I have been working on a project that involves photo-editing/manipulating satellite imagery in a similar vein to the NASA/USGS “Earth As Art” project.
I am using the Avenza Geographic Imager plugin which allow PS to handle .tiff files that contain spacial data.
The files I am working with are very large – at the moment I have ‘mosaic-ed’ together 15 ‘tiles’ of satellite imagery (Sentinel 2 imagery), resulting in a file that was 300 or so inches in height at 72 dpi. I reduced the file to 100 inches in height, but increased the dpi to 300. I am working at 16 bits/channel. It is saved as a .psb file, which is becoming difficult to save and perform tasks in.
I am working with a bunch of layers and layers masks to manipulate the colors of different parts of the image using ‘bands’ of the imagery to select what I want to work with (satellite imagery comes as a series of files of multiple greyscale bands that when combined create a full color image, different band combinations yield different colors). I have tried to merge layers when I can to try to reduce the file size.
I am ultimately going to have this image printed on canvas at 42 or more inches and 300 dpi, so I am hesitant to reduce the size and quality too much (I will crop the image when I am done and ready to export for printing).
I tried increasing the ram that PS can use to 78%, and selected ‘Optimize cache levels for huge pixel dimensions.’ (I have 16 GB of RAM, and currently 233 GB free on my SSD). I have been trying to close any other apps while working on it.
What advice to any of you have for how to adjust my workflow to make such files easier to work with?
Here is an in-progress screenshot. In my last work session, I had merged several layers that are now the ‘cloud effects’ layer, and I am realizing merging the layers took away the effects of different blending modes I had, so I guess that’s not a good strategy. In some tutorials, I have seen people just ruthlessly flatten their work as they go, but that requires really committing to things I guess. I will definitely flatten it when I am done.
What you described is a rare instance that calls for huge file sizes since you want to preserve every detail without any data loss.
Despite my questions about TIFF, the Avenza software, and PSB, the fundamental problem you’re facing is the enormous size of the files and not having enough memory or scratch space to handle them. Just off the top of my head, for what you’re doing, I’d be looking at 64–128 GB of memory, a 3–5 TB SSD, and a speedy processor.
In other words, a few thousand dollars will fix the problem, which probably isn’t the advice you wanted to hear.
Thanks! I’ve been considering upgrading my RAM for this project. But I have also run into these kinds of issues on other (less giant) projects too when I get carried away with adding effects layers. I suspect there may also just be smarter ways of working to avoid busting up against computer restraints? I wonder what other folks do in such situations?
On some of my larger renderings the smart objects became the biggest problem. If I couldn’t manage the editing and smart objects in one file, I broke it into pieces and then assembled it in another with flattened art. In other words I would merge the mosaic at the end… if you can.
Thanks! I think applying the same effects to each tile would be too tricky to acconplish, but I might try working with the different layer masks and effects separately.
In this particular image, wanting to isolate the clouds as layers overtop of an imported “digital elevation model” file prevented me from wanting to flatten the image too soon.
Well the masks in your screenshot are not particularly heavy or anything it’s just the overall resolution and layers. You can still save multiple versions of the layered file to be able to rollback to earlier choices, in that way you could still build modularly.
I don’t know if it will help, but I’d be tempted to restart my computer and then launch Photoshop only. Get a fresh slate with the computer so to speak. Other than that, max the ram and scratch disc space which you already know.
Okay, yeah I wasn’t sure how much impact the masks have. I thought about maybe ‘Apply Mask’ could help some. That good advice though, I think that might be a good way to go. I do tend to save new files when making a big change, but yeah maybe doing incremental files to then string together would help, and flattening once they are combined. The only thing is that the large files take up hard drive space, so I’m limited a little there too. Thanks for your ideas!
If you reduced 300"@72ppi to 100"@ 300ppi, you’ve unecessarily interpolated something somewhere. An unresampled 1/3 reduction in size would give you 216ppi actual. If you made it 300, then you’ve added pixels where there weren’t any.
Are you working in RGB or CMYK? Should be ProPhoto RGB if this is high art. RGB has one less channel making file size slightly smaller. It also supplies more color info to the output device rip. You can’t make a CMYK file into any kind of RGB file. It’s too late for that, but if your imagery is RAW, keep it in the ProPhoto RGB space or similar wide gamut RGB.
