Cinematography Mailing List - CML
    advanced

Tungsten Lamp Flicker Filming @ 96fps

Just when you thought it was safe to shoot any high speed frame rate under tungsten light... OK, here's a good one. A gaffer friend just did a job where they had a severe problem with flickering under only tungsten illumination while shooting at 96 fps. It caused a re-shoot.

Here is the situation :

There were shooting with a Panavision supplied Arri 435 at 96 fps and a 180 degree shutter. The set was a night interior where all the illumination came from in-shot practicals where the bulbs were either PH212, 150 watt photo enlarger bulbs, or ECT 500 watt photo flood bulbs, installed in the practical lamp fixtures in the set. The power was 110 volt AC, 60 HZ provided from an SMS motion picture generating plant. SMS is a Los Angeles based motion picture generator company. Some of the lamps were dimmed slightly using an SCR dimmer pack. The film is intended to be used in a TV commercial.

In the telecine video dailies there was a pronounced flickering, particularly in the areas of the set wall where the illumination was such that it provided a mid-tone exposure. It was also seen in the foreground on surfaces that were mid-tone exposures. They printed the negative to film after seeing this problem on tape and saw the same problem when that print was projected.

Several days later, they shot a test if they could recreate the problem, the only difference this time was the power came from the commercial power grid.

What they observed in the test via a telecine video tape was that yes indeed they again got flicker at 96 fps / 180 degree shutter. They got flicker regardless of whether the lamp was driven directly from the commercial AC 60 HZ power grid, or through the dimmer pack. They got flicker when the lamp was driven through the dimmer pack when the dimmer controller was at 100% and when it was dimmed to various levels. They got flicker.

The only time they didn't get flicker was when they ran lamps using pure battery power DC. Very interesting

They are going to re-shoot the job on in a couple of days using a DC generating plant. I warned him about some DC generating plants are really 60HZ AC alternators with a rectifier stack on the output which will yield pulsed DC with 120 pulses a second. This might yield the same flicker when the pulses momentarily drop to zero volts 120 times a second. He said that the technician at SMS generators here in Los Angeles was aware of that trap and said they would be providing a DC generator set that uses an AC alternator that puts out 480 Hz, which is rectified to DC and then run through a large bank of capacitors to smooth out the ripple. He said that they have checked the output of that DC generator under load with an oscilloscope and there is no ripple.

What we surmise about all of this is the photo enlarger and photo flood bulbs they were using have very small, lightweight filaments that are "over voltaged" to get the 3200° color temperature and relatively high light output. The penalty is these type of bulbs have a relatively limited life. We surmise that these filaments are lightweight enough to cool slightly and therefore dim slightly during the zero crossing point of the AC waveform, this yielding a "beat frequency" flicker, 120 light pulses per second against 96 camera frames per second. This has to be the answer, because they were only able to eliminate the flicker at 96 fps / 180° shutter when they switched to pure DC powering the bulbs .

I will let you know of the results of the reshoot.

I am posting this here in the hopes that people will file it away in the large memory bank labeled "The many things that can potentially bite you."

Bill Bennett DoP Los Angeles

A lighting trivia question...

Possibly some old timer in the New York area can verify this. I heard the origin of using bulbs with a lower rated voltage and over voltaging them to get more output and higher color temperature was started by New York crews "borrowing" bulbs used in their city's subway cars. Apparently the voltage used for subway car lighting is, or was, lower than the normal USA 110V household voltage, something like 90 or 100 volts. I guess the lighting in the cars is now all fluorescent, but it used to be tungsten way back when. Lighting guys realized that if they "borrowed" these bulbs from the trains and used them on the higher household voltage, they got a higher color temperature and more light, at least for a while. The bulb manufacturers then heard of this and started manufacturing similar bulbs specifically for photographic use.

Anybody know if this myth has any truth to it?


In response to Bill's very descriptive post...

Yes, the mass of the filaments is the issue here, and yes, what they are doing on AC is cooling down between voltage peaks and yes DC will solve it and yes SMS's rectified plants should be ripple free enough to solve the problem. In fact, if you consider that the ripple is 100% right now, you could have DC with a little bit of ripple and still be OK, but just about any 3 phase rectifier (such as the ones we used to use with Brutes ) would be OK...remember that the xenons we use and all xenon projectors are DC arcs running on rectified AC...and the arcs are much less forgiving re: flicker than even these tiny filaments.

