Jump to content

The correct way to expose for SLOG3 when using 8bit cameras


kidzrevil
 Share

Recommended Posts

On 10/29/2017 at 6:28 AM, kidzrevil said:

@maxotics idk man im just not seeing the problems you have experienced with LOG profiles shooting SLOG3. The crushed shadows were a stylistic choice but I will grade the image again. If you look at the clips under a waveform and vectorscope you will see that nothing is blown out. I made sure to expose below the clipping point to prevent any kind of overexposure. 

I chose to shoot this in LOG over REC709 for multiple reasons. 1) you are shooting at the native iso and also rate the ISO slower by over exposing. 3 stops overexposed at 800 iso is an image equivalent to one shot at 100iso 2) the cameras dynamic range doesn’t chance with any picture profile only the gain does. You will always get a better signal to noise ratio over a rec709 who’s noise floor is hidden in the shadows and will quickly appear when raising them in post. 3) sgamut3.cine works better with color than standard rec709 profiles because of how it handles saturation. After around 40 ire I believe colors don’t get more saturated they just get brighter like film. That color science is not present in any other profile. 4) Im not having the issue with loss of color data or banding because I raise the in camera saturation. SLOG3 white papers suggest to raise in camera saturation to emulate printed film. In post it preserves more color information. The SLOG3 & cineon white papers both state they were made to support 8,10 & 12 bit data so the 8bit argument when shooting log is null and void. Expose up to assign the higher code values of the codec to the image and you are fine. 

So in my experience there are significant gain in using LOG profiles over REC709 under the right lighting conditions. I will regrade this to look flatter.

I have never experienced the banding problem. Do you see it present in my footage ? Im not sure if I missed it but I didn’t see ang banding when I shot or graded this

Absolutely right. There is no exposure trick to get more color depth and this is why you raise the in camera saturation. Increasing the saturation holds onto more color information. SGAMUT3 is large enough to support this extra color info without blowing out a color channel. You can desaturate in post

Now if only I could apply this to my GH4 internal V-LogL - especially the part about no banding...

Link to comment
Share on other sites

EOSHD Pro Color 5 for Sony cameras EOSHD Z LOG for Nikon CamerasEOSHD C-LOG and Film Profiles for All Canon DSLRs
15 hours ago, Mattias Burling said:

Not to that extent. Not even close.

I rarely encounter banding on my A7s and A7sII. What were you shooting? How were you grading? I'd be interested to see examples of banding given I've not experienced it, yet you seem to find it so bad to the point of unusable...

19 hours ago, cpc said:

Actually, I can't think of any non-Arri cameras that don't need some amount of overexposure in log even at higher bitdepths. These curves are ISO rated for maximizing a technical notion of SNR which doesn't always (if ever) coincide with what we consider a clean image after tone mapping for display. That said, ETTR isn't usually the best way to expose log (or any curve): too much normalizing work in post on shot by shot basis. Better to re-rate the camera slower and expose consistently.

As you've sort-of covered... it isn't overexposure necessarily. Knowing the curve and where it places its values is extremely important when shooting log. Moreso when shooting 8-bit log. Blanket over-exposure won't necessarily give you better results. As for Arri, I regularly rate them at 400. 800 may give you the greatest spread of dynamic range, and results are generally acceptable. I find rating at 400 better. Knowing the curve and placing your exposure properly will give you even better results.

15 hours ago, kidzrevil said:

You want more shadow detail ? Expose up. You want more highlight and midtone detail ? Expose one stop over.

Have fun with constantly shifting depth of field and days (and therefore money) in the colour suite trying to get everything to match....

 

Link to comment
Share on other sites

1 hour ago, jax_rox said:

As you've sort-of covered... it isn't overexposure necessarily. Knowing the curve and where it places its values is extremely important when shooting log. Moreso when shooting 8-bit log. Blanket over-exposure won't necessarily give you better results. As for Arri, I regularly rate them at 400. 800 may give you the greatest spread of dynamic range, and results are generally acceptable. I find rating at 400 better. Knowing the curve and placing your exposure properly will give you even better results.

One discriminating characteristic of log curves compared to negative is that (on most of them) there is no shoulder. (Well, the shoulder on Vision3 series films is very high, so not much of a practical consideration unless you overexpose severely.) An effect of this lack of shoulder is that you can generally re-rate slower without messing up color relations through the range, as long as clipping is accounted for. Arri's Log-C has so much latitude over nominal mid gray that rating it at 400 still leaves tons for highlights. I don't think any other camera has similar (or more) latitude above the nominal mid point? Pretty much all the other camera manufacturers inflate reported numbers by counting and reporting the noise fest at the bottom in the overall DR. No wonder that a camera with "more" DR than an Alexa looks like trash in a side-by-side latitude comparison.

Link to comment
Share on other sites

 

7 hours ago, jax_rox said:

 

Have fun with constantly shifting depth of field and days (and therefore money) in the colour suite trying to get everything to match....

