Jump to content


Photo

High Speed Camera Color Flicker


  • Please log in to reply
17 replies to this topic

#1 Markus Ebert

Markus Ebert

    New

  • Basic Members
  • Pip
  • 2 posts
  • Cinematographer

Posted 30 October 2011 - 01:18 PM

Hello,

we shooted with the Weisscam HS-2 with 2000 fps and we have a color changing in the skin tones and on metal.
I think this called Arc Wander, is this right?

Does anyone know what the problem is? And is there a way to correct it in the Post?

Here we have a sample.



Thank You
Markus

Edited by Markus Ebert, 30 October 2011 - 01:19 PM.

  • 0

#2 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 30 October 2011 - 02:10 PM

Hello,

we shooted with the Weisscam HS-2 with 2000 fps and we have a color changing in the skin tones and on metal.
I think this called Arc Wander, is this right?

Does anyone know what the problem is? And is there a way to correct it in the Post?

Here we have a sample.



Thank You
Markus


2000fs is very fast and assuming you were using artificial lighting I'm guessing it might be an issue with the camera being able to see fluctuations in the lighting! I'm guessing there are either minute changes in the colour temerature of the light or even some kind of marginal flicker only visible at really fast. Not sure!

love

Freya
  • 0

#3 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 30 October 2011 - 02:23 PM

Hello,

we shooted with the Weisscam HS-2 with 2000 fps and we have a color changing in the skin tones and on metal.
I think this called Arc Wander, is this right?

Does anyone know what the problem is? And is there a way to correct it in the Post?

Here we have a sample.



Thank You
Markus


HMI's even flicker free can flicker & / or have ARC wander. Ideally you should use large tungsten units, the advantage is that you can dim them between takes, it's what I usually use for high speed. You need to use large units 10k or more & use multiple phases.
  • 0

#4 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 30 October 2011 - 07:05 PM

HMI's even flicker free can flicker & / or have ARC wander. Ideally you should use large tungsten units, the advantage is that you can dim them between takes, it's what I usually use for high speed. You need to use large units 10k or more & use multiple phases.


Hiya Stephen,
Can you elaborate slightly on this?

When you say multile phases do you mean like 3 phase power? Does that overcome issues with say 50hz cycles etc somehow?

Does tungsten light fluctuate on cycles at all?

Apologies if I'm not as clear as I might be, my keyboard is failing!

Thanks!

love

Freya
  • 0

#5 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 31 October 2011 - 03:07 PM

Hiya Stephen,
Can you elaborate slightly on this?

When you say multile phases do you mean like 3 phase power? Does that overcome issues with say 50hz cycles etc somehow?

Does tungsten light fluctuate on cycles at all?

Apologies if I'm not as clear as I might be, my keyboard is failing!

Thanks!

love

Freya


Bigger Tungsten (AC Powered) lights have more amps going through them so flicker less.
If you have 3 x 2 Kk's on different phases any flicker will be minimized, I have sucessfully used Brutes at 500 fps so the theory does work!
  • 0

#6 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 31 October 2011 - 06:19 PM

Bigger Tungsten (AC Powered) lights have more amps going through them so flicker less.
If you have 3 x 2 Kk's on different phases any flicker will be minimized, I have sucessfully used Brutes at 500 fps so the theory does work!


Ah! I think I'm starting t0 get it!
The lights are on different phases so the lights might flicker but while 1 light is on a down cycle the other might be on an up! This way they c0mpensate f0r 0ne an0ther? It evens 0ut any flicker?

S0 in yr examle y0u have 3 x 2 Kk's c0nnected t0 3 phase electric they g0 0ut 0f phase but all at different times! S0 the light remains even! :)

Is that right 0r am I talking n0nsense?

l0ve

Freya
  • 0

#7 Phil Rhodes

Phil Rhodes
  • Sustaining Members
  • 11934 posts
  • Other

Posted 31 October 2011 - 06:32 PM

This was often done in machine shops where fluorescent lighting was in use, before the advent of high-frequency ballasts - otherwise, you can get strange (and in the context of an engineering shop, potentially dangerous) stroboscopic effects where spinning pieces of metal appear to be stationary or have chunks missing out of them. When you have lighting across all three phases, you have peaks at three times the frequency (150Hz for UK mains, 180 for US) and a much higher minimum level of illumination.

Big tungsten lights flicker less because there is a comparatively large amount of metal in the filament and it simply retains heat long enough to even out the peaks and troughs of the mains. You could get a string of (say) 20 halogen dichroic lights at 12V 50W each and it would still be a 2K mains light with about four amps going through it, but because it's a lot of tiny filaments, they could conceivably flicker. Get one big 2K globe and it wouldn't, or at least would less so. And 2K isn't very big, as movie lights go.

