Best Practices. Obtaining Proper Exposure. Share your methods please.

[daniel caruso]

Thanks guys, and Noah. Your previous posts in this thread had some great insight to exposure, thanks for this input. I did some tests tonight and got some great results at 800 in gamma2 then RCX grading. Tomorrow I will try your suggestions and see what works. I did find that my digital meter(L758cine) looked better when metered for 320 on the meter and 800 on camera. Have to check into that still as well. Thanks again!

[Lliam Worthington]

Great post Noah.

[Will Keir]

You can always just apply a redlogfilm look to your footage later in just a few seconds using RC-X.

I never heard of anyone SHOOTING in relogfilm. Who would do this and why? I've heard a lot about people using relogfilm in Redcine x.

I would say get it by turning down your iso, not depriving the sensor of light just so you can shoot ISO800 underexposed and not use the cleaner top end of the sensor's range.

"depriving the sensor of light"

So are you saying the histogram will contain the most information when the aperture is as open as possible? So even in well lit situations, you want to open the aperture as far as possible and use ND filters to open it a bit more?

[Noah Yuan-Vogel]

I never heard of anyone SHOOTING in relogfilm. Who would do this and why? I've heard a lot about people using relogfilm in Redcine x.

Neither have I, but it seemed the user I was responding to was doing that. He did not really offer an explanation as to why, so I imagine it is possible it was just due to not fully understanding best best practices, which is exactly what this thread is for. On the other hand, I am strongly considering using Redlogfilm myself but with a corrective user curve (I plan to test this soon). Personally I noticed redlogfilm holds up a bit better as far as consistent dynamic range at all ISO modes, which does cause me to tend to lean toward it as a base for my footage. My concern with this is that without proper LUT support, the user curves that can be applied in camera (which only consist of 5 points) may not be powerful enough to de-log in camera, but I think it is worth a test to see if this is a good system for getting the most out of the system and seeing pleasing/accurate images while monitoring.

"depriving the sensor of light"

So are you saying the histogram will contain the most information when the aperture is as open as possible? So even in well lit situations, you want to open the aperture as far as possible and use ND filters to open it a bit more?

It depends on the scene, but generally assuming your camera is always capable of capturing 13stops (lets say) if your scene is contains only 9 stops, it is better data/image-wise to expose such that those 9 stops are sitting at the sensor's highest possible range (by opening aperture, shutter, removing ND etc) since then your scene is then represented using more bits in the 16bit container and more safe from being exposed to the noise level that increases toward the darker range of the camera. Potentially, in RED's 16bit linear REDCODE container, your 9stops would be represented in the value range of 128-65536 per photosite if you expose to the right by opening up as much as possible, whereas if you close down 4stops and use the lowest range of the sensor, it will be represented only with the values 8-4096 per photosite. One may look properly exposed at ISO 200 while the other looks properly exposed at ISO 3200. Of course this is a slight oversimplification in theoretical terms in an extreme range, and in practice it is not always useful to overexpose since you may need not have the lights or the stop or the control to know something hotter might not pop into frame during the shot that increases the range of your scene requiring more underexposure to avoid clipping. It doesn't usually make sense to shoot anything at ISO200 on MX and isnt really necessary since REDCODE gives you enough room that you dont really need to expose all the way to the right all the time.

But yeah generally, even in "well lit" situations, you'll get less noise and better use of REDCODE's bit depth by opening up and ISOing down but at the loss of highlight protection. It is a balance that is worth considering and understanding when you set your exposure. Then again if it really is "well lit" you'll be setting your exposure for your highlights and exposing to the right naturally anyway. The tradeoff is usually of greater significance as you work in lower lighting conditions or higher contrast situations.

[Will Keir]

Noah, are you using REDlogFilm or REDlog in Redcine x? REDcolor1 or REDcolor 2?

I haven't used REDlogfilm because of the milky look, but I have been hearing more and more that using it is the way to get the most out of your pictures. Do you know a good tutorial for using this method? Out of the box, usually REDgamma2 works great but I am happy to improve.

[Noah Yuan-Vogel]

Noah, are you using REDlogFilm or REDlog in Redcine x? REDcolor1 or REDcolor 2?

I haven't used REDlogfilm because of the milky look, but I have been hearing more and more that using it is the way to get the most out of your pictures. Do you know a good tutorial for using this method? Out of the box, usually REDgamma2 works great but I am happy to improve.

I'd stay away from the older curves (redlog/redgamma) since they do not seem to scale as well at various ISOs and may throw away some highlight detail. Older redgamma gets too clippy at higher ISOs and too muddy at lower ISOs, Redgamma2 seems to do much better in those situations.

REDlogfilm is great for people who like grading at least partly in log and it's great for getting the most out of your RAW for people using transcoding workflows. REDlogfilm also seems to me to hold dynamic range most evenly and reliably through all ISO settings but only by a little bit. It seems to me that for many people who don't like tweaking or don't fall into the first two categories of log grading or transcoding, redlogfilm may just complicate things and isnt necessarily something I'd recommend for everyone/everything because it takes a bit more work to make it look right in camera or in post.

[Sascha Rossier]

Thanx 4 this thread. Thanx for the enlightment!
Thinking back, all the problems i encountered in post with my footage suddenly make sense.

One further Question:
If you have a shot that - by its nature - surpasses the dynamic Range of the Cam and
lets assume HDRx is no option because of fps-restrictions on the Scarlet and lets further
assume there is no option to fill teh shadows with a bounce or a lightsource. Would you rather
allow to clip highlights or allow noise to be present. I know this is also an artistic
decision but if this was a binary decision like lets say: Plague or Cholera ... what would you opt
for?

[Noah Yuan-Vogel]

Would you rather
allow to clip highlights or allow noise to be present. I know this is also an artistic
decision but if this was a binary decision like lets say: Plague or Cholera ... what would you opt
for?

For me it really depends on the case and the content. Probably more often I would expose for highlights given the chance and let subject go into silhouette if necessary since that can look very nice in its own way, but there are plenty of situations I would just let it blow out especially if having well exposed talent is of the greatest importance for the particular content.

[Chris McKechnie]

Amazing thread!

[Chris McKechnie]

Ohh and I forgot to ask Noah...what about a severely underlit scene...do you keep ISO at 800, or do you push it up to let's say 2500 and crush the blacks in post then process it through something like neat video?