I've never owned an Apple device. I don't take photographs or video with my phone very often. But this video presentation was captivating. It was clear, concise, without any nonsense, and thoroughly interesting.
The guy in the video is Stu, not only does he have an impressive resume (https://www.imdb.com/name/nm0556179/ known for originality, e.g. he did Sin City's look), he is also the original author of MagicBullet, which is one of the most used software by people in the industry to do easy color work. If there's one person who knows about color work, LUTs in creative work, color encoding systems, etc., it's him so naturally he knows how to present relevant subject matter without nonsense.
> If there's one person who knows about [thing] it's him so naturally he knows how to present relevant subject matter
One lesson (no pun intended) of the academic environment is that no, it doesn’t work that way. Some people are subject-matter experts, some are brilliant expositors, but while you can’t be a good expositor without a decent knowledge of the subject matter, you absolutely can be a world-leading scientist and at the same time completely rubbish at explaining your ideas. (Some areas are better at making use of such people than others.)
Good exposition deserves additional credit on top of subject expertise, is what I’m trying to say.
oh thx for sharing his background. this highlights something that i repeatedly find so frustrating.
this guy knows what he’s talking about. he has authority in his field. he has proven both theoretically and materially that his wisdom is high quality.
yet, videos from people with far less expertise on the subject matter will likely drown his out. over and over again i’ve seen the top videos on so many subjects repeating just plain wrong information but the “creator” says it with confidence so people just eat it up. even worse, so many times i’ve seen people in the comments try to nicely point out where the “creator” went wrong, and people go after the commenter for daring to imply this person on the magic screen might be misinformed.
it’s such a shame that being higher in the search says almost nothing about your knowledge on a subject and only means you’re better at manipulating the algorithm rather than having actual higher quality information.
i really wish that our searches would prioritize quality of information. at this point, years later, i think it’s pretty clear that the “wisdom of the crowd will shoot the highest quality to the top” has been proven to not be the case.
also, i don’t want to imply that amateurs can’t have valid and interesting perspectives, sometimes they might, just that being higher in search means quite literally nothing on wisdom of a subject—especially if a grifter can monetize off a trendy nuanced subject.
anyway, this video was incredible information. thx for sharing his background.
I'm glad to know that the person behind this post have real expertise, because at first glance the headline is a bit clickbait-like. (The content, though, makes excellent points.)
Also the author of the excellent DV Rebel’s Guide book which was an indie film makers bible for a long time. Obviously the title gives away its age though.
Well, there was one bit of nonsense and I thoroughly enjoyed it. I'm referring to the Ren & Stimpy "Log Song" sound track to the video of the woman walking up the stairs:
I don’t know if it’s the same thing, but capture on my Nikon D7100 always felt more “manipulable” than capture on an iPhone or the like, I suspected as a downstream effect of using RAW format with a larger image sensor. Interpreting log through this understanding, it felt pretty intuitive reading through this post. I don’t know if it’s accurate, but it feels accurate…
One reason is also that phone cameras have many limitations, so to get good images they have to "cheat" to work around those limitations. Additionally they often apply filters to the images so it looks good "out of the box", like contrast, smoothing, sharpening. Those choices done for you mean you lose information to do better yourself.
It's not my primary job, but I do some pro and fine art captures as well as video compositing and photo editing– the pro phones definitely have their uses just as my little micro 4/3 and full frame DSLR do, both with still and video work.
With glass and a sensor that small they aren't for everything, but the days of mandatory compression, limited color depth, mandatory "enhancements" and all of that stuff are over. If I'm shooting a handheld gimbal video, of something like a person talking outdoors not in direct sunlight, I'm grabbing that iPhone Pro without thinking twice.
Well, except that log (as well as almost every other format : "gamma" refers to the exponent of a power function !) is the opposite of raw : raw is focused on fidelity to the number of photons, log and others on fidelity to how the (human) eye reacts to these photons :
This is almost entirely due to sensor size. The sensor on the iPhone is smaller than your pinky nail with pixels between 1µm and 2µm in size (depending on which camera is used), the Nikon on the other hand has pixels over twice the largest size on an iPhone.
This video is excellent. About halfway through I was thinking, "Oh so this is like RAW for video" and then seconds later he gets to explaining how it's not exactly RAW.
The concept of Log seems needlessly confusing from (still) digital image processing perspective, which I have some experience.
Firstly the name is called "Log" (for logarithmic) but isn't that what gamma does in color spaces like sRGB since forever? "Normal" video standards like BT.709 also have non-linear transfer functions. I don't get why "log" is stressed here. Maybe it just means a different/higher gamma coefficient (the author didn't talk much about the "log" part in the article).
And the main feature of it, at least according to this article, is that it clips the black and white level less, so leaves more headrooms for post-processing.
This is definitely very useful (and is the norm if you want to do something like, say, high quality scanning), but I failed to see how it warrants a new "format". You should be able to do that with any existing video format (given you have enough bit depth, of course).
For some reason you're getting a lot of wrong or just bad replies. But the answer to your question is yes both sRGB/gamma2.2 & log are non-linear, but almost in the opposite direction. gamma2.2 is exponential not logarithmic. As in, it's spending all its bits on the lower half of the brightness range, whereas log is actually spending more bits in the highlights.
