What about underexpose a tad bit?

[Emanuel]

Not exactly underexpose by default but some underexposure techniques can help. As for instance, from what I've seen the GH5 footage can really shine especially if you have the highlights under control...

 

What about your two cents?

[Thpriest]

i do the same

[OliKMIA]

Same here

[Shirozina]

What about just call it the 'correct' exposure?  On an internal codec the Zebras are linked to the green channel which means you can easily clip the red channel in skintone highlights without being aware of it so knocking the exp down a bit from an ETTR starting point is a good way of avoiding this. Then again ETTR is not my idea of a good way to expose non RAW images anyway - yes you avoid the noise but you also end up placing certain critical tones ( skintones) on different parts of the profile curve depending on your lighting contrast which will make shot matching in post more of a problem.

[Ehetyz]

This is taking me back to memory lane, lol

I used to underexpose back in the day quite a bit, mostly to protect the highlights, but also because imho on older HDV/HD cameras like the HV30 everything looked much more filmic when underexposed in the Cine gamma mode. I did this on the 550D as well, protected the highlights and underexposed by one stop. I got some pretty unique looking results that way.

This was a feature I did on HV30.

And this was on the 550D. Of course the subject matter in both is well suited for, well, a murkier look.

Haven't found any reason to do that with any recent cameras though, and with Blackmagics I've swinged to the other side and stick to ETTR most of the time.

[jonpais]

2 hours ago, Shirozina said:

What about just call it the 'correct' exposure?  On an internal codec the Zebras are linked to the green channel which means you can easily clip the red channel in skintone highlights without being aware of it so knocking the exp down a bit from an ETTR starting point is a good way of avoiding this. Then again ETTR is not my idea of a good way to expose non RAW images anyway - yes you avoid the noise but you also end up placing certain critical tones ( skintones) on different parts of the profile curve depending on your lighting contrast which will make shot matching in post more of a problem.

If you reduce saturation by -2, it can help avoid clipping the red channel. 

[tihon84]

I can say only about vlog:

you have to be near 42IRE using graycard. It gives you maximum DR. So, its underexpose - 0.3

it gives a little noise in shadows, but good highlights, DR, and colors

[Mat Mayer]

This is interesting and I assume because the GH5 has a higher bitrate and whatnot. Because under exposing on my G7 tends to ruin video by introducing macro blocking.

[Shirozina]

Just got my V-Log upgrade and did a quick studio test against Cinelike D (contrast -5) to see the DR difference and it's quite significant! Didn't realise the zebras max out at 80% in V-Log. Will have to test in the real world as that's a lot of tonality to expand out again on a highly compressed codec.......

[Benjamin Hilton]

This is something you really need to test for yourself.  Every camera will react differently to under/over exposure, and the variable of exposing with a hardware light meter vs histogram vs internal meter vs zebras vs false color will make a big difference.  

If you can, setup several scenes in the typical environments you will be shooting in; low light, back light, cloudy etc.; (It is probably best to have a person in the shot as skin tones are the most important thing to monitor.)  Using the metering system and picture profiles you will normally use, expose your camera in 1/3 stop increments from 3 stops under to 3 stops over.  In post grade these clips to normal and look through them all side by side.  Also test different white balances, ISOs etc.; doing this you will learn a lot about your camera.

I have been testing the GH4 and G85 in this way in many different environments and with several different picture profiles for quite some time.  I test, then use in in the real world, then test again.  After hundreds of videos filmed all around the world in some of the harshest of filming environments, I feel like I have really hit the sweet spot with these cameras. 

This is a quick project we just released yesterday shot with the GH4 and G85.  Filming and editing was done in two days; shot entirely in natural light run and gun. 

 

 

[Mat Mayer]

9 hours ago, Benjamin Hilton said:

This is something you really need to test for yourself.  Every camera will react differently to under/over exposure, and the variable of exposing with a hardware light meter vs histogram vs internal meter vs zebras vs false color will make a big difference.  