P
  • 0

#8 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 31 October 2011 - 06:41 PM

This was often done in machine shops where fluorescent

Big tungsten lights flicker less because there is a comparatively large amount of metal in the filament and it simply retains heat long enough to even out the peaks and troughs of the mains.
P


So it's kind of like they retain their light energy more and so don't cool down so fast, only in light terms, not in heat terms. Something like that?

love

Freya
  • 0

#9 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 31 October 2011 - 06:53 PM

This was often done in machine shops where fluorescent lighting was in use, before the advent of high-frequency ballasts - otherwise, you can get strange (and in the context of an engineering shop, potentially dangerous) stroboscopic effects where spinning pieces of metal appear to be stationary or have chunks missing out of them. When you have lighting across all three phases, you have peaks at three times the frequency (150Hz for UK mains, 180 for US) and a much higher minimum level of illumination.
P


(oh grief! don't tell me the return key is failing now too! At least Ive now found the accesibility options but it's very frustrating for a touch typist maybe it will teach me to be a bit less verbose! Guess I can pretend I'm using an ipad or something!) ;)

Anyway, I think that Stephen is saying that each light is running at a different phase but you seem to be suggesting it's more like those tellies that do 100hz or something? Or do u mean that the 3 lights would combine to make a kind of 150hz wave or something?

Apologies 2 u both if I'm not getting it.

love

Freya

Edited by Freya Black, 31 October 2011 - 06:55 PM.

  • 0

#10 Keith Walters

Keith Walters
  • Sustaining Members
  • 2219 posts
  • Other
  • Sydney Australia

Posted 31 October 2011 - 10:30 PM

The lights are on different phases so the lights might flicker but while 1 light is on a down cycle the other might be on an up! This way they c0mpensate f0r 0ne an0ther? It evens 0ut any flicker?
Freya


Yes and no.
Having 3 lights on 3 phases will theoretically smooth out the light output as you describe, except that that would only be true if all three lights' outputs were fuly superimposed.
In practice, because the lights have to occupy different physical positions, there will inevitably be "shadows" on the subject where the flicker cancellation is incomplete.

However, I don't think that is the phenomenon they're talking about here.
What has been described sounds more like what you see in a Plasma Ball, where you get long unstable streamers of "northern lights" discharges.
A broadly similar thing occurs in high-power discharge lamps, except that the instability is much more violent and rapid, so there is no obvious flicker to the naked eye. Unlike the flicker produced by the AC mains, this is more or less random "noise" so it's not obvious on ordinary film or video cameras, because the fluctuations have time to average out. It's only at very high frame rates that this becomes obvious.

The easiest way to avoid such problems is to go outside and shoot in the Midday sun. No flicker there....
  • 0

#11 Phil Rhodes

Phil Rhodes
  • Sustaining Members
  • 11934 posts
  • Other

Posted 01 November 2011 - 04:29 AM

It's time for visual aids:

Posted Image


Three phases. Three times as many peaks.
  • 0

#12 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 01 November 2011 - 01:41 PM

It depends what your sources are. Flourescents especially, and other discharge technologies, can have regular color variations within each cycle. Flourescents go yellow after the arc stops, because those phosphors glow a little longer. Standard speed shooting averages out several cycles in each exposure, so this doesn't show up. But at high speed, you're capturing the details of the cycle. Tungsten will go a little warm on the zero crossings, as noted, less so for the larger filaments. Back when we ran tungsten on the filtered DC that was provided for carbon arcs, that wouldn't have been a problem. But we didn't have cameras that fast in those days.



-- J.S.
  • 0

#13 Mitch Gross

Mitch Gross
  • Basic Members
  • PipPipPipPip
  • 2873 posts
  • Cinematographer

Posted 01 November 2011 - 03:16 PM

John is correct, that is the color shifting you see.

Many people try the different phase trick when shooting high speed to varying degrees of success. The real solution is as Phil states: a high wattage bulb just gets so incredibly hot that it doesn't have a chance to cool and dim before the next cycle rises in it again. So no flicker.
  • 0

#14 Keith Walters

Keith Walters
  • Sustaining Members
  • 2219 posts
  • Other
  • Sydney Australia

Posted 01 November 2011 - 04:09 PM

"Flourescent?" Is that the ones that produce that white, powder-ery sort of light? :rolleyes:
You mean fluorescent.

It depends what your sources are. Flourescents especially, and other discharge technologies, can have regular color variations within each cycle. Flourescents go yellow after the arc stops, because those phosphors glow a little longer.
-- J.S.

You can illustrate this by flipping a coin under an ordinary iron-ballast fluorescent. As the coin spins, you will see a slowly pulsating colour change as it reflects ("strobes") different parts of the illuminate/decay cycle.

Digital cameras set to auto colour balance will often show the same pulsation for the same reason.

Which is one reason Kino-Flos and the like always use electronic ballasts.
  • 0

#15 Chris Millar

Chris Millar
  • Basic Members
  • PipPipPipPip
  • 1642 posts
  • Other

Posted 02 November 2011 - 02:44 PM

nice trick.

