Mov vs dng (is there any video comparing the results of both workflows?)

[CHINOTAKER]

Hi I've being checking this forums for the last couple of years and I've always find the answers I was looking (thank you very much to all the ML team and the ML community by the way) for but not this time. I have shot many many videos with it and I am very happy with ML in general. Last weekend I shot a commercial and I started to wonder if I should try exporting dng's from Mlrawviewer to get that ''extra'' out of my footage. I've always export to proress444 and to be honest I can get pretty much the same ''quality'' that the Alexa or red gets in the same lighting situation (I have examples of that) to the point that the producer of the TVC couldn't tell the difference. I am an editor by the way

I've just always found that the extra work in the Mlrawviewer/dng process is not justify or is simply an intangible difference compare to the Mlrawviever/mov approach. I think now I should explain my workflow so it makes more sense:

1. shot 5d3 raw
2. convert .mlv to prores444 via mlraviewer (always s-log always manually ETTR)
3. offline in premiere with the proress444
4. xml to davinci
5. xml back to premiere with the graded footage in proresHQ
6. add graphics via AFX
7. export

So basically I wonder if I would get more dynamic range if I go to the extra process of working with dng's instead of straigh prores444 considering that I would use the same workflow offline in premiere/colour grade in davinci/final online and export in premiere

thanks in advanced

[Flocksock]

You will definitly get more dynamic range using DNG files. Not only more dynamic range.
with DNG you are able to shift color temperature and tint in Davinci Resolve (or AE). And also
more control over highlights and shadows.
My workflow:

1. MLRViewer to DNG
2. Open files in Resolve 12.5.
3. Using Rec709 for "contrasty" image... and the Scopes to set the color temperature and tint right.
4. After that switching to "BMC Film" which gives me a "flat look".
5. Adding contrast, stauration and dive into color wheels for grading. (Allways with vectorscopes / waveforms on)
6. After basic color grading... maybe add some luts + grain
7. Export (tiff sequenze for Premiere.. if you still need premiere)

Here is a very good Tutorial about Davinci Resolve:

Davinci Resolve + BMPCC grade + swearing
https://vimeo.com/153458368

There i found this "trick" to setting up the color temperature in "Rec709" and after that switch to "BMC Film".
Mov vs. dng. Simple do tests. Film something and than convert it in mov and dng. And grade your footage
as mov and grade dng files. than you will see the difference. and if you have to use "offline" files... get/inverst in a
faster PC. i7, 8 GB ram, 250 GB SSD + 2TB HD. Graphic Card (4GB Ram)

[Kharak]

Ok, so depending on what Log space you convert to, I found the Prores files to contain too much noise for my liking. But for s-log (1) i suppose it's okey as the curve is not too extreme, same for C-log (1).

I convert the DNG's to Log-c for maximum DR with Cinelog-c to log-c. With DNG's this works, but the prores files seem to fall apart in log-c space. Especially shadow detail takes a toll in prores.

But the prores files are great for proxy workflows.


Resolve plays back DNG's in real time with a descent GPU, so you should be golden
Via phone.

[Roberto Mena]

CHINOTAKER, Im glad you brought this topic up. I have found some interesting things in regards to DNG vs ProRes4444 color grading. I always shoot MLV with the ColorVision flat profile. In my opinion so far, with the extreme dynamic rage  experiments I have run with the 5Ti and 7D is that it looks like I get more dynamic range and better color corrected and graded images with DNG files converted to a to ProRes4444 movie file (S-Log or Log C) for interior shots lit with natural lighting and color grading the ProRes4444 video file directly. But I get better results color grading directly DNG's with exterior natural lit scenes. Everything run through MLviewer. This is so weird, because you would think color grading or correcting directly DNG files would get you the best results no matter with interior or exterior shots. Anyone else has found this to be true? I'm not saying this is always the case but I've noticed it enough to feel it is a pattern. Im talking about really pushing the DR for interior shots. I know that doesn't make sense and I am trying to figure out why that is. I almost feel like because Canon camera's DR is about 11.7 stops compared to the awesome DR of BMPCC's 13 stops that the ProRes4444 seems help hold tighter the image quality when color correcting or grading interior shots. I would love to hear other people's experience with this.

[Frank7D]

Roberto Mena, the camera profile you shoot with will not affect the DNGs in any way, at least not directly. If it causes you to shoot differently, however, (say, by exposing differently) because of how the image appears on the camera display, then that could indirectly affect the DNGs you get.