cpc

Hey Andrew, when you say that film simulation doesn't work in video mode do you mean the film grain simulation only or also the film color modes?

BM Pocket with an optical AA filter?

Looking forward to seeing this in IMAX2D this week.Curious if the "approximately 13% Alexa 65" is actually distinguishable from all the other 87% Alexa footage.

It should be noted that the Alexa would probably be the easiest to match with Kodak since it takes a lot of cues from it. Also, most of the film look colorimetry related properties come from printing the negative on release stock, not from the negative itself. The reasons for this are quite deep, but essentially they boil down to negative film being technically more feasible to design without excessive color shifts because the orange mask allows for some clever tricks. Positive/projection films on the other hand don't have this luxury.

Since the two sources in the video are not printed, it is much easier to achieve a decent match by feeding them both into a P3/rec709 output space. Still a great video though, and I do agree with the sentiment behind it.

Photoshop (which uses Adobe Camera Raw) will only have the Embedded profile available since it doesn't recognize Blackmagic cameras. ACR debayers in ProPhotoRGB, and then does another transform (to sRGB) on display and export. This is approximately equivalent to rec709 in Resolve.

Resolve will not debayer in BMDFilm unless you set it to do so. Photoshop will not debayer in BMDFilm, because it doesn't know what BMDFilm is.

If you debayer in BMDFilm color space using BMDFilm gamma you should be able to match ProRes/Film colors. These settings are in the Camera Raw tab on the Color page.

Resolve's manual is actually pretty good and informative. It is a good idea to have a look at the relevant sections.

SGamut isn't friendly to manual adjustments because of the heavily offset green primary. This is the reason Sony introduced SGamut.Cine which aligns much better with P3 and sRGB/Rec709.

Cinema, Pro and Movie (same as Stills or at least very similar, I think) gamuts have primaries aligned with Rec709/sRGB and are for this reason easily moldable with simple luma curve adjustments without introducing hue disbalance. Skin hues are subtly rotated between these (Cinema is a bit towards yellow, Movie - towards magenta, Pro in the middle; but these are subtle), and there is a small difference in hue compression (Cinema > Pro > Movie; Movie is the most compressed) but all three are fine. I've been shooting s-log2/Cinema, but wouldn't hesitate to use any of these.

It is really a question of how you interpret the color values. Since SGamut is the largest it may appear less saturated (= more subtle) even after treating it. But it certainly does have the problem (without proper remapping) of pulling skin towards yellow.

Film/Video setting doesn't affect raw. It is always flat (film) internally, but what you actually see depends on the raw processing settings in your software. What app do you use for processing raw?

I really like what 5d3 raw does with faces. I find it has just about the perfect amount of sharpness. And I've yet to see a better HD/2k camera for narrative in terms of image quality. Interesting that the 5d2 can be made to match this. Is 5d2 raw usable at iso 1600 with proper exposure?

That's a strange thing to say. I think Canon color science takes a lot of cues from film. Now of course these promotional videos are all graded so they are more representative of the colorist's inclination than anything.

The autofocus features are beautiful. Dual pixel AF was great on the prev gen C cameras, and this is another step in the right direction. I like how they are pushing this with face tracking, manual focus guides and the movable AF area. A dream for documentaries.

I think this is the most interesting camera of the recently announced/released bunch. Curious which features will propagate down to the next c100.

Hello!

I haven't slept much last night but here it is, the α7R II low light test! (i also posted my initial thoughts at the beginning but feel free to skip to the test if you get bored!) 

The results are very surprising, whereas during the day i couldn't really spot a difference in quality between the FF and S35 modes things are completely different when the lights go down: the S35 mode is vastly superior to the Full Frame, they are not even in the same league! I don't know if that is only the result of the better readout from the sensor in S35 or there is some noise reduction wizardry happening in real time but the output from the S35 mode is way cleaner.

It is not surprising and the reason is exactly the same as the reason FF mode has less rolling shutter than s35: the camera simply uses more photosites in s35 mode, hence less noise. 

This only applies to humans. We simply don't see color well at low light levels because of the way human vision works.

Cameras don't have this problem. What you perceive as lack of color in low light captured images is for a couple of reasons:

1) In naturally occuring low light situations the spectrum of the light is often overly monochromatic. For example, dusk light has lots of blue and tungsten has a strong red component. This offsets the image toward monochromatic.

2) Overly noisy image.

Anyhow, the A7s s-log2 lacks tonal precision even when shot at its minimal ISO (which is 3200) and overexposed for healthier signal.

Not enough of a difference for you to care then, I presume?

That's what I've observed in this synthetic case but I'm not gonna claim that it is universal law.  

Thanks.

Actually, my whole point with this paragraph was that even uncompressed 8-bit 4:4:4 log has its limits and they aren't particularly high.  : )

re: Cineon

I don't know what do you mean by R', G,' B' (gamma corrected values?), but Cineon encodes film densities which do represent scene light intensities logarithmically, the same way digital log curves do. It is indeed meant to be printed on print film. But the comparison is fair I think. If anything, one actually needs more precision for digital because print film grain, as fine as it is, still acts as dither.

