Copyright (C) 1986-2008 by Daniel H. Hudgins, All Rights Reserved.

No part of "This Web Site" (HTML document), including associated files, may be: distributed, sublicensed, transmitted, copied, archived, mirrored, modified, bundled, embedded, sold, given away, rented, loaned, or shared in any form without express written permission in a formal Vendor agreement contract dated and signed in ink obtained directly from Daniel H. Hudgins by registered postal mail. All agreements for permission to distribute expire after a period no greater than one year from the date of the signing of the agreement by Daniel H. Hudgins. See the current "EULA" for information regarding limited copying and storage for the purpose of "Beta Testing" "This Web Site."

To view or use the current version of this Web page you may need to reload or refresh the display of this page by your browser. Just clicking on the browser's [Reload] or [Refresh] icon may not be enough to insure that all of the page's most current contents have been cached and displayed. Some browsers may have additional commands to help display the page's most current contents such as: holding down the [Shift] key and clicking on the [Reload] icon, holding down the [Control] key and clicking on the [Refresh] icon, holding down the [Control] and [Shift] keys and clicking on the [Refresh] icon, pressing the [Control] and [F5] keys, pressing [Control] and the [R] key, or some other combination of keys or clicks. Check to see which commands your HTML browser uses to load the most current page contents into its cache and then to display them onto the screen.

This Web site is dedicated to the thousands of "users" of my programs, those who have helped test my programs over the last 22 or so years, and especially those who shared their experiences with me.

You must read this notice: This is a licensed Web site (HTML document and associated files). You must read and agree to be legally bound in contract by the Terms of Use and conditions given in the End User License Agreement ("EULA"), Legal Notices, Instructions, Warnings, Disclaimers, and all other text in "SECTION: 0" of "This Web Site" (HTML document and associated files) before reading or using any of the information, software programs, and or files, contained in, linked to, and or associated with, "This Web Site" (HTML document and associated files). Any use or "Beta Testing" of "This Web Site" constitutes your acknowledgment of your full agreement with the current End User License Agreement ("EULA") and your decision to have this current license supersede all prior and contemporaneous agreements and understandings. Information and files in "This Web Site" (HTML document and associated files) have been placed here so that long time users of "The Author's" programs DANCAD3D.COM (tm) , DANCAM.EXE (tm) , or DANPLOT.EXE (tm) could help proofread the text of the documentation files or screens displayed, and also help test data files, example files, and or any software programs that might be made available from time to time, to aid "The Author" in finding mistakes, bugs, and other errors, omissions, defects, mistakes, and faults. Everything in "This Web Site" (HTML document and associated files) is "Beta Test", "Beta Code", Experimental, Preliminary, requires proofreading, or is being evaluated for possible revision, and is NOT warranted to be free of defect. To help "The Author" report any bugs, foul-ups, defects, or mistakes that you find, see "SECTION: 8" for instructions. "This Web Site" (HTML document and associated files) and all other files and programs by Daniel H. Hudgins are made available "AS IS" without warranty of any kind express, expressed, or implied. All offers and specifications are subject to change or discontinuation without notice of any kind. Please read "SECTION: 8" of "This Web Site" (HTML document and associated files) before trying to contact "The Author."

SECTION: 4.10.30 is for some information associated with illustrations and screen shots relating to commands in the CAD program's Kinema Edit list command. See also the program files in the current distribution of my programs, the other parts of this HTML documentation, and the current On- Line version of this Web site for information more specifically about particular features in my programs.

The descriptions associated with the illustrations in the sections and sub-sections of SECTION: 4.10.30 relate to the version revision used to make the illustrations. The appearance of the results you get in the version you are currently "Beta Testing" may be different from the results shown in the illustrations since the illustrations may have been made with a different program version or revision. The illustrations may have been edited in various programs for conversion to files viewable through the HTML code and WEB site on a browser, and may look somewhat different than the original program output, e.g. sharper or softer, lighter or darker, in the JPG files.

Any comparisons of my programs or methods to some others is only given as a vague generality of my opinion and is not intended as a recommendation or reference to any particular products, or methods, always make your own evaluations and comparisons before taking any action.

To demonstrate the principle color correction, a.k.a. color grading, color timing, "printing lights", CC filters, command tools in the CAD v3.7 program's Digital Cinema command Edit list a "home made" motion picture film scanner was used to scan a scrap of LAD 35mm color negative (might be duplicate negative) film that has a standard reference image on it used for motion picture laboratory process control. This special reference control image is made by Eastman Kodak (tm) by exposing their master film in a camera with a combination of swatches of known reflectivity and a woman's face and hair. The transmission density values from the various reference patches on the processed LAD negative are measured by using a densitometer instrument after the film comes out of the film processing machine and if the density measured is low the temperature is raised, the speed reduced, or replenishing volume increased, and the opposite changes are made if the density measures too dark. Once a properly processed negative is generated at the film lab customers film is run through the film processor with another short piece of exposed LAD un-processed film on the tail to check what adjustments of the processing machine are required for the next batch of film to run through the processor. The exposed film rolls spliced up for processing would have a piece of un-processed LAD film spliced in about every 10000 feet or so depending on the volume of film to be processed and the volume of solution used in the lab, so that the LAD control density values can be checked by measuring the gray patch in the corner to get the Red, Green, and Blue density values (see the Kodak (tm) H-61 documents for more details).

