This is a very simple tutorial. Its purpose is to show how to work with Cineon files and combustion. Generally speaking the rule of thumb is, that you should do all your color correction in
Log space and the majority of your image transformations in linear space. Some image transformations are acceptable in Log space such as using the reveal or clone tool to remove dust and
scratches, other jobs like compositing should always be done in linear space. For more information on this, I recommend this book “Digital Compositing for Film and Video ” by Steve Wright.
The average size of one 2K 10bit cineon file is just over 12 Megabytes. Needless to say that’s a bit of a problem when it comes to writing a tutorial. So for this tutorial we will be using
the Kodak Digital LAD test image. Which can be found at : http://www.kodak.com/US/en/motion/support/dlad/ Please
download this image before beginning the tutorial and place it in the tutorial directory.
Okay lets go:
Start by opening up combustion if it’s not already running.
Now click on the FILE menu and choose NEW. (CTRL+N)
This menu will open:
Make sure that the following settings are correct.
Name: <- you can name your composite or leave it as untitled
Format Options: Film Full Aperture (2K)
Bit Depth: 10bit
Now click the FILE menu and choose IMPORT FOOTAGE. (CTRL+I)
Drive to the tutorial folder and select the dlad_2048x1556.cin file. Then choose okay.
Your workspace window should look like this:
And your view port should look something like this:
Use the in or out on the view port so that you can see the whole frame area as show above and make sure your display quality button is set to best.
Notice that the shot is washed out looking. It looks very low contrast. Now let see what happens when we choose to display all the information sent to this view port with a view LUT.
Click on the WINDOW menu heading and choose VIEWLUT and then choose the number 1. (note each one of the numbers listed in the window relates to a store in the view LUT control panel. This is
a very handy feature. This allows you to have various LUT’s and switch between them by choosing the appropriate number from this menu.)
Now if you don’t see the following window click on the view LUT tab to display the view LUT in your control panel area. If you don’t see a View LUT tab the select WINDOW: VIEW LUT: EDIT LUT
from the menu bar or hit the F10 key.
You should see this:
Click graphic above to see larger view
Change the Method setting from “Straight” to “Film Look”
Then choose “Log” as the Data Source. (Note the data source button appears after you change the method setting.
Leave the Print Film setting at 2383.
Okay look at the image now. It’s full of detail and contrast. The color may not be to your liking depending on how your monitor is calibrated, but you are looking at a representation of what
the log file will look like once it’s printed to film. There will be subtle to major differences in the color of the output film depending on your monitor calibration for instance and the
lights that the laboratory uses to print your film. If you are doing film work you will need to take the time to calibrate your monitor so that it is giving you a very close approximation of
There you have it. You are using a view LUT. The 10bit file is being displayed filtered through the view LUT. No change
has been made to the underlying footage and when you render it out you are rendering the original footage without any conversion taking place. For color correction this is the only way to go,
but what about compositing? This is when we need to convert our cineon files into linear so that we can pull good keys, make image transformation etc.
Let's take a look at how we can accomplish a bit depth conversion without loss of color data. We will now convert this
image from a 10 bit logarithmic format into a 16 bit linear format by using the bit depth conversion operator. Remember we want to do all or our color correction in log color space and all of
our image transformations in linear color space.
Deselect the Use LUT button on the View LUT tab.
Next highlight the dlad_2048X1556 layer in your Workspace tab.
In the menu Select OPERATORS: COLOR CORRECTION: BIT DEPTH CONVERTOR
Select the Bit Depth Controls tab:
Click on the Settings button then change the Output Color Depth button from 8 bit to 16 bit.
It should look like this when you are done:
Now click on the LUT Editor Button and change the following:
The Convert Button should be changed from “Straight” to “Log to Linear”,
after changing the button it should look like this: (without my markings)
Click graphic above to see larger view
Notice that the “Reference White” is set to 685 and that the “Reference Black” is set to 95. This means that all the pixel values equal to or above 685 will be mapped to white, and that all
pixel values equal to or lower then 95 will be mapped to black. For many jobs this is going to work just fine, however by leaving the setting at these default values we are degrading the
file. Look at the histogram, see everything to the right of the white reference vertical line in the graph? That information is going to be clipped and lost if we leave our White Reference
settings like this. So change the Reference White setting to read 1023 for each of the color channels. Since the default is to have all three channels locked, you will only need to change the
value of one of the channels. Now change the value of the Reference Black to 0. Leave everything else exactly as it is. If you know the correct gamma correction value of you monitor setup
then change the Gamma Correction value to that number. In my case it’s .6 If you don’t know your Gamma Correction value then leave it set to 1.0.