Incidentally, this is more a function of overall filament mass than whether or not they are short life 3200k or normal household bulbs. You would have the same problem with regular household bulbs. This is always an issue on miniature shoots where we do any high speed work (for pyro or the mandatory docking sequences between two spaceships) We usually end up feeding the little practical lights built into the model with regulated DC power supplies to avoid this.

Regarding the "borrowing" of bulbs from the train cars, I am not sure about that one. A lot of trains that have incandescent fixtures use/used 48 volt to power them - two in series would give you 96 volts. I do know that the area lighting in the stations used to be tungsten (and still was in some stations when I first lived in NY) and the fixtures alternated between being 3 bulb rosettes and two bulb rosettes - 5 bulb rosettes in some places. They were using track power (nominally 600v. DC) and running 5 lamps in series. In order to minimize pilferage of these bulbs by the local citizenry, the bulbs and sockets were left-hand thread. No one would steal more than a few, since they wouldn't fit anywhere else. Autotransformers have been around for a very long time, and my guess is that the big move towards boosting came when Colortran came out with their booster transformers which were autotransformers with a series of push-button selectors on the top which would engage different taps on the transformer to allow different degrees of boost. I would never admit to doing this personally, :-) but I am intimately aware of numerous occasions where a colortran booster might have been wired into a ceiling lighting circuit at the house circuit breaker panel to boost a whole hallway full of lights.n I believe that Leonetti (Sunray) is selling boosters again, though I would imagine that they would no longer be autotransformers.

I was the gaffer on the Tony Silver, Henry Chalfont doc. "Style Wars", shot by my late partner Burleigh Wartes, and we did a lot of shooting in the layups with car batteries feeding those old cheap plastic sylvania sun guns that were crap but gave such a beautiful light, as well as a tweenie and some omni's with 12 volt versions of a DYS in them and beefier cables to minimize line loss. We also lit off the third rail, using 5 Lowel D's in series, which worked fine until one lamp blew and the DC arced across it, feeding overvoltage to the remaing 4, which in turn blew. We did a lot with Photofloods in series clipped to the hot rail.

I bet they wouldn't let us do it that way nowadays. Aww, jeeze sorry folks, I have been rambling on all day with a bunch of posts - I'll shut up now and let the grownups get back to it. I think this is a reaction to getting off Jury Duty - I have spent the last week locked in a government building with no one to talk to who knew about any of this stuff.

Mark H. Weingartner

Lighting and VFX for Motion Pictures


I, too, have experienced a similar situation. Shooting at 45 fps on a music video with little tiny tungsten bulbs, I noticed a slight flicker in the transfer. It was very slight and nobody else noticed it, or cared when I pointed it out, but it sure made me do a "note to self". I was very surprised, because I always just assumed that tungsten bulbs were never subject to flicker.

I figured that it was due to the small filaments in the bulbs. Unlike a big filament, which stays glowing for a while even after you turn it off, these were a whole slew of little 15W bulbs (like you'd have on a makeup mirror) that we had rigged on a long 1x4 piece of wood to make this beautiful soft strip of light. In the future, I won't use such little bulbs.

Phil Badger, gaffer, LA


Yes, we have seen a fair amount of this as well in various situations. Once in a night exterior shot in Pasadena. The practical carriage lamps outside the buildings (fed off city power) were flickering madly at 48 and 150 fps. We also saw it in an interior shot looking up into a ceiling with a large number of small christmasy lights.(also house power) Also at 150 fps. We have seen it in a few other shots mostly at 150. We chalked it up to the non crystal regulated city mains, but that never really made sense. It is good to have it so thoroughly explained here. As Bill B. so eloquently puts it, just one more thing that shouldn't go wrong that might.

-- Ed Colman -

SuperDailies Cinematograper Supervised Video Dailies

www.superdailies.com


Responses to this thread have been excellent -- pulse-free DC current probably the best solution. But if that power supply is not available you might consider this -- assuming 240vac split-phase house current is available.

Split your "whole slew of little 15w bulbs" into two circuits -- every other one alternately wired to the "A" or the "B" circuit. Power the lighting array using both sides (120vac phases) of the 240vac line. As half the bulbs decay the other half builds, smoothing out the flicker. Overall light output may be less, but not much.

I've never tried this, but theoretically it should reduce flicker for filming at any shutter speed.

On a somewhat related note, Bill Bennett talked about old timers "borrowing" NY subway bulbs. In my early days of industrial filmmaking, we frequently used the Colortran voltage boosters Mark Weingartner described. You selected the boost needed to create the color-temperature you wanted. Gave you 3200K or 3400K and much brighter light from ordinary incandescents. (Considerably shorter lamp life, too.) But for strong keys we used airplane landing lamps. Couldn't get subway train head-lights in Iowa.

