cantsin

I think we're completely misunderstanding each other. I'm just trying to drily sum up a few facts, mostly for the people who are reading this thread and using EOSHD as an information resource. This is not about you, and I am not even aware that I'd be swinging at anything.

A Blackmagic Micro or Pocket Camera records 10bit Log in ProRes, a GH5 records 10bit Log, even a GH4 and an FZ2000/FZ2500 records 10bit Log with an external recorder. We established that indeed, this is a suboptimal if not a bad case. Nobody argues with that. But this is a particular issue of Sony's and GoPro's consumer cameras that squeeze aggressive log curves into 8bit. It's not an issue of Log capture as such. No, a very practical issue. Affordable 10bit Log cameras do exist, some are even cheaper than Sony's A6300 and A6500. The simple conclusion is, and hopefully we can agree on this and rest this dispute: If you use a Sony 8bit consumer/prosumer camera (RX10/RX100, A7x/A9, A6xxx), better avoid sLog. If you use it, reduced color resolution/color banding will be the (sometimes severe) trade-off for capturing a wide dynamic range scene without clipping highlights and drowning shadows, or for achieving postproduction compatibility with Slog footage shot with Sony's 10bit cinema cameras. But this has been discussed on video forums for a long time (just saying). P.S.: And @Andrew Reid addressed this issue in his own way by developing more visually appealing Rec709 profiles for Sony's con/prosumer cameras.

To cut the whole discussion short: No, it doesn't capture less color in the middle tones if you have a 10bit or 12bit codec (because a 10bit codec captures 4x the color information and a 12bit codec 16x the color information of an 8bit codec). And only with such color bit depth, Log curves like S-Log should be used. Btw., this discussion is not about respect or feelings, but about facts. (All the more in the Donald Trump age.)

None. But scenes with only 7-8 stops dynamic range hardly exist in real life (not even indoors if you film a living room with a desk lamp on), only in studio settings where you light for 7-8 stops. In all other situations, you get clipping highlights and/or drowned shadows. Given that EOSHD consists of perhaps 90% filmmakers who never or rarely shoot in studio light, the limitations of Rec709 are a fact. In former (camcorder) times, it simply meant that run-and-gun video filmmakers had no choice but put up with the "video look" of crushed shadows and clipping highlights. Let me also quote your initial posting: Yes, but your conclusion only talks about display color, not scene color. It leaves out the fact that 8-bit Rec709 color and 7 stops of dynamic cannot deliver a complete representation of a scene. You make one fundamental mistake here: The human eye, because it's wired to the brain, does not see like a camera, so - sorry to be harsh - all our 8bit and 7 stops arithmetic is meaningless. When we look outside the window and see a street or a landscape, which typically has dynamic range of 14-20 stops in daylight, we do not see clipped highlights and drowned shadows, because our eye roams and adapts, and our brain creates a composite image. So, in order to create a photographic or filmic image that correctly represents a scene, we have to offer the same range of visual information to the eye, and not cut out shadows below and highlights above the 7 stop spectrum./ This assumption is only true for Sony's consumer cameras which, in SLog mode, indeed do what you described above. In the meantime, we have established that SLog2/3 was engineered for cameras that record in high-bitrate 10bit codecs. Yes, one does, but the only solution to that is to shoot in perfect studio light like Hollywood from the 1930s to the 1960s - and not shoot documentary, and not shoot in the cinematography established with the Nouvelle Vague and New Hollywood. As said, it is - it is about capturing the dynamic range of a scene as good as possible. Max, with all due respect and whatever you tried to get across - your initial postings did not make a clear differentiation between scene dynamic range and display dynamic range, between Sony's S-Log and other log curves, and between Sony's S-Log in the 8bit consumer cameras vs. the 10bit pro cameras.

