Monday, December 23, 2013

Using Picture Profiles on the Sony FS-700

 

I was recently at a holiday mixer for the San Diego MCA-I group and ran across a Video Gear customer that bought an FS-700, but hasn't had time play with the paint controls on the camera.  So Eric, this one's for you!

When you press the "Picture Profile" button on the camera a menu pops up with the paint controls for the camera.  Anyone serious about using this camera should know this particular menu in detail for best results!  Many of the titles are obscure or just not obvious.  Hopefully I can help you shed some light on how to paint this camera properly.

Black Level

The "black level" is a somewhat obvious control.  Oh sure...right after I claim some of the controls are "obscure" we start with a obvious one!  You can see the extremes of the black level shifts below after I installed the lens cap.  -15 drops the black level down and +15 brings it up.  I still think it's strange that you can drop below 0 IRE on the Samurai Blade luma scope.  I would have thought zero would be zero, but some of the picture profiles (standard, cine1, cine2) assume you want broadcast levels (black = 7.5 IRE).  The Blade luma scope must be looking for black = 7.5 IRE(?)



Gamma

Gamma is the characteristic tone curve of the camera.  Art Adams at Pro Video Coalition has done a more comprehensive study of the FS-700 gamma curves and I recommend that you read their article as well.  For this article I used an uncalibrated Macbeth color checker to generate the six step tone curve examples.  Art's use of the DSC chart in the PVC article is a much more accurate and usable representation of the gamma curves.  I just didn't have a DSC chart available to do this test.  I assumed the third neutral patch from the right was 50 IRE in all of these gamma curve measurements. 

I will typically use Cine4 for most outdoor scenes because it holds highlights and Cine3 for most indoor scenes because it holds shadows.  When in doubt on highlight retention I use Cine4.  Some of the gamma curves also lift the black levels or cap the white level at 100 IRE, which seems silly for a "cine" gamma.










Black Gamma

The black gamma controls the fall of the tone curve into black.  There are three ranges: high, middle, and low.  You can only choose one to control.  A negative level value drops the black level down and a positive level value raises the black level.  Below is an example where I took the cine4 gamma and applied a value of -7 to the "high" range of black gamma.  You can see how the two furthest right neutral patches on the Macbeth chart have dropped in IRE value.  This would typically be referred to as crushing the shadows.



Knee

Analogous to the black gamma setting is the knee setting.  This simply controls the gamma curve in the highlights.  You can set the "point", which is the inflection point on the gamma curve within the highlights from 105% to 75%.  Then the curve is either made more extreme or less extreme through the "slope" setting.  Below I show an example where I set the inflection point to 82.5% and set the slope to +3 to lift the highlights just a little.  Middle gray and the shadows are unaffected.  In retrospect I should have set the value to a more extreme value to make the example more obvious.





Color Mode

Color mode is just the color matrix of the camera.  It assigns the hue and saturation of each color.  To understand the difference here you really need to know how to read a vectorscope.  Each hue is represented by an angle on the circular diagram.  The saturation of the color is represented by the length of the vector from the center of the circle toward the circumference.    You see blue at just less than 0 degrees and red at around 100 degrees.  Lucky for us, the Samurai Blade draws the vectors with the colors they represent.  B=blue, MG=magenta, R=red, YL=yellow, G=green, and CY=cyan.  Normally with a DSC color chart the vector angles would line up with the target boxes for their respective colors.  The Macbeth chart isn't color calibrated, so all I can show you are the differences.

The "Level" setting tells you how far away from "Standard" color mode the respective color mode gets.  So when you have the color mode assigned as "Standard" it doesn't matter what value you assign to "Level".  If you have the color mode assigned to "Still", for example, and have the "Level" set to 8 then the color assignment will fully look like the "Still" colors.  If you have the "Level" set to 1 then the colors will be assigned mostly like the "Standard" color mode.  You can go in between the two color modes to any degree you wish, sort of like an audio mixer, but for color assignment.

From the vectorscope plots you can see how the "Pro" and "Cinema" color modes are somewhat desaturated compared to the rest.  "ITU709" has the most saturation.  Also note the differences in hue angles, particularly the yellow, red, and blue color assignments.  Most of the time I just use ITU709 because I know it provides nice results with mostly accurate colors.  With enough camera test time I might use one of the other color modes.








Color Level

Color Level is nothing more than saturation.  -8 is gray fully de-saturated and +7 is very saturated.  The comparison picture below shows the difference between the three levels of color.




Color Phase

The color phase control rotates the color assignments.  A value of 0 indicates that there is no color rotation .  In the pictures below you can see red/green/blue being rotated counter-clockwise when the color phase is set to -7.  This means that red becomes orange, green becomes more cyan, and yellow becomes more green.  Likewise when you set the color phase to +7 the colors are rotated clockwise and red becomes magenta, green becomes yellow-ish, and yellow becomes much more orange.

Generally speaking I leave this control completely alone.  You can get some psychedelic colors when you mess with this color control and that's not good unless you're making a movie where you want everyone to look like they have jaundice (color phase=+7)!  User beware.




