FS7: This grading thing (or why choose Cine IE or Custom mode)

haha, yes I didn't mean reflectors for use trying to burn the soul out of a horse, but rather if you're doing interviews or similar - Good to know though re horses being weird, and I'd love to have seen that 2k fly.
 
As Doug states Slog 2 has more information in the highlights.

I usually shoot in custom mode and use either Slog2 or 3. 2 if I have to protect very bright lights, 3 for more normal scenes. I wrote about my settings a while ago including all the settings with video to demonstrate how nice custom can look right out of the FS7.

Basically I crush the blacks a bit and up the saturation and sharpen image a bit, this makes the video files usable as is, but they can greatly be improved with grading.

Here is a link to that post:
http://www.dvxuser.com/V6/showthrea...ike-it&p=1986554276&viewfull=1#post1986554276
 
Slog3 doesn't clip any earlier than Slog2 but it does have fewer bits allocated to each stop above mid gray so the highlights are less well represented. Once displayed on rec 709 the same highlights are going to be well and truly compressed in the non linear region, so if they are going to be blown anyway, why use up your bit budget trying to protect them with Slog2?
 
Slog3 does clip sooner than slog2, at 94%. But this is because it's allocating for 1.5 stops headroom possibly for next camera with higher sensor dynamic range cameras.
 
I know that S-Log3 is designed to cover 16 stops, hence a 14-stop sensor won't get the most bits allocated on that curve.

The Dolby Pulsar specs say it's 600mc/m2, which is only about 50% brighter than a typical LCD. But I do hope that new displays capable of extreme dynamic range will arrive on the market. I get some benefit by turning the backlight all the way up, but the blacks go to heck.


I'm leaning toward NOT baking in the LUT. S-Log2 seems friendlier to grade, too. No interviews, just showing the horse. (presumably the horse is for sale). I'm thinking perhaps a low angle, looking up will make the animal look more impressive to buyers, while avoiding direct view of windows across the arena (hopefully).
 
The Dolby Pulsar is not necessarily about brightness, but available contrast. Some of that comes from getting the whites brighter, but there's a lot more to it than that.
 
basspig said:
The Dolby Pulsar specs say it's 600mc/m2, which is only about 50% brighter than a typical LCD. But I do hope that new displays capable of extreme dynamic range will arrive on the market. I get some benefit by turning the backlight all the way up, but the blacks go to heck.
Click to expand...
They are calling it 10,000 nits, many times brighter than typical LCD. Three video heavyweights from SpectraCal—(L-R above) Joel Barsotti, director of software development; Derek Smith, CTO and co-founder; and Stacey Spears, chief color scientist (and co-creator of the Spears & Munsil HD Benchmark Blu-ray)—geek out about high dynamic range, electro-optical transfer functions, bit depth, peak luminance, how video HDR is different from photographic HDR, color gamuts and volumes, different HDR formats, calibrating HDR displays, answers to chat-room questions, and more. Watch the podcast here. http://www.avsforum.com/forum/138-avs-forum-podcasts/2068898-spectracal-hdr.html
 
Mitch Gross said:
The Dolby Pulsar is not necessarily about brightness, but available contrast. Some of that comes from getting the whites brighter, but there's a lot more to it than that.
Click to expand...
Hi Mitch, It really is about the brightness in this case. Watch the podcast I linked above. Peak brightness to be clear although the Pulsar has more than just an intermittent capability.
 
Tom Roper said:
Slog3 doesn't clip any earlier than Slog2
Click to expand...

If you pull out a real waveform monitor you'll see that nothing can go above 93% with S-LOG3. Everything flatlines just above that point. On the other hand, S-LOG2 goes to 107%. And that difference really helps on outdoor shooting and any contrasty uncontrolled lighting situations -- such as the horse arena with lots of windows.

A nice thing about shooting RAW is that I can process my footage in post as S-LOG2 or S-LOG3 with no penalty, and a couple of years experience doing comparisons with the F55 has clearly shown to me which version of S-LOG is better for this type of shoot. If you listen to some people they'll tell you that both versions of S-LOG can be graded to look the same. That might be true sometimes, but not all the time. Otherwise why have two? There are some situations where one is a better choice than the other. This is a situation for S-LOG2.
 
Last edited:
Basspig.. I would advise you to interview the horse.. then you have it from the horses mouth..

Let the window,s burn out.. its not your subject matter.. and its not for broadcast I guess..
 
Hi Doug, yes I see your point and Dennis' too, Slog3 does clip earlier, at 93-94%. Thanks for pointing it out. Don't talk to me about raw, I am green with envy.
 
SLOG3 isn't clipping earlier, the gamma curve is simply different. You still have the exact same highlight latitude (assuming matching exposures). If you're clipping in SLOG3, you'll be clipping in SLOG2 as well. The difference lies in the allocation of data per stop, not in the distribution of those stops.

So SLOG2 doesn't give you more highlight latitude, it's giving you greater tonality WITHIN the highlight latitude you already have.
 
Tom Roper said:
Hi Mitch, It really is about the brightness in this case. Watch the podcast I linked above. Peak brightness to be clear although the Pulsar has more than just an intermittent capability.
Click to expand...
Thanks Tom, I've seen the presentation and in fact the Dolby system itself in demo. I do understand that the brightness is what allows so much functional dynamic range, contrast and appearance of resolving power. But it is not brightness for brightness sake. The system has to maintain blacks, prevent bleed through, have a fast decay and response time and other factors. So I was trying to indicate that there's a lot more to it than just brightness, otherwise one could use a stadium Jumbotron.
 
