Film-Tech Forum ARCHIVE: kw and Lumens

This is topic kw and 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=000043

Posted by Philip Borgnes (Member # 2383) on 05-23-2005, 08:54 PM:

Ok here is a basic question that I can't figure out.
What is the relation of kw and lumens?
Is a 7kw = 7000 lumens?

Thanks!
Phil

Posted by John Pytlak (Member # 331) on 05-23-2005, 08:57 PM:

It really depends on the efficiency of the lamphouse optics. Remember, efficiency of any light source is measured in lumens per watt. One lumen per watt is quite inefficient.

Posted by Philip Borgnes (Member # 2383) on 05-23-2005, 09:27 PM:

Thanks John,
So what would an efficient ratio of lumens per watt be?
10? 50? 100?

If a 7kw 35mm projector can illuminate a screen X by Y, what is the size digital projector (in lumens) needed for the same level of illumination?

Posted by Paul Mayer (Member # 355) on 05-24-2005, 01:10 AM:

On a unity-gain screen lumens to foot Lamberts is a one-to-one function. For every fL you want reflected back from the screen you'll need to apply one ANSI lumen per square foot.

[(Screen Width x Screen Height) x Desired foot Lamberts] / Screen Gain = ANSI lumens

Examples:

20x37' unity-gain (1.0) screen (1.85AR)
12fL for max white level (assuming 16fL measured open gate, 12fL is about what you'd get projecting clear film)

[(20'x37')x12fL]/1 = 8880 ANSI lumens

20x48' unity-gain screen (~2.39AR) with 12fL white

[(20'x48')x12fL]/1 = 11,520 ANSI lumens

For anyone that knows, has DC28 come up with a standard for D-Cinema luminance yet? AFAIK 12fL is what many people have been using, though I hear that Sony is using 14fL for their 2K SXRD projector demos.

Posted by Philip Borgnes (Member # 2383) on 05-24-2005, 01:39 AM:

Well Paul that is the formula and logic I have been following too. Take for example the specs on the Christie CP2000H. It is a 2K 3-chip DLP and says that its brightness is

quote:
14fl (at screen) for up to 75ft/23m wide screens.

At 1.85 that would make the height 40ft and 75ft wide = 2960 square feet(bigger than my house). To hit 14fl that would be the equivalent of 40,000 lumens from a 6.0kW CDXL lamp.

One more thing. Is the 7kW in 35mm as efficient as a 7kW in DLP? or 6kW?

Using this logic then is a 7kW = to ~46,000 lumen? and a 4kW ~25,000? At bit more checking show that if a 7kW pumps out ~46,000, that would be 16fl coming off the screen. Isn't that spec. (SMPTE 196M)?

Posted by Michael Schaffer (Member # 1204) on 05-24-2005, 02:04 AM:

For DLP projectors, you only need to reach 12fL because the projector does not have a revolving shutter.
You generally have to use considerably bigger lamp sizes in DLP projectors of the first generation. You would normally need a 6k lamp where normally a 4k would be sufficient for 35mm. What the relation is with the most recent generation of DLP projectors is I can not say as I have not worked with these so much yet.

Posted by Paul Mayer (Member # 355) on 05-24-2005, 02:12 AM:

12fL is what people have been using for video since that's about what a film projector will put out with clear film in the gate, assuming that machine puts out 16fL with no film in the gate (SMPTE 196M).

I'd be willing to bet that that Christie spec assumes a 2.39 AR instead of 1.85. Try running your numbers again using a 'scope image at 75' width and see what kind of lumen numbers you get.

As for determining watts needed once you know lumens, there's so many variables that affect that calculation. Lamphouse alignment and efficiency, filters, lenses, stop-down rings, yada yada. So it's tough to come up with a consistent lumens-to-lamp watts ratio. It'll be different for each equipment combination.