But the simple truth is that you need extended dynamic range (and thus more than standard Rec709 capture) to record an image like this (where Rec709 may be the display color sapce): ...unless you use an incredible amount of reflectors and on-set lights as Sergio Leone's DP probably did for this scene. (In addition to using negative film stock with its much extended dynamic range over reversal/positive film. It also helps to think of Log video as an equivalent of negative stock and of Rec709 capture as the equivalent of reversal film.) Today, we can capture such scenes without artificial light thanks to the high dynamic range of modern cameras. But if the above picture would have been shot with an in-camera Rec709 profile, the house front would be black, the sky would be white and Charles Bronson's face would likely be unrecognizable. So this is why we need Log (if we don't use Raw capture) - even if we end up displaying the image as an 8bit Rec709 or RGB in 8 stops of dynamic display range as above. And images like the above, without cut-off shadows or clipping highlights, are exactly what we associate with cinema-quality images as opposed to the classical video camera look.

No, because this mixes up display dynamic range with scene dynamic range. When we speak of the dynamic range a camera can record, we refer to scene dynamic range: For example, 14 stops from shadows to highlights in an outdoor daylight scene you record. A typical, good computer monitor has a contrast ratio of 1:1000, that's 10 stops (2^10=1024). What we do in video postproduction is to compress those 14 stops you recorded outdoors into the 10 stops of your monitor, or, in the case of 8bit Rec709/8bit RGB, the mere 8 stops supported by this display and color standard. In other words, we align the brightest stops of the motive with the brightest stops of the display, the lowest with the lowest, and the medium is where our gamma sits. So we'll roughly end up mapping stops 11-12 of the scene to stop 8 of the display, stops 9-10 to stop 7, stop 7-8 to stop 6, stop 6 to stop 5, stop 5 to 4 etc. all the way down to stop 1-2 to stop 1. That's exactly what a Log-to-Rec709 LUT will do. If the camera sensor dynamic range were identical to that of the display, or if we would only record within the same dynamic range as the scene, the highlights of the scene would clip and the shadows would be black. - If I'm not mistaken, you're caught in math trap where you base all your assumptions on the fact that the human eye can see about 16 million colors and that an 8bit true color display displays 16 million colors, too. So why do we need more than those 16 million colors? Because this number doesn't take into account the adaptation of the human eye (in combination with the brain). While the human eye can physically only see about 7 stops, it adjusts its iris so quickly (and we usually scan a scene with our eyes going from shadows to highlight) that our brain effectively "sees" a higher dynamic range.

Well, it's not that a sensor cannot "handle" a log curve, but the question is whether a log curve can handle/fit the dynamic range of the sensor. If the log curve has been designed for the highest-end, highest dynamic range sensor, it won't be the optimal fit for a lower-end/lower dynamic range sensor. For example, if a lower-end sensor has 2 stops less usable dynamic range (note the word usable... on a paper, a sensor can have a lot of dynamic range, but if the two lowest stops are more noise than signal, those stops don't really count) and the log curve records those 2 stops in one bit, you'll be throwing away one bit of your color depth. In the worst case, you might end up recording effective 8bit in log even when the codec is 10bit.