The greenish spots in the image may have been caused by scratches and or blemishes in the frame of the LAD reference film I had to scan. Because the color negative film has an orange mask scratches can come out odd colors since they are through an orange mask layer that has been shifted to neutral tone by coloring scanner the light source with Bluish filters. I did not retouch the scratches out of the scanned image since the LAD image is soft focus and the main things that seem to be sharp in the image to focus the film recorder's lens on would be the scratches in the LAD film's emulsion.

My current CAD v3.7 programs have GUI tools for adjusting the color correction of scanned images derived from color and black & white negative and positive images. Most of the time you would be shooting color negative film, and would be using these command tools to correct the color of every shot in the feature film project you would be digitally processing the frames of using my programs. To set up your scanner you should get some processed LAD film of the same type of film stock you will be shooting, or make a LAD type film for yourself by shooting someone holding a standard 18% reflectance Gray card, a standard 90% reflectance white card (the back of another Kodak (tm) Gray card), a standard Kodak (tm) gray step strip, the Kodak (tm) primary and secondary color strip card, and a dark black card (or a box painted black inside and not lit inside). If you make your own LAD type film, be sure to expose with absolute 3200K light on "T" type film (measure the line voltage and adjust the color temperature of the light with a variac autotransformer) and get the reflective reading off a clean 18% gray card without reflections from its semi-matte finish. If you are going to shoot through filters you may wish to make additional LAD type films using those filters (marked with an on camera note) since some filters are made with green glass rather than clear optical glass and may add a color cast that needs to be corrected for.

My programs have a "digital densitometer" built in that you can adjust the "probe" area with to measure the "digital density" directly off the screen. The "digital density" is measured as values from Black=0.000 to White=1.000 with 18% Gray card, or middle gray, measuring about 0.5000 on the screen (it may read another value in the version you have, the programs are being worked on, let me know if you have an interest in such things.) The monitor gamma used was about 2.4 in making these illustrations, that is about the "natural" gamma of many CRT monitors and CRT televisions and should be about standard for making DVD disks from the reduced 720x480 images taken from finished output image files for the film recorder and for editing the finished frame images in other graphics programs with that program's gamma correction set to a neutral value (i.e. gamma compensation set to 1.0 using a monitor with natural gamma of about 2.4). So you use the cursor keys to put the sample area ("probe") for the "digital densitometer over the gray card in the scanned image and make adjustments until the Red, Green, and Blue digital density (i.e. brightness) values are about the same. If you want your final images lighter or darker than normal your target value would be higher or lower than the normal target value for middle gray, maybe 0.6 for lighter or 0.4 for darker. When you output the final frame images in your 35mm film recorder using DANCINEL.EXE (tm), or some other program, you could adjust the gamma in that program to get the middle gray swatch on the LAD image to come out with a density of about 1.05 on the 35mm print stock exposed in the film recorder (or a little less dense since the movie theaters leave the lights on dim during the show now). Because the monitor in the film recorder and the high contrast of the motion picture print stock alter the tones in the final color corrected images, you will need to make adjustments of the exposure of the film in the film recorder to keep the White, Gray, and Black densities in the print stock correct. The White level is set in the print stock by the Black level on the monitor, if there is little highlight detail in the print, increase the film recorder monitor brightness so that there is no true black on the monitor (just dark gray) this will expose the clear parts of the print a little to get more highlight detail by darkening the highlight areas in the print stock. The exposure time in the film recorder is set to just get the white areas on the monitor to translate on the print stock to become almost full black on the print, about transmission density 2.8 or so depending on the print stock used and the developing time. The transmission density of middle gray of the print stock is set after the white and black are worked out by adjusting the gamma correction in the film recorder software to reach the target density of about 1.05 or so, check the print stock data sheet for the target densities recommended. The overall color balance in the film recorder can be set with orange filters or by using different exposure times for the Red, Green, and Blue exposures, i.e. shortest for the Blue exposure, Somewhat longer for the Green exposure, and longest for the Red exposure on 35mm color print stock (i.e. AGFA film (tm) Color Positive film type CP-30 (tm), http://www.luckyfilm.com.cn product Lucky Color Positive film 5244/2244/6244/7244, Fuji film (tm) products F-CP 3510 normal-contrast color release print stock, Eterna-CP 3513DI high-resolution color positive film, Eterna-CP 3521XD digital original color positive film, or Eastman Kodak (tm), products Kodak Vision Color Teleprint film 2395/3395, Kodak Vision Premier Color Print film 2393, Kodak Vision color print film 2383/3383 and such). If you are printing onto black and white print stock (i.e. Eastman Kodak (tm) Black and White print film 2302, or Eastman (tm) fine grain release Positive 5302/7302) only the Blue exposure is required (from a Panchromatic image color matrix mix), and you should focus the film recorder film camera with a blue filter over the eyepiece since the film recorder lens will focus a little off due to the chromatic aberration and your eye seeing green better than blue if a white light image is used to focus the film recorder's lens.

