FS7: Sony FX9 v Canon C500 mkII

I adjusted the exposure before assigning the Sony Slog2 to 709 LUT. After I used the LUT it resulted in a clipped image that I had to correct again. Or is the underexposed clip the problem as Barry said?

Yes as Barry says the underexposed clip is basically the issue. I generally find the easiest way out of this dilemma is to make sure you have a whole set of thirteen +/- 1/2 stop over/under compensation LUTs for whatever flavor LUT you are using. My starting workflow which I find gives me a pretty accurate point to start from is to bang on two extras nodes to start with. The first node 99% of the time I use for white balance if necessary. I now apply the necessary native "0" or "+ or -" compensation LUT to the third named "LUT" node to give me the closest to the correct exposure levels required to compensate for any over or under exposure on the clip in question. Then finally any lift, gamma, gain adjustments are made on the second named "LGG" node. This LGG node adjustment is usually not too drastic as you shouldn't have any major clipping going on with your third LUT node as you have already chosen the correct 0 or +/- 1/2, 1, 1.5, 2, 2.5, 3 stop over/under exposure compensation LUT.

Example: Many of us tend to over expose S-LOG 3 for example by 1-2 stops. Well if that is the case you you cannot use a LUT that is a native "correct 0" exposure LUT as it will be out by the 1 to 2 stops of overexposure used on that clip. You have to design a LUT or use a LUT that takes into consideration that the footage is already already over exposed, or under in your case. Hence the need for exposure compensation LUTs. Doing it this way I just don't see clipping on the third LUT node. If I do see clipping on that third LUT node then I know that I need to change that LUT for a compensation one to bring the clip level down. Then as I said you can do you final teaks up/down mid range etc on you second LGG node. Obviously there are a myriad of other adjustments and extra nodes can be used for a variety of grading reasons but the general rule is that all adjustments should be done prior to the LUT node as they are basically non-destructive. A LUT is effectively a scene file of sorts and any adjustments done after a LUT are destructive in as much as you are now adjusting within the narrower 6 stop or thereabouts confines of lets say a 709 LUT. Resolve lay out their workflow pretty clearly and in general I find this "flow" works in most NLE and grading software. Slightly different workflows for RAW and ACES but the general gist of it is to do all adjustment prior to final LUT, RRT.

Chris Young

+- LUTs.jpg
Nodes workflow.jpg
Resolve RAW workflow.jpgResolve ACES workflow.jpg
Grading Order of Operation in Resolve manual.
Resolve grading node workflow.jpg
 
Last edited:
Agreed. Ben T has done us all a great service in developing LUTCalc. Used it a fair bit. Another tool I find that is very powerful for creating LUTs and copying color palettes from other images to create LUTs is 3D LUT Creator. So quick to work with. Been using it for a few years now and it is one of my go to tools for fixing things like Chroma key spill.

https://www.youtube.com/watch?v=3tyYjsYUrww

Chris Young
 
  • Like
Reactions: jcs
Apparently I don't have the premium Linked service which is required to send a note. If anyone else does though feel free to drop Ben a note asking him to update LUTCalc with s709 if he gets a chance. (I don't have the foggiest idea how to use Javascript or anything like by the way).
 
It’s a little too early for me to be at my desk, and so viewing this on my iPhone, but 3 comments.

First: this is the c700, so the same sensor as the c300, not the c500II, in case that matters.
Second: according to Colin who compiled this, all cameras were aligned to one stop above their specific clipping points, so comparison wise, they are all clipping at the same point in this video but not necessarily at the same points relative to exposure. Geoff is aligning the cameras around a middle grey (ie exposure), which is probably a more normal way to compare them. By aligning around clip, we see the comparative effects of the clipping more starkly.

Last: as has been noted before, the c700 clips early in the yellows (Alexa does this as well), but otherwise the colors stay pretty consistent. The monstro is showing issues in magenta and skin tone patches, although I’d need to get a better look.
 
Last edited:
Yeah, that's the C700, not the C700FF. Completely different sensor.

Thank you, Jaime and Barry. Glad for that. I totally forgot for a moment about the other versions of the C700 body, still had the other CML test in the back of my mind.
 
Last edited:
Are the V2 LUTs for all of the Canons or are they customized only for the C500 MKII? I haven't been able to find a source that definitively says?

Canon puts the same LUTs on their download page for the Canon EOS R as they do for the cinema cameras so I am assuming that Canon has tried to make their Log profiles universal.
 
CLog and CLog3 are pretty easy to grade without a LUT with practice using curves to set gamma (may need another transform if using a non-709/2020 colorspace). Note Canon CLog and Sony SLog2 gammas are nearly identical.
 
Yeah, that's the C700, not the C700FF. Completely different sensor.

If you look at the card in the video, it says “C700 6K to 4K”.

The link to CML leads to the test with the C700FF

In the comments, there is mention of C700 LF.

That all made me assume it was the C700FF. Was there some other information that says otherwise?

Thanks.

Here is the link:
https://vimeo.com/283320303
 
Ah! My apologies. I missed the 6K to 4K in the description. It looks like that might be the C700FF, indeed. Sorry for the confusion.
 
Hi all! Just wanted to share this if anyone have not read and would find it interesting. The canon white papers regarding their sensor designs and cameras can be very informative, I find. Does anyone know if orher manufacturers have similar documents?

This one is particularly focused on dynamic range in c300mk2 and c700:

https://support.usa.canon.com/resou...70000/ART170227/en_US/Deep-Dive-HDR-Part1.pdf

https://support.usa.canon.com/resou...70000/ART170228/en_US/Deep-Dive-HDR-Part2.pdf

The one on the c700ff says its a similar sensor design as c700, but not if its exactly the same or not.

Ps. Im not taking any claims or measures for granted, esp. by the manufacturer. I have to give it to Canon though, they provide a lot of information on their design choices and how they arrive to their conclusions. Ds.

Cheers,
Edvard
 
Last edited:
Here is another video generally comparing cinema camera exposure ramps.
https://vimeo.com/333772649

And the C700FF by itself:
https://vimeo.com/331079485


And this is the video i post before, from one of the CML shoots.
https://vimeo.com/283320303

Helpful tests, James!

C700 FF does ok up to +5, expect similar from C500 II with similar set up/test. Would be helpful to see C500 II and FX9 in same settings especially at +5, +6. Expect similar relative performance as here (FX9 > C500 II for IQ, same as Venice > C700 FF); won’t know for sure until seeing such tests.
8FE639FC-68C4-43A8-98B1-0248AD84A394.jpg

Test info translated to English: https://translate.googleusercontent...700271&usg=ALkJrhi2OkSjWe_8rnTfdQ5OAey2UVwDuA

Amazing that the 14-stop ~10-year old Alexa sensor still wins over the new >=15 stop manufacturer rated cameras. Tests reveal the truth. Kudos to all who make them (even when mistakes are made: still helpful if folks can learn from them (how not to repeat them etc.)).
7AD3B88F-77A2-41EC-84ED-3295BFAB2457.jpg
 
Last edited:
Back
Top