Incidentally, a few months ago I was involved with tone mapping and grading this real-time rendered short:

It uses high resolution 8-bit textures, so the source image is effectively 8-bit RGB (4:4:4) and the engine works in high bitdepth precision and/or floating point (so negligible quality loss with lighting and processing). There is a (perfectly flat) log conversion operation from the engine linear space at the end of the pipeline for film out simulation purposes (simulated Kodak print). It is parameterized for around 12 stops. The grade is done over this perfectly flat curve. I found that somewhere around 12 stops is where you start to see issues if you are picky (and I am), even considering the pulled down whites in this grade and the overlayed grain.

As an a7s owner, with all due respect, I can say that you are wrong. There is huge difference between s-log2 and raw. Both color and image density are significantly worse with s-log2 compared to the raw stills. Also, a7s externally recorded 4k downscaled to 1080p is easily more gradable than the in-camera 1080p.

good examples here: http://blog.inventome.com/Blog/2015/2/a7s4Ktests/Sony-a7s-Exposure-and-Noise-Workflow-with-UHD-and-Odyssey-7Q-Plus

Sensor, noise, processing, etc surely matter a lot but downplaying the importance of the available tonal precision is pointless. Yes, you can have badly quantized and processed data encoded in 10 bits, and it will still be crap. And you can have well sampled, quantized and processed data encoded in 8 bits. And it will STILL be bad if the curve is too flat for the available space. Ideally, one needs well processed data in at least 9-10 bits for log encodings with extensive DR.

It is no coincidence that the father of all log curves Cineon log is 10-bit. Back then nothing was taken for granted, nor there was any prejudice about 8 bits, so this was extensively researched by Kodak. They concluded that 9 bits are likely enough but went for 10, since 10 fits better with computers; plus, a little bit of redundancy doesn't hurt.

And lets not forget chroma subsampling. After all 4:2:2 is twice the chroma info contained in 4:2:0, and 4:4:4 is four times the chroma info contained in 4:2:0. So not all 8-bit is created equal.

In general, wider color gamuts do help grading by having better color separation, but there is no point having a wide camera gamut if there isn't the color precision to encode it properly as is the case with S-Gamut and s-log2 in the a7s. Note that many digital cameras have native color gamuts wider than film leading to color precision smaller than scanned film when encoded with the same bitdepth.

sunyata's analogy is quite good. A small correction only: Prints don't have a linear representation of the scene light, not at all. Prints are heavily gamma corrected for projection in dark environments much more so than material meant to be shown on emitting displays. Print-through film curves (that is, scene-to-projection) typically have a gamma in the range 2.5-2.8.

Now first it is important where the "log" comes from. It is because humans perceive exponential light changes as linear changes. This is a logarithmic relationship. Hence, log. Log curves mimic this. Exponential scene light changes are recorded as linear changes. In other words, each increase of exposure with a stop (or doubling the light) takes the same number of coding values to encode, and not double the values of the previous stop (as do linear encodings).

There are a couple of technical benefits:

1) Much more effective and economical utilization of available coding space. This is the reason log curves encode wide dynamic ranges effectively in a smaller bitdepth. Cineon was developed to capture the huge DR of negative film in only 10 bits.

2) (And related to 1) Increased tonal precision in the dark parts of the picture, compared to a physically correct linear encoding (when using the same coding space).

Since sensors work linearly, purely logarithmic curves would waste some coding space in the blacks, because there is not enough density there. That's why practically all log curves are pseudo-log, with some compression in the black end. Arri's Log-C is probably the closest to pure log. Canon's C-log is the furthest away from pure log. The other reason is, as mentioned, mimicking Cineon. This is also, I believe, one of the main reasons all log curves have a raised pure black level. This mimicks the base density (D min) of film, as encoded in the Cineon curve to accommodate scanning film densities.

​You can increase the DRO setting if you are going to use Autumn Leaves for grading. DRO +3 flattens the low-mid range of all the styles.

It doesn't have more highlights, whait it does is pull mid grey lower and clip signal at IRE 100. The lower midgrey point may create the illusion of more highlights latitude, but it is the actually opposite. The other three Cine curves go to IRE 109. Cine2 is supposed to be used "as is" and be broadcast safe (hence clipped at IRE 100).

I don't think any peaking setting will work well with slog. Peaking relies on contrast and contrast is very low with s-log. One way to sort of work around this is to pre-focus with a contrasty PP and switch to s-log for recording.

Seems to be clipping quite a lot?

There is no reason to ever use Cine2 if you are doing grades. Cine2 has no superbrights.

​Not gonna happen because the m43 lenses can't conver a sensor that big.

Actually, it is highly unlikely the A7rII sensor ends up in a video camera.

The A7sII (whenever this comes) sure, but the 42mp A7rII sensor? Not really. It would be a huge waste of sensor area, since the full frame video mode is by Sony's own words not as good as the s35. It simply doesn't make any sense (yet) to use that many megapixels for 4K video.

​Well, I have the A7s and s-log2 is unusable in 8-bit if you care one bit about image quality. Plastic skin everywhere once you restore contrast; simply not enough tonal precision in 8 bits for the flatness of this curve. Besides, the s-log2 setting doesn't use a huge chunk of the coding space, this way further screwing the tonal precision.

Canon c-log is a different story. Its curve is not log really, just log-like in the upper end. And it has less DR than s-log. Much more 8-bit friendly.

Codec is important but not as important as the bitdepth. Try working with 16 stops log in uncompressed 8-bit and see what happens even if there is no codec at play.