All Activity

Sure, but you can accomplish a similar thing by using S35 lenses on FF. Or S16 lenses on M43. You pointed to a bunch of other incorrect things too like DOF equivalence as well. And again, 16/17:9 is "true" open gate on many cinema cameras. But it's also not true that every lens made for FF has extra character when you use it on 44x33mm. The Canon EF 85mm f/1.4L covers the entire GFX sensor and has minimal character all the way to the edges. The tiny Canon 40/2.8 pancake similarly has minimal character while covering 44x33 pretty well. Fraser wanted to use specific vintage lenses that had more character near the edges of the image circle. Certain FF lenses, yes. And the eterna 55 doesn't use the high MP of the 44x33 sensor for anything. Just like the GFX 100 II, your options to use the full sensor width are 4K with decent RS and mediocre DR (up to 60fps), 4K with strong RS and good DR (up to 30fps? Not sure of the max, but less than 60), 4K open gate with decent RS and mediocre DR (this differs from GFX 100 II), and 5.8k 2.35:1 with strong RS and mediocre DR. 8K goes to a crop really similar to full frame on a 24x36mm sensor and also has strong RS and mediocre DR. I haven't, at any point, said that the Eterna isn't a completely invalid camera with no uses. What I am saying is: 1) There is no intrinsic "medium format look" 2) For a vast majority of use cases, the less expensive V-Raptor XE with a 41mm wide sensor, good DR, and a global shutter will likely be chosen over this one by higher-end owner-operators (those who don't just buy an FX9 (cheaper yet) or C400 (even cheaper) - which, realistically, is most of them). If you prefer the Eterna, you're not wrong. You're welcome to use any tool that you like. I might rent it myself if a project came along where it made sense. I just think that the number of sales that aren't to rental shops will be really low.

Somehow I missed this promo short film, its kinda brilliant in how it showcases a lot of the features in a pretty creative narrative, really liking the overall image from this new sensor:

@eatstoomuchjam Funny enough, the Greg Fraser quote you referenced actually underlines my point. He’s saying outright that moving to a larger format opens up creative possibilities precisely because you’re seeing parts of the lens you wouldn’t on a smaller sensor. The lens itself hasn’t changed, but the relationship between the sensor size and the lens’ image circle absolutely changes what’s captured. That’s all I was getting at: bigger formats don’t alter lens physics, but they alter what part of the lens’ projection you get to play with, which translates to a different look/feel on screen. This is similar to when you have open-gate (true 3:2 or 4:3, not just 17:9), you’re also opening up different aspect ratios and lens visibility. Whether you see the extra image area as flaws or character is subjective, but the distinction Fraser is making is exactly the one I was trying to point to. I know many GFX users like to put certain FF lenses for identical reasons, and the GFXRF & Eterna 55 use that high MP large sensor to open up multi aspect ratios which is why I don't think its irrelevant to the discussion.

Sorry, but this sentence makes no sense. A bigger sensor doesn't change how a lens behaves. The lens always behaves the same and projects the exact same image circle. And "cinematic" is a meaningless term so "cinematic signature" is equally meaningless. This is untrue. Assuming that you mean compared with S35, to be more specific, FF gives shallower depth of field for equivalent framing at the same focal range and aperture. You can get an identical image by using a wider S35 lens with a bigger aperture. This is exactly what focal reducers do when they focus the image circle of a FF lens down onto an S35 sensor. In fact, when using a focal reducer, a Komdoo or Komodo-X has slightly less DOF at equivalent framing than a natively FF camera (the equivalent crop factor at that point is something like 1.05x). This also becomes largely an academic distinction if you don't insist on shooting fast lenses wide open 24x7. And even if you do, do you need shallower DOF than Army of the Dead, that Zack Snyder movie from 2021 that was shot entirely with the Canon 50/0.95 dream lens wide open? This is true, but irrelevant to a discussion of whether there is or isn't a medium format look. It only reveals more lens character in the sense that for some lenses, you see the worse parts of the image circle outside of the standard 35mm film size. In some cases, it also just means you can't use the lens at all - for example, my Noctilux-M 50/1 only barely covers 24x36mm and already has dark corners and edges. On a GF sensor, you just get the edge of the image circle surrounded by blackness. That lens has plenty of character already on FF. On film, yes, but this is related to the inherent resolution limits of film. On the 100 megapixel GF sensor, this is technically true vs 35mm format cameras that have 61 megapixels, but it's largely an academic distinction that is barely noticeable in practical terms. But when you're using line skipped/binned 4K off of that sensor, you have less smooth falloff than off of a 35mm sensor recording 8K. This is also academic and can barely be seen. At the same focal length, sure. Luckily, we can change lenses. It would be true to say that you can capture a different image with the same lens as a smaller format. So if your goal is to get a different look out of your vintage Nikkor 200mm lens and if that lens has an image circle big enough to nominally cover a sensor that's 44mm wide. If you're shooting 16/17:9, will that difference be substantially different than the image from a V-Raptor XE with a 41mm wide VV sensor (that costs $1,500 less)? No. That or if 3:2 capture isn't a concern, go rent the Ursa Cine 17K 65 with a 51x24mm sensor. Yeah, footage from the GFX 100 II looks nice too. But so does footage from the GH7, a camera with a much smaller sensor. For video, I prefer footage from my UC12K LF to what I get from the GFX 100 II.

