How noticeable is Log latitude really?

[Marcus Durham]

A practical question.

We all shoot in a variety of circumstances, for example mixing indoor interviews with outdoor cutaways.

Obviously using S-LOG for indoors work is pointless and a waste of time when there aren't enough stops in the shot to make it worthwhile.

But those outdoor cutaways could be S-LOG. Anyone mixed the two in a production?

I'm doing this kind of work later in the month so I may take a few backup shots on S-LOG and see how it pans out (having first taken the same shot in my usual custom Cine gamma).

[Cliff Totten]

Well? Yes, it's true that SLOG indoors is far less necessary, especially if you control the lighting. But there can be some exceptions to that if you consider shooting inside and possibly protecting the bright scenery outside a window. This part of a shot could otherwise be blown out by rec709. SLOG can also capture speakers using projection screens that burn rec709. One time I masked a static PowerPoint and graded that differently from the speaker. SLOG made that possible.

Generally speaking I think you are right about not needing SLOG indoors too often. Lighting your subject in front of a window is no doubt the only way you want to go if that's possible.

I have mixed rec709-ish and SLOG shots together with no problems.

CT

[Marcus Durham]

Cheers.

Yes I was thinking of controlled and properly lit and setup situations. If you've done the job properly you aren't going to have the range there to need S-LOG.

Also worth commenting that the Cine Gammas are very good and by having a couple of custom presets you can cope with many situations where you may not need or want to go full S-LOG. I have a few (all labelled on the base of my camera) I can flick between.

[Nigel Davey]

Hi Marcus, are you happy to share your Cine Gamma settings if they differ from the presets? Np if not, I just like to try out new profiles on lower importance projects.

[Marcus Durham]

I've been using the Abel Cine and Light settings on recent projects and they seem to look pleasing enough:

Sony FS5 Scene Files | CineTechnica

There's a few others gained from here which give pleasing results.

[Nigel Davey]

Thanks I'll give that a whirl.

[Piotr Wozniacki]

Originally Posted by Cliff Totten

LOL,....well Doug, I wouldn't go "that" far are to call call people "fools" if they shoot SLOG on 8 bit.

When watched on my big 55" screen, Doug's grades show exactly the same flaws in grading somebody else mentioned: grass is unnaturally radiant green, plus in most of his samples it is full of noise.

I'm new to Slog (my FS7 camera is just 4 months old), but I already learnt the real benefits of shooting Slog (and mine's 10-bit footage, not 8-bit like that of the FS5) only shows when the scene was heavily back-lit, and I absolutely required to bring out the shadow detail while not blowing the bright background (usually, the overcast but bright sky). And grading in Resolve, after several initial trials I completely abandoned using LUTs (as the color I'm getting with them is unnatural, noisy in shadows, and reveals a lot of strange artifacts as soon as I try to further tune it up using other grading tools). Instead, I'm using RCM - once I applied the right RCM input transform, I never looked back to LUTting!

Speaking of those high-contrast, back-lit scenes that really justifies using Slog: the additional problem arises when such scenery changes to "normal" while I'm shooting for the same project, and I'm getting this uncomfortable feeling that:

- if I stay in Slog for consistency purposes, Slog use for scenes looking great with "normal" (lower DR) settings (i.e.HG/Rec709) will only bring unnecessary complications in Resolve

- switching to HG/Rec709 will bring consistency issues between clips (sure - correctable in Resolve, but at the cost of additional time and effort).

