MatchLight IMS User Guide
MatchLight IMS is a self-installing, stand alone, Windows, Mac and Linux application, designed to enable the automated generation of LUTs from 1 or more images pairs.
Applications range from simple LUT generation from a before/after image pair or image sequences, to the generation of LUT and ACES IDTs to match different cameras.
In very basic form, to use MatchLight open up the 2 images you want to process via the 'Browse' buttons. Hit process, then save the LUT.
MatchLight, exports LUTs in .mlc format, which are MatchLight 3D cube files. This format needs to be loaded into LightSpace CMS to be re-exported in any required LUT format. LightSpace LUTs is the minimum license required for LUT conversion.
In both the Raw Image and Graded Image file selectors you can use the left and right keyboard arrows to step through the sequence of frames, after first double clicking on one of the images to change keyboard focus to image control.
You can zoom in and out with the page up and down keys. Home resets the view.
You can pan with shift + left mouse (then move the mouse).
Just left mouse, while moving the mouse, gives a pixel sniffer, which shows raw and processed values.
The top cube apply button (the colour cube logo) switches between Raw Image source and the same image with the extracted LUT applied.
The bottom cube apply button switches between Graded Image and the Raw Image with LUT applied.
The Pixel Box control sets the size of the square used to average the image values. The default is 2x2.
If this is too big the extracted LUT will blur the colours, giving a de-saturated LUT. If too small, the LUT will be inaccurate and 'noisy'.
The Border L, R, T & B controls control the edges of the bounding box that defines the area of the image to be processed during the extraction process, in percentage. The default 2%.
This allows for image cropping for film scanned images, burnt-in text, etc.
The Trim control sets the minimum distance between extracted points. The default is 5%.
If this is set too small, the extracted LUT will be unstable and noisy. Too big and the LUT will de-saturate the image.
The Reject button, when on, causes a second pass to be performed. The default is 'off'.
The first pass averages the src->dst point values. In the following second pass the average is obviously already known. Therefore MatchLight can then re-calculate, rejecting src->dst values that are too far from the average.
You can set the rejection level in the GUI. The default is 100%, which keeps points that are between the old average + 100% of data max and old average – 100% of data max.
In an 8 bit system this would be old average +- 255. If this is too small, too many points will be rejected, giving poor results.
The Freeform-clip, Freeform, & Uniform drop-down menu options set the way the Cube is extracted from the image pairs.
Freeform-clip is the default, and it retains the hard clips in the cube.
Freeform ignores the hard clips as it generates the cube.
Uniform uses regression to generate a matrix, then populate the cube.
One of the benefits of MatchLight is its ability to process image sequences, not just single images, to generate enhanced accuracy LUTs.
To open an image sequence you need to use a specific naming convention, and type the path directly into the browser path window (using 'Browse' initially to select the first file in the sequence and then editing the final part of the path as required).
Where F = first frame number; L = last frame number; P = numerical padding (default is 6); I = frame count increment (default is 1)
For example, C:\Work documents\MatchLight\Test_Images\Stills_NO_LUT_New.1-3;4;1.tif, for a sequence of three images individually named:
The ;P;I component is optional if the image sequence has a numerical padding of 6, and a numerical increment of 1... The ;I gives the frame increment,1 or 2 or 3 etc. The ;P gives the numerical name padding, so 4 is for a name sequence ending .0001.dpx., .0002.dpx.. etc.
Please note the 'dot' (.) before and after the numerical component of the image name. This is important!
To use the keyboard arrow keys to navigate backwards and forwards within the image path name window double click the cursor on the text to change keyboard focus to the text string.
A matching sequence of images is required for both the Raw Image, and the Graded Image.
Left mouse clicking on the image, and moving the mouse while holding the left mouse button down, will bring up a pixel sniffer, with the X/Y coordinates and pixel colour values being displayed at the bottom of MatchLight.
Supported Image Formats
MatchLight supports DPX, Cineon, TIFF 8 or 16 bit, GIF, TGA JPG and EXR file formats
The images need to be the same resolution, and close to being pixel aligned.
If using two images that are not from the same identical source (pre-grade and post grade, as an example from the same source, vs. the same scene from two different cameras) an external image aligning/warping tool would be required to align the two image as perfectly as possible.
Additional Technical & Support Info.
German language version of this page: displaycalibration.de