4 ms·
The same thing will likely be true of raw files, as we get better (and probably machine-learning based) algorithms for decoding those into standard image format
by lambdaone 2y ago
The same thing will likely be true of raw files, as we get better (and probably machine-learning based) algorithms for decoding those into standard image formats. I wonder what fraction of projects currently preserve their raw files?
- ethbr1 2y agoAt some point you're extrapolating new information, not unpacking encoded existing information.
- ISV_Damocles 2y agoThat would also be true with analog film when you start reaching the film grain size.
- ethbr1 2y agoI'd be curious about the fidelity floor in analog vs digital optics. I'd guess(?) that you might be able to do more information reconstruction from analog + lense parameters + film parameters than digital + lense parameters? Simply by virtue of digital being quantitized at some point. (But signal processing is far outside my area of expertise, so honestly curious)
- surfingdino 2y agohttps://www.adox.de/Photo/films/cms20ii-en/ https://www.adox.de/Photo/films/cms20ii-en/ is the current max res standard for bw film. Try getting that from your monochrome Leica
- surfingdino 2y agoTopaz AI does is with some degree of success, but it starts to fall apart fast. You can use it for creative projects. Nick Knight used it to great effect when he was working on https://www.youtube.com/watch?v=nvaU7rtV9LI https://www.youtube.com/watch?v=nvaU7rtV9LI but please remember that the final images had a lot of human tweaking and painting, so no, AI will not turn your iPhone snaps into a Dutch master's painiting.