Can anyone tell me what am I actually doing when I free transform an image in Photoshop? Do I re-size an image or do I re-sample an image by doing so? I know there is difference between those two terms but I dont know how do they relate to free transforming an image.


2 Answers 2


You're resampling.

Whichever way you transform it, simple scaling or distorting, you are fitting a shape into a new pixel area compared to the original. You have greater or fewer pixels compared to the shape you started from, so photoshop has to do some guesswork to decide precisely how to best add or remove pixels to try keep relative sharpness. This is a lossy process, the original is not preserved.

Simple test - take a screenshot of this text, shrink it, then re-expand it…

enter image description here

We over-pushed the resampling until nothing remained legible.

  • 1
    Yeah I guess I get it know. I played around with the dimensions of a rasterized fragment of the text and it just came out basically as one grey pixel at the end (as opposed to resizing it when the fragment was converted to smart object in the beginning).
    – TSOB
    Commented Jan 5, 2023 at 11:51

Free Transform, resamples the transformation. Photoshop interpolates the existing pixel data to suit the transformation. In other words, Photoshop adds or removes pixels to fit the transformation.

If you transform a normal layer... Photoshop interpolates the layer when you commit the transformation. If you then transform that layer again, Photoshop again interpolates the layer when you commit the second transformation. Each an every transformation results in additional interpolation.

Normal Pixel Layer

Data 1 -> Transform -> Resample (Interpolate) -> Data2 -> Transform -> Resample (Interpolate) -> Data3 -> Transform -> Resample (Interpolate) -> Data4

The caveat here is in Smart Objects. Smart objects are not resampled. They are resized, then the preview is interpolated for display. The original Smart Object data is still the same. This is why working with Smart Objects can be beneficial.

If you transform a Smart Object layer.. Photoshop resizes the smart object then interpolates for the preview. If you transform again, Photoshop goes back to the original Smart Object data, resizes it, then interpolates for the preview. No matter how many times you transform a Smart Object layer, it only gets interpolated once because each transformation references the original pixel data before interpolation.

Smart Object Layer:

Data 1 -> Transform -> Resample (Interpolate) -> Data1 -> Transform -> Resample (Interpolate) -> Data1 -> Transform -> Resample (Interpolate) -> Data2
  • Ok thanks for this excellent explanation. Everythings so much clearer now. One more question to the above Smart Object Layer => as far as I understand Photoshop still "adds" new pixels in case I enlarge the picture? So the resampling algorithm is still involved (as explained in the equation above). Its not only the resizing that takes place, but the picture ADDITIONALLY gets resampled after the resize. This implies that if I want to ONLY re-size the picture I should simply do it from the Image Size window in Photoshop (resample option unchecked). Cheers
    – TSOB
    Commented Jan 5, 2023 at 11:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.