LED lights for filmmaking

Hello,

I recently saw an advert for some LED lights for filmmaking. They looked
perfect - very efficient, flicker-free, dimable from 0-100% etc. The
problem is that they're extortionately priced. So now I want to make my own
LED lights for use on film...

Has anyone tried this? What should I be careful of? Can I vary the colour
temperature of the lights by pushing more or less current through the LEDs?

Even better - does anyone know of any LED lights suitable for film that I
could buy off the shelf here in the UK?

Here's my dream LED light:

- dimmable from 0-100% with no change in colour temp (ultimately I'd like to
build in a remote control so I can change dim the light whilst I'm looking
through the viewfinder on my camera)
- cheap!
- highly efficient
- stable and predictable colour temperature (it would be very cool if I
could change the colour temp with a switch... my research into LEDs so far
has hinted at the possibility of changing the colour temperature by
increased in the current).

Please do let me know your thoughts - any leads you can give me will be very
well received

Thanks,
Jack

 

Have you priced white and/or blue LEDs? I'm all for rolling
my own equipment (working on a couple of ideas for compact
flourescent), but the price and efficiency of white and blue LEDs
isn't there yet, IMHO.

You haven't mentioned what KIND of lights you are talking about.
A small "fill" light that fits in the accessory shoe on top of your
camcorder is one thing. A 500W or 1000W equivalent elipsoid
or fresnel is something else again. And a big soft-light is yet
annother thing. IMHO, only a small, camera-mounted fill light
might be practical at this time (if you have deep pockets).

I would doubt it. LEDs don't produce light the same way as
filament-based lamps. Even the range of output is relative
limited. Full-range dimming can be done only by pulsing the
LEDs and adjusting the pulse-width. If you do this at a high
enough frequency, it appears to be "flicker-free".

I have seen some theatrical lighting equipment that claims to
have LED sources, but as you say, they are scandalously
expensive.

With the price of the LEDs, the cost of the dimming parts of
the circuit (power transistors, heat sinks, etc.) will seem cheap.

Don't hold your breath. At least for a few more years.

I seem to recall something about big heat-sinks on the backside
of the LED sources, and even fan cooling. Doesn't bode very
well for efficiency if all that power is going into useless heat.

Doesn't seem likely. OTOH, people even make full-color image
displays with LEDs by close-spacing red, green, and blue LEDs
and pulse-width modulating them to produce any color you want.
"white LEDs" are sometimes nothing but red, green, and blue
LED chips in the same package. The colors combine to make what
appears to be "white" to us exactly the same as color displays
CRT, LCD, etc.)

In the last year or so, there has been a lot of discussion of white
LEDs and the circuits used to drive them from batteries over on
the newsgroup Dunno
about any of the other newsgroups you have cross-posted this to.

At least this is my perception of the current state of the art. I
would be happy to learn that there is better news.

 

LEDs as film lights? Sounds highly intriguing. I thought that LEDs tended to
produce fairly monochromatic light, whereas you'd want broad-spectrum white
light for filming. Also, I can't imagine LEDs producing sufficient light to
illuminate something more than a few millimetres away!

 

You're forgotting how frightening the higher-intensity LEDs can be.

A couple of years ago, a story turned up about an LED-based theatre
light that generated relatively little heat and was tuneable in colour
(presumably three lighting components in varying amounts). If it was
tuned well to your sensor it could provide light that looked /white/
but unlesss the individual LEDs were quite broad band, you'd still get
crap colour rendition of things withing the field of lighting.

I really wonder what the current state of the art is for these things.

B>

 

Generally no. No.

Not a problem.

Extreme problem. Expect to pay well over 3 quid a watt, in quantity.

Nope, about as good as halogen bulbs, maybe a bit more if you go with
mixed colours of LEDs, but then you run into new problems.

It alters, but generally not in useful ways.

Wait a decade or two.