See the links below to download the LAD scans to explore and experiment with the color correction tools in the Kinema Edit list command, and to use in calibrating your home made film recorder. My film recorder operating program DANCINEL.EXE (tm) has a built in test pattern for alignment, but the LAD image might be used to supplement that in setting up and aligning your film recorder.

This is a un-cropped positive with rough color correction of the LAD reference negative used in the screen shots below for showing the color balancing and correction tools available in the Digital Cinema command Edit list in the Kinema sub-menu.

If you cannot get a piece of LAD film, you can make one for yourself. It is best to use a real person for the flesh color sample, since the camera film reacts differently in the IR and UV invisible light than your eye does and those rays can affect the color, real flesh is better as a reference of the way the light reacts to the colors and light on the film image.

You can download both 2048x1536 (computer 2K) and 1600x1200 cropped positive versions of this LAD image to use with DANCINEL.EXE (tm) in your homemade film recorder. The large gray patch on the lower left side should measure about 0.85 to 1.05 density on the color print stock. The old target density of the print film was about 1.05, but since the theaters leave the lights on, dimmer half burned out Xenon lamps are used in place of bright fresh carbon Arc lamps, and the print stock does not have a good black sometimes it may be better to make the prints a little lighter than the old control targets.

Click on LAD2048P.ZIP to download, 6456150 bytes.

Click on LAD1600P.ZIP to download, 3983125 bytes.

You would use the 2048x1536 resolution with a CRT monitor in your film recorder since that is probably the highest resolution you can get the video board and monitor to work at using the VESA BIOS calls.

If you are using a LCD monitor in your film recorder you may only be able to get the video board and monitor to work at 1600x1200 so you would use the 1600x1200 version of the LAD image for test exposures. A LCD monitor operating at 1600x1200 may actually give higher resolution than a CRT monitor operating at 2048X1536 since the CRT monitor's electron beam is not small enough because if it was it could "moire" with the shadow grill inside the tube.

Some other reasons you would prefer a LCD over a CRT monitor might be:

For other resolutions in your film recorder you can resize, letter box, or crop the 2048x1536 image to fit the screen resolution. Regular Digital Cinema image size 2k is 2048x1556 which many video boards do not support, and does not come out 4:3 ratio if the pixels are 1:1 ratio, which would make all images in the 2048x1556 resolution distorted on a normal 4:3 monitor, so for practical reasons you would probably be using computer 2k resolution of 2048x1536 with my programs and in your home made film recorder. Also 2048x1536 scales directly to 1024x768 which is used by my GUI tools since many video boards have up to 2007 supported the computer 1k video mode in 24bpp or 32bpp whereas only a few video boards support computer 2k in the required VESA BIOS calls, and computer 1k is faster to redraw the screen, but gives a undistorted image at quarter resolution. The true resolution of computer 2k CRT monitors might be close to computer 1k resolution even when working at computer 2k resolution, but you would see less digital artifacts shooting the frames at computer 2k resolution since the scan lines blend together better.

To get the recommended target density measurements visit the Kodak (tm) web site and read their documents H-61 et al. since each film stock may have a different mask, bias, or other factor to compensate for.

The color correction shown in this illustration is a compromise since the image on the LAD negative exhibits several issues, the illumination is uneven being brighter on the left side, the color of the image is uneven, the reference Gray swatch in the lower left seems to have a different white balance than the stepped gray scale patches so it is impossible to reconcile the two white balances, to get the best flesh tone it seems better to have the balance go a little yellow in the highlights, the Black card on the left looks a little blue, and the black patch on the right looks a little red so there does seem to be a need to look at the overall effect and make a judgment as to what looks good and how much highlight and shadow detail to keep, since after all you will be printing images of actors and not test charts of obscure reason. No doubt the color masking makes the intrinsic lighting and pigment faults in the LAD film more obvious. There may also be some issues with the printer used to make the LAD films, or at least the sample I scanned, since the illumination of the printing, and or camera exposure, seems to be uneven, as does the color of the light used. The image is also very soft, as if the optical printer used was slightly out of focus to mask the grain in the image or just out of focus for some other reason. You can see the resolution of my scan in the scratches picked up when you look at the full resolution NIP file with the 1:1 zoom image probe in the Sharp/Soft filter. It would be nice to make a better reference film using modern film stocks in the camera off a live subject, rather than printing from old color separations as I guess the LAD films have been. Kodak (tm) has a new Digital LAD image file, but it may lack the more important and useful features of the old film LAD image for contrast control by omitting some of the Gray steps and the near Black, Black, near White, and White areas. To correct the colors in photographed shots by color masking (chroma) you need primary and secondary colors shot on the camera film under the same lights as the actors, not color patches added digitally on top of a scanned image. Anyway, the Digital LAD file would not help set-up the scanner, just the film recorder, since you need a camera original negative film image to scan.