Here is how it should look now:
Click graphic above to see larger view
You may have left your Gamma Correction at 1.0, in any case you should have noticed that now the image in your view port is very dark and doesn’t represent what we expect to see on film. If
you look at the graph it’s easy to see why. The majority of our colors are being in represented in the lower 20 percent or less of the color space we have at our disposal. Here is the really
important part. As long as you remember these settings when your do another bit conversion back from linear to log, your file will not suffer any color data loss. But wait you ask “It’s too
dark I can’t see anything!” Here is a good way around this. Add another Bit Conversion Operator. This one will counteract the effects of the first one, turning our linear file back into a log
file. We can then use a view LUT to represent the composite in our view port correctly.
Highlight the Dlad_2048X1556 layer and select OPERATORS: COLORCORRECTION: BIT DEPTH CONVERTOR. Under settings change the Output Color Depth to 10 bit.
Click on the LUT Editor button and change the Convert Button from “Straight” to “Linear to Log” now Change the Reference White setting to 1023 and the Reference Black setting to 0. If you
changed your Gamma Correction from the default then set it to the same value you used earlier.
Now your workspace and controls should look like this:
Click graphic above to see larger view
Lets rename the Bit Depth Converters with more descriptive titles.
Right click on the Bit Depth Converter(2) and rename it to BDC_LIN2LOG
Right click on the Bit Depth Converter and rename it to BDC_LOG2LIN
Your workspace tab should look like this:
Next lets import another cineon layer so that we can do a simple key.
Select the Composite – Untitled layer in the workspace tab as shown above then:
Select FILE: IMPORT FOOTAGE and choose “Test_Cin_0001.cin” from the tutorial directory.
Your workspace tab should look like this now:
Position the Test_Cin_0001 layer so that it covers the Kodak Logo. To position the new layer “Text_Cin_0001” so that it covers the Kodak logo in the lower right, I used the layer transform
controls and I set my X-Position to 662 and the Y-Position to –453. You could also just drag it over with your mouse so that it covers the entire Kodak symbol.
With your view port set to Composite View (Normal) it should look like this:
Note: this new layer is also a cineon file. It’s from a feature film. I cut it down in size so that it wouldn’t be so big for the tutorial but it is also a 10 log cineon file. It’s pixel
dimensions are much smaller than the 2048X1556 however.
Next lets select the Test_Cin_0001 layer in the workspace tab and drag it down the hierarchy so that it’s just below the dlad_2048X1556 layer.
Your workspace tab should now look like this:
Now that the Test_Cin_0001 layer is below the dlad_2048X1556 layer it is now longer visible when your view port is set to Composite View (Normal).
Lets change our view port layout to two up. Click the View port layout list and select the two up button:
You should have two View Ports showing now.
Lets check the View LUT settings for each View port. Right click on the left View port and select the View LUT heading, make sure that number one is checked from the sub menu
On the right View Port lets keep our View LUT settings set to None, by doing the same procedure as on the left but making
sure None is checked instead of 1.
Now with the left View port selected so it is active (has the grey highlight around it) lets double click on the
test_Cin_0001 layer of the workspace tab.
You should see the layer in your left view port. This layer looks a bit green so lets color correct it a bit. Remember
since we are using a view LUT and our monitors are calibrated we should be getting a fairly reasonable idea of what this will look like filmed out. So lets adjust the color.
With the Test_Cin_0001 layer highlighted in the workspace tab add a color correction operator by selecting from the menu
OPERATORS: COLOR CORRECTION: DISCREET COLOR CORRECTOR.
Now select the discreet color corrector operator and with the Color button selected and the Master button selected double
click in the strength field and type 12. (this isn’t a color correction tutorial so this much will have to do. I’m just illustrating a point here is all)
Next lets Double Click on the Composite-Untitled Layer of our workspace tab to bring the composite layer into our left
Now lets highlight our dlad_2048X1556 layer and add a Discreet Keyer operator, by selecting OPERATORS: KEYING: DISCREET
Let drag the Discreet Keyer so that it’s below the BDC_LOG2LIN Operator and above the BDC_LIN2LOG operator. Your workspace tab should look like this:
Next lets pull a key from the orange area of the Kodak symbol.
Now with the left view port active double click on the Composite layer in the workspace tab to load it into the view port.
Next select the Discreet Keyer layer in the workspace tab under dlad_2048X1556 layer.
You should see the discreet keyer controls tab open, if it’s not click on the tab to make it visible.
Leave the Keyer in RGB.
Click on the eye dropper next to the Key Color and select the orange area of the Kodak logo from the left view port. Next Click on the Matte button and double click on the gain field. Set the
value to 280. Click on the Shrink check box and set the value to 1.0.
(note: there are many ways to pull a key I am using this one only to illustrate a point)
You have just made a composite of two cineon files. Notice that the only file converted from log to linear was the file with the Keyer applied. This is the image in your right view port. If
you were to render this file out to a cineon sequence you would do it straight and not apply any LUT to it as the composite image is in 10 bit log format.
There you have it. You now know how to convert a 10bit log to 16bit linear file and back again.
I would like to thank Ken La Rue and Lee “Rod” Roderick in helping me to proof this tutorial.
---Alan Edward Bell
Discuss this technique in the Combustion forum