Smart De-Interlacer by Richard Rosenman

Smart De-Interlacer - v1.3, © 2008 Richard Rosenman Advertising & Design. Release date: 07/16/04, Updated 09/28/08.


Simply unzip "" and copy "deinterlacer.8bf" to your "\Photoshop\Plug-Ins\" folder, or whichever plugin folder your host program uses. Load your program, open an image, go to the plugins menu and select the plugin.


This filter performs various de-interlacing procedures.

Field: This selection specifies which field to process.

Paintout Color: This color is used for various actions such as 'RGB Paintout' and 'Alpha Paintout'.

Action: This selection specifies what action to take for field processing.

Custom Parameters: When action #15 is selected ("Custom"), these variables are user configurable. Otherwise, they are disabled.

Help: Displays this help.

Reset: Resets all values to default state.

Realtime: Applies all effects (in realtime) on 'mouse drag' instead of on 'mouse release'.


RGB Paintout: Paints the specified field using the specified paintout color.

Alpha Paintout: Paints the alpha of the specified field (Except background layer) using the specified paintout color.

Duplicate Upper Field: Copies the upper field into the specified one.

Duplicate Lower Field: Copies the lower field into the specified one.

Interpolate Upper and Lower Neighbors: Interpolates the specified field between the upper and lower pixels.

Interpolate (x,x,x): Performs convolution interpolations on the specified field based on a x,x,x matrix. (Please see "Convolutions" below more information).

Realtime: Applies all effects (in realtime) on 'mouse drag' instead of on 'mouse release'.


Smart De-Interlacer's convoluted interpolations are based on 3x3 matrixes. This means that each pixel around the currently processed pixel is added together (according to their specified weights) and then divided by the sum to obtain an interpolated pixel value. Because pixels cannot be taken from the currently processed field, the top 3 pixels are taken into account, as are the bottom 3 pixels. The middle 3 pixels are each assumed to contain a weight factor of 0. To understand the labeling of interpolated convoluted actions, let's use an example. The action "Interpolate (1,1,1)" means top left pixel (weight of 1) + top pixel (weight of 1) + top right pixel (weight of 1) + 0 (weight of 0) + 0 (weight of 0) + 0 (weight of 0) + bottom left pixel (weight of 1) + bottom pixel (weight of 1) + bottom right pixel (weight of 1) are all added together and then, in this case, divided by the sum of the weights, 6.

Many of the convoluted interpolations are similar to one another but there are subtle differences. Others can contain negative values which may produce a sharpened interpolated field. It is recommended that you try several variations and magnify the results to understand the differences between them, as some interpolations may be more suitable for a particular image than others. Generally a matrix of (1,10,1) will give you a smoother interpolated field than (1,1,1) because the convolution greatly favors the top and bottom pixels.

Action #15, "Custom", allows the user to specify each of the available convolution pixel weights. This is useful when you wish to create your own custom interpolated fields.




1.1 - Fixed preview window flickering when using trackbar sliders.
1.2 - New GUI design.
1.2 - Windows Vista Compatibility.
1.3 - Added reset button.
1.3 - "Custom Parameters" window now displays "ON" or "OFF".
1.3 - Preview image now scales correctly to zoom level.
1.3 - Preview no longer flickers on button / slider adjustment.
1.3 - Ability to use up to 3 GB of RAM under 32bit Windows (2000, XP, Vista) and 4 GB of RAM under 64bit Windows (2000, XP, Vista).

Next Post »


Click here for comments
19 July 2016 at 21:27 ×

Amazing post and i'm glad to see you post . Thank you so much .

22 September 2016 at 20:09 ×

Love your post and glad to see here .

Thanks for your comment