Difficult choice it is in such situation. Therefore I totally agree with someone else's statement that Slog material will make for more future proof archives of my footage (even though, as my main "delivery" setup is my own large SUHD, Quantum Dot Samsung TV already capable of "sort of" HDR and with color gamut close to Rec2020). I even tried grading my Slog3 footage in HDR and rendering to Rec2020, and while the result certainly isn't as gorgeous as true, Dolby Vision pictures I once saw on an industry-grade, HDR Sony monitor - in certainly looks better on my Samsung than even the best HG/Rec709 shot material... But that's jus me; for the reason some of you guys may remember - I'm now a one-band freelancer shooting/editing/grading/delivering "crew". so I can afford all kind of experimenting :)

Piotr

[Doug Jensen]

Originally Posted by Piotr Wozniacki

When watched on my big 55" screen, Doug's grades show exactly the same flaws in grading somebody else mentioned: grass is unnaturally radiant green, plus in most of his samples it is full of noise.r

Not on my monitors and scopes. But nevertheless, that is just a matter of taste. The saturation could easily be dialed down, or any number of other changes made, if someone prefers a different look. And this one of the big advantages of LOG and RAW -- nothing is locked in at the time of shooting. It would be great if you can post some examples of your work to illustrate how you prefer a finished grade to look.

[Piotr Wozniacki]

Originally Posted by Doug Jensen

But nevertheless, that is just a matter of taste. The saturation could easily be dialed down, or any number of other changes made, if someone prefers a different look. And this one of the big advantages of LOG and RAW -- nothing is locked in at the time of shooting. It would be great if you can post some examples of your work to illustrate how you prefer a finished grade to look.

While I agree the saturation of a particular color might be considered "a matter of taste", Doug, the noise inherent to Slog images (somehow particularly prone to be present in those single-channel-saturated areas, like green grass or deep blue sky) - is NOT. What's worse, it's not easy to get rid of w/o special tools (Resolve Studio's own de-noising algorithm is not particularly effective, while the Resolve version of NeatVideo is very expensive and makes processing extremely slow - like 1 fps (Sic!) - w/o pre-rendering of course)...

And believe me - without those specialized noise-reduction tools, getting rid of the noise in darkish but saturated color areas is virtually impossible, in spite of the Resolve grading tools being very flexible with 10-bit Slog footage. Plus, the trick Alister is preaching (to overexpose Slog shooting by 1-2 stops, so that you can bring the levels down in Resolve thus masking the noise in shadows) is not very effective, as my many trials have proved that the picture suffers from the penalty of losing highlights, while the noise is still there in the shadows - only slightly mitigated by shifting the levels down).

So frankly, it's the noise (and the blocky artifacts, created around the noise particles "objects" (speckles) by XAVC codec - I wrote about them in another post of mine, when I wasn't yet aware the source of those artifacts is indeed the Slog- related noise), that presents the greatest challenge in grading Slog footage - even in such a potent tool as daVinci Resolve. Therefore - as I said before - I'm limiting my use of Slog solely to the super-contrasty (usually back-lit) scenery - and unlike someone said, I shoot Slog as often indoors as I do outdoors, as back-lit situations happen equally often in both situations (think of shooting against the window in a naturally lit room). Even the first Sony demo clips, published in order to show how the huge DR Slog can help you convey through entire workflow from acquisition to delivery, were actually shot indoors. Like the video with 2 women, getting ready inside their Asian room, with beautiful, full -sunshine bright, outside scenery visible through the room exit:

PXW-FS7K (PXWFS7K) : Specifications : United Kingdom : Sony Professional (under "Videos & Updates" tab)

But - in spite of all these shortcomings - I find it a great opportunity to learn and play grading my Slog footage in Resolve. While I can understand all the reservations, expressed in this thread by people who - unlike myself - are under pressure of time, their colorist/grading colleagues, or Clients - I'm totally free from such pressure entirely (being the camera operator as well as the colorist/editor myself, and delivering to myself and my own particular monitoring/viewing hardware) :)...

Piotr

[Doug Jensen]

Well, you've done your testing and come to your own conclusions and that is what everyone should do. Fortunately, my testing, results, and opinions don't match yours at all. And judging from all the great work I see from others who have used S-LOG on their productions it is clearly better than you give it credit for. You are welcome to come to your own conclusions. But if S-LOG was as bad as you make it out to be I'd sell my cameras tomorrow and buy something else instead. Fortunately, I am quite happy with results I get on REAL shoots and will be keeping my cameras and continuing to shoot S-LOG with no qualms or reservations about doing so.

