Photoshop supports any size layers. Layers can have any aspect ratio and size and are not limited in size to canvas size. When you scale up a raster layer in size its done using interpolation. If you save a layered file format all layers pixels need to be recorded in the layer file. Pixel data is the largest data part of image files. Scaling Layers up in size can greatly increase the size of a layered image file. The same document saved in an image file format that does not support layers will be much smaller. For only a single composite layer the size if the documents canvas will be saved.
Hello, If I'm not mistaken, the transparent areas are more compressible than the non transparent ones, given the latter have also RGB values.
I would think that two areas that are equal in size where one area is all transparent and the other area is all white 255 255 255 RGB that these areas would compress to about the same amount of data. Details show through different value RGB values and will not compress as well as a solid white wall. Not all images are compressed. File sizes differ greatly depending one image file format save.
Thank you for responding. I always thought that the canvas would limit how much info a layer could hold. If an image is used on a layer of an 8x10 PS document... and is large enough to fill the layer, why should scaling it up make the document file size larger if the the same amount of pixels fill the same layer? Afterall, what color the pixels are, (fuzzy or sharp) shouldn't make the file size larger if the total amount of pixels that fit on a layer is the same. Does PS count the rest of the scaled image (that goes beyond what the canvas shows) as part of the file size?
Additionally, if a 300 dpi shot is dragged into a 72 dpi PS doc., does the hi res shot drop to the target PS file's dpi?
Given the fact that the content outside the visible area of the image is saved (it is often called "big data"), unless you trim or crop it out, yes, I expect that data to have some "weight"...
About your second question, yes, it's the number of pixels that matters. Try it with a document of the same pixel amount, but different resolution, drag one on the other, it'll be the same, as Photoshop renders a pixel at a pixel.
It's still strange... I deleted the big data on the scaled up layer so just the visible layer was filled with image (nothing left over), then I dragged a low res shot in as another layer(scaled to fit layer). The layered file size went up from 101M to 106M. So the new layer weighed in at 5M it seems. Yet the layered file size drops from 106M to 5M when I delete that first layer which I trimmed to fit the layer. ! How can one layer of pixels in a file weigh more than another?
Out of curiosity, do you have a RAM shortage (e.g., resulting in slow system performance)? If not, worrying about how Photoshop allocates RAM to maintain its data structures, while academically interesting (well, I find it such), probably doesn't tell us all that much that's truly useful.
If you ever want to understand more about how Photoshop works under the covers, and you're a dyed-in-the-wool geek like me, a look at the Photoshop SDK can tell you volumes. Writing plug-in software isn't for the faint-of-heart, but it will tell you tons about how things work. Funny thing... I didn't used to think in rectangles...
Plenty of RAM. Thanks! This whole question came about yesterday, when I had to send a job to a client and was surprised at how large the packaged InDesign file was (including links, which I created in low res). I guess all the heavily scaled up photos (with portions that were still weighing in, BUT NOT SEEN beyond the canvas) were confounding me.
Yes, the [ ] Delete Cropped Pixels setting of the crop tool can result in interesting effects. In addition to the storage space, many (most? all?) filters actually work on the data beyond the visible edge of the canvas.
I've found that I've ended up turning off that feature and just being very sure about what I'm doing when I crop.
I NEVER crop the extended pixels out of a doc I'm working on. Too many times has the client came back and changed a size and say, maybe go horizontal instead of vertical, where if I'd cropped them out, I'd be back to square one. Having it, I can just chnge canvas size and go from there. IF a client requires a layered file (which I rarely will ever send) I'll save a "cropped" version for them to knock down the file size.
Our differences in needs with Photoshop of course lead to different choices.
But beyond that I find it always possible to go back to the original image if one needs more - assuming you keep your raw files. An added bonus is that it just may be possible to get a better image if one goes all the way back to the source data. This could be true, for example, if one were to have prepared the image using the raw converter in Photoshop CS5 initially. The converter just keeps getting better and better. I was just noting this morning how much better a particular night shot I had done back in 2011 came out with the newest version of the converter with the PV2012 development process.
Yes the main reason these days to upgrade to the latest version of Photoshop is the get the ACR updates. Still CS6 was a ruff and trying upgrade and I think the Adobe Cloud is gray and I don't mean it has a silver lining. I'm happy here on the ground.