FS7: Severe Pixelation in Center Crop Mode

I confirmed it on Yc waveform scope. An older LUT I created last summer goes to 109 IRE, the new ones I'm creating all hit a ceiling at 100 IRE. So it's not the NLE. It's the LUT.

Yes, some of my older LUTs clip the top 3 stops of highlights. I found that by inserting ProcAmp ahead of those LUTs, I can recover highlight information. But I am trying to create new LUTs that encompass more dynamic range, yet don't require additional plug ins to get the levels right.

So far, I have not been able to replicate the earlier LUTs' highlight capabilities in terms of using the full 0-255 range for digital media. I just need to find out which parameter gave me that extra headroom.
 
if you have REC709 anywhere in your pipeline, it will limit values to no more than 0-255. As of late, I've been using HLG instead of REC709. HLG has the potential to deliver output in either REC709 or HDR. Looking at HLG processed footage on a WFM will show very large data swing values, all in floating point, so you can dial it back to anywhere from 100 nits on up. Be aware that if your camera is delivering in REC709, you're already limiting your data swing. What this means is that you need to deliver in RAW to save all that HDR color space and dynamic range.
 
Last edited:
The problem is, I forgot how to make LUTs that use the full 0-255 range. Now I'm getting 7-235 out of my LUTs. I accidentally created some LUTs last year that utilize the full range, but I have no clue as to how I did that.

I'm shooting Slog3 and using the LUTs to convert the gamma to something distributable. I'm getting a nice look out of the last few LUTs I made yesterday, only thing is peak whites are limited to 235. If I could use that extra range, I would be happy with the resulting LUT as my all around default technical LUT.
 
You may want to consider a workflow that doesn't use LUTs. I've found DaVinci Color Managed workflow to be much more flexible and gives me considerably better results than using an input LUT.
 
I find DaVinci to be a whole new learning curve, lacking in full screen video monitor without an expensive investment in their hardware, and lacking in some of the features I need for multicamera concert recording editing.

I have no problem with the LUT workflow. I think it opens new possibilities for controlling color in dimensions not available in either DaVinci or Premiere. I just have to learn what each of LUTCalc's parameters do. I wish there were a standalone version for Windows, so I didn't have to sneakernet the LUTs from my internet-connected PC to my NLE.
 
Hi All,

I've just uploaded a new online build to LUTCalc Testing, which should hopefully help out when using the base Rec709 gamut.

First off, I've changed the 'Hard Clip' option so that it is now a drop down. I know, more complicated ;-). But actually, the idea is that now you can choose whether black, white, both or none are clipped. Furthermore, the default is to clip black only. This should reduce the chance of narrow gamuts such as Rec709 pushing saturated colours from wide gamuts into the superblacks, with whatever results or artefacts. If you need the unclipped maths, the option is there to turn it off, but I think that this default should be better and easier for most people.

Whilst on the clip issue, if the output is set to 'Data Range' and you have any clipping set, a little tick box will now appear marked 'legal'. This will make the clipping restrict things to legal range. The charts and preview will now reflect clipping choices, so it should be clearer when you try it than when I try to explain it! :-)

Now, on top of that, I have had a stab at what I'm calling 'Gamut Limiter'. This brings up a dropdown marked 'Display' with two options - 'Rec709 / sRGB' (which share primaries and white point) and 'Rec2020 / Rec2100' (likewise). This attempts to limit the saturation within the chosen displays vectorscope range, whilst maintaining luminance. It is applied after all other corrections and transformations, so the most significant effect would be to bright colours at luminances above white clip. Strictly that stuff is by definition outside of the colour space, but this way there is still image data, but hopefully a reduced chance of weird saturated colours.

This last one is something I have just come up with to try to maintain LUTCalc's separation of gamma and gamut within colourspaces, and I have only tested it out on the preview defaults (Rec709 Gamut and CIE 1932 combined with the saturation adjustment providing as much saturation as I care for). With some other stuff, notably batch creation of sets of exposure adjustment LUTs there has been a fairly substantial amount of coding done, so this is definitely LUTCalc testing (!), but I would appreciate any feedback, even (especially) if it draws my attention to something I've broken.

I'm going to kick the tyres whenever I have a chance as I'm keen to get out a proper update, but the day job is looking likely to get busy, so I'll take any help I can get. Thanks,

Ben
 
Ben, You've been doing a wonderful job on this 'labor of love' and I'm sure that a lot of us appreciate the hard work and time you've put into coding LUTCalc.

As with any program as powerful and rich in parameters as this, the learning curve is steep for many. Although I'm getting the hang of many aspects of LUTCalc, there remain some mysteries I have not been able to solve yet. One such yet-to-be-discovered is finding out what tick box determine the luma range. For the life of me, I don't know how I created a 0-255 range last year, and this year, I'm testing this and that parameter, but have not found the magic accident that I had last year, enabling more than 7-235 range. Perhaps you could offer a hint as to what controls that (I thought it was the Data/Legal options, but that turned out not to be it).

I'm trying to make at least one solid SLog3 technical LUT that captures all 14 stops and makes it available to me in Premiere. I still find a couple more stops of higlight by putting a ProcAmp plugin ahead of the LUT and I can't figure out why my whites still clip at 100 IRE on the scope. Last year, I created some LUTs that went the full 109 IRE. Can't figure out how! Suggestions?
 
You must log in or register to reply here.
Back
Top