[Cliff Totten]

Originally Posted by Piotr Wozniacki

While I agree the saturation of a particular color might be considered "a matter of taste", Doug, the noise inherent to Slog images (somehow particularly prone to be present in those single-channel-saturated areas, like green grass of deep blue sky) - is NOT. What's worse, tt's not easy to get rid of w/o special tools (Resolve Studio's own de-noising algorithm is not particularly effective, while the Resolve version of NaetVideo is very expensive and makes processing extremely slow - like 1 fps (Sic!) - w/o pre-rendering of course)...

And believe me - without those specialized noise-reduction tools, getting rid of the noise in darkish but saturated color areas is virtually impossible, in spite of the Resolve grading tools being very flexible with 10-bit Slog footage. Plus, the trick Alister is preaching (to overexpose Slog shooting by 1-2 stops, so that you can bring the levels down in Resolve thus masking the noise in shadows) is not very effective, as my many trials have proved that the picture suffers from the penalty of losing highlights, while the noise is still there in the shadows - only slightly mitigated by shifting the levels down).

So frankly, it's the noise (and the blocky artifacts, created around the noise particles "objects" (speckles) by XAVC codec - I wrote about them in another post of mine, when I wasn't yet aware the source of those artifacts is indeed the Slog- related noise), that presents the greatest challenge in grading Slog footage - even in such a potent tool as daVinci Resolve. Therefore - as I said before - I'm limiting my use of Slog solely to the super-contrasty (usually back-lit) scenery - and unlike someone said, I shoot Slog as often indoors as I do outdoors, as back-lit situations happen equally often in both situations (think of shooting against the window in a naturally lit room). Even the first Sony demo clips, published in order to show how the huge DR Slog can help you convey through entire workflow from acquisition to delivery, were actually shot indoors. Like the video with 2 women, getting ready inside their Asian room, with beautiful, full -sunshine bright, outside scenery visible through the room exit:

PXW-FS7K (PXWFS7K) : Specifications : United Kingdom : Sony Professional (under "Videos & Updates" tab)

But - in spite of all these shortcomings - I find it a great opportunity to learn and play grading my Slog footage in Resolve. While I can understand all the reservations, expressed in this thread by people who - unlike myself - are under pressure of time, their colorist/grading colleagues, or Clients... I'm free from such pressure entirely (being the colorist/editor myself, and delivering to myself and my own particular monitoring/viewing hardware) :)...

Piotr

Remember cassette tapes? Remember how we used to record our audio levels has hot as we could before we clipped or saturated the tape? Why did we do this? Because analog tape has a certain noise floor and recording our signal has high above that noise as we could improved our signal to noise ratio.

Image sensors are analog devises with a noise floor too. (they do get to an ADC quickly with Sony EXMOR tech but not before they pick up their noise floor...like all sensors) So, when shooting SLOG (or even Rec709) you need to think "signal to noise ratio" when shooting. You dont HAVE to expose +2 stops over, but if someone doesn't, they could be leaving lots of UNUSED headroom at the top of their scopes and when they go to normalize it, they are just going to pay a high price in bringing up noise to compensate. This sounds odd to say but even in rec709 the principle still applies. If your intention is to grade it yourself and your scene allows you to get away with letting in more light on your sensor and over exposing without clipping something important, you could cut 3db or more of noise in post by bringing your levels down.....yes, even in rec709! But yeah,...you have allot less headroom to work that with than SLOG and your odds are much less likely. But some scenes could work that way for you. Watch out for any knee that your rec709 might have up there.

