Retro pixelate game : better to upscale before or after

I'm in a Classic Construct mood these days so here is a new question

My work in progress is all in old school pixels.

I make my sprites in photoshop at 1:1 then import them in Construct and upscale them by modifying their width and height x 2 with sampling set to "point".

It works fine but I was wondering what is the best to do : my way or upscaling them in photoshop before importing them in construct.

With my way of doing it, I think it take less memory than importing all sprites scaled x 2 ... but I'm maybe wrong... and the thing is : it would be much more convenient to upscale all of them in photoshop with a script instead of manualy modifying all the height and width after.

Thanks in advance!

// izioq //

R0J0hound Any thoughts ?

Photoshop has a scale nearest neighbor option. Use it for a nice "pixelated" upscale. Upscaling in C2 will only make it more blurry.

Thanks for your reply!

However :

1/ I'm using Classic Construct, not C2.

2/ I know how to upscale it properly in photoshop without blured it. My question is more about knowing if it's better (for performance and userfriendlyness) to upscale in photoshop before importing them in CC or to upscale them directly in CC

I thank you again for your quick reply

Cheers!

Upscaling is in CC not differend then in C2. Plz prove me wrong.

If only Performance counts (and you prefer blurry), big sprites use memory, scaling will use a just little cpu, and a almost no gpu. So i suggest you use sprites with a widht of 1 pixel and a heigh of 1 pixel, upscaling them to any proporties.

I'm not saying that it is different, I just wanted to pointed that I'm using CC and not C2, just to avoid any mistake

Thanks for your explanation about cpu and memory ! ... and in CC if you set "sampling" to "point" in the properties, upscaling them will not make them blurry but pixel perfect... even in C2 I guess.

Cheers!