September 4, 2014
Lots of little fixes. The ACES article has been updated with feedback from Jim Houston, Academy co-chair of the ACES project (thank you!). Addressed an issue with the domain name so you may have to re-link the site's RSS. Added "Galleries" for my photo projects. And a lot more. Now that all this is done, standby for something completely different... ;)
Alexaremote Control Unit and Transceivers for 2 Cameras
September 4, 2014
Remotely access the Alexa's menus and functionality with this product.
I spent a few weeks with a two-camera Alexaremote kit courtesy of the developer, Balint Seres. One of my goals as a DIT is to accomplish as much in the camera as I can. If there are exposure issues, get it on the lens. Color temperature problems? Fix it in white balance. I prefer to leave as little to chance as possible and have the images come on in their own looking as "right" as possible. For the past few years I've been achieving this by using this device -
The Arri RCU-4, which is essentially the interface side of the Alexa camera detached. It allows full remote camera control but it happens over a bulky, finicky, expensive proprietary cable. I've been building this cable into my BNC looms but it requires a bit of extra muscle from whoever is cabling and it's delicate so prone to connection related problems. Yet still despite the inconveniences, it's allowed me to work the way I want to work.
The Alexaremote accomplishes the same thing but over the air! (note the unit can also be "hardlined" to the camera with a simple ethernet cable)
Here I am interfacing with an Alexa camera, full control, completely wireless.
Whereas with the RCU-4, one unit is required per camera, the Alexaremote can store and switch between up to 10 cameras! Switching between them is fast and seamless. Each camera requires it's own radio transceiver which plugs directly into the interface port which conveniently also provides power to the unit.
The transceiver is unobtrusively velcroed to the top of the camera under the handle.
Getting the Alexaremote and the camera talking to each other is fairly simple and involves inputting the camera's IP Address into the unit. Note the Alexa will occasionally change its IP upon reboot so multiple addresses may need to be stored for a single camera.
Alexaremote Menu Demonstration
The time I spent with the Alexaremote was positive. The setup was easy and unit responsive. It's construction is very solid so it should be able to take the inevitable beating. I was using it for hours on end with only two of the Sony NP type batteries and was surprised at how power efficient it is. In my experience the range was reliable 50-100 feet with it losing link more often at distances beyond that. I have heard of users putting higher gain antenna on the transceiver side to improve range. I was not able to test this myself.
Beyond access to the normal functionality of the Alexa, the remote has a unique Intervalometer feature that allows you to program timelapses and unusual increments into the camera.
The Alexaremote is available for purchase now. Balint Seres can be contacted at info@alexaremote.com
/ END
May 30, 2014
On negativespaces.com you’ll find numerous articles on the subject of monitor calibration. I’d say that of all the topics I’ve covered since 2007, this is the one that I've emphasized the most. Accurate monitoring is of paramount importance to the work we do as cinematographers, photographers, and technicians. Without a monitor we can trust, we don't have much as imaging professionals. Among the various display technologies currently available, OLED's have emerged as the gold standard for this.
I’d now like to dig a little deeper than I have in the past and present advanced information on OLED image processing and 3D LUT (Lookup Table) based calibration options for them. This article was made possible through collaboration with Bram Desmet from Flanders Scientific, Inc. (FSI) as well as input from Sony Broadcast OLED product manager, Gary Mandle. It's heavy on the subject of 3D LUT based calibration as this offers more control for the end user. Because FSI’s CM250 and CM172 professional OLED monitors allow for direct 3D LUT implementation, this article is somewhat specific to them.
Disclaimer - I've owned and operated a good variety of product from both Sony and FSI. I'm in neither one camp nor the other and both company's have excellent offerings for the professional video community.
My previous writing on this topic detailed the calibration process for Sony OLED monitors using White Balance Adjustment. To recap, this process uses Sony's Auto White Balance Adjustment software and one of several supported probes to either manually or automatically adjust the monitor’s RGB gains and bias to arrive at a chroma-free white and dark gray point.
Correct monitor white balance should ensure the reproduction of a neutral gray scale and accurate colors. If it doesn't, then there are calibration issues that cannot be resolved through white balance adjustment alone. In this case, a custom calibration 3D LUT could offer a solution. The advantage of 3D LUT based calibration is that gives the user total control; not just of white balance, but gamma and color gamut as well. This allows for greater user customization and can remedy display issues beyond what can be done through simpler means.