Really funny indeed, it seems with long takes the only way to trim and save only important parts of footage is through Resolve’s Media management. Tried to rename clips to R3D first in CFExpress, but exporting trimmed parts from Resolve did not work. Compared one old clip and other than R3D’s chroma NR there was not much difference to NEV. Have to test this more with clips that are saved straight from card as NEV.

Technically a sensor by itself doesn’t have a “look,” but bigger sensors change how lenses behave and how we can use the image circle, which is where the cinematic signature comes from. FF already gives shallower depth of field for equivalent framing and a FoV closer to classic 35mm photography. Open gate modes on 3:2/4:3 sensors take that further by giving you the full vertical readout, letting you reframe, use anamorphic, or extract stills without losing resolution. Medium format pushes this even more, however Fuji’s GFX sensors aren’t “true” MF like Alexa 65 or Hasselblad backs, but it’s about 1.7× the area of FF, which is enough to reveal more lens character, smoother falloff, and a slightly different perspective than FF or S35. So when you pair the sensor with fast primes and open gate capture, the format really does offer creative possibilities that smaller sensors simply can’t match. I'm loving what I'm seeing from the Eterna so far by the way!

It would be funny if turns out RED and Nikon couldn't push Davinci developers to address the issue with nraw rendering and decided to fix it themselves by a metadata trick.

Are you seriously asking me to explain the thought process of other humans? I'm pretty sure I'm not qualified for that. None of those films were filmed entirely on medium format. Are you suggesting that when you watch them, you're suddenly jolted out of your seat when The Joker switches from Alexa 65 to Alexa LF? Or that you can even tell? Note that at no part of that does Greg Fraser say "I wanted the medium format look." Instead, he's talking about how much he liked the look he got from using lenses designed for smaller formats. It is objectively true that the designers of those lenses never anticipated that the outer edges of the image circle would get used at some point. This seems like a pretty based and objective take and a reason that somebody might choose to use a larger sensor. I like cropping, it has fantastic dynamic range, and some of the first-party lenses for the system (particularly the 110/2, 250/4, and 500/5.6) are among the best lenses I've ever seen. I'll turn this around and ask you these questions: 1) What do you think medium format look is? 2) Is there a FF look vs an S35 look and does a speed booster give S35 the FF look? 3) Is a sensor size that's just as close to 35mm film as it is to traditional medium format film going to give a medium format look or a full frame look? Because for photos, at least, 44x33 gives a total area of 1452 where 24x36 gives a total area of 864. Meanwhile, 6x4.5 (56x42mm realistically) which is the smallest medium format film size has an area of 2,352 and 6x7 (56x72) film dwarfs it at 4,032. 4) To turn around the question above, if there is a specific medium format look, why do tentpole movies like Mission Impossible which have effectively unlimited budgets use smaller formats, even for some of their big, sprawling epic shots? (And yes, MI Rogue Nation used Alexa 65 for the underwater scene, but the rest was shot on smaller sensors IIRC)

With this aggressive price, in the next few months we'll see if it's capable to shake a little the market. Next iteration of Z9 will probably incorporate some of these features.

I think this camera looks pretty great for the price. Its still out of my price range, unfortunately, I’ll have to stick to an iPhone 17p to get my raw kicks. But I can’t think of anything in the price range that comes close. (Is the Panny S5 IIX also considered a “content creator camera, as well”?)

Yes I agree. I’m ok with this strange little ZR as a sort of 1st gen beta to roll out the new cross pollination. Smart to make it $2,000 because it’s still very capable but it will allow them to test out all the features on content creators, and then hopefully roll it into pros later.

Except it does not work if you import the NRaw clips from CFExpress to Resolve timeline, cut and export the parts you want to save via Media Management and then rename .NEV to R3D in Finder. it only works if you copy the clips straight from CFExpress to some Finder folder, then rename them, and after that import them to Resolve. It seems Resolve media management is doing something while exporting the .NEV files.

Just curious, if there is no such thing as a medium format look, then why do a lot of the great celebrated filmmakers choose to shoot on it when going for big tentpole films? The Joker, Dune, Revenant, Dark Knight, etc… Why go through all the trouble of extra weight, crew, expense if there’s nothing to gain over Super 35? Surely they believe there is a benefit. Why do you own a GFX 100ii?

If you put a card from a Z8 NRaw recording in a Z6iii does it play back? And vice versa.

Maybe NRaw from the Z6iii onwards is Redcode?

So is NRAW really REDCODE or is REDCODE really NRAW, I am confused now.

People with some memory will find strange he using Fujifilm cameras for stills in the movie, when he always showed that Olympus / OM cameras is what he use when fishing...