It's no different than recording your cassette tape levels that only peak at -12db. You'll be forced to turn the amp volume higher on playback and that raises the noise floor. (tape hiss)

So with SLOG-2/3 I often tend to press it aggressively to the right (histogram speaking). In SLOG-2, I might even press my super whites right into clipping at 109 and then back them off to 107. I mean,...it totally depends on the scene but the bottom line for me is to let in as much light as I can, maximum signal to noise ratio and the most DR for shots that need SLOG. (more light on a sensor = less noise)

For me, doing it this way, noise is less of a problem as it gets shoved down on rec709 conversion. Others?...well, they might scream "foul" and say not to break any "rule books". Oh,...I grade what I shoot so I'm not handing to any unsuspecting client or coworker.

SLOG has allot of headroom and no hard knee to deal with at the top. People don't have to use all that headroom, but it's truly a waste if they don't and the less headroom they use, the more noise they will have to deal with in post.

As far as green color shift? Yeah, I used to fight with that in SLOG. I too sometimes get radioactive, nuclear glowing green grass. Grass and leaves are something the human eye is used to seeing in nature everyday so when it's "off" it stands out much worse that if someones green shirt is off. So yeah, I'll pull back on green saturation and the lightness of the channel. Also, shifting the green channel phase (tint) will also help.

Oddly enough. I tried shooting a green Listerine bottle with my FS5 in S-LOG 2&3 with all S-GAMUT flavors. They all turned that green bottle "blue" or "teal". I then tried it again but with my A7s-II....same thing. Tried with my A6300....same thing. All cameras S-Gamut's shifted green in the blue direction.

I then decided to switch over to the "PRO" Gamut with S-LOG 2/3 but pulled back the saturation to -8 in camera. Oddly enough, all three cameras color were spot on accurate, before and after the grade. No shifting or individual channel saturation was needed. (just and overall saturation adjust)

I'm still experimenting with SLOG and "PRO" -8 color gamut. For years I have always used S-GAMUT with SLOG....but so far so good.

[Piotr Wozniacki]

Originally Posted by Doug Jensen

Well, you've done your testing and come to your own conclusions and that is what everyone should do. Fortunately, my testing, results, and opinions don't match yours at all. And judging from all the great work I see from others who have used S-LOG on their productions it is clearly better than you give it credit for. You are welcome to come to your own conclusions. But if S-LOG was as bad as you make it out to be I'd sell my cameras tomorrow and buy something else instead. Fortunately, I am quite happy with results I get on REAL shoots and will be keeping my cameras and continuing to shoot S-LOG with no qualms or reservations about doing so.

All I said has only been intended to (partially) confirm what some other participants in this interesting thread have said; I've never dismissed or denied advantages of shooting/grading with Slog entirely! So believe me - I'm not going to drop using it when (basing on my test) I assess Slog would be advantageous... And even less likely will I sell my wonderful FS7 for this (or any other) reason!

Originally Posted by Cliff Totten

I then decided to switch over to the "PRO" Gamut with S-LOG 2/3 but pulled back the saturation to -8 in camera. Oddly enough, all three cameras color were spot on accurate, before and after the grade. No shifting or individual channel saturation was needed. (just and overall saturation adjust)
.

Interesting technique, Cliff (particularly not fiddling with individual channels, and yet getting intended results of mitigating those offending channels in the first place) - I will certainly try it out myself! Thanks for sharing;

Piotr

[Nigel Davey]

Interesting video that compares 709 vs Log in places:

FS7 and FS5 - How well do they get along? on Vimeo

Not sure I always like the graded Log, but it does looks less 'video' than the 709 on my Ipad. But perhaps the 709 could have been graded more in that direction, ie that sepia'esque filmic look.

[Piotr Wozniacki]

Thanks for sharing that, Nigel!

I'd be most interested in knowing how you "delog" your footage shot with the "PRO" Gamut with S-LOG 2/3 but with saturation pulled back in camera; are you using LUT (if so, which one), or DCM in Resolve?

Thanks,

Piotr

[Nigel Davey]

I didn't shoot it Piotr. I just saw it pop up on one of the four FS5 Facebook groups I follow. Given my original post/request for any direct rec709 vs. Log comparisons, I thought I'd share it.