WHAT THEY MIGHT HAVE LOOKED LIKE
This is a new section we have had some considerable debate over whether to include or not.
It concerns experimentation with AI (Artificial Intelligence) to use its processing power to interpret what some of the images we have in our collection in questionable condition might have looked like in their original form.
There’s a moral dilemma attached to this, not the least of which is how accurate can AI be in reproducing something for which no authentic reference exists?
By definition what we are giving AI to ‘restore’ is in bad or compromised condition and details may not be clear, so none of the ‘restorations’ are claimed to be 100% authentic, simply impressions of what the image MIGHT have looked like.
Indeed we are more than conscious that there are fine details in these ‘restorations’ which AI simply cannot interpret, these imperfections don’t need to be pointed out to us – we are aware of them, eg small lettering.
Ultimately the charm and beauty of old and faded artifacts of originals has a merit of its own, these reproductions are never intended to replace that authenticity.
We have deliberately not put originals side-by-side with these ‘restorations/reproductions’, if you wish to compare old with new, please go to the main section and cross refer by alphabetical order.
These examples are for interest only, each image is watermarked ‘AI Enhanced’ and they are for reference – and absolutely not for reproduction.
Note – All interpretations are generated by Google Gemini Nano Banana 2.