 

Never been a problem for me. I’ve been freelancing for years and never ran into that problem on a paid project. Thats a novice issue like shooting on auto. It shouldn’t take you more than a day to color grade a 10 minute long project. No one should have this problem if we are doing this professionally for corporate or consistently making indie films. Please don’t make this sound harder than what it is, you can potentially steer someone away from a school of thought that works for them. @jax_rox

5 hours ago, Rodolfo Fernandes said:

Great job @kidzrevil do you mind sharing your S-LOG3 settings with us, or is it just 'straight out-of-the-box' 

Sure ! Detail -7 in manual mode with 7 crispening, 0 limit,b/w balance to type 1,v/h ratio to +2 and saturation turned up to 14. I also use Dali’s mustache to transcode to 1080 444 10bit. The transcoding step is very important, Adobe Premiere is not truly downsampling the image just scaling it to fit the project settings !

Link to comment
Share on other sites

On 10/31/2017 at 8:17 PM, cpc said:

I don't think any other camera has similar (or more) latitude above the nominal mid point?

I've yet to encounter a camera that properly stacks up against the Alexa when it comes to DR. Maybe the F65. Possibly the new Varicam (though I've never shot with it). 

That said, changing your ISO on the Alexa does affect your dynamic range above middle grey. As your grey point shifts, you get slightly less above and slightly more below.

22 hours ago, kidzrevil said:

Never been a problem for me. I’ve been freelancing for years and never ran into that problem on a paid project. Thats a novice issue like shooting on auto. It shouldn’t take you more than a day to color grade a 10 minute long project. No one should have this problem if we are doing this professionally for corporate or consistently making indie films. Please don’t make this sound harder than what it is, you can potentially steer someone away from a school of thought that works for them. 

I'm personally not sure how you could possibly match two shots (say a dialogue scene) where one person was shot at T2.0 and the other at T5.6. 

I don't know what level or what kind of work you do, but if you're paying a colourist $500+/hr you don't want to have to have them adjust every single shot in your entire piece to make it look right or your Producer isn't going to be particularly happy. 

Light your scene to a stop and stick at that stop. You may rate your camera for over-exposure but you should rate for consistent over-exposure.

If you're talking about event coverage, for example, then sure you're probably going to have to expose each shot differently. However if you decide to over-expose, you should aim for a consistent amount. Over-exposing two stops on one shot and four on another is only going to cause you headaches. Why spend a whole day grading 10 minutes of footage when you could do it in half that and end up with the same result (and a happier Producer)..?

Link to comment
Share on other sites

6 hours ago, kidzrevil said:

@jax_rox again I have personally never encountered that problem or know of anyone who has but I am more than willing to keep your advice in mind for my upcoming shoots. Where do you think I should lock my exposure when shooting SLOG ? Should I go 1 or 2 stops over ? 

Which part have you not encountered? Mis-match of dept of field..? If you shoot the way it appears you are, i suspect you probably have but it either didn't matter (as it wasn't, say, a film) or you didn't notice it, or you're adjusting your exposure with ISO rather than aperture (though this would create a mis-match in noise).

As for exposing Slog, it depends on which version and with which gamut. SLog2 and SLog3 place their mid grey tones in entirely different spots. You should know this if you intend to shoot with either and you should know what they are and you should use the one that suits you and what you're shooting best. Know your curves.

 Overall, the best thing you can do is exposure tests. Go out and do tests. Expose 1 stop over. 2 stops over. 3 stops over. 4 stops over. Etc etc. come back into Resolve and grade it and see what you like best from the standpoint of colour, noise, dynamic range etc.

If you do your tests properly and extensively, you'll also begin to see where exactly you lose detail. Where things start to clip. Where you're no longer happy with how things look. This is all subjective, but it gives you your personal rating of that camera. You start to understand where you can expose things before you lose them. You end up knowing exactly how many stops you feel comfortable using, and exactly where you want to expose to ensure the best image. This is what I mean when I say shooting Slog shouldn't be a roll of the dice. If you test appropriately, you will know, firmly, exactly where to expose everything and how it will look. 

I can't tell you what that point is, because it's subjective. I like rating the Alexa a stop slower because I've done tests and I like the image better and I know I lose dynamic range above middle grey by doing so. But I know exactly where I lose that detail so I can expose as I need to. Others rate it at 800. Others aren't fussed by the extra noise, or prefer to have the extra dynamic range. I prefer to rate the FS7 1.5-2 stops over. I rate the F5 2/3 of a stop over. I rate the F55 1 stop over. Some agree, others don't. Point is, I've tested the cameras and that's where I like my exposure to sit, and where I'm happy with the image I get out of it. There's no reason you can't do the same, and you will end up with better, repeatable results.