If you can, setup several scenes in the typical environments you will be shooting in; low light, back light, cloudy etc.; (It is probably best to have a person in the shot as skin tones are the most important thing to monitor.)  Using the metering system and picture profiles you will normally use, expose your camera in 1/3 stop increments from 3 stops under to 3 stops over.  In post grade these clips to normal and look through them all side by side.  Also test different white balances, ISOs etc.; doing this you will learn a lot about your camera.

I have been testing the GH4 and G85 in this way in many different environments and with several different picture profiles for quite some time.  I test, then use in in the real world, then test again.  After hundreds of videos filmed all around the world in some of the harshest of filming environments, I feel like I have really hit the sweet spot with these cameras. 

This is a quick project we just released yesterday shot with the GH4 and G85.  Filming and editing was done in two days; shot entirely in natural light run and gun. 

 

 

Any chance of sharing the settings you used for that video or in general?

[Orangenz]

17 hours ago, Benjamin Hilton said:

This is something you really need to test for yourself.  Every camera will react differently to under/over exposure, and the variable of exposing with a hardware light meter vs histogram vs internal meter vs zebras vs false color will make a big difference.  

If you can, setup several scenes in the typical environments you will be shooting in; low light, back light, cloudy etc.; (It is probably best to have a person in the shot as skin tones are the most important thing to monitor.)  Using the metering system and picture profiles you will normally use, expose your camera in 1/3 stop increments from 3 stops under to 3 stops over.  In post grade these clips to normal and look through them all side by side.  Also test different white balances, ISOs etc.; doing this you will learn a lot about your camera.

I have been testing the GH4 and G85 in this way in many different environments and with several different picture profiles for quite some time.  I test, then use in in the real world, then test again.  After hundreds of videos filmed all around the world in some of the harshest of filming environments, I feel like I have really hit the sweet spot with these cameras. 

This is a quick project we just released yesterday shot with the GH4 and G85.  Filming and editing was done in two days; shot entirely in natural light run and gun. 

 

 

Nice video but zero calorie post. You could write horoscopes. You are basically saying shooting everything 18 times and keep the best take, you face different light environments, you can use different cameras, you can use different ways to light meter, you can use different isos, and you can't actually help anyone because they have to work things out themselves despite having shot some films outside the USA in the rain. 

@Emanuel Lift the shadows by exposing to the right if you can as this will reduce noise in post and you can drop them back down. If you are not seeing noise then lift in post to your hearts content. Use zebras to not blow the highlights. If the highlights are the main subject then don't have them at 100 like normal, especially if they are someone's face. 

[Benjamin Hilton]

I am sorry for not being clear, but this is not what I am saying.  I am saying take the half a day to test each camera you will use yourself with the metering method you will be using in the field.  Then when you are in the field, you will know exactly how to expose with your camera so you can get it right the first time.  

[HockeyFan12]

I just wanted to add that if you expose cameras like the C300 and F5 with an incident meter at their given ISOs (850 and 2000, respectively) the image is a lot darker than most of what we see online (ungraded sample footage, that is). Given your profile picture, I assume you're using an incident meter, but I think people have gone a little too crazy with ETTR and not using a meter and that's responsible for some of the "video look" stuff you get, particularly with cameras that struggle with chroma clipping. (Weirdly, I see DPs underexposing the Alexa all the time, and it's the one camera that has its saturation vs IRE curve just about perfect. I think I just see a lot of bad exposures!) 

I do think the F5 looks terrible exposed at 2000 ISO, though. :/ And haven't used the GH5 but the GH4 seems to respond to over and underexposure better than Sonys at least.

[Benjamin Hilton]

12 hours ago, Mat Mayer said:

Any chance of sharing the settings you used for that video or in general?

I would be happy to share my filming workflow for the G85 and GH4.

White balance:  This is a major key.  When the white balance is set too warm with these cameras it can be very difficult to pull back in post due to shadow contamination and red channel clipping.  Setting your white balance leaning toward the cool side can greatly improve your skin tones. 