 

FOLLOW UP settto uk.tech.broadcast, and not 9 groups

Phil: The original poster may be confusing on-Camera LED assemblies
designed for on-axis chromakey (in green or blue or I think white** at IBC
this year) with full-spectrum lighting, which event he white ones don't
supply, sincle they are effectively 'fluoescents' in their light output:
eg starting from an basic excited UV via doped layers or coatings which
emit at selected visible colours (like fluorescent tubes) to give a
'balanced' white - as long as you are not then using that light to analyse
the colour composition of a scene 9as in a camera! ((unless exactly
marching the RGB filter - but even then, it is unlikely to have any
spread, but be 3 or 4 monochromatic sources.))

**used with directional (scoth-light?? or similar name) reflectant
backgrounds, giving good chromakey or matte separation of subject to
background - the efficiency coming also from the directional nature of the
reflective material.
(Which is a couple of hundred for a sqaure metre or 2?)

 

Colour temperature's one thing; what about the heat problem? You'll be
no doubt bunching these hi-output LEDs together in some sort of
reflective enclosure and they're bound to get very hot when running at
any useful level of brightness. Plus the hotter they get, the more
current they tend to pass (if not limited) which makes them hotter
still and makes for the difficulty in setting a constant level of
brightness. Can they really cope with the intensity of being used
closely bunched together even if you can control the brightness
satisfactorily?

 

Richard Crowley wrote:

This is not the only way: I have previously built a voltage controlled
current source for driving LEDs, which allows linear, full range
dimming. However, for high powers, this is rather harder to do without
the pulsing approace, due to the heat dissipated by the active devices.

Which should be quite easy. LEDs can be driven to very high frequencies
easily. I've transmitted a few watts (infra red) at 20KHz (using a
current source driver) with no trouble. I could have gone higher, but I
settled on that frequency.

LEDs are currently rather expensive, at about £10 per Watt.

At 90% efficiency, a 1KW source will give off 100W of heat. Consider how
hot a 100W lightbulb becomes. You don't want your LEDs getting that hot.

I believe that some white LEDs operate by giving off blue light anc
converting some to green and red using phosphors. Others work in teh
manner you describe. I would expect that to be more efficient.



-Ed

--
(You can't go wrong with psycho-rats.) (er258)(@)(eng.cam)(.ac.uk)

/d{def}def/f{/Times findfont s scalefont setfont}d/s{10}d/r{roll}d f 5/m
{moveto}d -1 r 230 350 m 0 1 179{1 index show 88 rotate 4 mul 0 rmoveto}
for /s 15 d f pop 240 420 m 0 1 3 { 4 2 1 r sub -1 r show } for showpage

 

Those red or cyan washes used on the glass parts of the 'Watchdog' set
are from LED sources. The fun bit is that the brightness is controlled
by high frequency pulse-width modulation, so Daniel may have issues with
anything fitted with a shutter, hint-hint !

Cheers,

G.

 

Brilliant, thanks loads for all your ideas.

There's a fantastic thred over here:

http://www.dvxuser.com/cgi-bin/DVX2/YaBB.pl?board=Links;action=display;num=1071261354;start=0

Thanks,
Jack

 

I saw a Beyonce video ("Crazy in love" maybe) with a rooftop, twilight scene
with what looked like flat discs of white leds maybe a foot in diameter, a
bit like continuous ring-flash giving flat lighting.
I don't imagine the "non-flicker" is possible -about the only way to dim
LEDs is to use PWM - just hope the frequency is very much greater than your
line rate.
I'd agree that such lights are not likely to be cheap - there were probably
several hundred LEDs per disc/
My 2 cents

P.

 

Surely if you pulse the LEDs you'll get patterning on the picture due to the
sequential nature of the scanning. For example, if you pulsed the LEDs at
15625 kHz (the line frequency of British TV) you'd get one half the picture
black and the other half white, with the proportion of black to white
varying with the mark:space ratio of the driving signal (ie the brightness
of the light). The only way to avoid this would be to have lots of LEDs all
firing from the same signal but each with a different phase difference.
Maybe the best frequency is one that is a long way from a harmonic of the
line frequency, so that any remaining patterning looks like a random dot
crawl over the picture rather than a static pattern.

Or do LEDs have a fairly long persistence - ie does the light continue for
some time after the driving voltage is removed? If so, the effect of varying
the mark:space ratio would be reduced if the frequency was chosen to be high
enough to avoid significant flicker.

