FX3 - did it lose slog2 with firmware update?

[ahalpert]

Can anyone with an FX3 confirm if the slog2 option was removed when they released the timecode firmware update?

I have one and I never updated because I thought this was the case and I use slog2 a lot. Today I wanted to try out the new Sony Monitor and Control app but it wouldn't work with the old firmware on my camera.

[Doug Jensen]

Yes, it lost S-LOG2 several updates ago. RIP.

http://thenewcamera.com/why-sony-rem...rmware-update/

[ahalpert]

Yes, it lost S-LOG2 several updates ago. RIP.

http://thenewcamera.com/why-sony-rem...rmware-update/

Thanks. Maybe I can find another gamma I like as much...

[Doug Jensen]

I haven't.

[morgan_moore]

Slog 2. Profile of Kings.

[Doug Jensen]

One LOG to rule them all.

[cyvideo]

Thanks. Maybe I can find another gamma I like as much...

I have one client who has always shot in SLOG2 and now finds he doesn't have it on his FX6. He also has a bunch of his own fave SLOG2 LUTs he wants to use. He is now shooting SLOG3 Cine and, Abe, I don't know what you are editing on but in Resolve I just use a CST (math transform) between SLOG3 Cine and SLOG2 S Gamut. Being a mathematic transform, there is nothing lost in quality going this way. Add a few little tweaks on the LOG wheels and usually a minor tweak on the Curves and I have a totally good SLOG2 image to toss his SLOG2 LUTs onto.

Initially to confirm how this might work I shot a bunch of the same scenes in SLOG2 and SLOG3 on one of my cameras and used the above workflow, and it worked out very satisfactorily. Seeing that SLOG2 is good for 13 stops, 14 according to A Chapman, I don't understand why Sony dropped it in their later cameras. Which at best on test going by CineD's tests the FX6 delivers 11.4 stops at a signal-to-noise ratio of 2 (SNR = 2), the most important reading, and 12.6 stops at SNR = 1. Both figures well inside S-LOG2's handling capability, be it 13 or 14 stops.

Chris Young

[ahalpert]

I have one client who has always shot in SLOG2 and now finds he doesn't have it on his FX6. He also has a bunch of his own fave SLOG2 LUTs he wants to use. He is now shooting SLOG3 Cine and, Abe, I don't know what you are editing on but in Resolve I just use a CST (math transform) between SLOG3 Cine and SLOG2 S Gamut. Being a mathematic transform, there is nothing lost in quality going this way. Add a few little tweaks on the LOG wheels and usually a minor tweak on the Curves and I have a totally good SLOG2 image to toss his SLOG2 LUTs onto.

Initially to confirm how this might work I shot a bunch of the same scenes in SLOG2 and SLOG3 on one of my cameras and used the above workflow, and it worked out very satisfactorily. Seeing that SLOG2 is good for 13 stops, 14 according to A Chapman, I don't understand why Sony dropped it in their later cameras. Which at best on test going by CineD's tests the FX6 delivers 11.4 stops at a signal-to-noise ratio of 2 (SNR = 2), the most important reading, and 12.6 stops at SNR = 1. Both figures well inside S-LOG2's handling capability, be it 13 or 14 stops.

Chris Young

Thanks Chris. This is for my low-maintenance non-LOG shoots. So, saturated and accurate color. Slog2 with the blacks crushed a bit in-camera. It's always too flat except in harsh daylight. But it's easy to grade with quick contrast adjustments.

I basically either use that profile or slog3/sgamut3.cine. so I'd be looking for something else that's easy to grade. I've always had trouble with the highlights when grading slog3 gamma from scratch.

And I need a profile I can set in-camera on my a7siii and a7iv since you cant load in LUTs on those. (Can you on the current FX3 firmware? I dont even know.)

I already have the Monitor+ app and am largely happy with it so I think I'm just going to let it ride on slog2. I've gotten positive feedback from the profile.

It also looks like the latest FX30 firmware update removes the percentage tracker from the battery icon if you're using 3rd party batteries. You can still use them, you just won't know how much charge they have left. I use a lot of 3rd party batteries, so that would be another concern... I'm not loving Sony firmware updates right now.

https://www.sonyalpharumors.com/new-...ttery-support/

[Andy9]

I've always had trouble with the highlights when grading slog3 gamma from scratch.

Do you mean without a LUT / colour transform?

[ahalpert]

Do you mean without a LUT / colour transform?

Yes. I had been sometimes using an slog3 based profile with a similar treatment to the one I use with slog2. Something about the highlights is kind of compressed. Even when I got it to look about the same as my results from the slog2 based profile I didn't think it looked as good

[Andy9]

Yes. I had been sometimes using an slog3 based profile with a similar treatment to the one I use with slog2. Something about the highlights is kind of compressed. Even when I got it to look about the same as my results from the slog2 based profile I didn't think it looked as good

Were you dissatisfied with the results from LUTs / transforms or have you just always done it that way? If you've got a look that you like (and you clearly have) then there's no arguing with that but have you ever tried to recreate that look post-LUT / transform? I don't mean post in terms of position in the signal path, I mean temporally.

