Film-Tech Forum ARCHIVE: Formula for video lumens

This is topic Formula for video lumens in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=000106

Posted by Richard May (Member # 2627) on 05-04-2006, 08:57 AM:

What would the formula be to determine lumens for indoor video presentation? Would it be the same as 35MM? The formula I mean. [Big Grin]

Posted by John Pytlak (Member # 331) on 05-04-2006, 09:06 AM:

With a matte white (gain=1) screen, to get 12 footlamberts on the screen, you need about 12 lumens per square foot of image area. But take "lumens output" ratings with a grain of salt.

Posted by Richard May (Member # 2627) on 05-04-2006, 09:36 AM:

Thanks John. So in general, is there a way to find out how many lumens you would need using an XGA LCD projector?

Posted by John Pytlak (Member # 331) on 05-04-2006, 09:49 AM:

If you want a 10 x 8 foot image to have 12 footlamberts on a matte white screen, you need at least 10 x 8 x 12 = 960 lumens. But again, the "lumens" are not aways as stated, and decrease as the lamp ages.

Posted by Barry Floyd (Member # 385) on 05-05-2006, 11:09 AM:

So I would need 18,000 lumens to light a 25x60 screen? How to lumens convert and / or compare to a wattage comparison?

Example: 18,000 lumens = ??? watts?

Posted by John Pytlak (Member # 331) on 05-05-2006, 11:32 AM:

quote: Barry Floyd
So I would need 18,000 lumens to light a 25x60 screen? How to lumens convert and / or compare to a wattage comparison?

Example: 18,000 lumens = ??? watts?

Yes, 18,000 lumens will light a 25x60 foot gain=1 matte white screen to 12 footlamberts.

The relationship between lumens and watts is not quite that simple. Lots depends on the efficiency of the optical SYSTEM: reflector and its alignment, heat filters, shutter, aperture size, lens, port glass, etc. For example, the f/number of the lamphouse and the lens are critical, whether you are using a two or three blade shutter, and the angle of the shutter opening, do the lenses and port glass have anti-reflection coatings to improve efficiency, etc. For digital projection, great strides have been made from the first models of a few years ago, such that luminous efficiencies are finally approaching that of film projectors.

Remember the old "rule of thumb" regarding lamp wattage for a typical 35mm setup: about 5 watts per square foot of "scope" image area to light a matte white (gain=1) screen to 16 footlamberts in the center. For a 25 x 60 foot screen, you are in 6000 to 7000 watt territory.

Posted by Michael Schaffer (Member # 1204) on 05-05-2006, 07:38 PM:

If you want 15 fL, does the number in the formula simply change to 15? The other way around, if you divide the (theoretical) lumen by square feet of the screen area, do you get the (theoretical) fL reading you should expect?

Posted by Steve Guttag (Member # 268) on 05-05-2006, 09:34 PM:

The current target is 14fL for Video, BTW.

Posted by Frank Angel (Member # 248) on 05-05-2006, 11:48 PM:

How come? I thought the reason they chose 16Fl for film was because any brighter and flicker becomes obvious and annoying. There was some testing done a few years ago with brighter images, I seem to recall up to 22Fl, with good results -- the observers thought a brighter picture looked better, except I don't remember how they said they dealt with the excessive flicker that would cause.

So since video has no shutter flicker, you would think they could push the screen brightness as high as they could go, limited only by the amount of brightness the video projectors can produce without setting the nano-mirror modules ablaze. Why did they settle on a figure lower than that of the film specification?

Posted by Steve Guttag (Member # 268) on 05-06-2006, 08:57 AM:

Orignally they went for 12fL since it was compared to a film projector with exposed film running in the gate...that is the brightest scene on a DLP running 12fL equalled a film projector running the same scene if the film projector projected open-gate white light at 16fL.

To my eyes they didn't look equal with the film projector having the edge. I think they moved a bit too quickly to come up with a single measurement that would explain it. How does one's eye integrate the 48-flashes as compared to a meter? Does your eye see the film projector as two 32fL flashes where as the light meter will average it to 16fL? I'm not convinced the eye sees it as 16fL but a bit higher.

So, in a sense when the target was raised to 14fL, I think in perception, the two images started to look more identical. I would consider 12fL to be a minimum spec when the lamp is at the end of its life.

Another limitation in raising the video white light level is that the black level rises too so as technology improved, the ability to put more light out also improved without negative consequences. People really should realize these absurd contrast ratio specs video people put out are just that absurd...there is no 2000:1 video projector out there. Sure, they may be able to put up a black screen and then a white screen in a controlled environment and claim that...try putting up a checkerboard patter that is roughly 50% black and 50% white and read it...it will be down in 100s (and not high 100s either).

If you have the black levels to do it, raising ones white level is the only way to increase contrast since it sets one of the numbers used in the ratio.

Posted by Michael Schaffer (Member # 1204) on 05-06-2006, 11:06 AM:

I don't think we see two 32fL flashes because of the inertia of the visual systems we have. I don't know what the "official" reaction speed of the eye is, I think it is actually a little faster than 48Hz - otherwise we wouldn't see any flicker at that freq -, but I think it is actually light level dependant, among other factors. I guess at those levels it's somewhere around maybe 60Hz because you really don't see much if any flicker on 3-blade shutter projectors at 72Hz.

Posted by Mark Gulbrandsen (Member # 72) on 05-06-2006, 11:31 AM:

quote: Michael Schaffer
I don't think we see two 32fL flashes because of the inertia of the visual systems we have. I don't know what the "official" reaction speed of the eye is, I think it is actually a little faster than 48Hz - otherwise we wouldn't see any flicker at that freq -, but I think it is actually light level dependant, among other factors.