Exposure:  The way I set exposure on fast shoots is with the camera's built in spot meter; also sometimes with false color when I can get it.  I have found exposing the key light on skin tones to 1 stop under (3 bars on the meter) to be the sweet spot for healthy looking, non plastic, skin tones. 

Picture Profile: (This one might be a touchy one:-) I use the natural picture profile.  I have put in many hours extensively testing natural vs cine like D in different lighting situations and natural wins hands down for me in daylight.  

Reasons for this:  Cine Like D actually doesn't have that much more dynamic range than natural.  It can appear that it does have more range, but this is simply because the camera drops the exposure by about a stop when you switch the profile to Cine Like D.  After you bring the exposure back up 1 stop to match natural, you actually have about 1/2 stop less room in the highlights then with natural;  you do gain about 1/2 a stop of shadow detail though.  Natural on the other hand, gives you about 1/2 a stop more highlight detail and a much gentler highlight roll off; at the sacrifice of losing about 1/2 a stop of shadow detail.  For me personally, highlight roll off and skint ones are much more important than shadow detail. 

For the reasons above I shoot with these picture profile settings:

Natural

Contrast -2 (Any lower and skin tones start to lose the natural look)

Sharpness -3

Noise Reduction -5

Saturation -2 (This greatly reduces clipping in the red channel)

Curves at +1 Shadows, +0 highlights

iDynamic set to Low (This give a little bit of lift in the shadows without introducing noticeable noise)

 

Those are my daylight settings.  Because Cine Like D handles shadows so much better, I shoot with it when I get in a low light situation.  This gives much cleaner high ISOs than natural.

 

Like I mentioned in my first post, these settings are my personal choice after a lot of testing and filming.  I would encourage anyone who shoots on a regular basis to get and test them for yourself with your camera, lighting, and lenses.  Find your sweet spot and run with it:-)

13 minutes ago, HockeyFan12 said:

I just wanted to add that if you expose cameras like the C300 and F5 with an incident meter at their given ISOs (850 and 2000, respectively) the image is a lot darker than most of what we see online (ungraded sample footage, that is). Given your profile picture, I assume you're using an incident meter, but I think people have gone a little too crazy with ETTR and not using a meter and that's responsible for some of the "video look" stuff you get, particularly with cameras that struggle with chroma clipping. (Weirdly, I see DPs underexposing the Alexa all the time, and it's the one camera that has its saturation vs IRE curve just about perfect. I think I just see a lot of bad exposures!) 

I do think the F5 looks terrible exposed at 2000 ISO, though. :/ And haven't used the GH5 but the GH4 seems to respond to over and underexposure better than Sonys at least.

I agree.  Each sensor is different.  Some sensors love over exposure, while others like the Panasonic cameras, really don't like it.  ETTR with the GH4 in my opinion, can easily kill skin tones with red channel clipping. 

You can see in some places in this video here, especially at 3:36, where I accidentally ETTR ed with the GH4; skin tones are kind of one golden block:-)

 

[Mat Mayer]

Thank you @Benjamin Hilton this is all great information. I too use Natural but just wing it with the settings based on the scene and the lens. I have been having problems when underexposing. Currently trying to edit a very dark scene which is taking forever, so next time I will try Cine D to help with the shadows. I have not tried the curves yet so will also put that up. I feel like once I upgrade to a higher bitrate camera like the GH5 it will be much easier to avoid artefacts and macro blocking and whatnot from stretching the codec too much. 

[Benjamin Hilton]

1 hour ago, Mat Mayer said:

Thank you @Benjamin Hilton this is all great information. I too use Natural but just wing it with the settings based on the scene and the lens. I have been having problems when underexposing. Currently trying to edit a very dark scene which is taking forever, so next time I will try Cine D to help with the shadows. I have not tried the curves yet so will also put that up. I feel like once I upgrade to a higher bitrate camera like the GH5 it will be much easier to avoid artefacts and macro blocking and whatnot from stretching the codec too much. 

Definitely will help.  The key is to only expose by 1 stop, 2 stops and it gets noisy really fast.  I have a GH5 coming in tomorrow, looking forward to testing 10 bit V-log! :-)