Someone on a professional retouching board I belong to, asked a question below. The majority of this post is my response. It is reposted here because the board is for members only and links will not work for non-members.

The posted question was: Does anyone have any tips for the best way to upsize an image for billboard use?

Here is my response:

I do a lot of outdoor campaigns, (or a mix of outdoor and print) so I run into this issue all the time. Worse, many print places that handle large outdoor pieces will establish some crazy guidelines for the artwork they accept. 200 dpi at size for a 14′ x 48′ billboard. Things like that.

In addition almost everything I do is composited, based upon 3-5 photos (sometimes a lot more), so the artwork is a bit more “created out of thin air” than what many retouchers deal with. Most of the outdoor work I do are ads for television, typically 2-3 big heads (or bodies) and some kind of background. Because of this the techniques I use may not work for you particular situation.
(As a curious aside, most television clients refer to outdoor campaigns are Out Of Home, or OOH. Why, I do not know)

So here’s the rules I have slowly discovered over the years:

1) Work at the native resolution of your original images, if at all possible.

This is the most important part of working a campaign of large outdoor pieces. The original artwork will be at whatever resolution it was shot at. This is your starting point, and if it was shot well, and high enough resolution at the start, the only point. For me this means I construct everything at the resolution of my worst photo (or if it is especially bad at some mid point). For the purposes of discussion, we’ll call this selected photo our ALPHA. All images will revolve around its resolution.

In practical terms this means retouching the native images as large as they will come out of RAW, and making sure the work (masks, color corrections, retouching, etc.) will transport well to your composited final files. I’ve found Smart Objects do this well, but not always.

If you doing a campaign (as opposed to a one-off) then build your first composited final piece at whatever resolution your ALPHA photo works to. (if you’re doing a one-off, then assume only one final composition)

This sounds complex, but actually it’s pretty straight forward. I’ll take the initial file (for me this is always a lower resolution comp provided by a design firm), flatten all the layers, and then drag this flattened piece on top of my retouched ALPHA. From that point I simply scale the flattened comp to the correct size so it matches my ALPHA. I find it easy to INVERT the image, and set the OPACITY to 50%, before scaling. Others do other tricks. The important point is to make sure you know EXACTLY what percentages the image was scaled.

For the purpose of discussion, lets say your final composited file needs to be scaled 254.82% to match the size of your ALPHA.

Assuming that the final file is built to mechanical specs in terms of length, width, and bleed (you did build to spec, right?), then all we have to do is change the resolution of the final file to match the ALPHA. Lets say your final file is 150 dpi. Length and width do not matter as we are not going to change them. To scale your 150 dpi file by 254.82% all you have to do is multiply the two numbers, and divide by 100. So 150 x 254.82 = 38223. Take that number and divide by 100 (38223 / 100 = 382.23). So the final resolution of your file will be 382.23 dpi. Set your image size to this (change the dpi ONLY, not anything else), wait for the file to rez up, and then save it off.

Now your ALPHA image should drop right in without resizing.

2) Build your background elements, and any other photos, to the size of your ALPHA.

Once you have them all together make sure you test for accuracy and fidelity. If you have several composited image you may have to tweak them to get them to appear the same level of sharpness. This is why my ALPHA is often my worst file, because everything else will look better.

When the file is done/approved, put it in your mechanical, (Most of my clients dump a flattened TIFF from the final file into an InDesign mechanical. You might do it differently), and you’re good to go.

3) Send the file away to the printers.

On occasion, I’ll get a printer that insists that their specs are the only “correct” ones, and demand the file at “their” resolution. (Never mind that a photo shot at 300 dpi will never be any better than 300 dpi.) To fix a file for them I add a final step. I flatten everything in my final file I can get away with (which is everything that is not type or vector, but excludes any noise layers) and then rez the flattened file up to the specs they require. If there is a noise layer on the top of the composited file, then I rebuild the noise at the final resolution (as opposed to rezzing it up). This will make the printer happy and save you lecturing them on the stupidity of assuming higher resolution is automatically better.
Now all of this came about because there is a lot of confusion out there about resolution and what it means. As I mentioned above, a 300 dpi image is never going to be better than at 300 dpi. Yes you can build it at 600 dpi, or more, but then your going to be doing all kinds of tricks with it to get it to be sharp, and the file is going to take almost twice as long to build.

You clients may or may not know this. They may be under the false assumption that higher rez is always better. Its not. For instance, a blurry shot at 300 dpi is just as bad as a blurry shot at 600 dpi, with the possible exception that one can retouch a 300 dpi file much faster than a 600 dpi one. So uprezing a 300 dpi blurry shot to 600 dpi, just to retouch, is a waste of your time.

There’s another thing. Most outdoor is not printed high rez. I’ve seen billboards printed at a 25 line screen. Do you have any idea how big the dots are at that size? More over, most outdoor is not viewed close up. Most billboards are 20 or 40 feet above ground. No one is going to walk right up to the canvas that high off the ground and say the image is a little soft. All I’m saying is that “sharpness” is a relative term. Chasing a mythical “perfect” point is a waste of time, especially if the original photos were never sharp to begin with. Its better to balance the image as good as you can get it, and let the RIP handle the rest.

Which brings me to my final point. Most outdoor images are sent from your mechanical to a RIP, and then to a printer. The RIP will automatically uprez your file to the proper size for the output device. As long as the type elements are still sharp (read still in postscript, hence building the mechanical and all possible type in InDesign) then the final image should hold up well. It makes no difference if YOU uprez up your file, or if the RIP uprezzes it, with the possible exception that the RIP will probably do a better job at keeping the file consistent.