With regard to the appearance of the color balance in the movie theater, the color temperature change of the projection lamp due to age and power supplied, the dirt and smoke on the projection booth window, the color of the "dim" lights left on in the theater, the type of screen installed, and many other factors such as the color of the upholstery on the seats reflecting light back on the screen all contribute to the perceived appearance of the image. Luckily if the image on the film is not too far off color at the reel changes, the audience's eyes will get adjusted to the off balance color print in a few minutes and probably not ask for their money back. You can look at the print through Red, Green, and Blue filters to see if all the patches of the gray step look to have similar spacing with none gone from the highlights and shadows, the Blue filter shows the Yellow dye image, the Green filter shows the Magenta dye image, and the Red filter shows the Cyan dye image, all of which should look to have about the same contrast and gamma for a black and white or gray scale image. Printing a gray scale photograph on color film can actually be a better way to set up the color balance than using a color image since getting a black and white gray scale image neutral (without any colored tint) requires closer balance than what makes for a tolerable color image, especially now since movies are being printed with highly off balance tints such as sepia, greenish, and bluish, pretty much anything goes, so if the audience does not know what tint you were going for they will probably accept anything that is not too wild, just say what you got was intentional because you will never get anything on film perfect.

I might mention that the Xenon arc lamps may be weak in the Red so when you look at the images on processed print stock exposed in your film recorder over a light box or over a white paper it may be better to have the print tinted a little yellow to orange since the rendering of the actors and sets may loose some warm colors on projection. Also judging color over a fluorescent light box is problematic since the fluorescent lamps or CFL lamps can have uneven spectrum and give off balance viewing. The color of the lights in the room you use to adjust the Grading on your computer can through off your eyes color balance, and should match the monitor color balance. LCD monitors may be a problem for making the color Grading adjustments since when you move your head around the apparent brightness can change due to the viewing angle, so it may be better to use a high quality CRT monitor to make color Grading adjustments in my or any other software.

I should mention that the shutter in your film recorder's camera should not have any play or backlash since when operating the film recorder's camera in stop motion single frame mode the shutter blades of a dissolving shutter can wiggle and not always stop in the same place introducing micro variations in the exposure time for each exposure that result in flicker in the processed motion picture print stock. The flicker problem is made much worse now because the print stocks have a higher contrast and amplify the exposure variations by perhaps three or more times. If you are making a printing negative in your film recorder you may not see the flicker until the prints are struck off of it. Be sure to watch the processed print in a real 35mm projector on a screen to look for flicker, since viewing on a flat bed, or from a Telecine, may not give you the right viewing conditions to see such flicker issues. Try to keep the shutter in your film recorder's camera as tightly locked to the animation motor as possible, or add a capping shutter to the camera to control the exposure after the camera's shutter has opened and come to rest.

See the video about building a film recorder, click here for links to video C0020020, and see the HTML pages about building a film recorder, click here to to SECTION: 5.1 Index.

This is a un-cropped negative scan without any software color correction of the LAD reference negative used in the screen shots below for showing the color balancing and correction tools available in the Digital Cinema command Edit list in the Kinema sub- menu.

If you cannot get a piece of LAD film, you can make one for yourself. It is best to use a real person for the flesh color sample, since the camera film reacts differently in the IR and UV invisible light than your eye does and those rays can affect the color, real flesh is better as a reference of the way the light reacts to the colors and light on the film image.

The light house, exposure, and filtration need to be adjusted to compensate for the orange mask in the masked motion picture stocks. Many camera negative stocks have a color mask to compensate for the color shifts and de- saturation, but when you shoot the color negative with a DSLR you will want to have the light very blue-green to compensate for the orange or pink cast of the negative or duplicating film stocks. You may need to raise the light source to a color temperature above 7000K with CTB type color correction gel filters. When you have the light the right color gray objects in the image should look gray in the scanned image both looked at as a negative or when inverted to be a positive. You can use the digital densitometer in the Levels command to measure the gray patch in the LAD image to help you adjust the light house filtration. It is best to get the image scanned as close to correct before making digital alterations to minimize the distortions introduced by having part of the scanned image out of the range of the cameras sensor.

