Tom Morrow

My rule of thumb is to put gels on the tungsten source instead of the daylight source when trying to achieve a particular balance.

Yes gels will tend to lower the cri if used on non black body sources because they are introducing further nonlinearities into the spectral response. How much this is an issue will depend on subtle interactions between the spectral response of gel source and sensor so there is no formula. If you are adding ctb onto a daylight hmi I doubt it would be an issue. But I have seen horrible results adding cto to cheap daylight LEDs that looked ok without gels as daylight. They totally didn't match tungsten at all. Also mediocre results for adding plus green to low cri daylight to match green spiky practical flos. Even reducing green spikes in high cri practical Flo lights with minus green gels has given inferior results versus skipping the minus green and balancing to white in post. This is why I avoid gels on daylight sources.

I have found a loose connection around the VEAM connector on my Arrisun 1200 plus head; when I wiggle the cable from the head to the connector, the HMI sometimes blinks off and tries to restart. I'm pretty sure one of the solder joints inside the connector has failed, so I'd like to take off the connector and resolder it. But I can't figure out how to disssemble the VEAM connector which looks like the left one in this pic: http://www.mldvideo.com/wms/api/image/product/2336/2286/600/600/image.jpg I can screw off the tension relief on the end but the body stays in place. It looks it might take an expensive spline tool that I don't have. I might be better off just bringing this into a service center, but I thought I'd see if anyone knows a DIY way to open this connector?

Ordered a sel/xs lamp from b&h. The lamp that I received said 1200w/sel but didn't say xs anywhere on it. I suspect that the xs is not supposed to show up on the lamp. I also learned that uvs is replacing xs so the next lamp I buy will be a 1200w/sel/uvs. Uvs means lowered uv output. Anyway I now cannot reproduce the flickering issues with the old or new lamps now. I am not sure if the flicker just got burnt out or what but the flicker that was occasionally very bad is now not in evidence. I am guessing the capacitors are marginal enough that I should probably replace them but for $750 or so that Arri said new caps go for I figure I might as well wait and not replace until the flicker comes back.

As I read ville's post he was thinking of shooting in daylight camera balance for the flos then correcting the 3.2k incandescent (high cri means no green spike to me) on faces to white with ctb in post. I think where David was going was that just shooting with camera at incandescent balance would achieve the same effect more simply.

I got an answer to my question about the different lamps: The SE designation stands for ‘single-ended’, while SEL stands for ‘single-ended lamp’. Those are just the terms used by each manufacturer to differentiate their products. The SEL/XS however does have a greater lamp life of approximately 1000 hours due to the eXtreme Seal technology to help withstand excess heat. On the other hand, the SE lamp only has a lamp life of approximately 750 hours.

As I understood it, Ville was talking about shifting the talent's face with a CTB type blue shift in post. It sounds like he was using the word hue in the colloquial sense of changing the color, not the more precise way that hue means moving along the magenta-green axis to camera folks. Shifts along the Warm/Cool (Orange/Blue) axis are called changing the Color Balance, or Kelvin Shifts long the Magenta/Green axis are called changing the hue. But in art classes and computer graphics (including photoshop), Hue changes mean going around a color wheel in any direction to any destination (as distinguished from saturation and brightness axes). Sometimes terminology gets in the way of communication!

Sounds like a good plan. If your scene has both green-spike flos and incandescent light, you will need to correct the green at the flo source not at the camera or post. Otherwise you will accentuate the magenta of the incandescent if correcting with minusgreen in camera or post.

I see that most websites say the correct bulb for the Arrisun 1200 Plus HMI head is the Osram 1200W/SEL/XS. But the bulb that Arri recommends in their literature is the 1200W/SE, and the bulb that was in my head was marked Osram 1200W/SE, not SEL/XS. Does anyone know if the SEL/XS is the same as the SE, and if not what is the difference between the two? I sent this question to my dealer but figured perhaps someone on the net would know better.