Don Ver Ploeg Kodak consultant Rochester, NY

In a former life I hot-wired many things that moved and some things that didn't...


Question: aside from staying away from those bulbs, is there a method for checking up front if a given bulb with a given power supply will flicker? Bill, will the B & S flicker meter work? Any other method besides shooting a film test?

Regards,

Marc Shipman-Mueller,

Technical Representative

Arriflex Corporation; 1646 North Oakley Ave, Suite #2,

Chicago, IL 60647-5319, USA

http://www.arri.com


There's a little handheld unit called, if I remember correctly, a Cine Check. Basically a small photo-tachometer in a little black case with a red L.E.D. readout in cycles per second. I don't own one, but borrowed one from an A.C. several years ago, and to my surprise I was able to meter power line frequency to 1/100ths of a cycle by holding it up to a 40 watt household incandescent bulb. Evidence of exactly what this thread has been discussing, pulsing from low wattage fixtures at higher camera speeds.

Bulb was turn on, of course.

Jim Furrer Dark Street Films / VGG Systems, Inc. Lakewood, CO, USA

http://members.aol.com/jimfurrer


There's a problem with theory-- it has no influence on reality.

House wiring (240VAC) split into two legs is just two circuits of AC with equal and opposite polarity at any instant. Both phases cross zero potential at the same time. There is no overlap and no advantage gained.

Nor will the light output of the bulbs change. Your house has several AC circuits scattered around - some on one leg and some on another. You never get more light out of one bulb on one leg than on the other. If you do, you yell for the electrician- you've got trouble!

David Tosh

Engineer, Complete Post Hollywood, CA USA


I don't think this would be a solution to the problem because the 240v is created by one leg being at +120v while the other is at -120v and then they both cross zero volts at the same time, so they are in effect both "off" at the same time. You could accomplish this by going to 3 phase and putting 1/3 of your lights on each leg, but I think that would only help with any off camera lighting.

For off camera lighting, it might be easier to switch from your array made from a bunch of little bulbs to a kino or a larger tungsten unit (less susceptible to flicker) instead of taking the time to wire every third bulb on a different leg. Also, running three phase power (3 stingers from your nearest 3 phase source) to the array of little lights could be more cumbersome than it's worth. If you have practicals in the shot, they can possibly still flicker because they are individually on a single leg. If you split all of your practicals to 1/3 on each leg, each 1/3 of the lights could potentially (depending on frame rate) flicker as a group and still ruin the shot.

Which brings me to my question. With AC power, do we have to use the HMI safe speeds for the little tungsten sources, or are there other safe frame rates?

Paul Szopa DP / LA


>Question: aside from staying away from those bulbs, is there a method for checking >up front if a given bulb with a given power supply will flicker? Bill, will the B & S flicker >meter work? Any other method besides shooting a film test?

I second those questions. I'm still amazed about tugsten flicker. However, hile storing this info away in case of future problems as Bill Bennett suggested, I am getting the unmistakable feeling that something like this happened on a table top shoot that involved one of our cameras -- and that we could never find anything wrong with it -- which may mean there may be a perfectly good Arri III motor and boards gathering dust on our shelves.

Thank you, Bill.

Brian Heller IA 600


Unfortunately, this would not work. 240v services are "single phase" even though there are two hot legs. That is to say, there is a 240 volt AC sine wave you could observe with a scope hooked between the two hot legs. What you would observe between each of the hot legs and the neutral wouyld be two mirror image sine waves. Where the A leg is at +120v, the B phase is at -120V When A is at 0, Bis at 0. The 240 comes from going from A to B When A is at +120v, Bis at -120v and the total distance between them is 240volts. So what's the problem?

Lightbulbs are dumb. They don't know what direction the electrons are going at any different time (+ or-) they just know that as they rush through the hallway it gets hot - hot enough to glow. When A leg is at +120 and B leg is at -120, both bulbs will be glowing. When A phase is at 0 and B phase is at 0, both would be dimmer (cooling down) SO while the voltage curve looks like a sine wave, light output curve looks like a wavy line with peaks when the voltage is at + OR - 120 and valleys when the voltage is at 0.