People familiar with film images will notice that the frame line and sound track areas are white in the positive and black in the negative, the reverse of what a camera negative would have, this indicates that the LAD negative was printed and not shot from a live subject, apparently Kodak (tm) wants the frame lines reversed in tone to allow for adjustment of the color analyzer white level and to act as a D-max reference, where as the area around the woman's head is unexposed blacker-than-black to act as a reference for adjusting the printing density, note that the white patch in the upper corner is not as dark in the negative as the frame line, showing that the highlight detail is held, and the central black patch on the side is not as clear as the area around the woman's head showing that the shadow detail has been held in this scan.

This particular scan was made from nine exposures of the LAD 35mm negative frame held still frame in an Oxberry (tm) optical printer pin registered projector, a 105mm Printing Nikkor (tm) optical printer lens at f/8, and a Canon Rebel XTi (tm) DSLR camera in RAW mode making 9 exposures using the auto bracket feature, 3 exposures Red through a Wratten 25 filter, 3 exposures Green through a Wratten 58 filter, and 3 exposures Blue made through a Wratten 47 filter. The camera was set to Monochrome mode, with the color temperature set to about 7000K. The color separation exposures were Fused together into a 48bpp NIP (actual bpp in images would be less than 48bpp) file by using the Fuse option of the image Insert command in the Edit list. The light source was tungsten about 2800K filtered up quite a bit with a stack of Lee (tm) 1/4 CTB filters. Better results might be obtained with Wratten filters 70, 98, and 99 but at greater light loss. LEDs might also give a better light source if the right wave lengths are selected for the film maximum dye absorption peaks, but may not be bright enough to keep the exposure under a quarter second. The power supply for the light source should be filtered DC if the exposure time is under 10 seconds. Digital camera's tend to have higher noise in the blue exposure, so making the light extra blue and reducing the gain on the blue sensors may help reduce the sensor noise from the scan that is above the grain noise actually in the film. Do not stop the printer lens down past about f/8 unless the negative is very granny since doing so may reduce the resolution due to diffraction, at 1:1 the markings on the printer lens are about two stops larger open than the actual stop, in other words f/8 is really about f/16 and f/16 is really about f/32. You would think that using f/5.6 on the printer lens, i.e. f/11 in fact, would be best, but the film does not lay flat in the projector gate, so stopping down to the diffraction maximum limit is usually best to avoid fluctuating sharpness due to the film "breathing" as it goes through the projector head's movement from frame to frame.

The dynamic range of the negative scan made with the Canon Rebel XTi (tm) DSLR using fusion of three exposures in each primary color with the bracket set to about +/- 1/2 stop seems to encompass all of the density range on the negative and then some extra to spare. If you notice the overscan on the side of the frame without the sound track you can see a difference between the unexposed area caused by the side of the projector's gate and the D-max surround on the film, also the clear area around the woman's head seems to have some recording of the fog level. If greater dynamic range was needed the bracket of the exposures could be increased, and or, the number of exposures fused could be increased. Increasing the number of exposures to fuse makes the number of data bits in the gray scale increase generally making the gradation smoother, but at the expense of slowing down the scanning. You will need to work out the compromise between quality of the scan and total scan time required.

You can download the un-processed Fused scan of the LAD image to experiment with in my Kinema Edit list, this is a large 3906x2602 image stored in a 48bpp uncompressed image file format, my NIP file format, so will take a while to download using a dial up service. As an alternative you might save the small 512 JPG file from the link above and convert that into a resized BMP file to play with in my programs.

Click on LAD3906N.ZIP to download, 34423983 bytes.

You will notice that this scan is negative and flipped left to right, when the film is loaded into the projector with the emulsion facing the camera the image can be correct or flipped left to right depending on if it is an "A" wind or "B" wind image, that is an even or odd printed generation. So you need to go into the Alter configuration part of the Grading command in the Kinema Edit list and select flip X axis and negative source image options to fix the image so that it comes out the right way around and positive. All adjustments should be made with the image shown as positive, if you want a negative output image you should make it negative after making the adjustments as output negative or make it negative in the film recorder software like DANCINEL.EXE (tm).

If you wish to see how the Fuse command converts color separation exposures this download has 391x260 resolution copies of the color separation images used to make the 0.NIP file in LAD3906N.ZIP,

Click on LADFUSE1.ZIP to download, 7436063 bytes.

Be sure to read the README.TXT files in the ZIP files for more information.

After you use the Fuse option of the Insert image command to make a color NIP file from the color separation BMP files you can look at it in the Grading command's Levels #1 command. The Grading will also work with BMP source images in positive or negative but there will be less image tone data than in the Fused data stored in a NIP file.

If the image shows up as negative exit the Levels command back to the Grading sub-menu and use the Alter config command to change the KCC (Kinema Color Correction) config file for the selected key frame so that the axis are flipped and the image is rendered negative (negative of a negative gives a positive.)

The Alter command in the Grading command in the Kinema Edit list command can be set to flip the X axis and to invert the negative colors to positive images, which as been done here.