Thanks Matt. I went to my local supply house and spoke with the guy who services Arri lights. He said showed me the label of the moly lube that Arri has recomended to him and it looked like this: http://www.amazon.com/Sprayon-S00200-OZ-MOLY-LUBE/dp/B00L42R48E When I search for Sprayon S00200 I get this which I assume is the new label: http://www.sprayon.com/product-categories/industrial-lubricants/dry-film-moly-lubricant-aerosol-lu200 But he also showed me graphite powder and said that mixing it with isopropyl alcohol allows it to be applied without making too much of a mess. So based on that and Matt's post I bought some graphite powder at my hardware store and will try that. I've been continuing to test the HMI and ran it three hours with only one brief irregular flicker period horizontally, no flicker most of the time. But when I direct the beam straight upward toward the ceiling I get some visible regular flicker, which seems perhaps 30Hz or less. So my next step in diagnosing this is to buy a new lamp and see if the flicker goes away with that.

Using digital color balancing (in camera or in post) is becoming a common way of dealing with flourescent lights. In the early days most overhead flos had horrible green spikes and they were fairly consistently horrible in the same way. So a minusgreen filter was a good "one size fits all" solution. Today there are many different types of flos in use in commercial environments all with different color balances and amounts of green spikes. More and more often the green spike is becoming smaller as manufacturing techniques improve and more R&D is going into making flos look like incandescent spectrum. Correcting digitally allows you to dial in just the right amount of correction for how your particular camera's sensor spectral sensitivity reacts to the particular flos you are shooting under. And if you are shooting RAW you can wait for post. The downside to doing it fully digitally is that a green spike (or imbalance in Red or Blue) could exceed or approach the sensor's dynamic range limits and so not be fully correctible. And if you have mixed lighting you will have to do something to get them to match if that is your goal, no matter if the correction is in the digits or a filter in front of the camera. More and more the easiest thing to do is just shoot without doing any color correction and see if it needs it in post.

I guess I can answer my own question here. I read the HMI chapter in Set Lighting Technicians Handbook, and learned that yes magnetic ballast HMIs basically rectify the 60Hz sine power into 120Hz half-sine pulses. So no the flicker is not 60Hz but there is regular flicker. As to my irregular flicker problem, I did a lot of work and I think I solved it. - Determined that the green ground wire in the outlet cable had a break in it, and replaced the cable. - Determined that someone had swapped hot for neutral in the NEMA5-15 plug so it was wired unsafely, and fixed that while replacing the plug. - Noticed a 220 microF capacitor on the logic board was bulging out as if it might be ready to blow, and replaced that. - Blew and wiped dust out of the ballast wherever I could. None of these things seemed to affect the irregular flicker, so continued... - Unscrewed the reflector and blew and wiped dust out of the lamphouse - Unscrewed the bottom plate of the head and the transformer cover, and blew and wiped dust out, while jiggling terminals to make sure they were seated well. It seems that the blowing and wiping dust out was what did it. I ran the lamp for four hours and only towards the end am I starting to see just-barely-maybe-perceptible irregular flicker. I think it might be associated with the lamp orientation (nipple up IIRC). So I will go in with a shopvac and make sure all the dust is out and reverse the lamp. I suppose the high voltages inside the lamp socket especially could make arcing flicker, assisted by the dust of 20 years. Question: can I walk into my film/lighting supply house and ask for lubricant that will withstand high temperatures so I can lubricate the rails to stop the squeaking when adjusting between wide and flood, or is there some special sauce I should know about? And is there a suggested orientation for the lamp (nipple up or down)? I suppose it probably doesn't matter since it looks like these lamps can supposedly be used in any orientation from pointing straight up to straight down.

So here's my more general question: is it normal for HMIs to have visible ~60Hz flicker or is that a sign that the waveform going to the lamp from the mag ballast has an unusual/defective duty cycle with too wide a gap between pulses?