No, I did not mean to imply this at all. Sorry if it came across that way (but this is an easy misunderstanding on a forum such as EOSHD). What I wanted to get across is this is: - a Log function normally is made to measure for a particular camera sensor, its dynamic range and often also the bandwidth of the camera's codec. (A good example is Blackmagic which uses different Logs, and different corresponding LUTs for Rec709 conversion, for its different camera families.) - sLog2 and sLog3 were engineered for Sony's professional cinema cameras with their high dynamic range and studio-grade XAVC 10bit 400-600Mbit/s codecs. - sLog2 and sLog3 are also implemented in Sony's 8bit consumer cameras never mind the fact that these log curves are neither optimized for their sensors (which likely have 1-2 stops less usable dynamic range), nor meant to be stored in low-bitrate 8bit 4:2:0. - Even if Sony's consumer cameras had better codecs AND a 10bit pipeline (which would require a complete reengineering of the cameras, and likely wouldn't be feasible with today's technology when keeping the same compact body sizes - given the fact that Sony's cameras already have overheating issues with their current feature set), sLog2 and sLog3 wouldn't be the perfect log curves for them because, as stated, they aren't optimized for the sensors of these cameras. The same is true for the GH4/GH5: Since V-Log has been engineered for the Varicam with its higher dynamic range, the German site Slashcam found out that when using it in the GH4 and GH5, the upper bit in the recording (which stores the highest exposed part of the image) remains unused, simply because the GH4/GH5 sensor clips at that exposure. This effectively reduces the 10bit recording of the GH5 (or GH4 with an external recording) to merely 9bit in V-Log, the price you pay for Varicam compatibility. - The advantage of using Log curves engineered for higher-end cameras in lower-end (consumer) cameras is that the footage of both camera types can be easily mixed in the same timeline and graded with the same LUTs. So, in this sense, SLog2/3 and Vlog aren't engineered for maxing-out the video quality potential of cameras like the A6300 or the GH5, but to make them easier to use as c-cams/crashcams in projects where the larger cinema cameras are being used. - So, in the best case for DIY/low-end shooters, we would have Log curves engineered for "our" cameras, their sensor and the dynamic range of those sensors - instead of "alien" log curves that were engineered for different, higher-end cameras. - The only exception so far (aside from Blackmagic) has been Canon with its CLog curve for the original, 8bit C100/C300. It's a "light" log curve engineered to preserve the dynamic range of the sensor in these cameras and not to be destroyed/yield questionable results in 8bit recording.

The comparison doesn't really help because (a) both cameras only output 8bit via HDMI, (b) they have completely different log curves. Slog-2/3 was developed for Sony's top-of-the line cinema cameras (such as the F5), optimized for their sensors and codecs, and is implemented in the consumer cameras only as a compatibility function. (So that F5 operators can use an a6300 or A7s as a c-cam or crash cam and shoot footage that perfectly mixes with F5 footage. Same for V-Log as an option for using GH4/5 footage as c-cam footage in projects shot with Varicams.) Canon's CLog (not to be mixed up with the newer Canon Log....) is an entirely different beast because it was developed for the first generation of C100/C300 cameras and is optimized for their sensor and 8bit 4:2:2 codec. This also explains why it is a much less aggressive log curve than SLog2/3.

"Consumer Log" is the same thing since it's also log created from the raw sensor data. In the case of Sony S-Log and Panasonic S-Log, it also has a much bigger color space than Rec709/RGB. The real issue of "consumer log" is that when this signal, which requires high bandwidth encoding, is compressed into a small bandwidth 8bit 4:2:0 codec (as was the case with the Panasonic GH4 and still is the case with Sony's RX/a6x00/a7x cameras). Again, color resolution behaves equivalent to pixel resolution. To stay in this analogy: If you compress a 48MP picture (shot with an A7r) with 10% JPEG quality in order not to exceed 1 MB file size, you end up with an image that still has 48MP but looks terrible because of the extreme compression artefacts (see below; the original image is here: https://4.img-dpreview.com/files/p/TS7952x5304~sample_galleries/7076458101/9405098100.jpg). In such a case, it's better to shoot or store with a smaller pixel resolution but higher jpeg quality. The equivalent happens with color if a high dynamic range log image is stored with "consumer" bit depths and "consumer" data compression ratios - i.e. in a bit depth and codec that isn't sufficient for high-end, studio quality mastering, and completely inadequate for an extended color space such as s-gamut. In such a case, it's better to shoot Rec709.

No this is excellent and should explain the issue to everyone.

Btw., I would much prefer if this thread wouldn't use "S-Log" as a denominator of logarithmic color curves, but would generally refer to "log" instead. "S-Log" stands for Sony Log, as "C-Log" stands for Canon Log, "V-Log" stands for Panasonic Varicam Log and "C Log" for Arri's Cinema Log. Most log curves are manufacturer-specific because they're optimized for particular sensors and their dynamic range.