It seems that the early DLP machines did need a lot of light, e.g. 7K lamps when film in the same installation could get by with less. At least that's my impression from the few installs I've personally seen. With the latest machines I really don't know. I'm sure the guys with current install experience will comment in the morning. [Edit: Ah, I see Michael is up at this hour... [Smile]

]

As for me, it's past my bedtime. Gotta finish grading and get my grades turned in tomorrow. [Smile]

Posted by Philip Borgnes (Member # 2383) on 05-24-2005, 02:16 AM:

Thanks Michael, but that doesn't make sense to me.

As I understand it, fl is what is coming off the screen. It doesn't matter what the source was before coming off the screen. 12fl seems a tad low to me. I try for 14-16fl.

Your other comment about the first generation DLPs (I'm not sure what first generation is though...) is less effecient than the film counterpart. So how much bigger lamp would you need to project a DLP 40x75?

Thanks John, it is late for me too but still have some more editing to do...

If you change the height to 32x75 for scope, then for 12fl you would need 30,000 lumens. So in this case ~5000 lumens per 1kw.

What I'm looking for is a general understanding on what the ratios and relationships are between the two units of measure. The projectors I'm using are in the 5000 lumen range with is fine for my small portable cinema, but didn't understand how that light output relates to these 2k projectors and 35mm.

My observation is that the DLPs are becoming more efficient which raises another question. How does DLP efficiency improve? Lens? Colorwheel? Mirrors?

Posted by Michael Schaffer (Member # 1204) on 05-24-2005, 03:30 AM:

I think Paul has an additional good point because light readings of film projectors are taken with no film in the projector.
Even though the measurement is integrated, the revolving shutter has to be taken into account because a film projector image flashes on and off while a DLP image is subjectively much more stable to the human eye.
As Paul also pointed out, there are so many variables that you really can not develop a general formula easily. It really depends on the effectivity of the lamphouse and projector design, and that alone can vary dramtically between different manufacturers and model options.
The NEC iS15 we are currently running on a 38' wide scope screen is happy with just a 3kW lamp. However, we are not talking about a "conventional" cinema xenon lamp, but about a specifically designed lamp module built around that lamp type. First generation 1.3k DLP projectors needed a lamp size of about 5-6kW, so you can see that effectivity has indeed dramatically improved, though I can not say how much of that increase is also a function of the higher resolution.

Posted by Bevan Wright (Member # 1884) on 05-24-2005, 04:59 AM:

It appears SMPTE/DCI are setting the luminance bar at 14ft-l for D-cinema.

With DLP, the efficiency in 3-chip units has a lot to do with: the tilt angle of the mirrors (new chips are 12deg vs 10deg for the 1.3k), the size of the integrator (2k is larger than 1.3k) and the use of VERY short arc lamps. On 2k systems with short arc lamps the efficiency is about on par with 35mm, 6kW lamp can light up a 70ft screen (1.5-1.8 gain) with better uniformity.

Posted by Steve Guttag (Member # 268) on 05-24-2005, 06:47 AM:

Subjectively, 12fL on Video does not look as bright as 16fL on film. A light meter is integrating the shutter for film so you don't have a perfect correlation even though you could put a piece of processed "clear" stock in the projector and only read 12fL.

14fL Video sure looks closer to 16fL Film to me when evaluating the same feature. In fact, 16fL looks even better. 12fL definately looks too dim.

There is a certain "snap" as I call it, when the whites start looking good and I just don't see that on 12fL Video (or film).

This area really needs to be better researched to find out how humans process there various projection methods and how the brain interprets the shutter versus the video image...there is more here than merely pointing a meter at the screen and saying A = B.

Posted by Mark Gulbrandsen (Member # 72) on 05-24-2005, 10:05 AM:

Well there is flicker generated by both means but just at differing frequencies, the eye perceivves these different frequency durations as differing brightness levels because of persistance of vision.... the eye perceiving the longer durations pulses of light fomr a film projector as being brighter. In D-cinema the chip is substituted in place of the film so there is still a picture source so to say...... but what would be interesting would be to compare a 3 bladed shutter equipped film projector and a good DLP projector both set to the same f.l. to see if perceived brightness is still different at those same levels.