Yea, and I suggested considering NRAW normal as R3S LQ. But you have to be careful about base ISOs.

This seems to work both in NRaw High and normal, when renaming .NEV to .R3D in finder and then importing files to Resolve.

The Reddit source https://www.reddit.com/r/Nikon/s/v8tkSHdsZa

I've been musing how to proceed next, but figured out I have two challenges: - work out what lens characteristics I can emulate in post vs those that can't be emulated easily / practically - work out what lens characteristics I actually care about The reason the first one is important is that there's no point in testing how much vignetting there is between various lenses when I can simply apply a power-window or plugin in post and just dial in what I want. This will then leave the lens testing to compare the things I can't do in post, like shallow DOF etc. I figure there's no point choosing a lens with strengths I can emulate in post over a lens with strengths I can't. In my setup the real question is why wouldn't I just use the Voigtlander and Sirui combination, because it is relatively sharp anamorphic with reasonably shallow DOF, and then just tastefully degrade it in post. So, naturally, I asked ChatGPT which things were which.. ==== Characteristics You Can Emulate in Post (2D effects). These are primarily image-level artifacts that don’t depend on actual 3D geometry of the scene: Vignetting (darkening at frame edges) – trivial to emulate. Chromatic aberration (color fringing at edges, longitudinal CA is harder) – lateral CA is easy to add/remove, longitudinal CA (color fringing in front/behind focus) is less convincing. Distortion (barrel, pincushion, mustache) – geometric warping filters can mimic this accurately. Flare & ghosting – lens flares, veiling glare, rainbow ghosts can be faked, though authentic multi-bounce flare with complex aperture geometry is tricky. Bloom / halation – glow around highlights is very doable with diffusion filters or glow nodes. Contrast “roll-off” from flare veiling – can be mimicked with tone curves and haze overlays. Field curvature (apparent focus falloff toward edges) – can be approximated with selective blur masks. Edge softness – can be faked by blurring vignette. Focus breathing (slight FoV change with focus) – can be keyframed with a digital crop/scale. Bokeh shape overlays – you can fake oval bokeh, cat’s-eye bokeh, onion rings to some extent with blur kernels, though it looks “pasted on” if pushed. Characteristics That Are Hard or Impossible to Emulate in Post. These are fundamentally tied to how the lens bends and gathers light in 3D: Depth of field & true bokeh rendering – You can blur in post, but real optical bokeh depends on entrance pupil geometry, field position, occlusion, and interaction with 3D scene depth. You can approximate with depth maps, but never get the organic overlap of layers of out-of-focus light. True lens breathing with parallax – cropping simulates magnification change, but real breathing also subtly shifts perspective/parallax in the image. Perspective rendering (due to focal length + entrance pupil position) – can’t be changed in post without full 3D scene reconstruction. Aberrations that depend on 3D ray paths – e.g., coma (stretched points off-axis), astigmatism (different focus for tangential/sagittal rays). You can fake streaky stars, but not with the depth-dependent accuracy of real glass. Micro-contrast / “lens pop” – the way fine details render due to MTF, stray light control, and glass polishing isn’t really capturable in 2D. Polarization effects – coatings reduce or enhance reflections differently based on polarization of incoming light; impossible to recreate after capture. True specular highlight interaction with aperture blades – starbursts, iris shape, and diffraction spikes are fundamentally optical. Subtle 3D flare interactions – when a bright light causes ghosts that float relative to the scene geometry, which post plugins often fail to mimic convincingly. Sensor–lens interaction effects – e.g., color shifts or vignetting from oblique ray angles on digital sensors, microlens interference. ==== It's not a bad list, although some things are a bit mixed up. The things in the above that stand out to me for further investigation are: Longitudinal CA I wonder how much of the Tokina + wide-angle adapter look was this, it seems to have subtle CA though everything Experiment with oval inserts to get a taller bokeh shape, maybe between the taking lens and Sirui adapter ...and just setup some tests in post for barrel distortion, edge softness, resolution (blur and halation). The rest of the stuff above I either know about or don't care about.

It looks like it is. It also look like original RED RAW codec doesn't show any superiority over ProRes RAW codec https://www.youtube.com/watch?v=RtTRoOWaw-w With ProRes RAW footage in Davinci Resolve you can choose between many Log profiles including Panasonic V-Log, Sony S-Log3, Canon C-Log2 and avoid using Nikon NLog. In this case Nikon Z6 III for me is a better camera. It has the same sensor, but also great EVF, mechanical shutter, full HDMI port. Nikon Zr main selling point is R3D NE or RED RAW lite if we can call it.

The fact that NRAW High bitrate is exactly same as R3D NE, could be a hint

I'd like to see some deeper analysis on it. Could be that R3D NE is just high-bitrate Nikon raw and it could be that Resolve limits log3g10 decoding to files matching the r3d extension without checking the underlying raw format. Either way, that's pretty cool.

No translation, but you don't need to.