|Moonrise, resized for desktops - 1920 by 1080 @ 72 pixels/inch|
I have played with a couple of different formats for my published pictures (and the use of a logo).
My originals (in RAW out of my cameras) when normalised at 72 pixels per inch (normal PC screen resolution) are 17,280 by 11,520 pixels. That is 72" by 14.4" or 1.8m by 1.2m. That works out at an awful lot of pixels but seems to be about the standard original size.
Before uploading any of these, I 'shrink' them, and in some cases change the shape by cropping them.
There is not a lot of guidance about optimal 'shrunk' size although Google+ Help has posted guidance about its use of community photos, and i suppose it will depend on the picture. For example, some very grainy pictures i took of Jupiter last night will contain no more information than one at 100 pixels by 100 pixels, so if i want to use these, i normally would shrink them pretty dramatically (making a very small fast loading file) and just use them (if at all) on my website.
Looking at pictures being put up by the rest of the community, most look pretty good at 1920 pixels by 1280 pixels @ 72 pixels per inch. The publication of much larger versions do not make these better, and can make them very slow to load. Co-incidentally, this is also 8" by 5.3" or 20cm by 13cm (about 25% bigger than the old standard printed photo, although printed pictures look better at a much higher pixel per inch level - maybe 240 pixels per inch). This is a reduction in size of the original to 10% of its size. This also allows reasonable reuse on HD screen devices using 1920 pixels by _1080 pixels_, although that leaves a little vertical cropping. In the unusual circumstance where it would be nice to zoom into the picture, i would scale up from 1920 pixels by 1280 pixels by a factor of 2 or 4.
Most people posting do not need to worry about any of this (if i am posting from a mobile device, i probably do not get the choice to resize in any event and Google+ manages sizing nicely in the background).
In terms of my workflow, i would like to have images that reflect the work i put into them, but which load pretty quickly. Using Lightroom to manage the hundreds of thousands of photos in my library, i set up simple one button workflow processes that i can invoke once i get to the point i want to publish a photo (i adopted this a while back when i started to freelance concerts, and where i can shoot up to 10k images in a day, and need to be able to download then publish a handful of the best for media outlets within 15-60 minutes of leaving the concert).
I started to use a visible logo (a wolf) as a bit of fun (and because it ties into my tattoo, which i am proud of). On a more serious note, it also lets me search for use of my images (although this takes a bit of effort).
I am starting to find the logo distracting so i have removed it (and i have developed a new way of identifying and searching for my images that is blindingly fast and does not need a visible marker).
On a final note, i have been working with some old negatives taken by my grandfather many years ago. Lest it be thought that our digital advances long ago overtook the past, i have worked with a friend at one of the nearby universities to examine the surface of the negative. Firstly, in some cases the resolution is 100 times more detailed than ordinary digital photographs. Secondly, because the surface is not regular, some elements contain disparate information, focussing different features of the film and giving some 3D capability. The film does not contain enough coherent information to 'look behind' objects, but a well focussed picture is a trove of information.
Thanks to +Nina Anthonijsz for prompting this post :)