Okay, let me be the killjoy and party pooper... In all of the above videos, I see 'cinematic' light, framing and camera movement, but what I don't see is a 'cinematic' camera image - i.e. with rich, deep colors in subtle graduations and natural vs. artificial sharpness of the image. Of course, a $2000 GH5 is no $20,000 RED or $40,000 Alexa, but when comparing it to a $2000 Blackmagic Cinema Camera (to just pick a few random examples from Vimeo):

Sorry Max, but you simply don't get the issue. "Compression" can not just refer to data compression of a codec, but also to: - aspect ratio/perspective compression through anamorphics; - dynamic range compression through log profiles. Anamorphic aspect ratio compression can kill your image if you have too little pixel resolution - for example, if you shoot anamorphic in SD resolution and desqueeze, you end up with pixel mush. Dynamic range compression can kill your image if you have too little color depth - for example, if you shoot log in 8bit and desqueeze into RGB, you end up with color banding and maybe 6bit/7bit actual color depth in the final image.

I'm afraid you're confusing bit depth (in this case, 8bit with its 256 shades per color channel) with color profiles. I think it's easier to draw a simple analogy between dynamic range and image resolution: A Log-Curve does the same thing for dynamic range that an anamorphic lens does for image resolution: It squeezes/compresses a larger bandwidth of visual information into a lower-bandwith medium. (In the case of an anamorphic lens, a wider image into a narrower frame, in the case of a log curve, a wider dynamic range/palette of colors into a narrower palette.) Desqueezing the image can easily result in ugly artifacts if your resolution is limited: - if you record 2:1 anamorphic in 1920x1080 and desqueeze to 3840x1080, you will end up with blocky/doubled pixels on the horizontal axis. You wouldn't have those artifacts if instead you recorded in 4K for desqueezing into 3840x1080. - if you use a log curve to squeeze 12 stops dynamic range into the linear 8 stops supported by an 8bit codec, you will end up with a torn curves in the histogram and blocky color gradients when desqueezing the color back into 12 stops. You wouldn't have those color artifacts if instead you had had 10bit color depth for your log curve recording. This is why, with an 8bit camera (and supposing that it has perfect 8bit Rec709 color science), it's preferable to record scenes with a contrast ratio of 1:255 (=8 stops) or less in Rec709 instead of Log. In reality, consumer and prosumer cameras implement all kinds of crowd-pleasing processing gimmicks in their Rec709 profiles (such as artificial sharpening, denoising, color shifts and contrast enhancement) but have them switched off in their Log profiles, so that many people end up using Log profiles for a less artificial, less-consumer-videoish, more organic-looking image - although this may come at the price of color banding artifacts and generally less color depth. After desqueezing 8bit log into Rec709, you probably end up with 7bit or 6bit actual color depth, just as desequeezing 2:1 anamorphic from a 1920x1080 recording into 1080p will half your pixel resolution.

The problem really is that DSLR and even mirrorless cameras are still being designed to appeal to photographers coming from analog photography, using the same mode dials as on analog cameras, the same nomenclature (such as ASA/ISO instead of gain etc.) A few months ago, a friend of mine decided to upgrade from her smartphone camera to a 'real' camera because she was taking a lot of indoor photographs without flash and wasn't satisfied with the noisy images she got from her phone. Upon my and somebody else's recommendation, she bought an LX100. Although she knows more about photography than most people (being familiar with shutter speed, aperture and ISO) and generally tech-savvy, most buttons and the camera menu were completely cryptic to her, so I ended up spending half an hour to explain everything. Aside from the fact, to repeat myself, that neither the LX100 nor any other camera has an upload button for pictures to Instagram, Facebook or other social media. It's like as if today's TV sets would still require UHF and VHF channel tuning from you and wouldn't allow you to watch Netflix. Unless usability and user interface design (+ network functions) of cameras catch up with the year 2017, the camera industry will destroy itself. (Just like most of the car industry is now in the course of destroying itself by not switching to electrical cars, but rather manipulating firmware to fake low gas emissions.)