Color Depth

I'm going to punt for now, because I'll be talking about this control later in the article.  In short, this menu has individual controls for R/G/B/C/M/Y.  A negative value makes the color lighter and the positive value makes the color darker.  Generally this control does not change the hue, nor the saturation - only lightness.  This is very useful for matching colors between different cameras, even when the color assignments don't completely match.

WB Shift

This controls the white balance shift.  There are two types of white balance shift available.  LB-CC and R-B.

The LB-CC mode has two controls.  One for Color Temp, so you can shift the white balance toward amber (warmer) or blue (colder).  The other control is for the Green-Magenta balance, which is infinitely useful for imperfect sources like fluorescent bulbs (green) and LEDs (who knows what!).

The R-B mode has two simple controls for red gain (R-gain) and blue gain (B-gain).

When you select the LB-CC mode you can only use the first two controls.  When you have R-B mode selected you can only use the last two controls.  So if you're in R-B mode and try to change the color temp nothing will happen.

 
 
 
 
 


Detail

And finally...the detail setting. "Detail" is nothing more than a fancy way of saying sharpness.  The "Level" can be set from -7 (basically off) to +7 (over sharpened).  The "Level" settings correspond to more detailed settings you can set manually, but Sony doesn't tell you what each "Level" actually means in terms of settings. 

My recommendation is to always turn the in camera sharpening off.  Sharpening is something you do in post to make up for a lens/codec issue.  Over-sharpened footage is a tell-tale sign of poor footage capture and there's nothing you can do in post to remove over sharpening artifacts.  My two cents.  So I typically go into the "Manual Set" control and turn the settings to what you see below.  That way I'm not allowing the camera to mess up my footage before I get it into the edit bay!

Showing the difference on the web wouldn't be a good representation.  I recommend trying this control for yourself to see where you like it.  Some people like to set the level to -4 (Abel Cine profiles come to mind).



Now for an Example

I had two cameras with me today.  The FS-700 and 5D mark II (yes, I'm aware it's obsolete).  My goal was to match the two cameras as closely as possible using a standard picture style on the Canon camera and a plethora of picture profile settings on the FS-700.

First, let's start with the gamma curve and color assignments on the Canon 5D mark II camera.  Like most video shooters I use the neutral picture style and usually set the contrast=-2, saturation=-1, and sharpness=0.  This fits most general purpose video shooting.  It's not color accurate, but it does render skin tones nicely, saturates the colors correctly, and has a nice smooth roll off into highlights and shadows (at least by DSLR's standards).

Sorry about the mess of the luma scope.  I didn't have a good way to isolate the neutral patches.  You can see the patches at about 87, 78, 65, 50, 30, and 10 IRE from left to right.  Also note how far off the green and cyan hue angles are on the vectorscope plot.  This will be important later.




The default picture profile on the FS-700 is somewhat more contrast-y in the highlights than the 5D and the colors are assigned very different with the picture profile turned off. The colors also appear to be over saturated, which isn't good in general.  Notice how the shadows are lifted as compared with the 5D mark II.  The FS-700 tends to see pretty far into the shadows, often 2 stops more than the Canon DSLR camera.



So what I did to match the two cameras is use the following custom picture profile settings.
  • Always start with the tone curve!
    • Set the Gamma to Cine4
    • Set the Black Gamma to High and -7 to crush the shadows a bit more and match the shadow tones on the 5D mark II
    • Set the Knee to Manual at 82.5% and +3 to better match the highlights
  • Then move to the color controls
    • Set Color Mode to Pro with Level 8.  Note that the green and cyan hue angles are just way, way off but everything else matches within reasonable tolerance - most notably red and yellow with bound skin color.
    • Set Color Level +3 to add saturation because the Pro Color Mode is a bit desaturated
    • Ensure Color Phase is 0
    • Set Color Depth G=-3 (lighter) and B=+1 (darker).  Even though green is off in hue, some of that can be visually hidden by matching of the lightness of the color.  It's a compromise.
  • Turn off all in camera sharpening we can
    • Set the Detail to Manual and use the settings I mentioned before.
Finally, I ended up with the "matched" picture profile shown below.  The default FS-700 settings are on the left, the 5D mark II in the center, and the custom FS-700 picture profile on the right.  If you look carefully you'll see that both the green and cyan patches on the "matched" FS-700 picture profile are a bit more green.  This is why I made a custom picture style for the 5D mark II that better matches the ITU709 settings, but most people don't have my custom Canon picture style so this seemed like a more useful example.  As I mentioned before...compromise.  This is what Davinci Resolve is for!



Leave comments if you have questions or call the shop and talk to Michael or Dominique.

Thursday, December 12, 2013

Lighting for high frame rate cameras

Martin came to me with a customer problem earlier this week.  The customer bought the "low cost" Phantom Miro camera to do high frame rate photography.  Cool bit of kit that I hope to play with some time in the near future (*ahem*, Martin...).  The problem is providing enough light for the camera at such fast shutter speeds.  The customer already owns a tungsten 6k light, but even that's not enough.