I think you're mixing up OOTFs and EOTFs here. sRGB or HLG can refer to either the stored gamma, but more often means the EOTF "reversed" gamma that is used to display an image. When we refer to "log", this is almost always means a camera gamma - an OOTF. So the reason it's "in the opposite direction" is that it's designed to efficiently utilize bits for storing image data, whereas the EOTF is designed to reverse this storage gamma for display purposes.
As you can see from the graph in [1], Sony's S-Log does indeed allocate more bits to dark areas than bright areas. (Though the shape of the curve becomes more complicated if you take into account the non-linear behavior of light in human vision.)
> This is definitely very useful (and is the norm if you want to do something like, say, high quality scanning), but I failed to see how it warrants a new "format".
This warrants a separate answer. Cameras are getting to the point where they can capture far more information than we can display. Hence, we need a lot of bit depth to accurately store this added precision. But adding bits to the data signal requires a lot of extra bandwidth.
In principle, we should just store all of this as 16/32bit FP, and many modern NLEs use such a pipeline, internally. But by creating a non-linear curve on integer data, we can compress the signal and fine-tune it to our liking. Hence we can get away with using the 8-12bit range, which helps a lot in storage. With log-curves, 12bit is probably overkill given the current sensor capabilities.
There's a plethora of log-formats out there, typically one for each camera brand/sensor. They aren't meant for delivery, but for capture. If you want to deliver, you'd typically transform to a color space such as rec.709 (assuming standard SDR, HDR is a different beast). The log-formats give you a lot of post-processing headroom while doing your color grading work.
> Cameras are getting to the point where they can capture far more information than we can display.
Haven't professional-grade microphones been in a similar situation for decades now, or is it the magic of remastering that keeps recordings from the 50s sounding so good on modern speaker systems?
The transfer functions in your (rec.709) color space is non-linear indeed. However, the pixel values you store are in a linear relationship with each other. The difference between values 20 and 21 are the same as the difference between values 120 and 121, assuming an 8bit signal. I.e., the information is the same for all pixels. Further down the chain, these values are then mapped onto a gamma curve, which is non-linear.
What the "log"-spaces are doing is to use a non-linear relationship for the pixel values, as a form of lossy compression. If the signal has to factor through 8bit values, using a compression scheme before it hits the (final) gamma curve is a smart move. If we retain less precision around the low and high pixel values and more precision in the middle, we can get more information from the camera sensor in a certain region. Furthermore, we can map a higher dynamic range. It often looks more pleasing to the eye, because we can tune the setup such that it delivers a lot of precision and detail where our perception works the best.
In short: we are storing (8bit/10bit) pixel values. The interpretation of these values are done in the context of a given color space. In classic (rec.709) color spaces, the storage is linear and then mapped onto a non-linear transfer function. In the "log" spaces, the storage is non-linear and is then mapped onto a non-linear transfer function. In essence we perform lossy compression when we store the pixel in the camera.
> The transfer functions in your (rec.709) color space is non-linear indeed. However, the pixel values you store are in a linear relationship with each other. The difference between values 20 and 21 are the same as the difference between values 120 and 121, assuming an 8bit signal. I.e., the information is the same for all pixels.
Difference on what scale? ... because (hint hint) it's not number of photons that hit the sensor. Nor is it photons emitted from the display.
The truth is, it's a linear measure of the voltage which drives the electron beam of a CRT. Not a very useful measure anymore, but we've encoded this response curve into all of our images and, now, this is proving to be a mistake.
Working with images would be so much easier if we stored values that represent linear light (i.e. proportional to photons entering/leaving a device) with no device curves baked in. Log formats do this, but because the order of magnitude of light is more important than the absolute value, it takes the log of the value. It's a more efficient use of bits in the storage / transmission.
No, gamma2.2/sRGB is how the pixels are stored on disk, not linear values. Linear is almost never used except as an intermediate for processing where you can throw lots of bits at it (eg, fp16/32 on a GPU when applying effects or whatever)
The difference is how the curves prioritize what bits to keep. Rec709 sacrifices the bright end to keep more detail in the darks, whereas log is more like linear perceptual brightness. So it'll have less low light detail but more bright detail by comparison
I think I get what you mean (in term of implementation), but can't we just alter the transfer function further so there are more values used for the mid-range colors?
The two-step process you said (the storage is non-linear and is then mapped onto a non-linear) is basically equivalent to a singular transfer function which is the combination of two curves, since the sampling process itself is lossy.
RAW formats on digital cameras are also storing data in a log format. RAW conversion process is normally converting that to a color space along with (for most cameras) doing the De-Bayer algorithm.
The built in converter that produces JPG files in the camera does this too.
Our eyes perceive light as linear when it's really logarithmic.
There is really no difference between video and still here, it's just that it's more normalized at the consumer level to deal with RAW formats at this point for stills.
> There is really no difference between video and still here, it's just that it's more normalized at the consumer level to deal with RAW formats at this point for stills.
Fun trick: convert ProRes RAW to cDNG. You now have 59.94 raw DNGs per second to choose a photo from.
> but I failed to see how it warrants a new "format". You should be able to do that with any existing video format
It's about support.
The .zip format supports LZMA/ZStandard compression and files larger than 4 GB. But if you use that, a lot of software with .zip support will fail to decompress them.
The same way with log. While in theory you could probably make .mp4 or .mkv files with H264 encoded in log, I bet a lot of apps will not display that correctly if at all.
>> ... in DaVinci Resolve ... choose Apple Log for the Input Gamma ...