What is the situation with xenon tubes used as TV lights? Are they pulsed
with variable mark:space ratio or are they on continuously with the current
varied to vary the brightness?

 

On Wed, 20 Oct 2004 17:22:23 +0000 (UTC), "Deep Reset"

[snip]

As has been stated before, LEDs can be dimmed with a continuous
current. In this case there is no flicker from the LED. The trick is
to make a variable current source with high efficiency. This is done
all the time by using a PWM circuit inside the power converter and
then smoothing (filtering) the current before it is applied to the
load. You get a continuously variable, high efficiency current source
that has very little modulation on its output - and hence no flicker.

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

 

A switched-mode current source can use relatively small inductors to
smooth the current to something approximating DC.

 

Trimmed crosspost abit because server will think its spam...
Hi from sci.engr.lighting

Efficent ,er, um, see rhose MR16s your using at the moment, theyre probably
abit more efficient..

LEDs?

Not really, whites because theri blue with a phosphor go a bit of angry blue
when you overdrive them though....

Sure I saw Kino Flo with some protype Luxeon LED film lights a while back,
but then they make fluro compact film lights which fro general light are
probably still a better bet.

Thats kind of doable , but current white LEDs are basic phosphor wise,with a
high colour temperature and not great Colour Rendering Index.
RGB colour mixing does not give a good white , some units have started
using RGB and Amber to warm it up a bit.

cheap or bright?
cannae have both

If you don`t need alot of light, say an LED macro ring it is efficient for
the purpose, as replacement Redhead, not yet.

Going cooler CT wise in white,
but cooking LED in practice,
lowering efficiency, LEDs hit sweet spot at exceedingly low currents
Negative temp co-efficient means output goes down as heat goes up
Lowering lumen maintenace , cooked phosphors and LED dice put out less light
as they age, hard life will age them faster

 

Described by one of the regulars on rec.video.production a
year or two ago. He made a large disk that fit around the lens
and laid "light-rope" around in a spiral to create a large flat
light for a special shot.

 

Have you seen many light entertainment shows recently? LED based coloured
effects lighting is everywhere these days.

You can get strips of single LEDs in a row, single miniature parcan-style
LED lamps, side lit panels and rows of LED discs. All of these can be
cycled in colour under lighting controller control - usually using DMX
data - allowing all sorts of multi-coloured effects. Pulsar Light are the
main manufacturers I know of who do this stuff.

Steve

 

I've followed this thread with interest but not a little scepticism.
There seems to have been lots of references to high power *display*
LEDs. However, providing enough radiated light to give adequate
illumination for video filming seems to me a different issue. LEDs are
semiconductors, and semiconductors always have had serious problems with
heat dissipation and can so easily burn out (all those fans on the chips
in my PC - wow).

To get the equivalent of (say) a 500 watt flood in a LED (or even an LED
array) seems to be a bit of an extreme design concept and would seem to
have limited practical offerings compared with conventional lighting
units.

 

Touch an ordinary filament torch bulb while it is working, and then one of
those new high brightness white LEDs. Notice the difference. Filament lamps
are about 5% efficient, which means that for every 100W of light, you have
to dissipate nearly a couple of kilowatts of heat. I'm not sure what the
efficiency of LEDs is, but torches that use them run for months, so it must
be greater.

I see no reason why scaling this up for film or TV lighting use wouldn't be
possible, though producing acceptable colour rendition might be a different
matter, as the colour separation mechanisms used in photography assume a
uniform spectral characteristic of the light source.

Rod.

 

While I see from the link that they do indeed exist, I'd argue that
efficiency should be tempered with cost ($4000 + with a lamp life of
2000 hours). If someone can afford the cost of buying, then they can
sure afford the cost of the power to drive them

Laurence (with his theatrical antecedents) will perhaps have some
insight into the relative cost of conventional lighting units.

I see that the spec from the link gives 6000K, so I'd assume that colour
is not an issue. I also recall several years ago when tricolour LEDs
were used in reflective colour spectrometers for the print industry, so
I'd speculate that colour is an issue (or rather that any issues can be
relatively easily worked-around).