Mark

Posted by Philip Borgnes (Member # 2383) on 05-24-2005, 10:07 AM:

Then what this means is that the kw measure really has no meaning. Lamp/projection efficiency is increasing as Michael and Bevan say. Lumens, contrast, screen size and gain are the factors to look at, not the lamp size. I guess there is one critical component of lamp efficiency and that is how long it maintains a stable output. As I understand it even though lamp life might be 1000 or 2000 hours, its output degrades quite a bit by the time it reaches half-life.

Posted by John Pytlak (Member # 331) on 05-24-2005, 12:11 PM:

"Lumens per Watt" for various light sources:

http://www.ysartglass.com/zdbk/Bulbs/LB26efficient.htm

quote:
(An approximate lumen output is given for a Standard size bulb)
Edison’s first bulb 1.4 lumens per watt.
Best carbon filament.16 cp. bulb = approx. 50 lumens 3 to 4 lumens per watt.
Cruto’s synthetic carbon 4 lumens per watt.
Cruto’s deposited graphite 3 lumens per watt.
1905 Metallised carbon filament.
Annealed carbon by Willis Whitney. 4 to 5 lumens per watt.
1897-1912 Nernst lamp 6 lumens per watt.
1907 Squirted Tungsten
30 cp. bulb = 240 lumens 8 lumens per watt.
1910 Drawn Tungsten
32 Watt bulb = 320 lumens 10 lumens per watt.
1912 ‘Half-Watt’ (Coiled filament)
40 Watt bulb = 480 lumens 12 lumens per watt.
1933 Coiled-coil, 10-15% better than coiled filament.
60 Watt bulb = 800 lumens 13 to 14 lumes per watt
1959 Tungsten-halogen
40 Watt bulb = 800 lumens 20 lumens per watt.
1986 Modern bulb (standard)
60 Watt bulb = 960 lumens 16 lumens per watt.
After 86 ??? to be determined 2003
Non-Incandescent Lighting
1898 Moore Tubes c. 8 lumens per watt.
1931 Low pressure sodium discharge 50 lumens per watt.
1938 Fluorescent (Calcium Tungstate) 30 lumens per watt.
1946 Fluorescent (Halophosphate) 60 lumens per watt.
1973 Three band 90 lumens per watt.
1986 Low pressure sodium discharge 200 lumens per watt.

http://www.megavolt.co.il/Tips_and_info/bulbs_at_glance.html

http://www.vigyanprasar.com/wos/electriclight.htm

quote:
Direct sunlight may produce illumination as high as 10,000 lumens per square foot. The maximum efficiency possible from radiant power is 680 lumens per watt.

Posted by Steve Guttag (Member # 268) on 05-24-2005, 12:26 PM:

To some degree, you are right..."Wattage" has always been a poor method of light output. There is a correlation between wattage and light output when all else is equal but when there is no frame of reference it is indeed meaningless...beyond figuring out how much electric power to supply (and pay for).

A proper lumen output is always preferred to a mere wattage rating.

However, the SMPTE really needs to step in on such things for projection equipment...contrast too, if not more so. Left to their own devices...manufacturers will always make their systems work better on paper than you will get in real practice. With contrast ratios of 1000:1 that the vidiots spout you'd think that we would have perfect blacks in theatres by now...since the theatre can't get above 500:1 (with very rare excptions).

Same with lumen output...ever read a video projector's lumen output versus what the manufacturer claims? It will almost never reach it in actual practice and it will only approach it with a specific lens set to a particular setting (be it zoomed in such a way or at a particular focal distance). I've measured 4000 lumen video projectors with a single zoom lens that measured about 3850 lumens at its best and 3300 lumens at its worst...all with the same projector, same lens (at the brightest point on the screen).

There really needs to be some standardization on how things are measured (film and video) in such areas so there is some meaning to the readings.

