Levin Liebig

I usually just use toilet paper or something similar to change the bulps. The important thing is that no oil of your skin comes in contact to the glass.

Most redheads have a grid in the front preventing shards from the bulb flying into your actors face. (That sounds really bad :huh: but that is not likely at all.)

One more important thing is to remember to relief the strain on the cable. So if someone trips over the cable, the lamp will not fall. It will also help to prevent the cable braking where it enters the case of the lamp.

If you choose to underexpose your picture, remember to always put one spot of brightness into your frame. This could be a practical in the background or a rim/shoulder light or something else.

Rather than ETTR for every shot you shoot.. with the EI mode (exposure index) I think is !.. in the camera you can then change the ISO, from native.. in combination with a LUT in the VF... you can uniformly "over expose" from the native ISO.. (same as you might rate film a different ISO on a light meter).. I wouldn't ever go more than 2 stops.. but alot of people will always over expose their log by 1 stop..to raise the noise floor.. then you are doing it for all your footage at a set uniform level.. not changing it for every single shot which will be a nightmare to grade.. e.g. on the Sony f5.. in Cine EI mode. native ISO 2000.. with MLUT 709 (M means monitoring ).. change ISO to 1000.. VF goes 1 stop darker.. levels go down .. open iris to compensate back to correct levels.. or just by eye.. but your native LOG stays at 2000.. so you are over exposing everything by 1 stop.. you loose one stop in the highlights but with 6/7 stops over grey you can usually afford to do this.. and you have raised the noise floor by 1 stop..

 

That sounds like a good aproach, Robin. Expose LOG with a rec709 LUT and using the ISO for "over-exposing". Thank you.

Its simple to trouble shoot this, but chances are 800 ISO on the camera is not 800 ISO. Camera manufacturers definitely lie about ISO so they can say clean 2000! when your really looking at 800 on other camera. Get a second meter if you want to be sure and take incidents and spot readings together to see if yours is off but I doubt it. Check the camera next, expose with the camera first. Get your f stop with the waveform and grey card to match 34% and compare the f stop the meter, adjust the iso on the meter until it matches the camera f stop. It might end up being 400 or 500 iso, if it is to get the exposure your are after you have to rate the camera lower on your meter. Usually the camera manufacturers' suggestions are super low for log middle grey to promote high light retention specs, but the darks will be muddy. I wouldn't trust that either. Any time you have to lift digital your not in the best place. For cinema a lot of cinematographers like white skintones at 45-55. If you like it higher and you have to lift your image to get there from 34% grey, and you're loosing some fidelity. You might need to aim for 43% for middle grey in that case. Trust none of the specs of the camera, trust only the sekonic as a bench mark, you might find that 200 iso is the look your going for. Best of luck.

This might explain our results.

Use Zebra or WFM.. all digital camera,s have these.. TBH I usually just go by eye.. but Im not shooting big feature films .. but guys like Deakin's .. do it too.. if your monitor is calibrated and you know the grey /white levels of what ever MLUT your using.. you can really work off it .. or even the EVF if its a good one.. you have to know the levels of your LUTs obviously.. thats why 709 is the easy one.. I guess it also depends how much 709 you have already shot..

 

But its not the only way.. I find it easy and I would say its probably the easiest if you have never used LOG and just want to get shooting..there is alot of voodoo about LOG but its not rocket science at all.. its pretty straight forward.. dont be scared of it.. its pretty forgiving actually in the high lights, its the shadow that will give you trouble .. the opposite of standard gammas where the golden rule was to protect high lights... .you can learn all you need to know in an afternoon on the inter web..

 

If your used to meters then you should stick with them.. some people do.. thats why camera,s have EI mode and ISO readings I guess too..but 709 LUT ,zebras for 70% skin.. 90 % white .. grey % isnt hugely different really.. its really white that is dragged way down from 90 to the 60,s.. and yes your LOG non LUT footage will definitely look under exposed..its meant to.. thats why I wouldn't really work off it solely .. the hardest thing is often finding a grader who knows how to work with LOG.. you cant just do it on a NLE.. you need something like Resolve.. than your shooting of it !.. alot of editors are being made to grade too.. which was easy enough on standard gamma curves.. but log will behave differently and can end up being butchered .. have a look at Roger Deakin's site he talks about shooting LOG ..

