Accurate Image Manipulation for Desktop Publishing

Transfer Functions filter

DeepSpace Filter
Filter Name Transfer Functions
Compatibility Adobe Photoshop 5.0 and later.
RGB Workspace This filter does not require a specific RGB working space, works correctly in gamma spaces also. Image has to be in the RGB working space (so that embedded profile is the same as the working RGB space profile)
Description Useful in creating density masks and in dynamic range conversions.
Installation Download and copy all files from the zip archive into the /photoshop/plug-ins directory.
Setup

Open a higher bit-depth image (or convert a 8-bit/channel image to the 15-bit mode, then enter to the filter dialog and press Alt+g to open the Define screen settings -dialog.

Enter the system gamma space that is the gamma value from the Adobe Gamma utility (or from other system calibration utility) not the gamma value from the RGB setup.

This one-time-only step is needed in order to have the image in the preview box to appear somewhat correctly (the preview-box when the image is in 16-bit/c mode is not color-managed due to the bug in Photoshop).

Selection Dropdown

Visual Density *

This is the same as the -Log(image) with the slider set at 6.64-bits (100:1 range). Good for creating a general density mask. The * tries to indicate that the slider is not functional with this selection. This creates usually an image that appear for the vision to be a perfect negative image.
Exp(-image) Exponential conversion. Good for creating a stronger overall density mask.
-Log(image) Logarithmic conversion. Good for creating a general density mask.
1/image Inverse of the image. Good for creating a very strong deep shadow mask for strong bit-depth adjustments. Also, approximately inverts a transparency (positive) scan of a negative film.
-Log(1-image) Logarithmic conversion, down-curve where the steep portion of the curve is in the highlight. Usable for simple bit-depth compression.
1-Exp(-image) Exponential conversion, up-curve where the steep portion of the curve is in the highlight. Usable for simple bit-depth expansion.
Slider Bits(.01):

Dynamic Range of the space where the conversion is done in bits, from 1-bit to 15-bits (slider values are multiplied by 100 since they do not accept desimals).

The image data is first converted to the selected range, then the selected function is applied and finally the data is scaled to the full output range.
Notes

Visual Density assumes the usual 100:1 range, it most often is good. This selection is there only because I usually use this particular range so I need not to set the slider every time.

These functions are strong, higher bit-depth data "holds" much better, the plugin however does work in the 8-bit/c also but there will be strong quantization. There is nothing much to worry about the quantization when the functions are used for creating masks (since they will be blurred) but it is beneficial to convert an 8-bit/c color image first into the 16-bit/c mode.

I do not provide the grayscale ouput since the plug-in does not have access to the working space gamut values. Accurate way to convert to grayscale mask is:

  1. Convert to grayscale mode.
  2. Convert to RGB mode (since the plug in only works in RGB mode).
  3. Apply the plug-in function.
  4. Blurr it.
  5. Convert to Grayscale.
  6. Convert to 8-bit/c.

but the following does not make a large error:

  1. Apply the plug-in function.
  2. Convert to Grayscale.
  3. Blurr it.
  4. Convert to 8-bit/c.
Author Timo Autiokari.
Homepage Accurate Image Manipulation for Desktop Publishing
Usage Right Free.
Distribution Right Distribution is prohibited.
Filter Engine FilterFormula v 1.13, by ATS/Graphics (extended version for 16-bit images).