Smart Object, Image Resolution and Printing

[master_experts98A4]

Hello,

If I was to size down an image from 72dpi to 300 dpi, then
convert it to a smart object and house it into a photoshop document of 300 dpi and enlarge it, will the image quality be lost when printing the enlarged image?

Thank you.

[Mylenium]

Transforms operate on the absolute dimensions and the actual sampling happens in the parent space, meaning the containing PSD. If all you did is really just change the DPI without actually reducing the pixel dimensions, then you have not lost any info. It just affects some of the internal calculations when converting the DPI at the document level. Generally it seems an odd thing to do, though. For multiple output resolutions one would set up a consistent master document and create versions of it by duplicating the image and resizing those. I'd only follow your workflow if I wanted a specific look based on how some functions and filters respond to DPI.

Mylenium

[master_experts98A4]

Thank you.

I do photography and I have some images in jpg format rather than the cr2 (raw) format.  So the resolution is 72 dpi.  I downsized the image by changing the width and the height which results in higher resolution.  I can edit the image smart object from there. I'm not too worried about functions and filters.  I want to make sure resolution is not lost (print quality) while printing at larger sizes at the end. 

Thanks again,

Deepinder

[Conrad_C]

---

@master_experts98A4 wrote:

I do photography and I have some images in jpg format rather than the cr2 (raw) format.  So the resolution is 72 dpi.  I downsized the image by changing the width and the height which results in higher resolution.

---

Did you downsize by changing Image Size values with Resample disabled? If so, resizing probably wasn’t necessary if the pixel dimensions did not change. You could have simply brought it in as a Smart Object and done the same resizing later.

What are the original pixel dimensions of the image? For example, if it is 6000 x 4000 pixels, you don’t need to worry about what it says for ppi at the beginning. Just resize it later, as a Smart Object in the Photoshop document.

For example, if you want 6000 x 4000 pixels to print at 300 ppi, it must be printed at no larger than 20 inches (6000 pixels divided by 300 ppi) by 13.33 inches (4000 pixels by 300 ppi). Smaller than that is OK, because any smaller dimension in inches will make the effective resolution of the Smart Object more than 300 ppi, limited by the ppi of the containing Photoshop document.

That it said 72 ppi at the beginning doesn’t matter. Again, ppi only has meaning if you state both pixel dimensions and dimensions in inches. If it is 6000 x 4000 pixels, then it can only be 72 ppi at 83.33 inches by 55.55 inches. You are probably not going to print that large, so you ignore that 72 ppi reading as irrelevant, and go on to recalculate ppi based on how large you are actually going to print the image in inches.

[Conrad_C]

It depends on how much the Smart Object is scaled in the second document.

PPI only means something when you also state the size in inches. For example, if the Smart Object is originally 300 ppi at 4 inches wide (100% scale), then at 2 inches wide (50% scale) it is 600 ppi (pixel density is doubled), and at 8 inches wide (200% scale) it is 150 ppi (pixel density is halved).

So if you bring in a Smart Object that is 300 ppi at 100% scale, its ppi resolution will be unchanged if you print it at 100%, its resolution will decrease if you scale it up, and increase if you scale it down.

But when scaling down, it can never exceed the ppi value of the Photoshop document itself. If the containing Photoshop document is 300 ppi, the print resolution of any Smart Object scaled down cannot exceed 300 ppi, even though the effective resolution of the Smart Object itself might be higher.

[master_experts98A4]

This is helpful info.  So if we resize smart objects, we can potentially lose quality?

Thank you,

Deepinder

[TheDigitalDog]

Resizing, by reducing the number of pixels does lose quality compared to the original.

Reducing the PPI without, does nothing. This is only a metadata tag to 'calculate' a 'size' based on that tag and the number of pixels you actually have.

A 1000x1000 pixel document at 100PPI/DPI is identical to a 1000x1000 pixel document at 10000 PPI/DPI. The calculated 'size' based on this is either 10 inches or 1 inch but both have the same number of pixels and you can enter any PPI/DPI metadata value and end up with a new 'size'.

Work in pixels, kind of ignore PPI/DPI.

[Conrad_C]

---

@master_experts98A4 wrote:

This is helpful info.  So if we resize smart objects, we can potentially lose quality?

Of course, as my examples show.

Again, you will get 100% of the original quality if you print it at 100% of its dimensions in inches when at the required ppi resolution. If you scale it up beyond that, the same number of pixels get spread out over more area (lower pixel density). Although the original quality is not technically lost, you will lower the effective resolution by scaling it up, and that can lead to an apparent loss of detail in the print.

The benefit of a Smart Object is that it does preserve original quality no matter how many times you transform it, but you are always still subject to the limitations of the original data in the Smart Object.