I’ve published a technical review of Fuji’s recently re-released Neopan 100 Acros II black and white film.
Do you shoot video with a Canon EOS DSLR or Canon EOS mirrorless (RF or EF-M) camera? Do you primarily shoot footage using a picture style commonly known as Prolost Flat? Do you edit your video in Adobe Premiere Pro? Have you struggled to get your footage to look reasonably good or behave correctly in Premiere’s Lumetri color panel?
I have something for you. It’s free. You’re welcome. Download it here, then read about the what’s, how’s, and why’s below.
Video input LUTS that actually work and are accurate
I’m actually a little amazed nobody has seemed to do this before. Of course, everybody has their own way of color grading Canon EOS video footage, but there is much confusion on the internet about how to deal with the video footage in modern NLE video editing suites for one very simple reason: Canon EOS DSLR video footage does not conform to Rec.709.
Yes, it has the same color primaries as Rec.709, but that is it.
There is nothing else about Canon EOS DSLR video footage that looks even remotely close to Rec.709. If anything, it’s awfully close to Canon LOG believe it or not, assuming of course that you’re using the Prolost Flat picture style.
You see, what is going on is Canon EOS DSLR (including EF-M and newer RF mirrorless) cameras actually record video with an OETF of sRGB. Kind of. Sort of. Actually, not really.
It records video using Canon’s in-camera picture styles. These picture styles are designed to render a pleasing jpeg image if viewed with an sRGB monitor. The reality of the matter is the camera is capturing over 12 stops of dynamic range (on newer Canon cameras) and jamming that into something that looks good in sRGB.
This by itself isn’t so much of an issue, the issue is:
- There is more than one picture style
- You can edit the picture styles
- If you make the Prolost Flat picture style you have a lot more dynamic range than Rec.709 can display.
- It’s not Rec.709
So what to do?
This is actually pretty straightforward. Adobe Premiere Pro (and Lumetri by extension) internally are very strict Rec.709 video platforms. They are Rec.709 video through and through, at least as of early 2020. You can’t change it, configure it to something else, or do anything about it. If you want them to work correctly with your video footage, it needs to be Rec.709 video footage.
Plain and Simple.
So, do a little testing, characterize the dynamic range of Prolost Flat and work out what actual RGB video code values map to scene referred linear light, then takes those code values and map them to appropriate Rec.709 values in a tidy little LUT .cube file.
What do you get?
The zip file contains two .cube files. They are both technical input LUTS designed to be used as input LUTS in the Lumetri Color panel in Adobe Premiere Pro. They are named and labeled to make them easy to tell apart.
The first LUT is a full range LUT. I recommend this LUT to be used most of the time because it linearizes the full range of what the camera can capture in Prolost Flat so that the rest of the Lumetri Color controls after it actually work correctly, and you therefore have the most flexibility in dealing with your video footage. This LUT makes Rec.709 video code values all the way up to 120 IRE (because that is what the camera is actually capturing in dynamic range) on the top end and does not dip below 0 IRE on the bottom end. How to actually deal with this much dynamic range during the video color grading process is a subject for another blog post, however, this LUT gives you as much in-camera video capture and post processing flexibility as what is afforded by using Prolost Flat as a picture style. You can expose in-camera however you want, confidently knowing that the Lumetri Color controls will work the way you expect them to once you apply the input LUT.
The second LUT is almost exactly the same as the first LUT, but has one change. It keeps the video code values in the Rec.709 broadcast safe legal range. If you use this LUT, you should take great care to expose correctly in-camera. This LUT introduces a knee in the highlights to roll off the upper highlights and specular highlights so that no matter what you do in-camera, you will never exceed Rec.709 broadcast levels in Lumetri after this LUT is applied. If you’re shooting in a studio with very controlled lighting and you’re not really going to get nutty with the color grade in post processing, this LUT is a very simple and fast way to make Rec.709 compliant footage that looks really good.
Using these LUTS may not look very good unless you’re actually looking at your video footage on a Rec.709 display. If you look at it on an sRGB display, it won’t look the same. Again, this is for making in-camera footage Rec.709 compliant so that Adobe Premiere Pro and Lumetri Color will work the way they’re supposed to. You still have to conform your video footage to whatever your output is, be it Youtube, Netflix, Vimeo, or actual network broadcast TV.
These LUTS are standard .cube files, and therefore will probably work just fine with other color grading suites like DaVinci Resolve to conform your Canon EOS video footage into something that those suites can then work with as well. I don’t use those other suites, so do this at your own risk if you do use something other than Adobe Premiere Pro.