Indeed, it just sounds to be a different choice of the curve, perhaps more suited for the HDR capabilities available today.
PS: I did not read the article in detail. My first reaction was to just search there for 'gamma' in the article to see how 'log' is being compared to it.
> Standard iPhone video is designed to look good. A very specific kind of good that comes from lots of contrast, punchy, saturated colors, and ample detail in both highlights and shadows.
I remarked to my wife showing me a video recently that you could tell it was taken on an iPhone, I don't think it's just the 'punchiness', for me the main thing is the way it seems to attempt to smooth out motion - the 'in' thing seems to be to sort of spin around showing what's around you while selfie-vlogging and tik-tokking and what-notting, and iPhones make it look like you did it with a steadicam rig that's not quite keeping up.
Another thing they've done more recently is HDR video (to my cave man brain, this means brighter brights).
They've paired this with much higher brightness on the screens, which makes the videos look much more realistic. I first noticed this on my M1 Pro screen, which absolutely blew me away (1600 nits peak brightness).
That's the biggest telltale "filmed on iPhone" trait I'm noticing right now. Yes, you can create HDR videos in other ways, and I'm sure it will be more popular on other platforms soon.
Someone used an iphone to record their desktop screen playing call of duty and the top comment on Reddit was how it made the game look Disneyesque, a spot-on assessment.
> I remarked to my wife showing me a video recently that you could tell it was taken on an iPhone
It's also relatively understood that certain camera companies (Nikon, Canon, Sony, Fuji) have a certain 'look' to them in how they process the raw image sensor data to generate a JPEG (there's a differences in the final colours).
I know exactly what you mean by this! I can always tell if it was taken on an iPhone -- not that it looks bad, or anything, but there's always a few little cues that make it obvious. As you mentioned, I think the motion is a large part of it.
To add, a few generations ago hand held video shot on iPhones was not (or hardly effectively) stabilized. But now iPhone have good stabilization. I think the tradeoff (the too-smooth motion thing) is worth it.
If I was a prosumer/hobbyist video equipment company, I'd be terrified about what Apple does next. They already have significant penetration into the editing market (both with Final Cut, and codec design), they control a number of the common codecs, and they have _millions_ of devices in the field along with substantial manufacturing capability. The cinema end aren't in trouble yet IMO, but the rest should be concerned...
This is just the mop-up operation. The only products left are going to be super-telephotos for live sports (sales: a hundred a year, if that?) and 4K+ IMAX digital cine cameras.
Not even close. The pocketable point and shoot cameras? Sure. DSLR’s? Not a chance. I’ve gone the upgrade path from a canon 6D to 5D4 to R6. The R6 especially is phenomenal and there isn’t a single phone that can even try to come close to what it can accomplish even in “auto” mode.
Smartphones changed the market such that people who just want to shoot good photos of their family don’t need to buy expensive cameras anymore.
But photography with dedicated cameras is alive and well, and won’t go anywhere anytime soon even as these phones get better and better.
The super telephoto market is alive and well, and Wildlife photography in particular is a big contributor to this. When Olympus released their 150-400mm (300-800mm full frame equivalent) super telephoto aimed at wildlife shooters, it was sold out for almost a year.
For me, the new iPhone means I can shoot B-roll footage that looks great, but this will not replace my main camera anytime soon. It’s currently far more viable for high quality video than it is for high quality photographs.
maybe for the average consumer. but how many professional photographers do you see using an iPhone?
sensor size matters for low-light stuff too. sure, an iPhone can do a pretty good job at taking several pictures over say a 2s. exposure, but there _will_ be artifacts in the shot as there isn't physically enough light to form a legible image regardless of post-processing.
this is just one of many reasons why digital cameras are NOT at the brink of collapse yet.
Looks like they killed cameras with built-in lenses. Cameras with interchangeable lesnes, which would've been used by the pros, have kept their market share identical if not grown a bit.
For taking photos and sharing them in the digital only space, sure I’ll buy that for the regular consumer. Making prints will expose all the small sensor flaws that exist quite quickly. I know it gets better every year but I used my phone camera (14 pro) to capture a few important shots that I would do anything to go back and had on a full frame sensor or film for.
Well, Apple seems very cozy with blackmagic design but that doesn’t mean that they aren’t going to be sherlocked. Apple already has offerings in all these categories, it’s just that the markets are different and will stay different some time more because of workflows and laws of nature but the laws of nature don’t seem as safe anymore.
Currently, the best editing software for social media appears to be CapCut as its ease of use for the power it provides is miles ahead of anything else.
Cell phones already killed compact point-and-shoots with their small lenses.
But DSLR's with their large lenses aren't going anywhere, because physics. If you want to capture high-quality footage under low light conditions, or work with a variety of lenses, the tiny aperture on a phone is never going to be enough.
The RICOH GRIII/GRIIIx is still alive and kickin'. Really nice sensor, super sharp lens, and small enough to fit into a pocket. I don't really use my iPhone camera anymore.
Setting aside the existing pro video pipeline... it's for this reason that if I were Meta, I'd be terrified about what Apple does next, from the perspective of photorealistic capture for VR/AR applications.
Even if the next-gen Oculus had parity with Apple Vision Pro in how vividly they could display VR/AR content - and I doubt this is the case - only Apple (perhaps among any company in the world) can mirror that with a battle-tested supply pipeline for creating pro-level video capture equipment and for integrating those sensors into consumer devices at scale.
