The following images show what can be achieved with no unpleasant bending or exercise whatsoever… (\u00ab\u00a0…wrestle poodles and win\u00a0\u00bb Viv Stanshall<\/em>)<\/p>\n\n\n\n <\/p>\n\n\n\n In truth there’s not a HUGE difference, and certainly nothing at this scale will show up (click on the little ‘+’ button and enlarge it as much as it allows) but the noise (albeit only ISO 200) has been radically ‘smoothed’ and if I were attempting a very large print, it would be the way to go.<\/p>\n\n\n\n Here’s another example:<\/p>\n\n\n\n I’m not much of a ‘post treatment’ junkie – principally ‘cos I’m not very good at it – but I can imagine this software coming into it’s own when I need to produce very large copies of my images. The ‘over-sizing’ program Gigapixel AI makes a very good job (I’ve only tested the 2x computations) of the images – keen to try a couple of images at large scale with and without the software changes, just to see how the quality stands up.<\/p>\n\n\n\n One of the risks of ‘AI’ is that the smoothing applied smooths to the point of creating weird artefacts which pollute the image – training is going to be necessary here…<\/p>\n","protected":false},"excerpt":{"rendered":" I must admit I didn’t think I’ve ever be singing the praises of a piece of software with AI (machine\/automatic learning) but I am slowly coming round to the fact that TOPAZ might actually have something here. I waited until a suitable moment (A: I had some spare dosh, and B: there was a VERY … <\/p>\n![\"\"](\"https:\/\/www.iangrandjean.com\/blog\/wp-content\/uploads\/2022\/12\/Roquelaure-1-scaled.jpg\")
<\/a><\/a><\/a><\/a>