You're right ! You might see if anyone is still around at Showscan. They did psychological studies regarding all that to help them determine what frame rate/light levels were required for the most realistic looking images. I'd say they were dam successful at achieving that!

Mark

Posted by Steve Guttag (Member # 268) on 05-06-2006, 12:04 PM:

I think it is much more complex than that. 3-wing shutters even give the appearance of more resolution and contrast when in fact the film is less stable with 3-wing shutters.

My perception of brightness changes with screen size. I can accept lower light levels on very large screens yet find low light levels very objectionable on smaller screens.

Just as with Mr. lines/mm I don't think a single number is sufficient to qualify the appearance...it is but one piece to the puzzle. I think more study really should be done one how people perceive projected images. A big objection I have with the current digital images versus film based images is that digital images always appear very lacking in depth...film based has an inherent depth to it. Lines/mm alone nor illumination explains that.

Posted by Mark Gulbrandsen (Member # 72) on 05-06-2006, 12:21 PM:

quote: Steve Guttag
I think it is much more complex than that. 3-wing shutters even give the appearance of more resolution and contrast when in fact the film is less stable with 3-wing shutters.

All the high speed pull down movements I've worked with over the years give a similar effect to that. An apparent increase in sharpness not to mention the oodles of extra light. Its very apparent in the ending titles in a film.

Mark

Posted by Tim Reed (Member # 137) on 05-06-2006, 10:25 PM:

quote: John Pytlak
With a matte white (gain=1) screen, to get 12 footlamberts on the screen, you need about 12 lumens per square foot of image area.

Measured at what distance from the screen? Footlamberts are not a measure of intensity on the screen, but rather of light reflected from the screen. By definition, it involves distance from the reflecting surface (or am I missing something). Without factoring in distance, I don't think a formula can be used to calculate reflected light readings.

Posted by John Pytlak (Member # 331) on 05-06-2006, 10:28 PM:

No, footlamberts is a measure of the light reflected from a surface (like a screen). It is measured with a "spot" luminance meter, as specified by standard SMPTE 196M.

You may be thinking of how distance affects the light from a "point" light source, with intensity (e.g., footcandles) following the inverse square law. With focused light sources, the inverse square law does not apply.

Posted by Tim Reed (Member # 137) on 05-06-2006, 10:33 PM:

Yes, but when the light is reflected from the screen, it is no longer focused. It is incident.

I know if you take readings at different distances, you will get different values of fL. The specification states (or, at least, it used to in my day) that readings be taken from the center of the viewing area.

Posted by John Pytlak (Member # 331) on 05-06-2006, 10:37 PM:

The reason you get different values has more to do with the uneven gain distribution of some screens, rather than distance. Hence SMPTE 196M specifies measurement location.

Measured from the same angle, you should get the same "footlambert" reading regardless of distance, if you are measuring the same area of the screen.

Posted by Tim Reed (Member # 137) on 05-06-2006, 10:47 PM:

Well, this is a new one on me. Where can I find more info or a dissertation? Obviously, I have missed something somewhere in my early education.

Btw, is that in the old National carbon books?

Posted by Steve Guttag (Member # 268) on 05-07-2006, 10:24 AM:

The key here is the same area of the screen. The light meter is supposed to read the same area of the screen but in fact will read more or less depending on the distance since 1-degree, is 1-degree...the further away you go, the more screen you are going to read. As the screen surface is almost never matte-white in theatres (based on actually measuring the screen, not on the manufacturer's names) explains why the measured location can indeed affect measured reflected light.

About the only screen manufacturer that I've measure that hits their screen gain is Stewart which is 4-5 times the cost of conventional screens from Hurley, MDI, Technikote or Harkness. I remember measuring a "matte-screen" at 1.2 gain...not very matte to me. The manufacturer explained it was a "high-gain matte white"...which is like saying it is a bumpy smooth surface. In truth, everyone wants more light and by having a brighter screen they were going to get more sales.

Posted by Tim Reed (Member # 137) on 05-12-2006, 04:28 AM:

quote: Steve Guttag
the further away you go, the more screen you are going to read

I was mulling this over as I was out on the truck this week. Thanks, Steve.

John, any comments to add? Not that I can't take your word for it, but I guess I would like to know more of where this is "written", if you will (i.e., technical literature). [Smile]

Most of what I learned in this area as a teen was from the aforementioned texts. Maybe I have mis-interpreted other things.

Posted by John Pytlak (Member # 331) on 05-12-2006, 07:53 AM:

Some links to light measurement tutorials:

Electro-Optical Tutorial

quote:
The diagram gives just the candela values emitted from the lamp. The designer needs to translate this into light energy falling on an object at any distance from the lamp. The energy density striking an object is given in lumens per square meter, generally known as lux.

International Light

quote:
Online Inverse Square Law Calculator
Use this calculator to find out how an irradiance/illuminance measurement taken at a particular distance from a source will change as the distance to the source increases or decreases.

Oldenburg Tutorial

quote:
Quantity Measures

Luminous flux is commonly called light output and is measured in lumens (lm).
Illuminance is called light level and is measured in footcandles (fc).
Luminance is referred to as brightness and is measured in footlamberts (fL) or candelas/m2 (cd/m2).

Posted by Tim Reed (Member # 137) on 05-12-2006, 08:47 AM:

Jeepers, you're good! I was hoping you'd come back with something like that.

Thanks, John! I appreciate your time. [Smile]