I'm assuming that the customer bought the new Miro M320S which maxes out at about 1540fps in HD and has a native ISO of 1100 - oddball ISO rating, huh?  Given these specs, which lights does the customer need to buy?  That's actually fairly easy to calculate.  A while back M. David Mullen, ASC published a simple formula which he uses to order up lights by using the foot-candle photometric data of the fixture.  Good thing I published this to my personal blog just as a way to keep track of the formula!

foot-candles (approximately) = 25 x (T-stop)^2 / (ISO * shutter time)

So let's say the customer wants to operate the lens at a typical T/2.8.  We're using cine lenses here; none of that f/# funny business!  I'll assume the camera can operate with a near 360-degree shutter speed, but in truth I have no clue since I don't have any experience with the camera.

FC (approximately) = 25 x (2.8)^2/(1100* (1/1540)) = 267 foot-candles

Let's say the customer wants to use a more comfortable to focus T/4 with a typical 180-degree shutter angle.

FC (approximately) = 25 x (4)^2/(1100* (1/3080)) = 1120 foot-candles

From here it's pretty easy to know what lights to order since most vendors, like Arri, publish photometric data for their fixtures.  The Arri 200W HMI provides approximately enough light to do what the customer needs, maybe even with diffusion involved, however that assumes a short throw distance.  The customer will have to do his own research with photometric data when the light-to-subject distance is better understood.  A Film Gear or Arri 575W HMI might be a safer, more versatile purchase.

The only real variable is how much light loss comes from different types of diffusion and bounce.  That's something for a future blog article.

There is an easier way to do a back-of-the-envelope calculation when you need something quick and dirty.  The rule of thumb is ISO 100, 180-degree shutter angle at 24p, T/2.8 equates to 100 foot-candles of light for proper exposure.  So for instance if you have an Arri Alexa with a native ISO of 800 then you only need 1/8th that amount of light, or 12.5 foot-candles.  If you want to operate the lens at T/4, then you need 200 foot-candles because that's half the amount of light/one stop less making it's way through the lens.  Easy, yes?

Saturday, December 7, 2013

Introduction to the Indie Dolly

Last weekend I had a chance to take out the Indie-Dolly for a roll on a short film.  This very portable dolly comes in a couple cases.  First each set of four three-foot track segments comes in a nylon bag.  In the shop we have two straight track bags and two curved track bags. Each bag of curved track (not shown) covers approximately 13-feet of travel along a 10-foot radius (quarter circle).  The straight and curved track can be mix and matched, as we did on our production shown in the pictures at the bottom of this article.  The dolly itself comes in a medium sized Pelican case (not shown).

 

The track gets unfolded during assembly - something Michael quickly corrected me on during my first assembly attempt!  Notice how the support bars go straight across, whereas they are in an "X" configuration in the carrying bag?  The track is simply joined by a series of plastic pins, however these pins are just for alignment and have a semi-loose fit.  In order to get the track to hold together you use a series of Velcro straps.  I only show one Velcro strap for demonstration purposes, however during normal use each track joint would require this strap.




If you need to fit through a smaller area, like a door way, the track can be moved diagonally.  Note that this shortens the effective travel length, but the dolly can accommodate the track width change to a reasonable extent.


As anyone who has operated a Dolly knows, uneven floors can be a real pain.  The Indie Dolly kit comes with both shims and mid-span floor supports...you know, just in case you have a fully loaded Panavision camera and don't want the track to sag.  In my use, I found that I didn't really need the mid-span supports even with my beastly O'Connor head and Miller legs, as well as a fully loaded FS-700.

 

The dolly itself is pretty easy to assemble, so I thought it was unnecessary to give step by step instructions.  Each roller leg (skateboard wheels and silver metal shaft) inserts into the Dolly base.  You then attach the riser and saddle as well as mount a push arm.  That's about it.  No biggie, right?  In my use, I didn't use the chair nor push arm so all I had was step #1.


Finally, you'll want to mount your tripod to the Indie Dolly.  The dolly rents with two types of tripod leg inserts, but the most common one (shown below) is the Manfrotto spike.  Even my Miller tripod has this exact same leg mounts.  First, remove the boots from the tripod, then mount each leg into the dolly and strap over the rubber tie-downs.  It's a solid mount and I didn't experience any sort of play between the dolly and tripod, even when operating the camera at the tripod head while moving along the track.


Here are a couple behind the scenes production pictures from our film showing both the straight and curved track in action with the Sony FS-700 and Atomos Samurai Blade.  Note that sandbags are a necessary accessory to keep the dolly from rolling away on a slope, like at our location.  Also note that the dolly itself takes up three feet of track, so to move three feet you'll need two joined pieces of track.

 

If you have any questions about using or assembling the dolly call Marc, Michael, or Dominique at the shop. Michael gave me the two minute tour and that was about all I needed to use the dolly.  As far as travel goes, in general the kit should fit in the trunk of an average sized car.  Dollies are great at getting those wonderful cinematic traveling shots, so I recommend trying them at least once.