How Important is Learning the Science of Light

[ahalpert]

Just coming back to kino flos for a moment as a possible reason people may have gotten more used to shooting daylight balance in studio, I found a thread about 3200 vs 5500 bulbs where people seem to agree that the 3200 were crap: https://cinematography.com/index.php?/forums/topic/64378-kino-flo-colour-innaccuracy/

 

[CharlesPapert]

I don't remember having issues with 3200 Kinos nor them being avoided on many shoots. The first time I can remember seeing sets lit for 5600 when there was no matching the existing sources required was with the RED One. Reading that thread did remind me of the magenta cast with the Diva style lights. I had a Gyoury light kit (let's see who else here remembers that!) with the double-sided Diva style tubes and it had the same issues. Usually had them wrapped in 1/8" green to combat that.

 

[Run&Gun]

Yep, the Diva’s and BarFly’s, which used much smaller tubes than the larger fixtures, suffered from a noticeable magenta cast. I don’t miss dragging mine around. The two light Diva 400 kit was a little larger than a full-sized Thermodyne horizontal camera shipping case. And weighed even more. Now I can carry two foldable 2x2 full color LED fixtures in a milk crate. I can also fit Aputure 600 series kits in a milk crate. Couldn’t do that with my 800 HMI’s.

 

[paulears]

There is light for illumination, and there is light for art. Physics is always handy and helps you know when to ignore it!

 

[morgan_moore]

To me it is scientifically obvious that daylight, tungsten and halogen make for nice skin tones.. which rely of reasonable reproduction of deep reds.

Graph.. science.

I assume you will all be looking at the performance around the 700 mark and dont need this image explaining?

 

[CharlesPapert]

I do need that image explained in the following way: which emitters are being measured for the two LED's represented? I see that it is on the website of a residential/commercial lighting manufacturer. Those don't represent the spectral graphs of a tuned film industry unit. Nor does the fluorescent unit graph represent Kino Flo tubes.

Graphs only go so far when one is talking about a subjective medium that was processed through a camera sensor. If I were to show a blind test of two headshot images, one lit with tungsten and the other with a late model LED unit, shot under identical lighting levels and matched in color temperature, would everyone feel confident they could pick out the tungsten one? What if you picked wrong, would you still sing this song? Hey, that rhymes.

 

[Doug Jensen]

I do need that image explained in the following way: which emitters are being measured for the two LED's represented?

YES!! I was about to say the same thing. What brand and model of lights? You can't just lump all LEDs together and make a blanket statement about them.

 

[morgan_moore]

Of course. I feel this is an old image and tech moves on... but AFAIK the deep reds are still poor on many led fixtures.

Any way the question is 'is science releavant' to our work. You chaps have looked at these graphs, understood them and questioned them.. which all demonstrates basic scientific skill

Using the scientific method of googling it doesnt look good for the aperture 1200 fixtures.

Maybe the $$S led lights are better

 

[CharlesPapert]

I'll double down on this: this science is maths and thus incontestable on paper, sure, but the ultimate result of this is human perception and thus not solely scientific. What counts is what the image looks like to our eyes.

HMI's are imperfect spectrum, as are fluorescents. Yet we shot with them for years and no-one seemed to be able to pick out those scenes vs tungsten. I watched Roger D. light a shot in a hallway using a bank of Image 80's. They were good enough for him! In the past 10 years people have pronounced many films gorgeous even as they used Skypanels as keys, which are by today's standards measurably inferior to current units.

Sam, if you were willing to take that blind test I laid out to you and you picked the LED light over the tungsten, would you be willing to modify this belief system, or would you still adhere to it?

 

[Run&Gun]

I'll double down on this: this science is maths and thus incontestable on paper, sure, but the ultimate result of this is human perception and thus not solely scientific. What counts is what the image looks like to our eyes.

HMI's are imperfect spectrum, as are fluorescents. Yet we shot with them for years and no-one seemed to be able to pick out those scenes vs tungsten. I watched Roger D. light a shot in a hallway using a bank of Image 80's. They were good enough for him! In the past 10 years people have pronounced many films gorgeous even as they used Skypanels as keys, which are by today's standards measurably inferior to current units.

Sam, if you were willing to take that blind test I laid out to you and you picked the LED light over the tungsten, would you be willing to modify this belief system, or would you still adhere to it?

I’d slightly modify that and say it’s how they look to the camera. And then how that image looks to our eye. I’ve seen many LED’s that looked off to the naked eye, but fine on camera. Of course, in a perfect world, they look indistinguishable from the real thing to both our eyes and cameras.

 

[ahalpert]

I'll double down on this: this science is maths and thus incontestable on paper, sure, but the ultimate result of this is human perception and thus not solely scientific. What counts is what the image looks like to our eyes.

HMI's are imperfect spectrum, as are fluorescents. Yet we shot with them for years and no-one seemed to be able to pick out those scenes vs tungsten. I watched Roger D. light a shot in a hallway using a bank of Image 80's. They were good enough for him! In the past 10 years people have pronounced many films gorgeous even as they used Skypanels as keys, which are by today's standards measurably inferior to current units.

Sam, if you were willing to take that blind test I laid out to you and you picked the LED light over the tungsten, would you be willing to modify this belief system, or would you still adhere to it?

