Regarding the size of the whole project, the bigger the better. The reason behind that reasoning is simple: it is often hard to make everything accurate to the last pixel (or at least it takes lot of patience, time and experience), so if you work on a big canvas, once resized to DPC standards, your masks and technical adjustments will look flawless. Of course, working with large resolutions require fair amount of RAM on your computer, as layers can stack up quickly. I usually use 9000x6000 pixels for my projects, and Photoshop often uses 40-50Gb of RAM, while project files are around 4-5Gb on disk. Reasonably, try to work on a canvas at least 3 times larger than your final file (for example 3600 pixels on the largest side for a DPC project).
About the size of individual images, first, you have to be sure about how each photo is going to be used. If chances are you are going to resize/deform a given layer a lot, convert it first to a "Smart Object" (right-click on the layer, the option is just there). This will tell Photoshop to remind about the original photo, so that you don't lose details (even if you first make it very small, then very large again: details will reappear). Of course, if you know a particular photo is going to be used just for some detail in the background, there is no need for that.
Once you positioned and resized each of your photos on the canvas, zoom at 100%. You will probably see that the photos you made smaller are much more detailed than the big ones that you didn't resize much. In order to make everything appear smooth, I advise to actually blur a bit the photos you made smaller, so that they are not much more sharp and detailed than the rest. It helps mixing things together.
Hope this helps! |