If, on the other hand, you have 3 phase power, where the three legs are out of phase by 120 degrees, you would get relatively uniform light output if you distributed the bulbns equally on all three phases, BUT, if the sources were in frame, they would still pulse in the shot. Apogee and I both used three phase power to power VHO (Very High Output) flourescent tubes on old fashioned magnetic ballasts to light blue screens. The individual tubes would pulse, but the overall output was constant with four tubes per leg in each fixture.

>But for strong keys we used airplane landing lamps. Couldn't get subway train > head-lights in Iowa. I remember my mis-spent youth in theater and rock 'n roll where we used ACL's, which were 28volt aircraft landing lights (5 in series) for those big "rays of light behind the band" look. There attraction was that since they were low voltage, the filaments could be lower resistance for the same wattage, which meant that they could be shorter, which meant that more of the filament could be in the focus of the parabolic reflector, so the rays coming off of thereflector were closer to parallel than the same reflector with a 120v. filament (These were PAR 64's.)

Mark H. Weingartner Lighting and VFX for Motion Pictures


Jim's memory is correct. The Cine Check will read the line frequency of a an incandescent lamp by reading the fluctuations in the lamps intensity. This applies to 10k lamps as well as 15watt refrigerator bulbs. We just re-tested this hypothesis with an in line frequency meter. Both the freq meter and the Cine Check gave the same readings and changes.

Unfortunately, the Cine Check will not tell you if the lamp or lamps in question will cause visible flicker at 96fps or at any other speeds. Don't know yet about the B&S meter. Could it be that camera speeds have become so precise that in certain circumstances their accuracy can work against what we're trying to do. Should there be a crystal speed disable switch?

Brian Heller IA 600


>...A lighting trivia question. Possibly some old timer in the New York area can verify >this. I heard the origin of using bulbs with a lower rated voltage and over voltaging >them to get more output and higher color temperature was started by New York >crews "borrowing" bulbs used in their city's subway cars...

In an old textbook I used in the '50s "THIS IS PHOTOGRAPHY" by Thos. Miller and Wyatt Brummit, both of Kodak, the authors claim that "someone took a 100-watt lamp of the type used in railway trains, where the voltage runs at about 64, and subjected it to the full impact of ordinary domestic current--110 volts, more of less. The surprised railway train lamp lighted up furiously, delivering an intensity of light approximately equal to that of an ordinary but expensive 500-watt lamp. Naturally, it burned out in an hour or two...." They claimed that the photoflood lamp that resulted from this discovery came along about 1931, at the same time as the introduction of panchromatic film.

There you have it, folks--straight from the Great Yellow Father, the ultimate authority!

We still have several of those old Colortran transformers, including a couple of small 4 light units from the early '50s that predate pushbuttons (toggle switches.) We also constructed our own autotransformer units that provided continuous variable voltage from 0-210v. We used them for strip lights back in the old ECO days, powering 13 - 150w PAR 40 Reflector floods on a bank.

And yes, our Cinematography Electronics Cine-Check meter reads the AC frequency on Tungsten lamps as readily as it does HMIs.

Wade K. Ramsey, DP

Dept. of Cinema & Video Production

Bob Jones University Greenville, SC 29614


Right you are, Mark! My theory in this case may be one of the reasons why I'm still simply a shade-tree electrician. Thanks very much for sorting out the phases and for your detailed and informative explanation. I'm sure others found it helpful, as have I.

Don Ver Ploeg Kodak consultant Rochester, NY


There is no reason every DP should have to no every bit of electrical theory - that is, of course why one hires gaffers.

On the other hand, the very jobs where one needs to know all that stuff are often those jobs that do not have the $$$ for you to hire people of the neccessary experience level to cover you on all the details or to hire the sort of equipment that would let you avoid certain types of pitfalls.

It's a bit of a catch 22

In my case, a background in theatrical lighting, Rock 'n Roll, and physics kept me from screwing up too much as I moved up in the world as a gaffer in NY, the land of tie-ins on the 32nd floor. I taught a lot of people how to tie in and how to recognize what it was that they were looking at, but there is no clear path to gaining this knowledge out there. Film schools concentrate on film-making, not the intricacies of electical theory, practice and wiring. The unions are making an effort to train their new and existing members with regard to safety and practices, but most people build their reels using non-union personnel. There is nothing wrong with that, but where do the young incoming craftsmen learn? One used to apprentice with established craftsmen, but I see people gaffing with very little time spent on other people's sets. This is not a condemnation - I started off myself as a "big fish in a little pond" having been a department head for nearly all of my jobs in the industry. I don't have answers, but I wonder whether anyone would take the time to take advantage of classes in electicity for motion picture and televisoin if it were being taught.