Lookup Tables aren’t used solely for calibration and are actually an integral component of image processing in all professional monitors. An order of operations is employed between signal input and the final display. Somewhere in the signal chain a 3D LUT is used to transform the monitor's native, wide gamut to an output specific color space such as Rec. 709. All manufacturers do this a bit differently so there is no universal image processing sequence.
For example -
Sony uses a 1D LUT for each red, green, and blue color channel at the end of their processing chain to ensure uniformity from panel to panel.
In regards to Sony Trimaster EL OLED Monitors.
FSI's OLED monitors accomplish screen uniformity in a different way. Instead, at the beginning of the chain there's a panel-specific 1D LUT for white balance and gamma. This is followed by a panel-specific 3D LUT for color gamut. These user accessible LUT's combine adjustments to ensure panel uniformity along with transforms for specific color gamuts.
Sony OLED's utilize 3D LUT's as well but users can't access them beyond basic white balance adjustments. For example, when Rec. 709 color space is selected on a Sony OLED monitor, a 3D LUT is being used to transform the native gamut to Rec. 709 but only white balance can be adjusted within this selection. The user has no real ability to adjust other aspects of the calibration. The only option for control beyond this is to put the Sony monitor into wide color gamut and then create a custom calibration 3D LUT to be implemented with external hardware. Despite this, Sony OLED's, BVM series in particular, have a strong track record of stable and accurate calibration for different color spaces achieved through white balance adjustment.
FSI OLED monitors can be calibrated with white balance adjustment but also have a lot of other options including user access to the various LUT's used in the signal chain as well as directly loadable custom calibration 3D LUT's.
For simplicity's sake, this article uses Rec. 709 for the example as it's the most widely used and well understood color space. DCI-P3 would not be as universal an example because it uses the different white point of x 0.314, y .351. Thus a monitor calibrated to DCI-P3 would not be useful for monitoring HDTV images and vice versa.
CIE 1931 chromacity diagram showing Rec. 709 Gamut
So what do we know about monitor calibration?
We know that a monitor is deemed calibrated when it's reproducing a neutral, chroma-free grayscale and accurate colors. This is done by ensuring the monitor's white point and RGB primaries are the same as specified within the color gamut (aka color space) in which we're working. A gamut is represented by the CIE 1931 chromacity diagram (above graphic), a two dimensional chart describing a wide color space divided into coordinates. As in the example above, the limits of Rec. 709's gamut are defined by the triangle; any color outside this cannot be reproduced. The three corners of the triangle define the RGB primaries, that is where 100% Red, 100% Green, and 100% Blue reside. The Red primary is at x .64 y .33, Green Primary at x .30 y .60, and Blue Primary at x .15 y .06. It also defines the 100% white point of this gamut at the coordinates x 0.3127, y 0.3290 (shorthand x .313, y .329). If our monitor is measuring a 100% white test signal at these correct coordinates, then it is at least properly white balanced.
But what if it isn't? What if we white balance a monitor and it's not quite hitting the targets. Or maybe the Red, Green, or Blue primaries aren't hitting where they should. The only real solution is to create a custom calibration 3D LUT that specifically addresses these issues. A white balance adjustment is great place to start but in practice may not be able to address specific problems you may discover.
It's very important to note that despite Rec. 709's white point coordinates of x .313 y .329, it's now recommended that for OLED Judd Modified Color Matching Function (aka Judd-Vos CMF) targets of x .307, y .318 be used instead. This CMF is adjusted from the CIE 1931 Standard Observer and is intended to match an OLED monitor to our standard reference, a D65 targeted CRT monitor. This is a complicated topic largely having to do with the phenomenon of Metamerism Failure which has found that different technologies will look different even if they measure exactly the same. This is a shortcoming of the CIE 1931 color science model to predict accurate metamers between all light sources for all viewers. This is not a specific fault of any probe or display, just color science limitations when dealing with devices that have different spectral distributions.
Because of this, the Judd-Vos targets are now widely viewed as being the "correct" ones for OLED displays despite there being no industry body adopted specification or even a recommendation for this. For the purposes of this article, the Judd-Vos targets of x .307 y .318 will be used.
Here's where we get Flanders specific as these monitors are able to directly load user created calibration LUT's.
FSI CM250 OLED Monitor
There's a lot that happens between plugging a video cable into the monitor and what you ultimately see on it. Every manufacturer handles image processing differently but in the case of the FSI OLED monitors, this is signal chain.