While I'd leave the branding to better minds than my own, I'm bullish about Apple's ability to make an "iPad Vision Pro" that has two cameras at eye distance apart as well as laser rangefinding. That would allow for binocular Apple Log capture, and with the advent of Gaussian splatting for point cloud rendering, and increasingly better generative AI for inferring colors/textures of occluded points in the point cloud, you could have professionally color-graded interactive scenes. All that's missing is better ergonomics for this workflow in DaVinci Resolve etc., and one imagines Apple's war chest could go a long way towards incentivizing this.
Apple products will be where high-quality VR content is created, proofed, and consumed. They're not rushing because nobody else has a hope of being close.
> Apple products will be where high-quality VR content is created, proofed, and consumed
Eh... I think this is a bit like claiming "Apple hardware will be where high-quality video games are created, proofed and consumed" in the 90s. You'd be right given the circumstantial evidence, but the industry doesn't really want to play ball. Besides AppleTV and Apple Arcade, I don't think Apple has a huge constituent of VR developers the same way they did for Apps. Simply building the infrastructure won't be enough.
And plus, I don't think your argument rules out a world where Vision Pro is used for recording and Quests are used for content consumption. Unless Apple intends to reinvent the wheel again for 3D video, nothing should stop it from being playable (or at least portable) on other platforms. For most people, the $3,000 price difference will probably not be worth the recording capabilities (especially if they already own an iPhone).
Anything could happen, I guess. Even if you're entirely right though, I don't think Content Creation will be the killer app you think it is. Games and video moves millions of units; productivity and capture are both unproven features poised to reprise Hololens' failure.
I think what is terrifying is that they're better enough to kill the details.
Sort of a contrived example... you're a pro and let's say you NEED a headphone jack, but apple just killed headphone jacks.
But more indirectly, they killed the lower volume, higher margin folks with alternatives that offer a headphone jack, and maybe even XLR headphones and microphones.
An analogy might be tesla giving you a 90% better car experience, except they have killed off the dashboard. (now they've killed control stalks like PRND and turn signals)
I don't know. If I were a pro like that, I would just have a dongle for that headphone jack. It would probably also have a better DAC than iPhones ever had.
A big factor in buying a "pro" camera is controls. It's really difficult changing the focus in a controlled way while shooting with a phone. While in theory you can imagine Apple giving you API control for that and hooking it with an external focus pulling device, it's still a sub-optimal solution.
Cinema is safe from the optics point of view as achieving some of the effects of large sensor + large lens is impossible with a phone size camera. But Apple has it cracked and they could easily crush that market. They have a great sensor with enough resolution per inch, great dynamic range, ability to produce lenses with super low defects and have enough processing power. Sensor wise ~100 megapixels should be enough to replicate fine grain of a good movie film and iPhone 15 sensor's dynamic range of 12-14 stops is on par with film already.
And it turns out, Sony already has significantly better sensors in their cinema cameras than what they sell to Apple, which is why almost half of Oscar and Emmy nominations are filmed on Sony Venice. Even the low end of their camera line is extremely capable, e.g. the FX3 which shot The Creator.
Apple is far away from actually competing in that market.
The market for 'actual' pro & prosumer cameras and such is pretty tiny. I think they'll be pretty safe for quite a long time.
But they have pro video editing features! Yes, but it's a subfeature of their general platform, so they can 'count that low' for a hardware feature like that, that will also be useful for their entire userbase since everyone takes videos with their hardware and watches video on their devices anyway.
I don't see this as a high risk. Rather adding log just increases the choices for different types of filmmaking and adding it to the professional workflow. The iPhone hasn't killed the professional stills camera market, instead many photographers have implemented it as a supplemental tool into their work where a larger camera is not suitable.
At the end of the day, a professional camera with good glass and a larger sensor gives you much more control of an artistic vision, and better performance in challenging imaging conditions. I've never seen a cellphone camera that gets anywhere near the image quality I can get from my mirrorless camera in still images. This is likely to be even more apparent in video production due to how the cameras are used.
They owned the editing market with Final Cut and completely dropped the ball with Final Cut X. To the point they had to start selling the old version back. Then Premiere came back from the ashes and took the throne.
One of the most glaring mistakes Apple did to its Pro market. And they did quite a few.
It's always surprised me that there's not more interest in log-scale/floating-point ADCs built directly into camera sensors. Both humans and algorithms care a lot more about a couple-bit difference in dark areas than light, and we happily use floating point numbers to represent high-range values elsewhere in CS.
There was a company that did this circa 2003 - SMaL. Their "autobrite" sensor is built to capture log-scale natively. They've switched owners twice since then, but it seems like they're getting more traction in car vision systems than in professional video.
Blast from the past. You could do things like that [0] in 2005... Autobrite was the first thing to go out the window, once we knew how the chip worked we did or own exposure control...
From an analog design perspective, I don't think that makes sense. Not that I'm an analog designer, but I worked closely with them as a digital designer on CMOS camera sensors.
You're already extracting as much information you can from the analog signal on the least significant bits. It's not like designing a log-scale ADC lets you pull more information from the least significant bits. So you don't really have anything to gain. Why make a more complicated analog circuit to extract less information? It's generally better to let the digital side decide what to keep, how to compress the signal, etc.
And I should mention that CMOS camera sensors can often do a lot of digital processing right there on the chip. So you can do log-scale conversion or whatever you want before you send the data out of the CMOS camera chip.
