I've often seen people flatten out a video to nothing and then correct it in post. They'll use D-Log color profile, -3 contrast, -3 sharpen, -3 saturation. Then they'll stretch the footage in post. Shouldn't the goal be to get as much range covered from the blacks to the whites without clipping? That way we don't have to artificially stretch the footage. Maybe I am misunderstanding, but the footage is compressed... and then you turn all constrast down essentially placing everyting right in the middle..... then you stretch the darks down and the whites up... then you'll get noise and poor color depth... right?
I guess my list of questions goes like this (keep in mind I plan to color grade):

1.) Should I go for super low contrast or find a middle ground?

2.) If I shoot with D-log, do I even need to turn down contrast?

3.) D-log, D-cinelike, or none.

There is no simple answer to this.  Your aim should be to capture an image that has detail in the blacks/shadows whilst not blowing out the whites/highlights, i.e capture the full dynamic range of the image.  In many circumstances, especially when shooting in daylight, a "normal" exposure will not be able to achieve this as the dynamic range of the scene will be higher than the exposure can cope with.  To counter this and thus to improve dynamic range a log exposure can be used that effectively pulls the highlights down and pushes the shadows up.  In doing so the exposure gives more protection against loosing shadow detail and blowing out highlights, post processing is then required in order to manipulate the image to something more akin to what you expect to see.

The decision as to using D-log or not and then following on from that what contrast etc should be used really does depend upon the scene that you are shooting.  The higher the dynamic range of the scene and the more detail you want to capture broadly speaking the flatter the exposure you will need to capture.  Best thing to do is just try out the profiles and settings in different lighting conditions and see what works best for you. For example on a really grey day and with a scene that has little contrast there is probably little to be gained in shooting D-log, when the sun is out and the sky is blue D-log (or one of the other flat profiles) will almost certainly be the way to go.

"then you stretch the darks down and the whites up... then you'll get noise and poor color depth... right?"

If you want to end up with it stretched, or modified in any other way then it is best to do it while it is still 10bit raw data, before it is encoded into 8 bit compressed data and all the non-visible detail is thrown away.  That applies to contrast, brightness, log profiles, colour, saturation, but not sharpening.  Any changes you make after it has been encoded will result in visible noise, the greater the change the more visible the noise.  With sharpening, it is best to do it as the last stage in the process, before the final encoding.

Wow! I like how responsive this community is! Now when you say to work why the 10bit raw file, do you mean adjust the contrast? As far I know the file written to the sd is 8bit right?

JasonT2013 Posted at 2017-1-6 14:27
Wow! I like how responsive this community is! Now when you say to work why the 10bit raw file, do you mean adjust the contrast? As far I know the file written to the sd is 8bit right?

Yes, the file written to the SD is 8 bit, what I meant was to get it as close as possible to the end result in the camera settings and avoid making big changes in the post processing on your computer.  By the time it is written to the SD card a lot of detail has been permanently lost.

The risk of increased noise is definately a factor with a log profile as what ever way you look at it the profile is lifting the shadows.  However, in general it is a price worth paying.  That said if you are capturing a low contrast image and not planning on doing something drastic with it in post and you want to minimise noise then you would probably be better off using a profile that gets you nearer to what you want in camera.

Great! What is the difference why dlog and dcinelike?

The dcinelike is, as far as I can see, a compromise between a flat log profile and a Rec 709 profile (none or truecolor or whatever DJI choose to call it).  Cinelike has more saturation and contrast than Dlog, in some circumstances it's probably not a bad option.  You don't say what drone you have but do bear mind that even the P4P has a relatively small sensor, although more than twice the size of the P4, and the data rates are not that high, although again better on the P4P, all of which means there is limit to have much post processing the files will tolerate.  As always the key thing and IMO more important that the profile you use is to get the exposure correct particularly in respect of the highlights, with the P4 for example I always try and set the EV to no less than 3.5 - 4.0 stops off the brightest highlight else they are lost for good.  With the P4P I would expect this to be better as DJI claim the camera has a much better dynamic range although I don't believe for a minute that in the real world that it has the 12 stops they claim.

If you are going to heavily grade your footage it's probably worth spending some time shooting some test footage of a relatively high contrast scene in the various profiles and then seeing how they hold up in post.  You can also dowload from the DJI website the Dlog to Rec 709 LUT and then compare that with a straight Rec 709 (none ?) profile straight out of the camera.

P3s is what I'm rolling with at the moment. Hopefully going to the p4 in the near future! I'll definitely  do some test footage. I'll try the lut as well. That may simplify my post process significantly if the lut can "normalize" the footage for me.