1. Signal Input >
2. User Loaded "DIT LUT", a 3D LUT using FSI's Color Fidelity Engine (CFE2) (optional and intended to be a "Look" LUT) >
3. 1D LUT for gamma and white balance and includes specific calibration for the panel. Can be turned on or off. A custom white balance adjustment affects this 1D LUT. >
4. Calibration 3D LUT for color gamut. It takes the panel's native, wide gamut to a target color space and includes specific calibration for the panel. Can be turned on or off. Preloaded for Rec. 709, EBU, SMPTE C, DCI-P3, USER 1, USER 2, and USER 3. When the monitor is in Wide Gamut mode, this LUT position is bypassed. Any of the USER positions can hold a custom calibration 3D LUT from CalMAN or LightSpace software which is implemented using FSI's Color Fidelity Engine (CFE2). >
5. Final Display.
Because any of these LUT positions can be turned on or off, a custom calibration can go in one of several positions. The order of operations is thus highly customizable which allows you to do technical transforms and calibrations at a higher degree of precision than simply having to concatenate everything into a single LUT. This flexibility also allows the user to implement "Look" LUTs without having to profile the monitor.
As detailed in my previous articles on white balancing Sony OLED monitors, the FSI's can be adjusted in the exact same way. First, input a 100% white test signal into the monitor and using your probe of choice, adjust its Red, Green, and Blue gains until the probe confirms white is reading at the desired targets. Do the same thing using 20% gray test signal and adjust Bias until the same targets are hit. For Studio levels, our Luminance target for gain is 100 nits. Using the display gamma of 2.4, the Luminance target for Bias is 2.4 nits. These adjustments happen in the monitor's 1D LUT position.
FSI monitors include a 1D LUT to transform the standard CIE 1931 white balance to Judd-Vos Modified CMF targets. At the factory, FSI creates their Calibration 3D LUT using LightSpace software so if you were to put a new monitor into Judd Modified mode and then measure it, it should read at or close to x .307 y .318. If additional calibration is required, FSI recommends you do it with 1931 CIE Color Matching Function turned on and then switch to Judd Modified afterwards. This selection applies the correct Jodd-Vos white balance offset to the custom calibration.
Why would you need to create your own calibration LUT? When is a white balance adjustment not enough? White balance is the primary thing that drifts on a display so it's the most logical place to begin your calibration. You may however discover issues that can't be corrected with it alone. For example, if you were to measure a monitor with a given probe and determine that the Red primary was a bit under where it should be for the color space selection, you would have no real way to adjust it. With access to a 3D LUT based calibration, you can adjust where that point lies based on your reference instrument. Similarly even if your probe comes up with a perfect measurement, you may find these primary coordinates drift over time and with no access to a LUT based calibration, there's little that can be done about it.
Currently, these LUT's are made with third party software, either LightSpace CMS or SpectraCal CalMAN. These can be purchased bundled with the monitor along with different probe options directly from FSI.
The process for creating calibration 3D LUT's is specific to the software / hardware you're using. This article is not intended to be a LightSpace or CalMAN tutorial as that information can be found on those company's websites. As an overview, the process is done by putting the monitor into its native, wide gamut, profiling it using the software, exporting the resulting LUT from the program, and then finally loading it into one of the monitor's 3D LUT positions.
FSI monitors can also be automatically aligned using the Minolta CA-310 Colorimeter. No computer is required, just plug the probe into the monitor to automatically adjust the 1D LUT controlling gamma and white balance.
With the nearly instant pixel response time of OLED panels, the longer the interval between frames, the more they flicker. This is inherent to this display technology and causes lower frame rates such 23.98 to strobe and pulse exactly as they would on a CRT monitor. FSI OLED's have something in common with the Sony PVM series in that they require an additional step in image processing to minimize this issue. The Flicker Free mode on the FSI is a double pulse method, which is not quite the same as simply doubling the frame rate but is close to how the PVM handles this process. The problem of low level clipping and luminance shift when using Flicker Free mode has been resolved through firmware on FSI monitors as it has on Sony's new PVMA250 and A170 monitors. Sony’s BVM series OLED’s are driven with more sophisticated electronics that eliminate this problem altogether by displaying 23.98 at 72 Hz. Virtually all of the motion imaging related and signal delay problems seen on the PVM and FSI monitors are eliminated by this process. Another advantage of better hardware is less calibration drift and more stability over long periods of time.
It is recommended that all OLED monitors be measured and if need be, realigned at least every 6 months.
RELATED ARTICLES:
© 2021 Bennett Cain / All Rights Reserved / Copyright & Disclaimer