Mitch Gross said:
The Dolby Pulsar is not necessarily about brightness, but available contrast. Some of that comes from getting the whites brighter, but there's a lot more to it than that.
Click to expand...


Mitch, are we talking about the same monitor here? Googling Dolby Pulsar brings me this:

http://www.dolby.com/us/en/professional/cinema/products/prm-4220.html

They're talking about pretty docile luma levels. And only 1920x1080.

I've been aware of a need for more peak white levels for years, and have experimented with cranking up the backlight on footage shot low key to protect highlights. XDCam EX proved to be too noisy to do this effectively, but the FS7 is yielding much better results. But when I publish some of my test footage, other people tell me it looks dark on their monitors.

Over on a forum I frequent for home theater tech, an interesting article was recently written by one of my associates there: http://www.whatsbestforum.com/showt...nding-Ultra-HD-Blu-ray-Disc-and-TV-Technology
 
Last edited:
Grug said:
SLOG3 isn't clipping earlier, the gamma curve is simply different. You still have the exact same highlight latitude (assuming matching exposures). If you're clipping in SLOG3, you'll be clipping in SLOG2 as well. The difference lies in the allocation of data per stop, not in the distribution of those stops.

So SLOG2 doesn't give you more highlight latitude, it's giving you greater tonality WITHIN the highlight latitude you already have.
Click to expand...

I disagree. Real world experience says otherwise. I'm not going to waste time posting some examples, but I know for a fact that S-LOG2 is better for highlights and does not clip at the same point as S-LOG3.
 
Doug Jensen said:
I disagree. Real world experience says otherwise. I'm not going to waste time posting some examples, but I know for a fact that S-LOG2 is better for highlights and does not clip at the same point as S-LOG3.
Click to expand...

It is not clear what you are saying here. If you use the methodology of stopping down until you have no clipping then neither slog2 or slg3 clip (!) .. so to evaluate the two you would do that and then examine other facets of the image, like when the darks are lost in shadow or the gradability or tonality of the image?

S
 
I can't speak for Doug but personally what I'm saying when I say slog3 clips at 94% is open iris on any lens aimed at fairly bright lit white wall or light subject and slog3 clips at 94% IRE on your waveform monitor.

But in terms of tonality, gradability, etc. they are valid points. slog2 will give you more data and tonality but again it depends on the scene, how you're exposed, and what will happen in post or who will be doing it.

So many variables that to me picking between slogs2 is kind of like talking about who's car might be faster.

Who cares.

If you're shooting and need that much flexibility overall, then shoot RAW.
 
Last edited:
Personally I have no clue which is 'better' or 'better for what' - does anyone else have a clue? Can they show me with image?

To me 'better' is an equal allocation of data to stops because (I think) that this leads to the most flexible file (ie bendable in post) - the very opposite of the HG curves which, for example, bummed you if you let skin go beyond a certain level (into the knee)

The graph posted above shows Slog3 having that equal allocation so looke 'better' to me, ie the graph looks better.. but Im not 100% bought in to that. Certainly Ive been bemused by the clip at 93 which would appear to give an advantage to slg2.

So I can see theoretical pro's to both but have not done a real world (but well conceived) test. A well conceived test would not involve examining the highlight retention of both at a given Tstop as its sole metric.

S
 
Last edited:
morgan_moore said:
Personally I have no clue which is 'better' or 'better for what' - does anyone else have a clue? Can they show me with image?
Click to expand...

I can't offer an image yet as I haven't tested it but will offer two scenarios where you might choose SLog2 and SLog3.

Looking at the sample distribution it would tend to suggest shooting the sun sparkles on the sea would be best done with SLog2, a scene where there isn't much in the way of dark tones. A magic hour scene with long shadows with a mix of light detail and dark detail might be better in SLog3.

If the darker tones of SLog3 footage are noisy and exhibit compression artefacts then the benefits of the number of darker samples in SLog3 becomes moot. So ETTR and using NR might get full value out of SLog3.

For most scenes the goto tone curve would be SLog3 with the SLog2 being better under exceptional circumstances. Would you really notice though? It'll take pixel peeping an extreme grade to spot much of a difference and would depend a lot on how compression artefacts are distributed through the scene. You could easily foresee a situation where the posterisation from compression artefacts render the differences between Slog2 and Slog3 moot not the sample differences between each curve.
 
Another advantage of Slog3 is the color gamut that comes with it.. S gamut3.cine.. the primaries line up with Rec 709 and P3.. its very similar to Cineon and LogC.. and so you don't have the green "skew" to deal with.. a lot more LUTS can be used.. and you don't need the best colorist in the country to get your log footage into Rec 709 looking its best..

Nirv..Slog 3 is not more noisy than Slog2.. SNR ratio is the same...or displays compression artifacts.. ! just more data in the darker levels.. as per the graph.. so "appears to be more noise".. before a LUT is applied.. its actually giving you "better" more shadow detail..ETTR is a stills approach and works as its only one frame.. for movies this will give you a lot of problems in post.. correcting every single shot..
 
Last edited:
You must log in or register to reply here.
Back
Top