Sure, but if there's something superior about a fixture to the eye, the reason must be discoverable somewhere in the data comparison, even if it's somewhere we haven't thought to look (such as the relative measure of two different hues or the blues minus reds or something unusual we haven't thought of).

And lastly I'd say that if Sam picks the LED over the tungsten, there's nothing wrong with the tungsten but there's something wrong with his eyes. (OK, this last part is a joke.)

 

[morgan_moore]

Sorry I cant double quote..

CharlesPapert said:


I'll double down on this: this science is maths and thus incontestable on paper, sure, but the ultimate result of this is human perception and thus not solely scientific. What counts is what the image looks like to our eyes.

Yes. if you want the very best looking human skin tones, use tungsten. I make that clear in all my interview lighting workshops. But if you're willing to give up just a little bit of image quality, you get a ton of advantages from using high-quality dimmable, battery-operated, daylight-balanced LED lights. The trade off is well worth it to me.

Doug said that 'the best looking skin' is tungsten as if it were some kind of mystery sorcery or art.

--

I suggested that the reason that this is the case is clearly visible on scientific graphs the fully lit spectrum at the deep red end.

 

[morgan_moore]

A side note.. Im gonna be shooting some shots for a butcher soon. im aware that making meat attractive is not so easy

Steak, Mince whateever.

Firstly their own shots are really bad right now - hence the need for 'professional' shots. I know the butcher counter lights are very poor.

Im 'mildly stressed' at making the slabs of meat look attractive and thinking on wonder If I need to consider tungsten/32oo in camera.

I have my 800s in a box and a 350 and my dedos.

This test I happened to swipe a dedo as a hard backlight in a go for a hard shadow. M test is with cheap meat. The shoot will be with premium product.

 

[ahalpert]

the patties in the photo on the right-hand side look pretty good to me

 

[morgan_moore]

Thamk you. That is the thought of my “survey”

'Oddly it is a F11, which brings some lighting challenges and is interesting to one used to delivering 'cinema style bokeh'

F11 needs a clean sensor. Not my high point.

 

[morgan_moore]

Another side note. Unlike the folk here that 'grew up' on tungsten my career started in still using daylight temp 'strobe' flashes. Probably why im so happy with working at 5500k in camera.

I thought I would google the colour fidelity of strobe flashes. The junkshop clunker vivitar 285 which cost $20 and was a staple for me from 1993-2010 maybe.

It has good fidelity in the skin zone.

Im thinking 'project meat' should be lit with a base of my strobes maybe with a morning streak of (tungsten) Dedo

I have battery led Dedo too - known on set as 'the electric Dedo' (a phrase a minimal audience may find funny)

 

[ahalpert]

F11 needs a clean sensor. Not my high point.

Gotta stop packing your peanut butter sandwiches in your camera bag!

 

[ahalpert]

HMI's are imperfect spectrum, as are fluorescents. Yet we shot with them for years and no-one seemed to be able to pick out those scenes vs tungsten. I

I don't think this is entirely true. Maybe in terms of the total spectrum it is. But I'm watching the Sopranos and I notice magenta highlights on CUs. I'm guessing from a kino. And I remember watching Liar, Liar and thinking there were hints of uncorrected green in the courtroom scenes, I'm guessing from HMIs.

I just think there are far more important aspects of lighting design and beauty than if the color quality is perfect.

Conversely, there's a really nice presence and texture and color/contrast curve to the film stocks they used on Sopranos and a lot of digital stuff is kind of antiseptic by comparison. But by and large we won't be going back to film.

 

[morgan_moore]

Gotta stop packing your peanut butter sandwiches in your camera bag!

Did i tell you the one about the chef who gave me a chunk of roast venison.

I was at the Billingham bag stand at a show trying to persuade them to launch an SMM bag...

 

[Run&Gun]

I don't think this is entirely true. Maybe in terms of the total spectrum it is. But I'm watching the Sopranos and I notice magenta highlights on CUs. I'm guessing from a kino. And I remember watching Liar, Liar and thinking there were hints of uncorrected green in the courtroom scenes, I'm guessing from HMIs.

I just think there are far more important aspects of lighting design and beauty than if the color quality is perfect.

Conversely, there's a really nice presence and texture and color/contrast curve to the film stocks they used on Sopranos and a lot of digital stuff is kind of antiseptic by comparison. But by and large we won't be going back to film.

I think a lot of it was we just became accustomed to aspects of the "imperfect" light sources and accepted it and tuned it out. But also, they were not consistent. Pretty much tungsten is tungsten, but HMI's shifted noticeably as the lamps aged and flo's were subject to inconsistency in the tube batches(my presumption is the phosphorus coatings), just like LED's can look vastly different between manufacturers. So you could shoot something with one or both lighting technologies and everything looked great or it could look not so hot, depending on the condition of the lamps. The first set of tubes for my Kino's looked great. Appeared very clean, on cam. But I had replacement tubes(actual Kino branded tubes) that looked like complete garbage(so, so magenta). AND they had to be run in, before they hit the optimal light that they could produce. Again, don't miss Kino's one tiny bit, but HMI's, especially the big stuff, isn't going away. The bigger and more powerful you go, most of the LED advantages go away vs. HMI's, at least currently.