Thanks for that. Sounds like the issues you are talking about are generally high resistance from bad connections. I will try taking the lamp out and putting it back in the socket, to see if perhaps it had a bad connection. I do note that there is some pitting on the aluminum head which leads me to think these might have been exposed to rain at some point. BTW that link doesn't work for me on iOS but I assume I can see the same thing by going to the HMI lights with Magnetic Ballasts section in http://www.screenlightandgrip.com/html/emailnewsletter_generators.html What I got out of that is that it sounds like mag ballasts are supposed to output power pulsed at 120Hz to the lamp, probably by doing a full wave rectification on the AC 60Hz sine. I remember back in the old days of CRT monitors that I could see flicker at frequencies that many people could not. I can't remember if the cutoff where I started not being able to see flicker was slightly below or above 60Hz but it definitely was not anywhere near 120Hz. So I'm thinking perhaps part of the rectifier could be defective leading to 60Hz output instead of 120Hz. I suppose I could record some highly overcranked video to check that out.

I just bought my first HMI light, a used Arrisun 1200 Plus with magnetic ballast. I am seeing flicker with my naked eye which I wasn't expecting. I am curious if others have seen the same issue. Here is the note I sent to the experienced gaffer that I purchased it from. He said the light was about 20 years old and he has owned it the whole time.... I want to ask your advice about the Arrisun 1200 HMI that I just bought from you earlier today. As soon as I fired it up in my dark apartment I noticed that it was flickering in a way that was easily visible to the naked eye. The flicker was reminiscent of old televisions or old fashioned mag-ballast office flourescent lights, perhaps about 60Hz. I don't remember having noticed this with other HMIs I've seen on set, although most of them were electronic ballast. This persisted over the half hour I had the light on. Additionally there was a sort of irregular flickering every now and then sort of like a candle flickering in the wind which anyone would have noticed. This did show up on my camera tests too, although more subtly than to the naked eye. I'm guessing I didn't notice it in your shop because the sunlight filled in the dark-flickers and I was looking for banding in my test clips not actual flicker. ------ He wrote back kindly offering a refund if needed, and offering to let me test with different heads in his shop next week. He also mentioned that when he has seen such problems it was due to the circuit that it was plugged into. My apartment where I plugged it in was constructed about 10 years ago with all 20 amp circuits and outlets so I doubt it's the circuit itself, although I will have to check to see if it might have been on the same circuit with my computers and switch-mode-power electronics. I'll be doing more experiments Monday or Tuesday, and I'm open to any suggestions.

Thanks Jim for the nuanced discussion of what exposure means. I don't want to beat a dead horse, but want to close the loop and note that it was your comment: That led me to think you were saying that changing ISO didn't change the exposure. If I understand correctly, that statement could be rewritten as ... rather than reducing light on the scene, stopping down the aperture, or reducing the shutter speed, you lower the ISO of the sensor to darken it. Anyway the point is well taken that darkening by decreasing the ISO is, all things equal, better for noise reduction than darkening by reducing the light falling on the sensor.

Interesting distinction between exposure and ISO. That prompted me to look up the definition of exposure, which is basically the amount of light falling on the sensor. So technically, and as Jim used the word, changing the ISO does not change the exposure.

I prefer the predictability of not over lighting. That way you see what you are getting in the camera. It really comes down to how much you want to work to avoid noise. When I was doing still photography I would over light and reduce in post to get fine art super low noise. With the camera on a tripod and static scene over lighting was just a matter of using a slower shutter. But with video I don't because over lighting involves - extra time for bigger lights - easier to clip highlights - lower dynamic range possible. If you are shooting a dark scene there is often a candle or flashlight or other bright light source in frame and not blowing that out may be more important than shadow noise. Given that you only get one shot at the best performance with video I dont want to waste time obsessing with the waveform monitor. It really comes down to your skill and trust of your end to end workflow. You can over light and get lower noise if you really have everything under control but I am not at that level. Also reducing the lighting in post means you shift highlights that were in the nonlinear part of the curve down into the linear part of the output curve leading to more likelihood that things won't look the way they did on the camera monitor. If you are super knowledgeable about exposure curves you can work around or even straighten this out but for most folks easier not to try.