0

In music, if you like the frequency content of a sound you can run the sound through an advanced equalizer plugin (Fab Filter) to analyze the frequency content of the sound. Once you have that curve you can then apply it to any other sound that you like to adjust frequency content of the new sound to match the frequency content of the original sound. Is there an equivalent working with images in Photoshop? For example, I have this black and white image:recording_session_1

I like the "spectrum" of this image and want to alter the following image to match that profile: recording_session_2

Any thoughts? Thanks.

  • Can you add an example of how you want the second image to look like after adjusting the correct 'frequency'? On the first image you have a particular range of greys that were turned white: if you'd apply the same adjustment to the second image (doesn't matter how: color rangle, curves, gradient map), you'll have the same greys turned white. What needs to be analyzed? – Sergey Kritskiy Jan 29 at 9:11
  • Sound is temporal and much much simpler than a image data. The thing is you can do exactly the same thing to a image than sound, only its not as exciting. So because the signal is more complex same aporoach does not work as well as we have more types of filters than mere equalisation and convolution. – joojaa Jan 30 at 4:50
  • To add to what i said. Because the image is nontemporal we do not have a baseline for this analysis. So you would need 2 exactly same images before and after. In sound its much easier to correlate disseparate parts as i only have 2 dimensions to consider. But in a image i have 5 dimensions to match. – joojaa Jan 30 at 4:57
  • FabFilter can be used as an equalizer and as a multiband compressor-expander. That really affects to the spectrum of the sound, but the generation of new freq components is minimal because no distortion is ecpected. Your image example has serious amplitude distortion. Insert a curves adjustment layer and drag some twists to get the partial inversion effect as already suggested. People use it for weird, often metallic effects. You can feel this is like eq but in math it's heavy distortion, much more complex than fuzz. – user287001 Feb 1 at 18:16
1

There are actually a few ways to alter the overall tone of an image and then save that alteration for later application to other images.

A couple are....

Image > Adjustments > Curves...

enter image description here

or

Image > Adjustments > Levels...

enter image description here


You can define an adjustment and then save it as a preset to apply to other images if so desired.

| improve this answer | |
  • Thanks for the reply. This is not quite what I'm looking for. The Levels and Curves functions do not analyze the "frequency spectrum" of the image. They allow you to save a particular change to make but that change is not dynamic. In other words, that change is going to be the same no matter what image its applied to. – comp1201 Jan 28 at 22:06
  • I, personally, am not aware of a function to analyze the histogram of an image then make auto-adjustments based upon that analysis. It may exist, but not to my knowledge. – Scott Jan 28 at 22:42
  • however you can make a series of edits to an image then save out a LUT, which will come closer perhaps to the OP's requested function... – GerardFalla Jan 28 at 22:58
1

I think that what you are describing may be some form of FFT filter.

SO, I have a Fourier Transform aka FFT/IFFT plugin that unfortunately I do not recall the source, but it is very basic. I used it once to remove a paper texture from a scan.

I basically ran FFT on the image, then "inversed" that using IFFT without any edits. I then placed a copy of the image as the bottom layer, inversed the color, and then set the IFFT layer to "difference."

The end result:

enter image description here

And here is the IFFT result layer, with normal blend, half on nothing and half on a blue background (for visualization):

enter image description here

| improve this answer | |
  • aside from directly manipulating the FFT before inverting it (which is very touchy and requires more practice than I have), some control can be had by adjusting the levels of the top layer and its opacity etc. Also, edge detection and solarization can be used in addition. – Yorik Jan 29 at 23:06
  • though the system he uses in sound is just as likely to be a laplacian than a FFT – joojaa Jan 30 at 5:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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