Posted by Philip Borgnes (Member # 2383) on 05-24-2005, 12:27 PM:

John, you are a fount of knowledge!

This is a hoot,

quote:
"For comparison, an 1860 Oil lamp with a single flame produces approximately 2,500 lumens"

Now I just need to keep an oil-lamp on hand or two to back up my projector!

Steve, you raise an interesting point about something I just learned about - the difference between off-to-on contrast at the projector (typically the 2000:1 you see on specs) and the ANSI contrast (measured at the screen) which is more in line with your experience of seeing a 450:1 to 600:1. Environment plays heavily into ANSI contrast and is perhaps a more valid measure.

It seems this topic has wandered off a bit of the relation of kw to lumens, but it points out to me at least the need for SMPTE to help us out on this issue.

Posted by John Pytlak (Member # 331) on 05-24-2005, 12:55 PM:

Obviously, your quote from that source is likely a typo. I suspect an oil lamp produced only a few lumens.

My usual test of contrast ratio is to cast "shadow puppets" in the darkest areas of the picture, like a night scene. A distinct shadow in what should be a black area says the contrast ratio is mediocre.

I remember the comments about how truly black the dark areas of the Kodak VISION Premier Color Print Film 2393 prints were for "Minority Report":

http://www.film-tech.com/ubb/f4/t000334/p2.html

quote:
This was the 2nd time for me. An outstanding effort by Spielberg. The main thing I want to mention though is image quality. Aaron Haney said: "Also, the print that I saw had amazing contrast. There was a "feature presentation" snipe which ended with a fade to black just before the actual movie started. When the cut to the real movie happened, I couldn't believe just how much darker the screen suddenly became. That continued throughout the whole movie. Very deep blacks."

The print I saw today was exactly as Aaron describes. When the black leader of the feature started, it was nearly pitch black dark in the theater -- I actually thought the xenon had blown or something, the screen was so dark. Well, the entire movie was like that - incredible contrast, and very deep, rich blacks. I don't think I've ever seen anything like this before. And no, the screen is not under-lit. There was PLENTY of brightness (focus seemed perfect, and the image was steady too).

Long live film.

Posted by Larry Myers (Member # 738) on 05-24-2005, 03:00 PM:

Hi

I collect old oil lamps. The reason the whales were almost wiped out is because of all those old oil lamps. Whale oil burns very bright. I am not sure if it's 2500 lumens but it might be 250 lumens or 250 foot candles or the same as burning 250 bees wax candles at 1 foot away.

USG standard light level in a work place should be about 12 foot candles. A 250 foot candle oil lamp would give off 16 foot candles at 4 feet away. That's 250fc at 1 ft, 64fc at 2 ft, 16fc at 4 ft and 4fc at 8 ft. These old lamps worked very well although, with just plain jane oil, light output is now only about 1/2 of the old days.

Larry

Posted by Steve Guttag (Member # 268) on 05-24-2005, 05:48 PM:

One of the things that is has been in the works is a uniform contrast ratio measurment target that could be used for film and video...it currently is looking like a modified checkerboard pattern of black and white squares. It is modified so a phometer can fully real the black and white sections and to actually excite such things as lens flare, which really do lower the contrast ratio.

Note too, in a theatre, ultimately the limiting factor will be the ambient light of the theatre itself in order to provide safe visible exits and pathways to those exits.

But back to the contrast test pattern...I'd be quite surprised if any DLP projector even approached 1000:1 in contrast ratio even in a pure pitch black room. It is one thing to basically lens-cap the projector for one reading, and then measure the direct light and say I have a bazillion to one contrast ratio, it is another thing to put up both black and white targets and do the same thing. Panasonic, for one is boasting a 4000:1 contrast ratio...how do they do it? With a variable iris on the light source...that works on all on/all off...but it doesn't work for all on, AND, all off.

If the film people used similar measurement methods...we would take a light reading with the shutter open and then with the shutter closed...it would be a meaningless number...just like the current video CR numbers.