A minimal, run-and-gun rig with the BMPCC is possible. This is my setup: Always have 4 charged batteries with you + a small charger to plug into the wall somewhere if necessary. Original Nikon EN-EL20 batteries last substantially longer than off-brand batteries (including Blackmagic's own batteries); a fully charged battery should give you about 30 mins of continuous shooting. Use a 256GB Sandisk Extreme Pro card, with enough capacity to shoot 1:20 hours raw or 2:50 hours ProResHQ footage. Use one, compact, fast, affordable, run-and-gun lens; excellent options are (a) the Samyang 12mm/f2 MFT, (b) the Sigma 30mm/f1.4 (old version is fine) + Metabones 0.58x Pocket SpeedBooster, (c) the Sigma 18-50mm/f2.8 + Pocket Speedbooster. Always have an IR cut filter on the lens, have an ND filter ready to screw on if necessary. (Using the Manfrotto Xume magnetic filter system makes things easier.) Hoya is a recommended brand; B+W, Heliopan, Marumi are fine, too. Use one of the following three good, compact & inexpensive mic solutions: (a) a Rode Videomicro [the $60 mini mic, needs no batteries, works decently on the BMPCC], (b) a Zoom H1 as both an external recorder and an external mic+preamp wired with a minijack cable to the line-in of the BMPCC; (c) an Ohrwurm Audio binaural mic [fantastic sound with the BMPCC, since it needs minimal preamplification and sounds great in general]. Use one of the following compact and inexpensive stabilizers: (a) a vintage chestpod as they were manufactured in the 70s/80s for 8mm and 16mm cameras, for example by Braun and for the Krasnogorsk-3 camera; (b) the $30 Cowboy Studios Shoulder rig, probably the best handheld stabilizer there is, (c) the Cullmann Cruiser travelpod.

The Achilles heel of this phone will be Android, because it means that RED is putting itself at mercy of chipset providers and Google's OS updates. To give one example: Android 7.x no longer runs on Snapdragon 800/801 chipsets (used in many 2014 phones like the Samsung Galaxy J, Nexus 5 and Sony Z1). Even for its own high-end Pixel phones, Google only promises OS updates for two years. After that planned obsolescence, a phone basically turns into mobile security hole since an unpatched Android is prone to malware exploits. Even Fairphone, the only smartphone with a honest business model, got screwed by this because their hardware is based on Snapdragon 800; being stalled at Android 6.x, it still receives Google's security updates for that OS version, but will miss its goal of providing a long-life product. RED normally sells high-end products that can be used many years. Now it will be forced to release a new phone every 2-3 years, or stay on a rapidly depreciating Android version with both its hardware and its apps. In the worst case, people could end up using the RED phone only for controlling their cameras/camera modules, plus some other Android phone as the "real" smartphone. Or RED plays along and releases a new phone every 2 years.

(Voigtländer was actually a German company before the brand got bought up by Cosina.) This is a bit deceptive - in Austria (particularly in Vienna, a rich city) you have a huge collector's market for vintage cameras, and therefore a lot of high-end 2nd hand/analog camera shops. Since the major European photographica auction house (Westlicht) is based there, the customer base of these shops by far exceeds Vienna and Austria. Since we're both based in that country, let me just count the casualties of the last few years - and those were the best camera stores, particularly for specialist and good-quality second-hand gear: Foto Patent in Utrecht, Talens in Delft, Abro in Zaandam, Van der Waal in Rotterdam and Schiedam, Konijnenberg in The Hague, Parallaxe in Amsterdam. (In Amsterdam, the situation is still much better than the rest of the country with Nivo-Schweitzer, Den Boer and Fransen still in business.) Of the smaller chains, Foto Klein in the whole of South Holland went out of business last year, which included their Pro store near the Rotterdam Central Station. Kamera-express became a near-monopolist, buying up traditional, long-standing camera stores all over the Netherlands (FOKA in Rotterdam, in Utrecht, in Arnhem) to reopen them as cash-and-carry warehouse outlets.