The first thing you should notice is to look at the three separation images across the top of the screen, you should see that the Blue separation is higher contrast than the Green separation, and both the Blue and Green separations are higher contrast than the Red separation. To correct this situation you would narrow the spacing between the Black and White values for the Levels of the Green separation image a little, and more so for the Red separation image. Once you get the Black and White features in the three separations to match, you adjust the Gamma for the separations to get the Gray patches to have the same brightness value, and then adjust the Gamma for all three separations together (by selecting the lock color adjustments together mode by pressing the [+] key until indicated) until the Gray LAD swatch patch meets the target digital density.

Here you see the White and Black levels in the separations have been balanced, and the Gray levels have been adjusted with the Gamma settings, as was described just above.

The digital densitometer probe was adjusted to cover an area over the LAD Gray swatch patch in the lower left corner, it is hard to see when the gray level is adjusted properly since the XOR of middle gray is middle gray, but if you look closely you can see the probe outline since the film image is not perfectly uniform.

The "magic color balance widget" in the lower left corner should be compared to the example in the uncorrected Levels screen shots. The "magic color balance widget" works by having four nested square color patches, the outer ring and the inner square are reference target gray, the larger ring is the average color from the pixels in the digital densitometer probe sensing area, and the ring near the center square is the monochrome version of the average pixel values with the color removed. To use the "magic color balance widget" you resize the digital densitometer probe sensing area with the [1], [3], [7], and [9] keys then move the probe over a 18% Gray card in the shot. You then look at the "magic color balance widget" and adjust the Gamma for each primary color separation until the average color and monochrome rings disappear (look the same), indicating that the probe sample area is now balanced to neutral rendering. The numbers above the "magic color balance widget" are also helpful in seeing if the separation colors are high or low for their target value.

As an adjunct, the Output Histogram shows a peak for the colors in the result, so that you can adjust the Gamma for the separations to make the Red, Green, and Blue peaks superimpose showing equal brightness, and if you put middle gray on the center of the Output Histogram you have the gray in the middle, obviously.

With all of these Levels and Gamma tools and widgets you should be able to see that color balance is neutral even if you have a monochrome monitor, a broken monitor, or are color blind, just read the numbers and graphs, and match the color separation images for highlight, shadow, and middle tones on the gray scale images. At least you will be in the ballpark.

If the source material has the highlights or shadows clipped in one or more of the separations, the only way to achieve neutrality across the scale is to clip the other separations to match the worst one, or accept some drift in the shadows or highlights. If the tones are compressed rather than clipped in one or more of the separations you may be able to get a better balance by using the Curves tool rather than the Levels, so in such a case, only match the clipped parts of the image scale and do not clip more than that, then use the Curves to bend the image tones back to balance. If the source material has bad grain or noise in the shadow areas, you may want to clip the shadow area with the Black Levels and use the Curves to knock the shadow tones down so that the clipping is less obvious.

The digital densitometer probe has been adjusted to read the Gray LAD patch and the Gray scale steps. You see the peaks in the output Histogram for each of the steps line up somewhat. The White peak is a little low, and can be raised by changing the White level for each of the separations, or by using the Curves tool to bend the transfer curve up to expand the highlights near White without fully clipping all the near white tones, giving a more film look rather than the hard clipping of the Whites associated with video images. In other words it is better to leave a little too much highlight and shadow detail than to clip the tones as close as possible since the film stocks used in the film recorder and other output devices used for display will tend lose some highlight and shadow detail and if there is none there to lose you will lose those parts that you wanted to keep later down the line.

You can see the difference in the lightest peak in the output Histogram, the Red peak is right of the Green peak, and the Blue peak is left of the Green peak, which you should know from color theory would indicate a yellowish orange tint rather than neutral very light gray. In this instance, if you push the Blue and Green peaks up by moving the White clipping levels for the Green and Blue to the left the reflections on the woman's cheeks and chin tend to go pink or magenta, so it is best most of the time to have the highlights a little yellowish since the clear areas of the print stock can go greenish or pink from the processing stain rather than being clear film base as would be the case in black and white print stock.

The digital densitometer probe has been adjusted to read the near White, near Black, and middle Gray patches of the LAD image. The average color shown in the "magic color balance widget" is off center since the near Black swatch takes up a disproportionate amount of the probe sample area.

You can see that the near Black peak in the output Histogram display is in about the right place, and the middle Gray peak is in the center, but the near White peak is somewhat low and the peaks from the separations are not well aligned. The Curves #1 tool can be used to bend the curve to stretch the highlights, or you could clip the White levels lower and adjust the middle Gray back down with the Gamma values.

The individual transfer Curve for each of the three color separation images can be bent to correct for drift between the transmission of the primary colors through the color negative for neutral colored objects with different reflectivity. The digital densitometer can have its probe size reduced to read small patches in the image to match colors from one shot to another shot, or as in this example to read one of the Gray steps. Notice that while the "magic color balance widget" is off the target density since a lighter step is being measured, the color of the two inner rings is the same indicating that the color for the probe area is balanced to a neutral Gray color.

