TheRenaissanceMan

Totally agreed. I find Blackmagic to provide the most neutral starting point, with Panasonic (Neutral or V-LOG) and Nikon (FLAT) leading the mirrorless/DSLR space, respectively.

I don't know...it looks like improved color NR, not just compression. Probably processing and not an improved sensor, but if it gives us better shadows with less weirdness, then...*shrug*.

That's an interesting point. What programs are everyone using to grade? Would Resolve, with its 32-bit float color space, give better results than Premiere/FCPX?

That's wacky-looking...

Exactly. Shoot each camera for best results, not the most identical settings. They're different cameras with different strengths. I'd love to see a skin tone comparison with the A7S II shot Brandon Li style, with Autumn Leaves and mild CC.

Don't forget, when we're talking color and tonal precision in skin, we're comparing an 8-bit camera to a 10-bit camera. Not to mention that the A7S II is stretching a LOG profile across that 8-bit depth. Huge difference in codecs, as well. Sony's 100mbps XAVC-S codec isn't even as good as Panasonic's internal codec, let alone the C300 II's.

You might also try using an external recorder. Some of that "noise reduction" is just smearing from the H.265 compression. It might be contributing to the crunchy-looking detail as well. Bypassing the internal codec altogether and recording 4K Prores from the uncompressed HDMI might solve a lot of your image issues, and provide an easier workflow to boot.

Besides just 10-bit V-LOG tonality, do you find the difference in compression noticeable going from internal to external recording? I haven't done much work with the GH4, but people tell me the internal 4K suffers quite a bit in the shadows. Is the ProRes easier to manipulate/clean up?

There definitely is...you just don't share it. Important distinction.

The GH4 is just grainy in dark conditions, like the F35 or film. Once you accept that, you'll like it a whole lot better. Unlike A7X S-LOG 2/3, I've seen almost uniformly great results from V-LOG in samples from Vimeo and Youtube. Externally recorded and exposed properly, it can hang with--or even beat-- cameras in the the $7-$10k range. Keep in mind that 4K recorders will only get smaller and cheaper from here, and the ones we have now are already practical for most shooting situations.

This is different than I've seen on Blackmagic cameras. There, the blown highlights just turn black. Easy to fix with one click in Resolve. Here, there's this gunky purple transition from normal overexposure to black hole. I don't know how well Resolve will handle that.

On the big screen, I've found that lenses and grading make a far bigger difference in apparent resolving power than 4k. That's why Leica, Zeiss, etc demand such high prices.

That's how S-Gamut is supposed to work. It's a storage format, not meant to be graded by amateur colorists without R709/P3 LUTs. Every A7S video shot in S-LOG/S-Gamut I've seen that's been handled by a professional colorist has looked great, which tells me it's inexperience and lack of knowledge, not a weakness of the hardware. If you want something easier to grade with a simple curve or two, perhaps you should try the Pro or Cinema gamuts. Pro's given me great results in the two A7S shoots I've worked on.

How so? LOG is just one option in the technician's toolkit. It's not like using other gammas in low light eliminates every creative choice you have in low-light scenes. Also...calling out industry gurus for pixel peeping? Hey Pot....Kettle called...

Please. If it's good enough for Paul Thomas Anderson, it's good enough for our dumb asses.

I can just see it. I don't know what it is, but 10-bit always looks subtler and more natural to my eye. It also maintains a more pleasing, natural look when you apply some "style" to it in grading or push the saturation/DR recovery. Try testing internally vs externally recorded footage with some work done on them and you'll see more difference. I'm surprised you don't see any difference in the compression. Most reviews say the GH4's codec suffers from muddy darks, so you'd think an external recorder would help with pulling up and denoising the shadows without any weirdness. But I've only used one once and wasn't involved in post, so I can't say one way or the other.

Probably more of a codec snob.

People using the D750: Do you find it difficult to focus with no peaking and no magnification during recording? What's the highest ISO you'd consider usable? And how much "style" can you apply in grading before things get unnatural? The Nikons seem like pretty ideal cameras for me if only they had a more adaptable mount. Besides Leica R, there's nothing I can put on them that focuses the right way.

A stills body primarily targeted at video is a niche product. Not hybrid--that's common now--but primarily meant for video, with decent stills as a side bonus. Great for guys like us, but it leaves most stills shooters scratching their heads. Not to mention Sony's lossy RAW, abysmal battery life, and lack of tilt/shift lenses (adapted lenses are an option, but suffer from planarity issues that kill sharpness on such a pixel-dense sensor).

That's one of the big reasons I want to see NX1 footage shot with an external recorder, but for some reason, no one seems to have done it. At least judging by the lack of any test clips online.