This thread hits home like an eerie coincidence. Two days ago, I happened to find out that in my hometown... (a) ....the local consumer electronics superstore (Mediamarkt) had moved the camera department from the entrance area into the remotest corner of the ground floor, slashing its size by half in the process; (b) ....the biggest local second-hand outlet (CashConverters) moved its camera and lens department from a dedicated showcase near the entrance into a larger mixed electronics cabinet in the back of the store; (c) ....the biggest local camera shop had greatly reduced its stock on display inside the store, reducing the former tripod section (with complete Manfrotto and Sirui lineups) to only four cheap consumer models, tossing out a whole video camera accessory & light section in favor of flatscreen tvs; ...and all this 1-2 years after the majority of small, high-quality, owner-operated camera stores in my country (Netherlands) either went out of business or became swallowed by one big chain that only operates warehouse stores (competing with Internet shops) in industrial areas outside city centers. (I could also add the large duty free consumer electronics store inside Schiphol airport which used to have an extensive, open camera section that now has been reduced to maybe 8 cameras behind glass in some half-hidden shelf.) On top of that, shrinking camera sections were, in the past, compensated with growing action cam and drone sections. Now that the action cam market is (over)saturated and that drones sell less because of increasingly restrictive government regulations for their use, these shop sections are shrinking or disappearing as well. I don't think that classical cameras will completely disappear. But they will suffer the same fate as Hifi components/stereos in the past, ending up as niche products, likely with a deteriorating manufacturing quality on the low/consumer end and skyrocketing prices on the high end. [As it happened with CD players, turntables, stereo amplifiers etc.- you can still buy them, but if you're on budget, choice is rather limited, and the products are much more plasticky and cheaply built than their equivalents from a few decades ago.] Hifi components sold like hotcakes in the 70s/80s until cheaper, portable consumer technology killed off their mass market. In this sense, the story of smartphones marginalizing cameras just repeats the history of the Walkman killing stereos. (And later, mp3 killing off consumer-grade high-quality, uncompressed digital audio, whether CD, SACD, DAT or DVD-Audio. And now, Netflix and YouTube killing off BluRay and, very likely, 4K and HDR BluRay.) Call it as you will, but - speaking of consumer technology - we're in a race to the bottom. (Conversely, I agree with previous posters that it's completely ridiculous, if not outright insane from a marketing perspective that in 2017, you can't buy a single camera that lets you share a picture on Instagram, or a video on YouTube, with the touch of a button. Compare this to the software world, where even high-end/professional NLEs like Premiere, Final Cut and Resolve have YouTube and Vimeo rendering presets, in some cases also built-in YouTube and Vimeo upload functions.)

Yes and no, because again you disregard binning. Use this calculation example: A 48MP FF sensor has photo sites with 25% the size of a 12MP FF sensor. Let's assume that this reciprocally results in a 400% higher noise floor per pixel. However, if you use the 12MP image as a 4K video, each photo site becomes one display pixel. If you have a 48MP image and good signal processing, you'll bin 4 pixel into 1 - which will reduce the noise equivalently. Same is true if you print on a large format; the 12 MP image will have each pixel at 400% the size of the 48 MP image, which also means that the noise floor will get enlarged 400%, so the advantage evens out. In other words: One shouldn't consider noise floor per pixel, but literally the whole picture. The full well capacity advantage of a 12 MP sensor gets neutralized through its disadvantage of having fewer photosites. You could also compare it to rain (instead of light) falling into into two grids of vessels: Both grids have the same size, but one consists of 12 (4x3) vessels/compartments, the other of 48 (8x6). In the 8x6 grid, the individuals vessels have a lower capacity and will spill over sooner than the individual vessels in the 4x3 grid. But if you consider the two structures as a whole, both grids have the same capacity of holding water. No, because as soon as you blow up both images to cinema screen size, the lower noise advantage of the sensor containing the larger photosites will get neutralized by the fact that each individual pixel is being more enlarged and thus having its noise amplified. - Your model only works for cameras that produce their downscaled video image from line skipping rather than from binning. (Which used to be the norm for DSLR and mirrorless cameras until recently, so your model isn't completely wrong - it just no longer applies to most present-day camera technology.)

