Kinoflos and High Speed Flicker
Published : 14th April 2008
I've a test coming up but since I've already heard so much conflicting anecdotal opinion on this one I'm curious to throw it open to CML....
At what frame rates will kinoflos begin to visibly flicker?
The shoot I've got coming up will be using both the Phantom & Weisscam digital High Speed cameras and the desire is to get within the vicinity of 1000fps.
Despite working at that sort of frame rate before I've never faced this particular issue since I've always used nothing smaller than 10k tungsten/12k HMI and avoided problems with 50Hz that way.
The treatment calls for 100s of fluorescent tubes in frame. Obviously domestic units with conventional 230v/50Hz ballasts are a non starter. Since Kinoflo ballasts over-cycle the tubes at a much higher frequency I know they're safe up to 150fps but beyond that it's all guesswork at this stage.....
I thought it was the older 50Hz ballasts that had the flicker problem, and that it was corrected in the newer 50Hz ballasts. Contacting Kino Flo would probably be the best source. Are you actually photographing the bare tubes in frame, or are they in fixtures to provide illumination?
IA 600 DP, Los Angeles based
Your right, our newer ballasts are designed to be Flicker Free at most frame speeds. However, that is a relative statement. But, Just to dispel any misconception. ALL Kino Ballasts are designed to be flicker free not just the newer ones. Old ballasts may suffer capacitance degradation which could influence their flicker free capacity. However, when your talking the kind of high speed of 1000fps digital or Photosonics here's what I can recommend.
Flicker tests have been performed to 150 fps with good results. No shots have been done at 1000 fps. A test would have to be done to verify the stability. Do not try this shot without running a camera test before hand.
Please keep us informed of your test and subsequent digital capture and their results.
2840 N Hollywood Way
Burbank, CA 91505
Graham Futerfas wrote:
> I thought it was the older 50Hz ballasts that had the flicker problem,
That chimes with what I'd heard....
> Are you actually photographing the bare tubes in frame
That was the idea. 6' mega singles. A hundred of them.
For a variety of reasons - not least expense – the idea's changed so there'll be no kino tubes in frame now.
I'm still curious to know what the flicker issues would have been. Kinoflo's UK distributor didn't have a definite answer beyond stating that a new ballast was a good starting point, that it's 'not recommended when shooting above 150fps' and 'there may be safe 'windows' that are multiples of 50Hz'.
I have shot a test with a 4' 4-bank Kino on a Photosonics 4ER circa 360fps and had no evident flicker (all on 50Hz Power).
Dean Slotar | One8Six Cape Town
t +27-21-555-1780 | f +27-21-555-1828 | m +27-82-895-2620
Tom Jacob wrote:
>>Flicker tests have been performed to 150 fps with good results. No >>shots have been done at 1000 fps.
Given that you must compensate 7.1 stops to shoot at 1000fps (with a Photosonic 4C with a beamsplitter for reflex viewing), Kino fixtures do not have enough output to light the scene, especially given that the thread originally discussed a shot wide enough to see 100 6' Mega Singles (and now has dropped those globes because of budget). And don't forget that most of the lenses available for the Photosonic 4C have a minimum aperture of T4.5.
In terms of reflected luminance (in shot) of a bare globe without any ND, the globe would be underexposed and not create the brilliant effect one would want.
Look at http://www.sgps.net/lightbox_thmb.html (4th and 5th picture) to see a lighting rig one might need to illuminate such a wider shot proposed with 100 6' Mega Singles. From the look of it, looks like 20 20K Fresnels through a silk. Anyone involved with this shoot?
Chief Lighting Technician (IATSE 728) / Sometimes DP
Roberto Schein wrote:
>>Look at http://www.sgps.net/lightbox_thmb.html (4th and 5th picture) to >>see a lighting rig one might need to illuminate such a wider shot >>proposed with 100 6' Mega Singles. From the look of it, looks like 20 >>20K Fresnels through a silk.Anyone involved with this shoot?