Yes, there are some people dropping their Canon bodies for Sonys, but lots of them are still shooting with Canon glass. You think Canon doesn't make any profit there?

It can do 720p or so from the whole sensor, but it produces pretty bad moire. You could also go the Tragic Lantern route and just get a jacked up bit rate for h.264. If you're a stills guy, pair it with the 22mm and you've got the best, cheapest 35mm compact on the market.

Definitely get the groundwork shots to let your audience know what's happening and why. If they don't know what's going on, they might be intrigued, but they'll never be emotionally invested. Also, try and create more variety with your shots. Captured the last speaker from the waist-up? Get the next one as a head and shoulders shot. Got the last speaker from the left? Run over and cover the next one from the right. That variety, when edited together, will make the video more exciting, give it pacing, and create a sense that each shot is showing us something new (even if it's the same old crap). You might consider a livelier bit of background music, as well as a more definitive stopping point. And as someone else who's figuring out how to shoot corporate/wedding/commercial stuff...try not to use those chintzy transitions. They tend to read as cheap. Stick to straight cuts and dissolves as much as possible. That'll also help keep a consistent sense of pace. A dissolve moves you to the next shot delicately, the hard cut forces you there. Ebb and flow. Savvy? Can I ask what picture profile settings you used in camera? And did you use a lens with IS?

From Art Adams on Reddit: "So... let's see: I need to think how to say this simply. The sensor has little red, green and blue filters on the photosites. The wavelengths that these filters pass determine how the camera sees color. There's no way for them to perfectly match how the human eye sees color, so there has to be some tweaking. First, photons pass through the filters and the photosite counts the number of photons that hit it. That signal is then moved off the sensor where it is sorted into a color signal (either red, green or blue) and then amplified so it's strong enough to be manipulated. It's converted into digital bits from analog voltage. It then goes into the DSP (digital signal processor), where the magic happens: Each color signal overlaps the others. Red filters pass some green and a little blue, blue passes a little red and a bit of green, etc. You need that overlap to get secondary colors (yellows, cyans, purples) but you also can't make pure primaries with those other colors mixed in. There's a thing called the matrix that is a math formula that subtracts color channels from each other to make them more pure while also retaining their ability to render good secondary colors. Also, the combination of the dye filters, the sensitivity of the silicon and the spectral shaping filters on the front of the sensor (IR and UV cuts) create a color space that's unique to that design of camera. There's a lot of math involved in bending the camera's color space into a color space that we can see on a monitor or in a movie theater, like Rec 709 or P3. That all happens in the DSP. This is also where gamma is applied. Sensors see in linear gamma, which looks really dark on a monitor, so that's adjusted to look normal. More than that, it can be adjusted to create different looks depending on the setting: boosted shadows, compressed highlights, log/flat, etc. Normal displays are only designed to show six stops of dynamic range and most cameras now capture 10-14, so gamma curves shoehorn that extra information into a range the monitor can see. Somewhere in there you'll find white balance, where the red and blue channels are adjusted to match the strength of the green channel when viewing a white card reference. The dye filters used, where the colors in the dye filters overlap, how pure colors are, how much cross contamination there is between color channels, how the camera's color space is translated into a lesser color space, how the camera white balances... those are all color science. It's a combination of crazy hardware tricks and sophisticated math, and that creates a camera's look. Even in log or raw these characteristics transferred. Raw doesn't mean that you can do anything you want to the footage, you still have to process it somehow... and that's the color science. The interaction of the dye filters and dynamic range of the camera isn't changed based on how the data is stored, you just have access to it at a deeper level. Arri's look is very filmic. Shadows are most saturated, highlights are less saturated, highlights roll off very evenly, and the tone mapping is beautiful. Sony's look is a little more sterile somehow... less soft, less subtle color... unless you shoot in Cine-EI mode on an F55/F5/FS7, where they've done some nice filmic things. (Their previous color science, SGamut, did nasty things to yellows and blues. SGamut3.cine is very pretty, with much softer yellows and accurate blues.) Canon pushes reds toward orange to make flesh tone look better, and pushes greens toward blue to make them cooler. They don't go for accurate, they go for what they think is pretty. Panasonic is great with flesh tones. RED Dragon's color is very good. The M and MX were a muddy mess: color was awful under tungsten light, with pasty red faces and blueish greens. They were okay under daylight, but colors looked a little muddy. Dragon doesn't see much subtlety in flesh tones under tungsten light but it's way better. Color science is how a camera manufacturer makes its color unique and pretty. Each company has its secret sauce."

The Leica Q absolutely has a lens.