Exact Ae version/ Exact Pr Version?
are you working 32bpc in Ae?
is color management the same between the two Apps?
(In Ae you can access them in Ae's project settings)
latest AE & Premiere, on Mac, working in 32 bit in AE.
Color management - Premiere doesn't color manage, so I've turned off color management for both the clip and the project in AE.
I don't think color management is the issue, since I've gotten the files to match between the two programs except for when Lumetri is used.
I ran a regular test to see if I get consistent results. created an HD sequence applied the lumetri color and did a curves adjustment. then I copy pasted the sequence to Ae and the result look the same. if you want, you could upload a reduced Pr project with just one frame with the lumetri applied and I could test it in my system to see if I get consistent results Between the two Apps
Interesting - was your image sequence standard dynamic range or HDR? I was using the Lightroom style controls in Lumetri towards the top - blacks/shadows/highlights/whites/HDR specular highlight, that section.
Thanks for spending time on this!
Hi MikeCurtis_1138, I have a question regarding these HDR scopes. How do setup it to see the HDR highlights?. I have some Sony F55 Raw files that in the Raw Viewer I can see these peaks, like this:
But when I load these MXF files and see it in Lumetri (proper configured to see HDR, I think), I see this:
I can export in OpenEXR and MXF from Sony Raw Viewer but the resulting files are equal to the original MXF. So, how I set Premiere to show these scopes like the Sony Raw viewer?
Hey Antonio -
I just upgraded to the 2017 versions but haven't read up on if anything is different in all of this.
So: Thing I KNOW were true for the prior-to-2017 versions:
-Premiere will accept HDR footage in two formats - OpenEXR and PQ encoded JPEG2000. I don't know of a single tool to do the latter, so I went with OpenEXR.
-so I wouldn't expect the MXF to be supported for HDR work natively
Things I SUSPECT or THINK are true but can't say 100% definitively:
-it appears that anything other than the two formats listed above will clip to 100 IRE. I think that's what is happening with your MXF files.
-While you exported as OpenEXR, it appears that the export process you used in Sony Raw Viewer is clipping to 100 IRE. I see this problem OFTEN - a lot of programs that supported .exr do so in the standard 0-100 IRE range, even though the .exr format can support a much broader range.
SO - it would appear that you're handling the footage correctly within Premiere - there's no settings or color space anything to apply to make it HDR-y (phrasing?), but I think the challenge lies in trying to find a way (or more accurately a tool) to convert your MXF files into OpenEXR AND have the full extended dynamic range encoded in there and not clipped to 100 IRE.
So hopefully that helps to identify the problem, I'm sorry but I don't know a solution to get you there the rest of the way.
I'm going to start messing around with the freebie DaVinci Resolve from BMD and see if I can PQ encode JPEG 2000 in there. But of course, if I have that, why not just do my grading in there?
Hi Mike. The thing is that you can import a MP4 with HDR, ant it looks with the Scopes in HDR, look:
I have tried to export various type of EXR from the Sony tool, but I don´t find the exact type of export I need...16 or 32bits per pixel, or what type of data from the MXF loaded...
I have some tests made from either SDR and pseudo-HDR (as I cannot import direct HDR as for now), and while it appears so show it good, I think that I´m missing some workflow in how to make these HDR adjustments in Lumetri.
Hum, good point that export PQ J2K from resolve, try it and if works maybe try I too to see...
Also, I imagine that HDR in Resolve it´s more easy and convenient, isn´t?. At least to make some export to finally make a H265+HDR in Premiere in the last stage...
OK, that's news to me! I am guessing that is a new feature in CC 2017.
When I've done HDR in Premiere (2015.3), it only supported the two formats listed. And there was no setting to apply to the material - it just worked, dropped on any timeline.
Checking in v2017, there's no color settings in Properties, Interpret Footage, or Source Settings having to do with color, so it appears that it should "just work" or not based on the file that you bring in.
BTW - I have tested HDR HEVC/H.265 output and it WORKS - I can play it from a memory stick plugged into my Samsung HDR set and it looks great! Except I'm getting banding, but I suspect that is a Samsung issue and not an Adobe issue - I don't see that banding on my 8 bit computer display.
Not to ding Adobe (because if it did everything I wanted I'd gladly use it due to integration with the rest of the Adobe Suite - my workflow is After Effects heavy so I love it for that) - but Resolve is, um, FREE if you can live without the HDR scopes ($1000 for the version with that feature, unfortunately).
But PQ encoding your MXFs and exporting as JPEG2000 might be a way to prep the content in a way you like - and if you could normalize to the peak brightness you want, all the better.
Speaking of which - the current UHD Premium Alliance specification calls for a minimum of 1000 nits. I'm trying to figure out if I should aim for that or 2000 nits? Don't the sets adjust on the fly for this? Something more to test....
BTW- on your scopes, don't forget to set the Colorspace to Rec2020.
But PQ encoding your MXFs and exporting as JPEG2000 might be a way to prep the content in a way you like - and if you could normalize to the peak brightness you want, all the better
Oh, maybe that´s work!. It is possible to load these Sony F55 files and export in JPEG2000 PQ in Ressolve?. Will try to see if works.
Premiere can export in MXF OP1 J2K PQ, but Sony Raw Viewer cannot read it.
Another question is to make a HDR version of SDR material. I have done some test but I don´t know if I am at the right direction, because not HDR display attached in my system and I take the Lumetri Scopes as reference to see what happens. If you have some ideas about how to proceed in these situations, I will appreciate it, as I not have found any information via Adobe to make HDR adjustments, even with HDR material :-(
Sweet! Also, I checked and yeah, .mp4 HDR support is new - I got it to work on my system as well - that opens up some good workflow options - next on my To Try list is using .mp4 proxies to stand in for higher resolution EXR files (or whatever) so that I can have realtime performance. I can't get 24fps UHD playback AT ALL on my Late 2013 MacBook Pro (topped out/pimped out for what was available at the time), but 1080p it can do - so I'll be digging into Proxy Workflow stuff next.
"HDR version of an SDR material" - I think this is a dead end - you don't have the information in the shot you need to distinguish "pretty bright" from "super duper bright" - and stretching the dynamic range of SDR to HDR isn't the answer, even with a custom curve, for the same reason. There are some TVs that offer this mode, but they have the same problem, and the results look like bad photographic HDR - ick.
Yes, you are right, but this is the same question with the "Real 3D" and "Converted 3D" dilemma. There is good and bad converted 3D and the same with real 3D.
Of course, from a graded and finished shot you cannot go to nowhere. In fact I have done some tests and you cannot push the lights so much, because there is not information to play with.
But from a simple LOG image (not raw), you can tweak more it. Of course you don´t have the data of the highlights, but you have some information to play with it.
Now with the HDR ability of Youtube, loads of people want to make his own HDR version of his familiar videos, and we can see horrendous ones, but maybe some good ones.
I think that we are in a new era and the things are fast, very fast, and we cannot catch em.
Not quite - with coverted 3D, there's information that can be gleaned from neighboring frames to help the process. Plus, it is incredibly laborious and needs lots of "hand guiding" - I'd say HDR conversion faces similar challenges, but lacks any clues as to how bright things are from adjacent frames. And they're just offsetting existing data to make the second frame, not boosting brightnesses to levels that never were in the source. Short version - it is harder to do well.