These LUTS make the assumption that you’ve actually correctly set your Canon EOS DSLR or Mirrorless (RF or EF-M) camera up with a correct and proper Prolost Flat picture style. While other camera manufacturers (Nikon, Sony, Fuji, Panasonic, etc.) have in-camera neutral picture styles that you can modify in a similar way to Prolost Flat, they are not the same and should not be used with these LUTS. Those cameras should have their own LUTS made for those specific camera picture styles.
If there is enough interest, I can probably make a few posts or videos with how to deal with the post-LUT footage in Lumetri along with some best practices that I’ve come up with over the years.
OK, we’re going to get nerdy here. There are many different kinds of artificial lights, and then there is the light from our sun, commonly referred to daylight. If you want to read up on light color temperature, Wikipedia has a great article.
For artificial lights, they generally fall into a couple of different classes: warm, neutral, and cool.
Warm lights are lights that are tungsten, incandescent, candles, etc. Neutral lights are what you generally see in most retail stores, and nowadays usually consists of fluorescent or LED lighting. Neutral light usually sits between 3800K and 5000K. Cool lights are generally any light that has a color temperature above 5000K.
For photography, if you’re shooting ambient light, it’s going to be all over the place. If you’re shooting with continuous “Hot Lights” then it’s generally on the warmer side (2800-3200K), and if you’re shooting with the newer LEDs, the good ones are tunable, but otherwise they’ll be on the cooler side, and if you’re shooting with a speed light or studio strobe, it most definitely will be on the cooler side, usually at least 5500K, but sometimes upwards of 6500K depending on the quality and power setting.
This all leads to an interesting question: When shooting people, is there a best color temperature to use that renders more pleasing skin tones? Searching Google leads to lots of articles on how to light for skin tones, but very little in the way of whether to use warmer or cooler color temps for your lighting. With that being said, generally, pros tend to err towards warmer color temperatures (i.e. Tungsten) because it tends to look better. This is why much of our indoor lighting for houses is generally pretty warm.
I thought I would put this to the test and take some pictures using a studio strobe at it’s native color temperature, which is advertised to be ~5600K, and then that same strobe, but gelled with a CTO gel to get the color temperature down to roughly what a tungsten hot light would be, which is ~3200K, then compare the two and see what it looks like. Note: the color temperature is going to vary a little depending on the power level of the strobe, but it should be roughly correct.
So lets start with a standard bare strobe with no color correction:
Here’s the color checker chart:
So we can see we’re correctly white balanced for the light in our software, now lets see what yours truly looks like under the exact same light with the exact same white balance in software:
I look like a caucasian guy. If anything, my skin is rendering a bit on the flat or gray side partially because I’m one of those awful people that has a neutral skin undertone. I am however cursed with a pretty bad complexion, so there are parts of my face that are rendering as blotches of color. I guess this is what happens when you’ve been a life long sufferer of super severe eczema.
Let’s look at the same light, but gelled to 3200K:
Again, we can see that we are correctly white balanced in our software. Let’s see what I look like under this same light:
Hmm… that’s interesting.
Let’s look at the two color checker cards side by side. The one on the left is 5600K the one on the right is 3200K:
We can see that even though the the white balance between the two matches, the colors don’t exactly render the same way.
Let’s look at me side by side, again left is 5600K, right is 3200K:
Again, very interesting. Keep in mind some of the color you’re seeing is due to my very unfortunate complexion.
So does the color temperature of your light effect the skin tones in photography? I would say yes. Which one looks better? That’s really a subjective thing more than anything else.
The important thing to take away here is that you should be cognizant of the fact that some people might look better in warmer light and some might look better in cooler light and adjust accordingly.
Till next time.
Simple Film Lab is offering three new Black and White film tonal distribution profiles for black and white film sent in for processing and scanning.
The new tonal profiles are meant to be applicable to black and white film sent in for processing to normal contrast and is meant to provide a look similar to what you can get if you were to do an analog darkroom print using a split grade printing technique to punch in the shadows and blacks with a grade 5 filter and also control the rolloff of the highlights using a grade 0 or 00 filter.
So, let’s take a look. As a frame of reference, first a frame of Ilford HP5 in 120 format exposed (with an incident meter) to ISO 400 and developed to ISO standard 0.615 contrast and scanned in with a matching tone profile to make the straight line part of the characteristic curve land on Gamma 1.0.
Contrast wise, it’s pretty much what I’d expect of a linear black and white digital image, and is basically what you’d expect to see if you just did a straight print at grade 2.5 in a darkroom.
There isn’t anything special about it, and in all honesty, the floating point DNG that you get from Simple Film Lab is insanely malleable and can have a huge amount of dynamic range that you could recover any number of ways in either Adobe ACR or Adobe Lightroom.
So the first profile is called BWSG1, which is short for “Black and White Simple Grade 1”.
This tone curve starts with the standard normal contrast curve of the image above, but includes the equivalent of adding a couple of extra seconds of exposure at grade 5 if this where being printed onto paper in a darkroom.