Once the Levels # and Curves #1 tools have been used to adjust the tone values to achieve neutral color in the output image for neutral colored objects from Black to White you then need to consider the degree of desaturation of the colors. The color desaturation occurs because the three layers of the film are not just sensitive to their own primary color but also somewhat to the adjacent colors. Also the DSLR even when used with color separation filters will not divide the separation images perfectly due to the separation filters not being monochromatic, and the dyes in the film not being absolutely able to select just their separation image alone. The color Masking can subtract some of the crosstalk that causes desaturation but before you use the Masking you should try to use the Levels #1 and Curves #1 to get all parts of the image on as good a neutral balance as you can since the Masking will exaggerate any small imbalance in the colors and other tones.

You should not try to make one shot exposures off color negative using the Bayer filter array in the DSLR since the color quality will probably not be as good as when shooting separation exposures through color separation filters. Using a non-Bayer type monochrome digital camera or a Sigma SD14 (tm) type DSLR using a Foveon Inc. Foveon X3 (tm) sensor may give better results than using a Canon XTi (tm) type DSLR using a Bayer filter array, particularly with regard to the noise in the Blue image and the sharpness of the Red image.

The Masking tool is a supplement to the color mask incorporated in the color negative, duplicate negative, CRI, or master positive film stock, and is used to subtract desaturation and bring the colors back to what they looked like to the cameraman's eye on the set. You can also over compensate a little to partially compensate for desaturation caused by the print stock and colored filters used in the film recorder.

The Masking tool can also be used to add or mix colors to desaturate the colors in the image, make a monochrome image with various ratios of the primary colors to simulate different types of black and white film stocks or filters with black and white film, tint the image to give it a color cast, or create false colors by mixing the wrong input primaries to the output primaries.

The chroma Vector display shows the color of the pixels in the probe sample area without regard to the brightness of the pixels, for instance any shade of gray from black to white will make a spot on the center of the chroma Vector display if it is balanced neutral, and saturated colors plot along the color hexagon that runs between the Red, Yellow, Green, Cyan, Blue, and Magenta markers. Here you see several blobs in the chroma Vector display since the whole frame image is being sampled, with each point representing a different color of pixel in the image, the dark areas tell that that color is not in the image. If you load an image of a color chart that has all colors at all saturation levels you would see the full hexagon filled with dots.

Here you see that the probe area has been reduced so that only the primary swatch patches for Red, Green, and Blue along with a little of the dark Gray between them is sampled. Looking at the chroma Vector display you see the three well defined blobs for the Red, Green, and Blue primary swatch patches along with some other blobs representing the slightly off neutral color of the dark Gray lines between the primary swatch patches.

In addition to the chroma Vector display, you can look at a Waveform type display of the sample area within the probe. Here you can see that the primary colors from the primary swatch patches make three uneven peaks with ratios to the opposing colors of less than 100%. If the colors were fully saturated the proper colors would be at the top and the improper colors mixed in would be at the bottom, however that is not the case in the un-masked image data since the original primary color swatches were not back lit filters, were made of impure color paints, had reflections from their dull finish, got washed out with lens flair, and underwent desaturation being captured by the film stocks.

The Waveform display and the probe average values displayed can be used as an adjunct to the chroma Vector display to judge the neutrality of the color balance, or the degree of tint being introduced. This can be important to match the color properties from one shot to another, or to match studio process shots to location shooting.

Here you see the blobs representing the primary swatch patch colors have moved out closer to the edge of the chroma Vector color hexagon. The Green blob is not all the way out, but the Red and Blue blobs are out about as far as you would dare take them, actually a little too far out since the colors of paint on the primary swatch patches were not that saturated when viewed by the eyewitness on the set.

You do not want to see the outline of the chroma Hexagon by having the chroma Vector dots line up along the edges, seeing the edges of the Hexagon indicates that you have increased the saturation too much and you should decrease the amount of Masking applied. You can see this defect between the Red and Yellow markers in the chroma Vector display shown in this example, i.e. the line of dots along the edge of the chroma hexagon. To fix this problem reduce the amount of Green you subtract from the Red in the Red separation, and so forth.

You can compare the Waveform display before and after adjusting for the color Masking for the primary swatch patches to see that the spread between the correct color signal and the polluting colors has been increased by the Masking. Even after Masking the Green signal is not at 100%, nor should it be since the degree of Masking shown here is over extreme in order for you to clearly see the changes applied.

If the film I had was grain free and had even color properties the blob in the center of the chroma Vector display would be a small dot or not even visible at the very center of the chroma Vector display, however in practice the sample of the LAD film's middle Gray reference patch has numerous colors near Gray giving a blob of moderate proportions, along with a few stray dots caused by the off color blemishes.