[Tom Roper]

Yes. I had been sometimes using an slog3 based profile with a similar treatment to the one I use with slog2. Something about the highlights is kind of compressed. Even when I got it to look about the same as my results from the slog2 based profile I didn't think it looked as good

Doug Jensen has been saying for a long time slog2 is better than slog3. I agree with him. In spite of some opinions to the contrary from Alister, SLog2 just feels cleaner, and with extended highlights. I've always ever used only transforms not luts. I've had good results with SLog3 too but I prefer SLog2 better. But these comments are prefaced on the F55 and I don't know how applicable they are to the Sony mirrorless cameras which look different to me in my limited experience with them. For example, on the F55 you only have 3 WB options, no gain is possible, only Cine-EI is available whereas the mirrorless are different in those regards?

Edit: Custom mode on F55 does permit WB and gain adjustment.

[NorBro]

On a related note, it's likely no coincidence Canon's C-Log2 is also better (it is) than C-Log3 and is only available in the higher-end cameras.

[Doug Jensen]

Doug Jensen has been saying for a long time slog2 is better than slog3. I agree with him. In spite of some opinions to the contrary from Alister, SLog2 just feels cleaner, and with extended highlights. I.

Yes, absolutely true. One of the easiest ways for me to come to that conclusion was by shooting 16-bit RAW with my F55 and R5 (and later R7). The first steps in the Resolve workflow is to convert those RAW files to S-LOG3 or S-LOG2. Because it was RAW data, it could be processed either way and the footage would be exactly the same as if it had been shot as S-LOG2 or S-LOG3 in the first place. So, whenever I graded two versions of fhe exact same RAW footage (dozens and dozens of times) using both types of LOG, S-LOG2 always won. Sometimes there was no meaningful difference between the two, such as with an indoor lit interview., But everytime I had a preference for one or the other, S-LOG2 was always the winner. Not once did I find the S-LOG3 version better. Not only did S-LOG2 look better to my eys, it was always a hell of a lot faster and easier to get the look I liked.

Not only all that, I hate the way S-LOG3 clips at 94%. Now, in a controlled lighting environment it really makes no difference. But for landscapes, sports, wildlife, etc. that highlight clipping causes a lot of trouble.

S-LOG3 came around because it can be treated basically the same as Arri footage. Colorists in Hollywood didn't like having to treat S-LOG2 differently frm how they processed Arri footage, so S-LOG3.cine was born by Sony to makle those folks happy. Any LUT that works for Arri works for S-LOG3. So S-LOG3 is not "better" it is basically dummed down version of S-LOG when compared to S-LOG2. Also, Resolve has never had any built-in S-LOG2 LUTs, so a lot of inexpereinced users of Resolve had no clue how to process it properly.

I've pretty much discarded LOG3 on my FX6 and now shoot S-CINETONE 99% of the time -- with some grading in Resolve to polish it off. S-LOG2 is history.

[ahalpert]

Were you dissatisfied with the results from LUTs / transforms or have you just always done it that way? If you've got a look that you like (and you clearly have) then there's no arguing with that but have you ever tried to recreate that look post-LUT / transform? I don't mean post in terms of position in the signal path, I mean temporally.

Sometimes clients request that I not shoot in LOG, so I need to have something I want to use in that case. (And personally, I'm generally happier with the color I can get out of the Movie color gamut on my mirrorless cameras than I am with sgamut3.cine + LUT. So, if it's my choice whether to shoot LOG or not (which is rare) then I often don't shoot slog3/sgamut3.cine. It's hard for me to get as happy with the color and sometimes i think that parts of the contrast gradient don't look as nice either )

The slog2-with-slightly-crushed-blacks profile I use is not actually a WYSIWYG profile (except for very high contrast scenes), but I can't bring myself to throw away that much highlight information on a typical shot. When I grade it, usually I have to push up the highlights a lot and pull down the shadows a little. But nobody seems to complain about having to do that.

I think part of the reason clients sometimes request I don't use slog3 is because it can be harder to fix problems because the contrast/color info is stored less efficiently. Theoretically we have a lot of flexibility with the file, but I think it can be hard to make it look as good if you screw something up.

And finally, I think it's easier to make a big screw-up if you're shooting LOG and not monitoring properly. So, i had a wedding studio a couple years ago blast out an email to all their cameramen requesting slog3. The idea was that it would simplify their grading process. They reversed course about 2 months later and blasted out another email saying that no one should shoot slog3. My thinking is that it didn't simplify their grading. They still had to grade everything. But now some cameramen (especially the second shooters probably) were making bigger errors and those errors were harder to correct to begin with.

So, I think if you have a sound method for exposing and color balancing then it should be no issue to shoot slog3 and start with a LUT or color transform or whatever you want. Or maybe it's easier to fix these problems and get good results in Resolve rather than Premiere and the studio was using Premiere. I don't have Resolve experience so I can't say.

I still think that shooting slog2 + LUT would probably yield modest image improvements vs slog3 based on my observations. But slog3 is everyone's standard now.

And if you look at the history of it, it makes sense. It seems like slog2 was designed to optimize the output from Sony's sensor. Slog3 was designed so editors could use Arri LUTs they were already familiar with. In other words, slog3 was not designed to optimize signal processing.