Photoshop: Why does Image resizing small makes image pixelated?

  • 2
  • Question
  • Updated 8 months ago
  • (Edited)
When I resize an image smaller it becomes very pixelated. I thought this only happened if resize larger. How can I resize an image smaller (Ex, 3x5 inches) without losing image quality?
Photo of LissaM

LissaM

  • 2 Posts
  • 0 Reply Likes

Posted 7 years ago

  • 2
Photo of Chris Cox

Chris Cox

  • 20280 Posts
  • 813 Reply Likes
Which application are you using?

If you want to just change the size of the image, change the size and don't resample. But if you're resampling the image, you will lose some quality.
Photo of LissaM

LissaM

  • 2 Posts
  • 0 Reply Likes
Photoshop cs5. I'm taking a large image and sizing it way smaller, ie. approx 2x4 inches. I'm resizing it by ctrl T (transform), holding shift and dragging the corner in until it's the size I need to fit into my shape with a clipping mask. Is that why I'm losing quality. I tried resizing and not resampling, then I drag the image over to my other canvas where I need it which is 4x6 inches in size...then I resize by transforming and I lose quality. I'm stumped. Is it because I'm putting in on a 4x6 canvas??
Photo of Chris Cox

Chris Cox

  • 20280 Posts
  • 813 Reply Likes
OK, you are resampling because you have to in order to get the size in another document.

Check your preferences and see which resampling algorithm is chosen, it should be bicubic or bicubic sharper.

And, again, when resampling down you are always going to lose some quality because you have fewer pixels when you're done.
Photo of Brett N

Brett N, Official Rep

  • 2258 Posts
  • 114 Reply Likes
When resampling images, larger or smaller, you always loose some quality. This is natural as you are either adding or removing pixels from your image.

When you make an image smaller, it will become sharper and more pixelated, as the higher contrast edges are maintained at the cost of the less important filler information.

When you make an image larger, it will become more blurry. This is because you are adding pixels between the existing ones that are a created as a blend of its neighbors.

The actual quality of the transformation is determined by the algorithm used, as mentioned by Chris. When using the Nearest Neighbor option, resizing larger will create a pixelated image. But you may be thinking of a non-resampled increase in size, or a zoom, which really is just making the pixels larger on screen, thus more visible (i.e. pixelated).
Photo of Tom Sart

Tom Sart

  • 1 Post
  • 1 Reply Like
Brett, great explanation, thanks. So, if I should make a web banner (full web banner) at 468 x 60 (pixels I guess?), which apparently is standard size, do I have to find smaller pictures to have them sharp in such a small/normal format. Most pictures that I find online and want to work with is normally at least somewhere round 600-700 pixels and when I downsize to a 60 pixels in height format, the images becomes very pixelated (sharp). Do web designers start with smaller original pictures or why can I find sharp nice pictures in pretty much every web banner? Thanks alot for your input.
Photo of mark palmos

mark palmos

  • 3 Posts
  • 0 Reply Likes
Wow, years of using photoshop and I did not realise transforming a layer smaller resamples it and reduces it's quality... So if there is no way to use the mask tool with a smart object, and no way to resize without reducing quality of a non-smart object layer, how do you create a collage of images that need cropping/masking/resizing without losing quality?
Photo of Cristen Gillespie

Cristen Gillespie

  • 1259 Posts
  • 323 Reply Likes
If you are sure you want to crop your images, then you should do so before you bring them in. Perhaps work with a duplicate if you have reason to preserve the original pixels. Smart Objects don't play very nicely with non-destructive crops and can give you all kinds of fits, unless you're opening your Camera Raw cropped image directly into PS as the base Smart Object layer. After that, you can drag your image in from Bridge and they'll automatically be SOs. You can also convert after bringing them in any other way. There won't be any loss of quality in the original that's preserved in the SO if you transform them now. However, going from a very large image to a very tiny image may not look great. Most reasonable downsampling doesn't cause  a problem, though, if you haven't sharpened the image beforehand. If the downsampling is going to be rather extreme, I'd recommend doing it in stages, not all at once, though with any image, you can reach the point where you've simply thrown away too many pixels to have a decent image left.

As for masking, you can mask the SO layer (use that to crop, too, if you like), or paint a mask on a layer below and clip the SO layer to that mask. If you have something else you're trying to do that prevents you from using a mask with the SO layer, you'll have to be more specific. I composite images all the time and some of the work I do in advance of bringing it into the composite—not sharpening, as I mentioned— but I have no issues with scaling and transforming them at any reasonable size.

When it comes to sharpening, if you use a sharpening Smart Filter on the Smart object to target just the image, you can make sure the sharpening is the right amount for the image at the size you're presenting it.  You still have blend mode and opacity available to you through the smart filter. You can also clip a sharpening layer above it, such as High Pass, and then work with blend mode and opacity as with any normal image layer.

Does this help?
Photo of Marryam Uzair

Marryam Uzair

  • 1 Post
  • 0 Reply Likes
I want to set the image pixels 714x893 but everytime I do it and it gets stretched I want to make it proportional but exact 714x893 How can I do it without the LINK proportional tool,