Centering the blob representing the LAD film's middle Gray reference patch in the chroma Vector display tells you that the reference patch is being corrected for neutral color balance, which can be a help in correcting for film shot under light that was not at the correct Kelvin or came from vapor discharge lamps. If the blob for a 18% gray card in some film you shoot is very large you would try to center it on the marks. Make the probe area as large as you can and still fit on the Gray reference since a small probe area's sample will be skewed more by uneven texture of color in the film grain.

You can read the LAD film's middle Gray and the gray steps at the same time by making the probe area into a wide short rectangle. Looking at such a sample on the Waveform display shows you the purity of the balance of the various shades of Gray. Notice that the LAD reference patch is half way up on the graph, the slight slope down on the right is due to uneven illumination. The spacing of the steps is on a slight curve, this is in part due to the choice Kodak (tm) made in the selection of the reflection properties of the swatch patches in the Gray steps, and how the Levels and Curves were adjusted.

The Blue fuzz in the Waveform display comes from the heavy noise in the Blue exposure, which can be seen even in the Blue thumbnail separation image in the upper right of this filter's screen shot.

The misalignment of the Waveform for the primary colors associated with the White step patch indicates some drift from neutrality which could be somewhat compensated for with adjustments to the Levels and Curves.

As was noted above one pass through the chroma Masking was not enough to bring the Green saturation to 100% on the Green patch swatch, so this screen shot is taken from a second pass through the Insert command to do the chroma Masking a second time.

On the second pass the Green saturation is brought just about to the Green limit on the chroma Vector display.

By limiting the probe area to just the primary swatches you can see the blobs in the chroma Vector display more clearly at the extreme color points, with a few other spots representing the dark Gray lines between the primary swatch patches being a little off neutral at this time. The dark Gray could be brought to neutral by using the Levels and Curves but the paper used may have had a slight tint, and one needs to make the color balance look best for the actors and sets rather than just a test chart.

You can compare the Waveform here showing almost 100% saturation for the Green swatch patch with the two previous Waveform displays above for the primary swatches made using no masking and an intermediate level of Masking. Since the Green object photographed was not at 100% saturation, making it so through masking is not the proper way to do the color correction, this is just showing that it is possible to restore colors from a desaturated film scan.

The Sharp/Soft filter has a 1:1 zoom probe to allow you to find some part of the image that has sharp detail and view that to make adjustments to the filter's values. You can only judge a filter's settings by using a 1:1 image display since the effect of such filters is on the pixel level and the values selected depend on the size of the image, the viewing distance, and losses in the film recorder and motion picture projector being out of focus most of the time.

In addition to the 1:1 zoom probe there is a Waveform display that shows the Waveform for a line across the 1:1 zoom probe's sample area. An increase in the peaks of the Waveform indicates sharpening, a decrease in the peaks of the Waveform indicates softening, and a jagged appearance of the Waveform indicates film grain and or digital camera noise and would indicate the need for less sharpening for that separation color or more softening to blunt the noise. The Blue separation is particularly prone to high noise levels, so may require less sharpening or more softening than the Green or Red primary separation.

Here the Sharp/Soft filter values have been adjusted to have the filter be active but since this image was very large and fuzzy the effect is hard to make out except for the reflections in the eye.

The Sharp/Soft filter uses edge detection combined with unsharp masking to sharpen edge details without increasing the grain in undetailed areas of the image. The Bias values can be used to have the filter act on all of the pixels rather than just those near edges if desired, but when the net Bias is positive the film grain and blemishes will increase across the frame. The Bias can be adjusted slightly negative to have the filter soften the low detail areas of the image to reduce grain, while at the same time sharpen the high and small detail areas of the image to compensate for the use of high speed lenses and shallow depth of field a little. The character of the unsharp mask is best suited to processing camera negative shot with high speed camera lenses that have some fuzz around a sharp central image, rather than for processing copied film that was made in an out of focus optical printer.

The Sharp/Soft filter needs to have its values adjusted to different settings for making output frames at 720x480 pixels rather than 1920x1080 because you need to use a larger radius for the unsharp mask since the radius of the circle of confusion of the camera lens was in relation to the width of the frame, not the pixel size of the scan, therefore the unsharp mask radius needs to be larger for high resolution output images than it does for low resolution output images when the filter is applied at the output files resolution.

The affect of the filters on the image will depend on the way the output images are viewed, you cannot over sharpen an image and compensate for the movie projector being very out of focus, while if you use a large radius unsharp mask you might be able to compensate somewhat for a movie projector that is slightly out of focus if the viewer is near the back of the theater. As with many things, everything in moderation will generally be the best path to success, do not over mask, sharpen, or soften your film images or they may start to look like bad video. You should use the sharpest lens you can on your film scanner since sharpening a soft scan will amplify the DSLR camera's sensor noise, perhaps even far beyond the original grain noise from the film you are scanning.