I could see teaching something like that, but I have no idea where

Mark H. Weingartner Lighting and VFX for Motion Pictures


Sounds like you're describing a CineSpeed meter. I got mine online from Mole.com, in a situation where we had a mystery flicker that came and went despite different camera bodies and power sources, and I decided enough was enough.

These are fabulous little meters, though - you can meter HMI flicker, Tungsten line frequency (I agree with Jim's observation), camera speed (at least with a 180 shutter) and television/computer screen frequency.

George Hupka Director/DP,

Downstream Pictures Saskatoon, Canada


Marc Shipman-Mueller wrote :

>Question: aside from staying away from those bulbs, is there a method for checking >up front if a given bulb with a given power supply will flicker? Bill, will the B & S flicker >meter work? Any other method besides shooting a film test?

Marc,

Funny you should ask... On the test day, four days after the shoot day that was ruined by the unexpected flicker, the gaffer did check the PH212 lamps with his B&S flicker meter as they were being powered by 60 Hz commercial power. The reading was 5.9%. The B&S manual says the limit you should not exceed is 6 to 8% for tungsten lamps, so theoretically IF he had checked on the shoot day, he was barely within the limit specified by the manual.

I suspect the maximum allowable limit for tungsten lights should be lowered to around 3%, similar to the limit allowed for HMI lamps with the B&S meter. This gaffer says he rejects HMI bulbs that are greater than 2%, just to be safe. I think that you might need to be that conservative to avoid any flicker problems with tungsten lamps also.

Of course, you would have to shoot tests to verify all this though.

Bill Bennett DoP Los Angeles


Greetings,

-There is or was indeed an optical frequencymeter called Cine Check made by Cinematography Electronics. I wonder if this is still around. Any help where to get one of these appreciated? (Is Cinematography Electronics the same company is Cinematography Products?) I could do with one of these over here. Everything flickers.

-This said I have some low budget doco projects coming up (for a change :-)! ) Some china balls and other low wattage bulbs may be used. In order to avoid flickering (and no proper generator will be available) I wonder if I could connect this to a UPS or something similar to partially solve or anticipate any possible flicker. As filmstock is limited And + 25 fps will be done. As quite some gaffer experienced people around thought asking this here?

Thank you beforehand.

Regards Manny from Beirut (in Lebanon that is) AC


>There is no reason every DP should have to no every bit of electrical theory - that is, >of course why one hires gaffers:-) That should read: ...every DP should have to >KNOW every bit of electrical...

At least you "no" you made the mistake!

It's a bit ironic that so many film/video people obviously had figured they really didn't need to learn all that English stuff. They were going to communicate visually, not write treatises. Now they find themselves e-mailing the King's English as one of their primary means of communication and don't know how to spell/type/construct sentences, etc.!

For years I have taught our lighting students more about power distribution, cabling, tie-ins, etc., than they ever wanted to know. I have scaled that back a lot because, for the amount of time they have in class, they need to know more about lighting techniques than about electricians' duties. In view of the increasing code limitations on non-licensed persons doing anything other than plugging in, along with the decreased need for really heavy current in the non-theatrical areas we deal with, it seemed the better choice.

But it's good to know enough that you can size up a location and know what to have the electrician do for you, so we still deal with some of the practical theory.

Wade K. Ramsey, DP


15 years ago, I had a Cine-check that had problems, went back for repair, and never returned to my local vendor. I never replaced it, but went back to my home-made set-up, which has served me well. Marty Satloff at Empire exposure meter suggested (in 1983) that I go to Radio Shack and buy the [then] $8.95 "Solar Cell Experiment Kit". I put a 3 ft lead on the solar cell, leaving it in the clear plastic box in which it was sold, and plugged that into my Fluke 8060A multimeter - the only one that they made back then which had both true RMS voltage (for trimming dimmers) and freq. The meter would read the pulsed DC as AC and give me freq back. Dead accurate.

Dead cheap (well, not counting the $500 for the 8060A meter...but I already had that:-))

This would not replace a B&S, but I just realized as I type this that I could hook the the solar cell to my oscillascope and maybe...

...I will get back to you on this one:

Mark H. Weingartner Lighting and VFX for Motion Pictures


(Cine-Check)

There is a Web site for C.E.

http://www.cinematographyelectronics.com/index_gallery.html

Ian Green Vancouver, BC

Technical Service Manager

Panavision Canada Corporation


 


Copyright © CML. All rights reserved.