I’m afraid these Models were only designed for Low Resolution Digital images and not Analogue images

Topaz recommend using the Basic Models for larger images

I have put Topaz own information about these Models below I’ll also, post the link to the original document at the bottom

Recover

This model brings the best fidelity in upscaling for extremely low resolution images.

Use images that are 1000px and smaller to get the most optimized result.

Redefine

Use this model to take your imagination to the next level. You can prioritize realistic upscaling for either fidelity or creativity.

image

Make the most of Redefine when you come across the following type of images:

  • Low resolution images (either from crop or otherwise)
  • Blurry images
  • AI-Generated images

Are you aware you can also access the Legacy (previous) Models from version 6.3.3 which gives you more options to play with here’s a link

Finally here’s a link for you to contact Topaz Support directly

Topazlabs Support
support@topazlabs.com

Hope some of this helps