It might be possible that you could reduce the power consumption of a SAR (successive approximation) ADC by skipping AD conversion of less significant bits if you have signal on the more significant bits. But I doubt the power savings would be very significant.
A lot of the processing steps expect linearity and would have to be reworked for floating point or log scale data. Most HDR sensors are using some kind of logarithmic compression for sensor readout, but I've never really heard of a floating point adc. Google seems to suggest tbey're not readily available.
Yeah there isn't, it's not that simple, even at sensor level to have that much dynamic range
A linear ADC with enough range is usually fine, you can do the math later. But maybe for this case it needs a non-linear element before the ADC? (no idea how log recording needs anything at the HW level)
Is the quantisation error on a modern 14-something bit sensor really that big of a problem compared to something like the inherent shot noise of dark areas?
Some sensors do this internally, unusual though. The rest of the high-end ones apply curves manually in software directly at the egress of the sensor. The reason they don't in all cases is that it complicates black level correction, gamut transforms and demosaic operations (without some assumptions).
I feel like professional/prosumer photogs, aka the kind of people who buy fancy SLR cameras and serious business lenses, probably already know this stuff. I also suspect that the vast, vast majority of phone users just want subjectively good-looking photos.
Yeah, 100%. About 99% of the customer base just wants to take a good photo with the smallest effort possible. Which makes it even more remarkable the company cares enough to include this kind of functionality in a consumer product.
See but you don't necessarily want full range logarithmic.
If your darkest pixel is halfway down the ADC range, on linear you're throwing away one bit, on logarithmic you're throwing out way more bits. Just using higher bit linear ADC then converting it to logarithmic in post-processing seems more sensible. Hell you could even go signal magic and merge few photos with different shutter speed together to get most detail.
Also proper log-to-in converter like AD8307 costs like $20 so I'd assume doing that woudl bring the sensor price way up if you needed to have a bunch of them.
From my understanding, the ADC's are still fixed point and linear. Two (or more) then run in parallel over different signal levels to produce the 32-bit float output.
Encoding audio with different log-scale companding has been around for some time too (since the 1970's) with A-law and mu-law in G.711.
I didn't know it could record straight to USB-C storage! That gets rid of a major reason to spend crazy money on a 1TB phone, and it's definitely a game changer for anyone shooting 4K ProRes.
Afaik it’s actually not even possible to record that directly to the phone, it has to be into an external usbc drive. If I had to guess it’s probably because of overheating concerns with the high write rate.
The 4k60 ProRes mode is not available for shooting on the 128gb model until an external drive is added, but for any larger capacity iPhone Pro, the mode is available for shooting without the external drive. This doesn't affect the Pro Max as that isn't available in a 128gb configuration.
Notably, a similar limitation existed with the iPhone 14 Pro using 4k30FPS, at the time the reasoning was that it simply fills the device too quickly to be useful.
> With its high bit depth and dynamic range, log footage has many of the benefits of raw. But Apple Log is not raw, and not even “straight off the sensor.” It’s still heavily processed — denoised, tone-mapped, and color adjusted.
I wonder if this is because, at the end of the day, it is still a tiny little camera with a small sensor and small lens, and so with none of the processing magic the image would look pretty terrible under most circumstancdes.
RAW video isn’t like RAW photography. The sheer size of raw footage is insane - it’s normal for cameras to be unable to record RAW footage natively without an external recorder.
Thats not to say processing isn’t part of it, but even $2k mirrorless cameras don’t record RAW video internally.
The sensors and lenses are small, and now processors are very fast. And as it turns out, the vast majority of people do not want photos or videos that are "accurate" or "real". They want ones that look good.
So, processing has been the name of the game. It's all about making an image people will like to look at, regardless of how different it is from reality.
Probably true, and I enjoy using my phone for social snapshots. But the way in which everything ends up looking "iphony" annoys me for anything where I care about the image, and so I'm shooting more with my dumber digital cameras that are about documenting what was actually in front of me rather than spinning up a "good looking" image. I like being able to remember what a scene actually looked like, rather than what the AI in my phone thought it should look like ; )
Even high-end cameras that can shoot in Log, like the Sony A7 series, apply some noise reduction on their end. This is important for most compressed formats.
However, most people would be horrified to see how noisy and unsharp the images from top cinema cameras are when most post-processing is disabled.
Log seems like a strong reason to finally switch from Android to iPhone if you're a photography/filmmaking enthusiast like myself. The ecosystem is so much more mature and the gap seems to be growing not shrinking.
Android has Raw Video with MotionCam which also produces insanely good results¹ (even better than iPhone's ProRes video), but everything else just sucks.
> a strong reason to finally switch from Android to iPhone if you're a photography/filmmaking enthusiast
Correct me if I'm wrong but there's nothing stopping Android of supporting Log (or similar). I'm not a video engineer but it really doesn't seem that magic that it couldn't be supported outside of iphone 15, right? My guess that if this gains any real traction it'll show up in the next Android flagship.
I've been a long-term mcpro24fps and a user of Filmic pro before that. It's a great app, no doubt about it. The issue is not the app, but the OEMs who makes things difficult, artificially limiting the capabilities of the devices and even removing features in updates. Nothing is consistent and each device works differently from the next one, even from the same manufacturer. A long running joke in the McPro24fps Telegram chat is to never upgrade!