This is what it looks like:
Just like with adding a couple of seconds at grade 5 in the darkroom, the mid-tones and highlights are largely unaffected, but the blacks and shadows punch down a bit and make for a significantly more punchy looking image while still retaining the same contrast in the rest of the image.
The second profile is called BWSG2, and is a progression of BWSG1. In this tonal distribution profile, we keep the same look of BWSG1, but include the equivalent of a grade 0 rolloff on the extreme highlights. The effect here in this image is very subtle, but in images where there are lights in the frame or very high contrast specular highlights, it provides the same effect of burning in the highlights at the lowest possible contrast so that you retain as much detail in the highlight areas as possible when printing in the dark room.
This is what it looks like:
The difference in this image is admittedly extremely subtle, however, if you look closely, the cloud cover is not the same luminance level as the BWSG1 image because it is just at the bottom part of the grade 0 rolloff. So everything brighter than the cloud cover (i.e. light sources etc) would have a nice gentle rolloff to pure white.
The third profile is called BWSG3 and is a progression of BWSG2. In this tone profile, we start with BWSG2, and add the equivalent of even more exposure time at grade 5 so the increased contrast amount moves a bit up the tonal scale and starts to effect what would be zones 4 and 5 if this were printed on paper in a dark room. Likewise, on the highlight side, we move the grade 0 exposure down the tonal scale ever so slightly and introduce an equivalent of a grade 00 exposure at the extreme top of the characteristic curve.
This is what it looks like:
Again, the mid-tones are mostly intact, but the image is rendered in a significantly more punchy manner while still retaining very legible details in the important parts of the image.
It should be noted that if you want maximum post processing flexibility, these three tone curves may not be for you, as they bake in a specific look that is meant to look similar to what you would get with a well rendered analog darkroom print.
These tonal distribution profiles also are not as good of a fit for films that don’t have particularly flat characteristic curves. For example, JCH StreetPan has a very S-Curve shaped characteristic curve, and so these profiles won’t necessarily render the same way.
These tonal characteristic curves are also meant for films that are developed to “normal contrast” (i.e. ISO standard 0.615 contrast), and therefore won’t have the same visual effect with films that are normally push processed like Kodak TMAX P3200 or Ilford Delta 3200.
Since these are specifically developed looks, they are not available for free because every frame needs to be individually looked at and adjusted to provide the best rendering into the selected tonal profile, just like with doing an analog dark room print, where you have to adjust for every frame to best fit it onto the paper while still providing nice punchy blacks and lovely highlight rolloff to white.
So with that being said, I also want to provide affordable pricing, so if you’d like your black and white film that you send in for processing and scanning to have this look applied, it’s not a huge additional expense, so the pricing is an additional $5 per roll, and can be added to the straight scan through the deluxe scan options. I’ll be updating the downloadable order form to reflect the new profiles in the coming days, but if you want these looks sooner than that, just send me an email and request it and we’ll work out payment at that time.
In the future I’ll be adding a handful of new tone profiles that will provide a few more variations on what has been released so far, and to better address film developed to other contrast levels.
Till next time!
For this post, let’s talk about taking advantage of white balance to enhance your scene.
I’ll start with a mantra: White Light is a Lie. I’m a long time reader of Strobist, and if there’s one thing I’ve learned over the years it’s that white light has its uses, but largely is a subjective creative choice that can be manipulated in the camera, or manipulated with colored gels if shooting with a speed light or studio strobe.
Let’s take the image above. I was recently photographing an event and saw a group of youngsters outside playing hacky sack. It was evening and the sun had already dipped down under the horizon and they were trying to get the last game or two in for the remaining minutes that the light was still there enough to see by. It had been mostly overcast throughout the day and at this point, the sky had light cloud cover.
I took a quick step outside and grabbed a number of frames, the one above being the best one.
Once I got the images pulled into the computer, I initially did a white balance that looked like this:
You can see that it’s evening time, you can see the warm light from inside the building spilling onto the concrete in the foreground, and spilling onto the skin of the people in the frame.
Looking at the Adobe ACR data, the white balance was 8300 Kelvin. Very blue, very cool ambient light. When I did the initial WB, I pulled the reading off the concrete that didn’t have the warm light spilling on it.
On its own, white balanced like this, this could be a totally serviceable image, however, we can enhance the mood and feeling a little bit by adjusting the WB in ACR to something that is a bit more evocative of outdoors during the late evening.
The first modification, I set the WB to Daylight in ACR:
That’s better, and what it’d probably look like if you were shooting daylight balanced film, but not what I was envisioning. The Daylight WB setting in ACR is 5500K.