P.S.: But having said all the above, I'm with Mattias in one respect - that signal processing quality can make a huge difference. This is why pro-level camcorders like Canon's C series have much better low light performance than most DSLRs and mirrorless cameras with the same APS-C sensor size, simply because the C-Series has more powerful signal processing electronics with better debayering and noise filtering algorithms. The same is true if you shoot raw video. I found that shooting raw on a BM Pocket and using Neat Video in Resolve with highest quality temporal denoising, I can push the image to ISO 128,000 without falling apart. The video below was shot at practically no light. It holds up very well against an A7s video shot at ISO 128,000 (simply because the A7s has a much worse internal codec than the Pocket's raw, and much worse internal debayering and denoising than Resolve with Neat Video, never mind that its sensor is orders of magnitude more capable): versus the A7s in similar bad light settings: Yes - on top of that, one can compensate the sensor size disadvantages of MFT vs. APS-C and full frame by simply using faster lenses. If MFT generally has a ~2 stop ISO disadvantage versus Full Frame, a 25mm/f0.95 lens (like the Voigtlander) will neutralize that difference vs. a 50mm/f2 lens on an FF body or a 35mm/f1.4 lens on an APS-C body. On top of that, you'll also gain the same possibilities of using shallow DoF. (Conversely, a cheap $100 50mm/f1.8 lens on a modern FF body will give you roughly the same low light and DoF possibilities as a $700 25mm/f0.95 on a modern MFT body...)

What you write is only true for dynamic range (because of the larger full-well capacity of a bigger sensel). Noise and low light performance won't be affected because (a) the larger sensor still allows more photons to hit the sensor, no matter how coarse or fine the pixel grid, (b) binning the native sensor resolution to the delivery resolution (such as 4K) will reduce single-pixel noise. This is why there really isn't a dramatic difference in the low light performance between any of the present-day FF sensors. There's not even a dramatic difference between the A7s/II with its 12 MP sensor and A7R/II with its 42 MP sensor in regard to low light performance (in fact, the A7R has even better low light performance than the A7s because of its slightly newer sensor tech: https://www.dxomark.com/Cameras/Compare/Side-by-side/Sony-A7S-II-versus-Sony-A7R-II___1047_1035) (And aside from that, there isn't a single current MFT sensor on the market with a larger photosites than an APS-C or full frame sensor. The last one was in the Blackmagic 2.5K Cinema Camera which is no longer for sale. The BM Cinema Camera beautifully illustrates the point because it doesn't really have a better low light performance than any other MFT camera, but - thanks to the big photo sites on its sensor - a much better dynamic range.)

Yes, one can. Unless one believes that real life beats physics. (A bigger sensor lets in more light, just like an F1.4 lens lets in more light than an F2 lens.) They do have better ISO, but other factors (like DoF - as you pointed out - or quality of denoising/signal processing) can even out the physical ISO advantage.

No. Only if you compare older sensor tech with newer sensor tech. But even a 2005 full frame camera like the Canon 5D beats a 2017 GH5 very clearly in its low light capabilities: https://www.dxomark.com/Cameras/Compare/Side-by-side/Panasonic-Lumix-DC-GH5-versus-Canon-EOS-5D___1149_176 If you compare sensors from cameras built in the same time, the rule of thumb still applied that MFT has about one stop less low light sensitivity than APS-C, and APSC has about one stop less low light sensitivity than full frame (which corresponds to their sensor size, with full frame having twice the sensor surface of APS-C and four times the surface of MFT). If you don't believe me: https://www.dxomark.com/Cameras/Compare/Side-by-side/Panasonic-Lumix-DC-GH5-versus-Sony-A6500-versus-Sony-A7R-II___1149_1127_1035 Performance differences between same-size sensors of the same manufacturing generations aren't that dramatic. (And 90% of all camera sensors are produced or designed by Sony anyway, even Fuji's X-trans and - partly - Panasonic's sensors.)