Readit News
One lesson (no pun intended) of the academic environment is that no, it doesn’t work that way. Some people are subject-matter experts, some are brilliant expositors, but while you can’t be a good expositor without a decent knowledge of the subject matter, you absolutely can be a world-leading scientist and at the same time completely rubbish at explaining your ideas. (Some areas are better at making use of such people than others.)
Good exposition deserves additional credit on top of subject expertise, is what I’m trying to say.
this guy knows what he’s talking about. he has authority in his field. he has proven both theoretically and materially that his wisdom is high quality.
yet, videos from people with far less expertise on the subject matter will likely drown his out. over and over again i’ve seen the top videos on so many subjects repeating just plain wrong information but the “creator” says it with confidence so people just eat it up. even worse, so many times i’ve seen people in the comments try to nicely point out where the “creator” went wrong, and people go after the commenter for daring to imply this person on the magic screen might be misinformed.
it’s such a shame that being higher in the search says almost nothing about your knowledge on a subject and only means you’re better at manipulating the algorithm rather than having actual higher quality information.
i really wish that our searches would prioritize quality of information. at this point, years later, i think it’s pretty clear that the “wisdom of the crowd will shoot the highest quality to the top” has been proven to not be the case.
also, i don’t want to imply that amateurs can’t have valid and interesting perspectives, sometimes they might, just that being higher in search means quite literally nothing on wisdom of a subject—especially if a grifter can monetize off a trendy nuanced subject.
anyway, this video was incredible information. thx for sharing his background.
https://forums.pelicanparts.com/off-topic-discussions/301497...
I'm the (formerly) young guy who rented the 550 to the production team for a few days. Pretty fun story to be tied to the movie.
https://duckduckgo.com/?t=ffab&q=ren+and+stimpy+log+song&atb...
With glass and a sensor that small they aren't for everything, but the days of mandatory compression, limited color depth, mandatory "enhancements" and all of that stuff are over. If I'm shooting a handheld gimbal video, of something like a person talking outdoors not in direct sunlight, I'm grabbing that iPhone Pro without thinking twice.
https://prolost.com/blog/rawvslog/
Deleted Comment
Deleted Comment
Firstly the name is called "Log" (for logarithmic) but isn't that what gamma does in color spaces like sRGB since forever? "Normal" video standards like BT.709 also have non-linear transfer functions. I don't get why "log" is stressed here. Maybe it just means a different/higher gamma coefficient (the author didn't talk much about the "log" part in the article).
And the main feature of it, at least according to this article, is that it clips the black and white level less, so leaves more headrooms for post-processing.
This is definitely very useful (and is the norm if you want to do something like, say, high quality scanning), but I failed to see how it warrants a new "format". You should be able to do that with any existing video format (given you have enough bit depth, of course).
It actually looks more like HLG in this way.
https://www.artstation.com/blogs/tiberius-viris/3ZBO/color-s... has some plots of the curves to compare visually
I think you're mixing up OOTFs and EOTFs here. sRGB or HLG can refer to either the stored gamma, but more often means the EOTF "reversed" gamma that is used to display an image. When we refer to "log", this is almost always means a camera gamma - an OOTF. So the reason it's "in the opposite direction" is that it's designed to efficiently utilize bits for storing image data, whereas the EOTF is designed to reverse this storage gamma for display purposes.
As you can see from the graph in [1], Sony's S-Log does indeed allocate more bits to dark areas than bright areas. (Though the shape of the curve becomes more complicated if you take into account the non-linear behavior of light in human vision.)
[1] https://www.enriquepacheco.com/blog/s-log-tutorial
Deleted Comment
This warrants a separate answer. Cameras are getting to the point where they can capture far more information than we can display. Hence, we need a lot of bit depth to accurately store this added precision. But adding bits to the data signal requires a lot of extra bandwidth.
In principle, we should just store all of this as 16/32bit FP, and many modern NLEs use such a pipeline, internally. But by creating a non-linear curve on integer data, we can compress the signal and fine-tune it to our liking. Hence we can get away with using the 8-12bit range, which helps a lot in storage. With log-curves, 12bit is probably overkill given the current sensor capabilities.
There's a plethora of log-formats out there, typically one for each camera brand/sensor. They aren't meant for delivery, but for capture. If you want to deliver, you'd typically transform to a color space such as rec.709 (assuming standard SDR, HDR is a different beast). The log-formats give you a lot of post-processing headroom while doing your color grading work.
Haven't professional-grade microphones been in a similar situation for decades now, or is it the magic of remastering that keeps recordings from the 50s sounding so good on modern speaker systems?
But yeah, I've been wondering why nonlinear formats would use integer values for a while now ?!?
What the "log"-spaces are doing is to use a non-linear relationship for the pixel values, as a form of lossy compression. If the signal has to factor through 8bit values, using a compression scheme before it hits the (final) gamma curve is a smart move. If we retain less precision around the low and high pixel values and more precision in the middle, we can get more information from the camera sensor in a certain region. Furthermore, we can map a higher dynamic range. It often looks more pleasing to the eye, because we can tune the setup such that it delivers a lot of precision and detail where our perception works the best.
In short: we are storing (8bit/10bit) pixel values. The interpretation of these values are done in the context of a given color space. In classic (rec.709) color spaces, the storage is linear and then mapped onto a non-linear transfer function. In the "log" spaces, the storage is non-linear and is then mapped onto a non-linear transfer function. In essence we perform lossy compression when we store the pixel in the camera.