I then went to Tungsten WB in ACR:
Whoa. That feels like a bit too much blue. The Tungsten WB setting in ACR is 2850K. We need something between daylight WB and Tungsten WB in ACR. Good thing I shot raw.
For those of you who spend any amount of time shooting with flash indoors, you’ll know that many times, you have to CTO your flash to get it match the lighting indoors if you are blending the two. I primarily use Rosco lighting gels, and their range of CTO gels (1/8 to full CTO) with flash results in the following white balances in Kelvin: 4900K, 4500K, 3800K, 3200K, and 2900K.
When I shoot interiors that are lit warmly, I often end up with my white balance set at ~3800K. This still allows me to render the interior lighting to be anywhere’s from roughly neutral to a bit on the warm side (always a good thing if people are in your shots), and since a 1/2 CTO on a speed light or studio strobe lands at 3800K, gives me 3/4 and full CTO if I want to warm my flash, 1/2 CTO if I want it neutral, and 1/4 and 1/8th CTO if I want to cool if off relative to the camera WB. Many interior lights nowadays are generally 4000K to 2800K depending on the type of light.
Being that I was just inside shooting right before this image was taken, and actually had my camera WB set to 3800K, I decided to make my WB center 3800K. It’s right between 5500K and 2850K and falls nicely where a 1/2 CTO gelled flash would be if I wanted to do that, and if I wanted to, I could push my flash into CTB gel territory and match the color of the ambient light outside. That resulted in the image at the top of this page.
So what’s the takeaway? If your camera or post processing software supports it, don’t be afraid to explore the color temperature or white balance of the light in your image to enhance its mood and feel. You’d be amazed at the effect that it can have.
For the month of February 2019, Simple Photography Services is running a special. All new customers who book a portrait or headshot session with me will receive a 10% discount on their session when they pay for their session.
This is for new customers only who have never shot with me before and applies to headshots, individual portraits, couples portraits, and the kids/families/groups sessions. The sessions can be either in my studio, or on location.
Above is a recent project I worked on that I thought I’d share how it was lit since many people always ask me how I get my photos to look the way they do.
Aside from the color, this photo was actually lit very simply with two lights. Yep, two lights.
Let’s start with the fill light. It was a Paul Buff X-1600 White Lightning studio strobe behind the camera, pointed at a very large, very white wall with an umbrella reflector, so it was blasting light pretty much everywhere.
I gelled it with my stand-by Rosco Calcolor gels: 135 units of Calcolor Cyan, 75 units of Calcolor Blue. I metered it to f/2.8, so that it painted the entire scene with a super deep blue with a slight push to green.
For my key light, it was also a Paul Buff X-1600 White Lightning studio strobe, however, it was placed camera left, and had a standard reflector and 40 degree grid. I gelled it with 90 units of Rosco Calcolor Yellow, and 45 units of Rosco Calcolor Red. I metered it to f/8.0 and placed the light so that it would split the egg in half, one half a nice warm glow, the other half a deep cold blue.
Other technical data: the background is a Savage Thunder Gray seamless paper, the egg is sitting on a studio stool with a black table cloth.
I kept the Lightroom and Photoshop work to a minimum, set the white balance to 5500K, spot healed a few spots of dust/fibers, etc. Very minor tune ups, what you’re seeing in this picture is pretty much straight out of camera.
So how does this work? How do you figure out what strength gels you need where?
Again, this is pretty straightforward if you’re using a calibrated set of gels, which is why I love the Rosco’s Calcolor gel set so much.
So I wanted the overall lighting to push to yellow-orange on the highlight side, and push to blue-green on the shadow or fill side of the egg, both by equal amounts. Since I metered 3 stops between the two halves, all I had to do was figure out my overall proportions, then scale them. For the key side, I kept my scaling at 1.0, which meant my shadow side needed gels with 3 added stops of strength to match.
So, starting on the fill side, I knew I wanted 60 units of blue, multiplied by 3, makes 180 units of blue. In order to reduce how much magenta crossover happens, you need to put a fair amount of green in with the blue, typically half to three quarters. More than that, and it starts to look more cyan than blue. You can totally do that if you want, but for this image, I wanted more blue than cyan, so I went with 135 units of cyan, and 75 units of blue, which gives me a total of 180 units of blue (cyan = green + blue), and 135 units of green. Taking the multiplier into effect, this means 60 units of blue and 45 units of green.
OK, so on the highlight side: The fill touches everything, so just to get back to white, we need 15 units of yellow, and 45 units of red. But we don’t want white, we want yellow to orange, so on top of that, we need to add an additional 60 units of yellow, bringing our key side total to 75 units of yellow, and 45 units of red. This gives us a more green yellow than orange yellow, so we need to add just a touch more of red, say 15 units. This brings our total to 60 units of red, and 75 units of yellow.
And that is what you see in the image above.