Link to comment
Share on other sites

@jax_rox oh ok so Im fine then, no need to test. I think I got a handle on the limits of the camera when exposing for log. Exposing to the clipping point has been fine for me. Exposing to what the meter reads as +2 when using A priority gives great results for run and gun as well and keeps the highlights at a reasonable distance below the clipping point. Works great for consistent results especially when using the a6500 new average scene exposure metering.

Link to comment
Share on other sites

9 hours ago, kidzrevil said:

This was shot in slog2 with sgamut3.cine color matrix. It is overexposed by +2 and color corrected with colorista iv

Love what you're doing here, always find your threads to be the most interesting to read on here. Curious why you chose to go for -7 detail and then +7 crispening? I've never touched crispening before, and found that -7 detail was actually a little too much. I've since moved to -4 and found that my images now appear to be 'sharp'. I know -7 detail is best practice and all, but it was too soft for me.

Link to comment
Share on other sites

@Gregormannschaft thank you ! That means a lot to me. Im glad I can be of service to this community !!!

I use +7 crispening because it prevents the in camera sharpening from activating. I do this to avoid sharpening artifacts but you may be onto something. Sometimes I feel like the image is mushy at -7 and so I will try raising the in camera sharpness up and see where it leads me. 

Link to comment
Share on other sites

30 minutes ago, kidzrevil said:

@Gregormannschaft thank you ! That means a lot to me. Im glad I can be of service to this community !!!

I use +7 crispening because it prevents the in camera sharpening from activating. I do this to avoid sharpening artifacts but you may be onto something. Sometimes I feel like the image is mushy at -7 and so I will try raising the in camera sharpness up and see where it leads me. 

Interesting. That "mushy" feeling was exactly what I've always experienced. When I brought it up to about -4, I found a bit of structure was brought back to my images, especially helpful for uploading to YT or Vimeo. But your footage looks really nice, very organic and sharp. Perhaps on the A6500 it's not such a bad thing to have -7, as you're already downsampling 6k footage. On the A7SII you don't have that. 
 

I've also really enjoyed using Geoff's GFILM settings, might be worth a shot trying out for yourself. I'd be curious what you could do with it.

Link to comment
Share on other sites

@Gregormannschaft hmm maybe in camera sharpening isn’t the devil it was once made out to be. Have you tried raising it even higher than -4 ? Its worth testing to see what is the limit before sharpening artifacts become present. As far as the organic look to my footage I blame it on the lens that I used. These old nikkor pre ai lenses with a speedbooster brings a very unique rendering to the image

Link to comment
Share on other sites

41 minutes ago, kidzrevil said:

@Gregormannschaft hmm maybe in camera sharpening isn’t the devil it was once made out to be. Have you tried raising it even higher than -4 ? Its worth testing to see what is the limit before sharpening artifacts become present. As far as the organic look to my footage I blame it on the lens that I used. These old nikkor pre ai lenses with a speedbooster brings a very unique rendering to the image

I'm also using primarily vintage glass - I find that with 4K you want as much character as possible in the image, even if it's not technically perfect. So I have a few FD lenses, a lovely Zeiss Jena 35mm and the cheapest lens I have which is probably my favourite - a 50mm Zeiss Tessar f2.8 which I picked up for 20 bucks. So that's what I'm working with, and I've found that -3 is probably my limit in terms of in camera sharpening. I eventually settled on -4 as a happy medium for vintage glass. It could be slightly lower on the A6500 but might be worth a shot. Even -5 made a noticeable difference.

Link to comment
Share on other sites

@EthanAlexander it works very well for those who like to grade manually ! I do however prefer my modified slog3 with sgamut3.cine because of how it behaves with LUT’s. I am going to try it out with increased in camera sharpness and locking the exposure at +2 since I know people generally expose with the in camera meter. I noticed SLOG3 squeezes the entire histogram in the middle which contains the least amount of compression artifacts. 

Link to comment
Share on other sites

But you're losing 95-109 IRE with SL3. You can actually make a LUT (or find, or hell I'll send you one) that transforms SL2 into 3 after shooting, and it's the exact same for grading but you have more starting info because it originally clipped at 109. Compression artifacts IMO don't matter more than getting the most code values from the beginning. You're re-sampling down to low compression 1080 anyway.

Link to comment
Share on other sites

Figure I'd upload a few screengrabs from a recent project using the SLOG2 + SLOG3.cine combo we're talking about here. I used a DeLUT as a starting point to grade - most aren't great but I love Downton Alt. Tends to give very natural colours without doing too many weird things with grain or colour blocking. These were all shot using a Tiffen BlackPro Mist 1/4, which at times was possibly a bit much. But when it works I really love the effect it gives.

 

JJ BTS Screengrab2.png

JJ BTS Screengrab8.png

JJ BTS Screengrab11.png

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

  • EOSHD Pro Color 5 for All Sony cameras
    EOSHD C-LOG and Film Profiles for All Canon DSLRs
    EOSHD Dynamic Range Enhancer for H.264/H.265
×
×
  • Create New...