Difference on what scale? ... because (hint hint) it's not number of photons that hit the sensor. Nor is it photons emitted from the display.
The truth is, it's a linear measure of the voltage which drives the electron beam of a CRT. Not a very useful measure anymore, but we've encoded this response curve into all of our images and, now, this is proving to be a mistake.
Working with images would be so much easier if we stored values that represent linear light (i.e. proportional to photons entering/leaving a device) with no device curves baked in. Log formats do this, but because the order of magnitude of light is more important than the absolute value, it takes the log of the value. It's a more efficient use of bits in the storage / transmission.
The difference is how the curves prioritize what bits to keep. Rec709 sacrifices the bright end to keep more detail in the darks, whereas log is more like linear perceptual brightness. So it'll have less low light detail but more bright detail by comparison
I think I get what you mean (in term of implementation), but can't we just alter the transfer function further so there are more values used for the mid-range colors?
The two-step process you said (the storage is non-linear and is then mapped onto a non-linear) is basically equivalent to a singular transfer function which is the combination of two curves, since the sampling process itself is lossy.
The built in converter that produces JPG files in the camera does this too.
Our eyes perceive light as linear when it's really logarithmic.
There is really no difference between video and still here, it's just that it's more normalized at the consumer level to deal with RAW formats at this point for stills.
Fun trick: convert ProRes RAW to cDNG. You now have 59.94 raw DNGs per second to choose a photo from.
It's about support.
The .zip format supports LZMA/ZStandard compression and files larger than 4 GB. But if you use that, a lot of software with .zip support will fail to decompress them.
The same way with log. While in theory you could probably make .mp4 or .mkv files with H264 encoded in log, I bet a lot of apps will not display that correctly if at all.
Deleted Comment
>> ... in DaVinci Resolve ... choose Apple Log for the Input Gamma ...
Indeed, it just sounds to be a different choice of the curve, perhaps more suited for the HDR capabilities available today.
PS: I did not read the article in detail. My first reaction was to just search there for 'gamma' in the article to see how 'log' is being compared to it.
I remarked to my wife showing me a video recently that you could tell it was taken on an iPhone, I don't think it's just the 'punchiness', for me the main thing is the way it seems to attempt to smooth out motion - the 'in' thing seems to be to sort of spin around showing what's around you while selfie-vlogging and tik-tokking and what-notting, and iPhones make it look like you did it with a steadicam rig that's not quite keeping up.
They've paired this with much higher brightness on the screens, which makes the videos look much more realistic. I first noticed this on my M1 Pro screen, which absolutely blew me away (1600 nits peak brightness).
That's the biggest telltale "filmed on iPhone" trait I'm noticing right now. Yes, you can create HDR videos in other ways, and I'm sure it will be more popular on other platforms soon.
It's also relatively understood that certain camera companies (Nikon, Canon, Sony, Fuji) have a certain 'look' to them in how they process the raw image sensor data to generate a JPEG (there's a differences in the final colours).
I don't have an iPhone, I've just noticed this (perhaps it's more obvious to me because I don't have one) in others' videos.
EIS is not usually needed for video but maybe in some cases it’s used?
This is just the mop-up operation. The only products left are going to be super-telephotos for live sports (sales: a hundred a year, if that?) and 4K+ IMAX digital cine cameras.
But photography with dedicated cameras is alive and well, and won’t go anywhere anytime soon even as these phones get better and better.
The super telephoto market is alive and well, and Wildlife photography in particular is a big contributor to this. When Olympus released their 150-400mm (300-800mm full frame equivalent) super telephoto aimed at wildlife shooters, it was sold out for almost a year.
For me, the new iPhone means I can shoot B-roll footage that looks great, but this will not replace my main camera anytime soon. It’s currently far more viable for high quality video than it is for high quality photographs.
sensor size matters for low-light stuff too. sure, an iPhone can do a pretty good job at taking several pictures over say a 2s. exposure, but there _will_ be artifacts in the shot as there isn't physically enough light to form a legible image regardless of post-processing.
this is just one of many reasons why digital cameras are NOT at the brink of collapse yet.
Currently, the best editing software for social media appears to be CapCut as its ease of use for the power it provides is miles ahead of anything else.
But DSLR's with their large lenses aren't going anywhere, because physics. If you want to capture high-quality footage under low light conditions, or work with a variety of lenses, the tiny aperture on a phone is never going to be enough.
Even if the next-gen Oculus had parity with Apple Vision Pro in how vividly they could display VR/AR content - and I doubt this is the case - only Apple (perhaps among any company in the world) can mirror that with a battle-tested supply pipeline for creating pro-level video capture equipment and for integrating those sensors into consumer devices at scale.
While I'd leave the branding to better minds than my own, I'm bullish about Apple's ability to make an "iPad Vision Pro" that has two cameras at eye distance apart as well as laser rangefinding. That would allow for binocular Apple Log capture, and with the advent of Gaussian splatting for point cloud rendering, and increasingly better generative AI for inferring colors/textures of occluded points in the point cloud, you could have professionally color-graded interactive scenes. All that's missing is better ergonomics for this workflow in DaVinci Resolve etc., and one imagines Apple's war chest could go a long way towards incentivizing this.
Apple products will be where high-quality VR content is created, proofed, and consumed. They're not rushing because nobody else has a hope of being close.
