Upscaling within Photoshop

If anyone would have told me 10 years ago that we could upscale a 150px image such as this one:

to a 2400px image like this:

...I would have laughed straight in your face.

That we are pulling this off with free and commercial tools... Unimaginable.

Photoshop offers many kinds of upscaling, including AI upscaling, but all of them are unnecessarily hidden in multiple places in Photoshop.

Super Zoom can do the job, but how to do that is also unnecessarily obscure. To have Super Zoom upscale without cropping, the Output menu must be set to New Document. Like I said…not obvious to anyone.

Currently, you’ll find different upscaling options in these corners of Photoshop…

Traditional upscaling:

Image Size dialog box, Resample menu choices (several, including Bicubic Smoother, Preserve Details, and Preserve Details 2.0).

AI upscaling (newer and often better):

Neural Filters, Super Zoom; set Output to New Document to upscale the entire image without cropping, or set Output to New Layer to upscale the layer but limit its pixel dimensions to that of the containing document (crop).

Camera Raw only when opening an entire document (won’t work as Filter > Camera Raw Filter); in the Enhance command (hidden in More Settings menu, filmstrip menu, or by right-clicking image), use Super Resolution.

People would be more aware of these upscaling options if they were more discoverable at the top level of Photoshop.

I use AI (Topaz) for upscale / sharp / denoise. It's a whole different ball game.

Super Zoom doesn't need to crop, that's just a secondary function. At basic operation it scales the full original to 2x linear dimensions. One pixel becomes four.

You probably already know the following, so this is just general information for others reading this:

There is nothing magical about the 300 ppi number. It's a theoretical upper limit, beyond which no improvement is possible - but it's not a lower limit. You can print eminently sharp at lower ppi numbers, down to 200 ppi or so, on matte paper even lower - if the file is prepared and sharpened optimally.

The 300 number relates to a standard book/magazine halftone screen of 150 lines per inch (lpi). The theory is that at twice that number, it is no longer possible to distinguish individual pixels. They are totally masked by the halftone screen. So in other words - it's not even a sharpness limit, it's a smoothness limit.

And of course, for bigger sizes the ppi (and lpi) requirement drops rapidly because it's seen from farther away.

What I'm getting at is that upscaling isn't needed nearly as often as most people think. There is no reason to upsample just because ppi drops below 300. Upsampling never improves anything, regardless of the algorithm used, ai or not. On the contrary, it will often just introduce artifacts.

Upsampling is possible with the Image Size-dialog so please elaborate.