Sorry I am getting of topics here, but on further viewing the fixtures must be 12 or 18 HMI Fresnels (Silverbullets on top, LTM or Arri on bottom). If one was going to use tungsten, Maxi-brutes or Dinos would make more sense. Any one have experience with such setup that would like to shed light on the topic. Bill Bennett?
Given that you could distribute the ballast along 3 phases, would any flicker be hidden.
Chief Lighting Technician (IATSE 728)/ Sometimes DP
Roberto Schein wrote:
>> If one was going to use tungsten, Maxi-brutes or Dinos would make >>more sense.
Actually, 10 & 20ks would be better. Maxi-brutes or Dinos would flicker.
On a recent shoot with the Weisscam
I got burnt with one set-up this way.
A row of windows needed soft(ish) illumination coming through them but the location didn't afford the space to physically fit a large single unit plus diffusion outside each window. The compromise solution was to use a 12 light Maxibrute outside each one.
The hope was that the flicker that each 1k par bulb produced would 'average out' if they weren't on the same phase but there was still visible pulsing above 250fps.
Over the course of the shoot the most consistently flicker proof source was the 18k.
You should take a look at our Parabeams & the new "VistaBeam" Equal to a 6K soft. Very bright and flicker free. Should have the punch for high speed shooting.
Here's the link, if you would like to check it out.
2840 N Hollywood Way
Burbank, CA 91505
We tested today and found that a Kino Flo Diva-Lite 400 does not have discernable flicker at 1000fps or even 24,000fps. We shot the light with a Vision Research V10 and put two the 2 clips at:
Jeff Silverman wrote:
>>We tested today and found that a Kino Flo Diva-Lite 400 does not have >>discernable flicker at 1000fps or even 24,000fps. We shot the light with >>a Vision Research V10 and put two the 2 clips at:
Useful and timely info. Thanks for doing that and posting the clips.
Scopica Inc. / Scopica3D
>> We tested today and found that a Kino Flo Diva-Lite 400 does not have >>discernable flicker at 1000fps or even 24,000fps.
The 1K FPS didn't seem to play and the 24K FPS did flicker on the clip ??? Is that a product of the QT or??
ICG, New York
Looking at the original data file format of the 1000 fps clip looks exactly like it does in Quicktime (no variation, almost like it is not playing). There is no discernable flicker. If you study each individual pixel of the clip like we can do in the native software, you can see at most a random 3-4% variation per pixel per frame in the RGB levels present. This same variation is present even in the black areas of the frame to a slightly lesser extent and so I think the variation is more due to the CMOS imager than anything else. What I don't see at all is organized flicker. None of this is apparent directly visually.
The 24k clip does show extremely minor flicker. I measure the luminance variation in the center two bulbs at about 10%. The 2 center tubes vary brightness at different times than the outer 2 tubes. Playing the clip frame by frame I see a flicker of the inner 2 tubes which has a approximate frequency of slightly less than 3 frames at 24,000fps or about 7000-8000 hertz. Looking at the outer cool bulb the flicker seems to be a little over 3 frames per flicker but less overall variation. I can not determine the combined luminance difference of all 4 bulbs added together but there are frames where all 4 are close to peak and all 4 are close to minimum.
What would be interesting is a test of the overall light output itself rather than looking directly at the bulbs themselves. I think that would offer a more accurate representation of the combined effort of the lighting fixture. If I get another opportunity to test the light I will do this. While I have an few high speed cameras lying around, sadly I don't own any Kino Flos.
By comparison though, the overall amount of the flicker I see here is quite small when compared to other light sources we have been forced to shoot under.
In any case I am confident to say 1000 fps is safe.
Jacksonville, Vermont USA
Thanks for clarifying. Thanks for doing the test and posting the clips.
the dop shop
As an addendum to this thread, at the recent NYCDES show we ran a Phantom camera up past 1500p with illumination provided by a Kino Diva 400 and a Litepanel 1x1. There was no flicker visible from these lights. Tungsten and industrial flo's on magnetic ballasts did flicker noticeably.
Abel Cine Tech