Eh... I think this is a bit like claiming "Apple hardware will be where high-quality video games are created, proofed and consumed" in the 90s. You'd be right given the circumstantial evidence, but the industry doesn't really want to play ball. Besides AppleTV and Apple Arcade, I don't think Apple has a huge constituent of VR developers the same way they did for Apps. Simply building the infrastructure won't be enough.
And plus, I don't think your argument rules out a world where Vision Pro is used for recording and Quests are used for content consumption. Unless Apple intends to reinvent the wheel again for 3D video, nothing should stop it from being playable (or at least portable) on other platforms. For most people, the $3,000 price difference will probably not be worth the recording capabilities (especially if they already own an iPhone).
Anything could happen, I guess. Even if you're entirely right though, I don't think Content Creation will be the killer app you think it is. Games and video moves millions of units; productivity and capture are both unproven features poised to reprise Hololens' failure.
Sort of a contrived example... you're a pro and let's say you NEED a headphone jack, but apple just killed headphone jacks.
But more indirectly, they killed the lower volume, higher margin folks with alternatives that offer a headphone jack, and maybe even XLR headphones and microphones.
An analogy might be tesla giving you a 90% better car experience, except they have killed off the dashboard. (now they've killed control stalks like PRND and turn signals)
And it turns out, Sony already has significantly better sensors in their cinema cameras than what they sell to Apple, which is why almost half of Oscar and Emmy nominations are filmed on Sony Venice. Even the low end of their camera line is extremely capable, e.g. the FX3 which shot The Creator.
Apple is far away from actually competing in that market.
But they have pro video editing features! Yes, but it's a subfeature of their general platform, so they can 'count that low' for a hardware feature like that, that will also be useful for their entire userbase since everyone takes videos with their hardware and watches video on their devices anyway.
At the end of the day, a professional camera with good glass and a larger sensor gives you much more control of an artistic vision, and better performance in challenging imaging conditions. I've never seen a cellphone camera that gets anywhere near the image quality I can get from my mirrorless camera in still images. This is likely to be even more apparent in video production due to how the cameras are used.
One of the most glaring mistakes Apple did to its Pro market. And they did quite a few.
https://www.vision-systems.com/cameras-accessories/article/1...
[0] https://youtu.be/0UaTYX-ygG8?feature=shared
You're already extracting as much information you can from the analog signal on the least significant bits. It's not like designing a log-scale ADC lets you pull more information from the least significant bits. So you don't really have anything to gain. Why make a more complicated analog circuit to extract less information? It's generally better to let the digital side decide what to keep, how to compress the signal, etc.
And I should mention that CMOS camera sensors can often do a lot of digital processing right there on the chip. So you can do log-scale conversion or whatever you want before you send the data out of the CMOS camera chip.
It might be possible that you could reduce the power consumption of a SAR (successive approximation) ADC by skipping AD conversion of less significant bits if you have signal on the more significant bits. But I doubt the power savings would be very significant.
A linear ADC with enough range is usually fine, you can do the math later. But maybe for this case it needs a non-linear element before the ADC? (no idea how log recording needs anything at the HW level)
hardware accelerated HDR on cameras are commonplace these days, especially in dashcams and CCTV cameras.
If your darkest pixel is halfway down the ADC range, on linear you're throwing away one bit, on logarithmic you're throwing out way more bits. Just using higher bit linear ADC then converting it to logarithmic in post-processing seems more sensible. Hell you could even go signal magic and merge few photos with different shutter speed together to get most detail.
Also proper log-to-in converter like AD8307 costs like $20 so I'd assume doing that woudl bring the sensor price way up if you needed to have a bunch of them.
Encoding audio with different log-scale companding has been around for some time too (since the 1970's) with A-law and mu-law in G.711.
The 4k60 ProRes mode is not available for shooting on the 128gb model until an external drive is added, but for any larger capacity iPhone Pro, the mode is available for shooting without the external drive. This doesn't affect the Pro Max as that isn't available in a 128gb configuration.
Notably, a similar limitation existed with the iPhone 14 Pro using 4k30FPS, at the time the reasoning was that it simply fills the device too quickly to be useful.
I wonder if this is because, at the end of the day, it is still a tiny little camera with a small sensor and small lens, and so with none of the processing magic the image would look pretty terrible under most circumstancdes.
Thats not to say processing isn’t part of it, but even $2k mirrorless cameras don’t record RAW video internally.
The sensors and lenses are small, and now processors are very fast. And as it turns out, the vast majority of people do not want photos or videos that are "accurate" or "real". They want ones that look good.
So, processing has been the name of the game. It's all about making an image people will like to look at, regardless of how different it is from reality.
However, most people would be horrified to see how noisy and unsharp the images from top cinema cameras are when most post-processing is disabled.
Android has Raw Video with MotionCam which also produces insanely good results¹ (even better than iPhone's ProRes video), but everything else just sucks.
[1]: https://youtu.be/O5fnGDR4i9w?feature=shared
Correct me if I'm wrong but there's nothing stopping Android of supporting Log (or similar). I'm not a video engineer but it really doesn't seem that magic that it couldn't be supported outside of iphone 15, right? My guess that if this gains any real traction it'll show up in the next Android flagship.
But the fragmentation does work against it. If some company does it it would be limited to their camera app and their format.
Would be interesting if some company just decided to put c-mount on their phone so you could use actual proper lens...
On Android you have mcpro24fps app that supports multiple log profiles, shooting 10 bit video and more.