Yes I know that using a LOG pushes all values on the waveform down in comparison to Rec.709 but I really don't want my LOG footage to be underexposed because, as you know, this comes together with a lot of unwanted noise in the shadows. And from the test we made I can say the footage was so underexposed that we could not save it in Resolve. Thats why I wanted to find out how to expose correctly. Either ETTR or with a gray card. But the gray card did not answered our questions at all, hence I started this topic.

So you recomend to use a Rec.709 LUT on the monitor/viewfinder together with a waveform (also with the LUT applied.) and expose only using IRE values of 70% skin and 90% white? And does mean 90% white, when shooting regular Rec.709 (no LOG), that there is still some information left or is it already burning out?

I'm not sure if it's possible to show a waveform with a LUT applied to it. I have to try it the next time I have access to the camera. But according to you, I can use the LOG waveform on around 60% IRE for white?

Thank you for your replies.

There is alot on the inter web about exposing LOG.. and of course there is more than one way.. dont ETTR.. this is from the stills world and some of those guys think they should expose LOG that way too.. it will be a pain in the grade.. you actually want consistency,ie the opposite.. not every single shot cranked over to the right ..with your levels all over the place.. a uniform 1 or even 2 over exposure is quite often used with Log.. to avoid noise..(you usually have quite a bit of leeway in the high lights )..but Log can get noisy in the shadows due to not much data down there.. depends a bit of your scene.. its a way of lifting the noise floor..

 

Personally I dont use a metre for video.. the simplest way to do it is to apply a Rec709 LUT to your EVF and /or monitor .. expose for that at the usually 709 levels.. white 90..etc.. and your underlying LOG will be perfectly exposed.. you can sometimes switch off the LUT if you want to see how the high lights are keeping.. some camera,s have high/low functions to make that easier..

 

Okay I understand that I prevent clipping in the highlights and in the shadows with this approach. But how do I know if I have consistency in the midtones in different lighting situations for example in the skintones?

Wouldnt you have gotten 34% instead of 26% if you lowered the ISO on your meter? Maybe the camera isnt 800 ISO in C-Log, its more like 500 ISO?

As far as I know, canon says that the best setting for c.log3 is 800 ISO. That's why we set it like this in camera and on the seconic.

Hey!

Some friends and me tried to expose log with an 18% gray card.

We used a canon c300 mk2 with the c.log 3 cinema gamut picture profile.

Canon gives the following IRE values for c.log3: 18% Gray → 34,3% | 90% White → 56.4%

But before we tried to use canons values we tried to replicate this workflow to derive the values:

https://www.youtube.com/watch?v=WZb3u220EwU

So we set our camera and seconic light meter to 800 ISO and used the spot meter right next to the camera on the 18% Gray chart. We got a value of F 5.6⁸.

Our next step was to set the lens to a F-stop of nearly F 5.6⁸. The waveform gave us a value of 26%.

This is not only a big difference to 34,3%, the value of canon, the picture was clearly underexposed. :huh:

Of course, we asked us what our mistake was.

Was it the fault of the video? Is our gray card not standardized? Is our light meter broken? Is a F 5.6⁸ not a F 5.6⁸ on the lens we used, maybe its broken? Is the camera screwing us?

Then we stumbled over this article: http://bythom.com/graycards.htm

The author, Thom Hogan, tells us that cameras don't see 18% gray. Instead they see 12% gray.

This article confused us a lot. We don't know what to think about light metering. :unsure: What is true and what is a myth?

What do you think about all this? Are you noticing a mistake we made? How is your workflow for exposing log gamma curves? Do you use a gray card, a light meter, ETTR (expose to the right)? Do you use a Bt.709 LUT on your monitor/viewfinder and expose just with that?

And do you understand the 12% gray article and do you think it's true/false?

Thank you for your help! :)

Levin Liebig

Hello Anthony Liu.

There're lots of different gels available. Some of them with a blue-magenta tint. Also it is possible to use more than one gel on a light. Which gels they used exactly, is hard to tell. If you want to go for a specific look take a camera and try different gels and combination of gels.

Be aware that some effect gels might look absolutely fine with your eyes but not on the sensor/film material.