On the output part of it, have you talked to the person doing your printing onto canvas? It is very very rare to output a file above 200ppi. Actually 150 is considered optimal for most inkspewing technology. You may be working above the resolution necessary or even available to your output device. So-called giclee printing may (key word ‘may’) require up to 400ppi, but size and substrate matter. If your end goal is ~ 40" on a canvas texture, why waste time and energy working above 60"?
Don’t confuse dpi with ppi with inkjet on canvas. A 150ppi image (pixels per inch) printed at slower-than-grass-grows speed on a high end art printer at 1500DPI (ink droplets per inch) would look great (remember there is the texture of the substrate to deal with too.) You can supply a 300ppi .psb file, but if you pay the “gallery-quality” premium to do that, be sure they rip it that large. We don’t call .psb files “photoshop bloated” without reason.
Ohh thank you , that is helpful info. I have always heard 300 dpi for printing, but I hadn’t thought about there being different standards for printing on canvas. I guess I am a little fuzzy on understanding how reducing size of an image relates to the resolution. What would be the correct way to achieve a smaller size but with high resolution for printing?
I am working in RGB.
Good suggestion, I will ask my print shop what their requirements are.
I was working above the size of the canvas goal because I intended to crop the image before I was finished (but since I am combining different image bands that need to lay on top of each other, I couldn’t crop it until I was done editing).
With photoshop, images have a specific number of pixels when created.
If you took your 300" image at 72ppi (pixels per inch) and made it smaller, to 100" with resample UNchecked, you’d multiply that 72 pixels by 3 so at 100 inches your image is now 216ppi. Play with it in photoshop and see.
If you had resample checked, and made your image smaller, while making it 300ppi arbitrarily, you just made Photoshop interpolate (guess) what those extra 84 pixels should look like. That is not an accurate representation of the data you had originally. While photoshop is pretty good, upping the resolution by interpolation doesn’t always do anything good for the image. You might not notice a difference on screen. You might not even notice a difference in print. It’s that print part. If working with that many more pixels per inch is not going to get you any noticeable difference in print, why waste your time bogging your machine down?
Call the person doing your output. Find out what their optimal output resolution is at full size. Since you already are working larger and intend to crop (which is always a good idea, BTW) working at the resolution they want, or maybe a littler higher if the pixel info is there, is where you want to be.
Now that said, if you ever intend to output this photographically on say, a lambda photo printer (if you can find one any more) the resolution there would be 200ppi optimum - and for an art print opt for the slower output speed (some call it 400 speed or gallery output.) Since you might get an answer of 150ppi, you may want to work at 200 anyway so you have the photo option later.
If you intend to publish the image later, say in a magazine or book, an image created at 100" at 200ppi will have more than enough resolution to take it down to printable size later.
Thanks so much for that great explanation! My printer said for their machines they recommend having at least 150 DPI. I’m not entirely clear on how that relates to ppi? It is as simple as going under Image size and specifying dpi instead of ppi?
If I have a large image that is 72 ppi that I want to print, should I just leave it at 72 ppi (not sure what dpi that would make it), or should I try to increase it to 150 dpi? What is the best way to scale the resolution up without damaging the quality?
Sorry if this is redundant, I’m still a little confused!
People often confuse Dots Per Inch (DPI) with Pixels Per Inch (PPI). Even people who know the difference will often use the term DPI to avoid a long explanation.
PPI are the number of pixels in a linear inch when printed. If you print 200 pixels in an inch onto a piece of paper, you’ve just printed the file at a resolution of 200 PPI. If you spread that same set of 200 pixels over two inches, the pixels need to be spaced out twice as far, so the resolution decreases to 100 PPI.
So if you’re following me so far…
Pixels can’t be printed directly — they’re electronic bits of information on your computer. To print a pixel, it needs to be converted into a dot of ink. The rule of thumb (although not entirely accurate) is that using up two pixels per ink dot will produce the sharpest images.
In other words, if you print a 300 PPI file with two pixels per dot of ink, you will have printed 150 dots per inch (DPI).
Your printer likely was referring to PPI instead of DPI. So assuming that’s the case, they’re asking for a file that’s 150 PPI. However, a PPI measurement doesn’t do you much good unless it’s 150 PPI at the size it will print. For example, if it will print 2 x 3 feet, your printer is asking you for a file that is sized to 2 x 3 feet at a resolution of at least 150 PPI.
If your image when enlarged to that size is less than 150 PPI, it will do no good to resample it up to 150 PPI. All that does is spread the image across more pixels without doing anything to increase the sharpness of the image.