Same applies for spade type drill bits even under the sun.

If you spin them at a gazillion RPM and make the nice flat surface reflect light just so you get a much brighter direct image of the surface for a very short period of time compared to the rest of the cycle (i.e. a strobe) - you can very clearly read what ever is written or stamped on them...

So much so they look stationary (as referred to above in machine shops) - but in this case its to do with the object itself rather than any lighting hoo hah.

It wont work with diffuse light so well...

I'm guessing it's also how occasionally I've noticed wagon wheel effects on car rims with my plain old eyes (and no the cars weren't driving past regularly spaced barrier columns either) - there was me thinking I'd discovered my vision had a frame rate :blink:
  • 0

#16 Freya Black

Freya Black
  • Basic Members
  • PipPipPipPip
  • 4161 posts
  • Other
  • Went over the edge... Central Europe

Posted 04 November 2011 - 12:39 PM

It's time for visual aids:

Posted Image


Three phases. Three times as many peaks.


That truly is an excellent visual aid, you can see how the troughs are far more minor than they might otherwise be! Excellent! :)

Great explanations from everyone too, so a big thanks to you all (Stephen, Keith, Phil, John etc)!

love

Freya

Edited by Freya Black, 04 November 2011 - 12:42 PM.

  • 0

#17 Tom Guiney

Tom Guiney
  • Basic Members
  • PipPip
  • 25 posts
  • Cinematographer
  • Oakland, CA

Posted 07 November 2011 - 10:55 PM

Ah! I think I'm starting t0 get it!
The lights are on different phases so the lights might flicker but while 1 light is on a down cycle the other might be on an up! This way they c0mpensate f0r 0ne an0ther? It evens 0ut any flicker?

S0 in yr examle y0u have 3 x 2 Kk's c0nnected t0 3 phase electric they g0 0ut 0f phase but all at different times! S0 the light remains even! :)

Is that right 0r am I talking n0nsense?

l0ve

Freya



That's the general size of it. Each of the three phases reaches their maximum voltage in the cycle at a different point in time, and so at any given point, one of the three is at maximum output, one is partway down, and one is around minimum output. If you visualize a single cycle of a sine wave, and then make two copies of it and move one 1/3 of a cycle to the right, and the second copy 2/3 of a cycle to the right, that will be a visual demonstration of the output of your lamps over time.
here is an illustration:
http://bit.ly/sgl9Eh
threephases.gif

When the sine wave is at either its positive or negative peak, then the output of lamps attached to that phase are at their maximum. By powering three identical lamps that are set up next to each other doing the same job off the three separate phases, you never have too bad a drop in your output.

Apparently you can use this same technique with HMIs on magnetic ballasts to reduce flutter on non-safe frame rates, but it seems dodgy to me and I've never attempted it. Get aquare wave ballasts.

All this being said,
General rule from experience with high speed is to use the largest filaments that you can, ideally 20ks and never less than 5ks. The bigger filaments take longer to dim down and up and maintain a more constant output. Get 20ks. Or tens. Don't use multiple-1k filament units like maxibrutes or bargers, those little filaments flicker badly. You see it worst of all in practicals in shot with their tiny little 60w filaments.

Don't use HMIs for high speed even as high as 120, you could get pulse like what you showed us on your sample video. 120 is supposed to be safe, but I've seen pulsing with HMIs at that rate.

Tom Guiney
gaffer, dp
airboxlights.com inflatable softboxes for litepanels
lighting tips on twitter @airboxlights
  • 0

#18 Chris Millar

Chris Millar
  • Basic Members
  • PipPipPipPip
  • 1642 posts
  • Other

Posted 08 November 2011 - 08:37 AM

Tom, your graph isn't right...

(and following from that "and so at any given point, one of the three is at maximum output, one is partway down, and one is around minimum output" isn't quite right either)

See Phil's for a more accurate rendition of an ideal 3 phase pattern.
  • 0


CineLab

rebotnix Technologies

Visual Products

Aerial Filmworks

Ritter Battery

Gamma Ray Digital Inc

Opal

Wooden Camera

Willys Widgets

Metropolis Post

FJS International, LLC

CineTape

Paralinx LLC

Tai Audio

Broadcast Solutions Inc

The Slider

Rig Wheels Passport

Technodolly

Glidecam

Media Blackout - Custom Cables and AKS

Abel Cine

The Slider

Gamma Ray Digital Inc

FJS International, LLC

rebotnix Technologies

Aerial Filmworks

Ritter Battery

Rig Wheels Passport

Willys Widgets

Media Blackout - Custom Cables and AKS

Paralinx LLC

Abel Cine

CineTape

CineLab

Broadcast Solutions Inc

Wooden Camera

Tai Audio

Metropolis Post

Visual Products

Technodolly

Glidecam

Opal