8K vs 4K TVs: Double-blind study by Warner Bros. et al reveals.....

[Misha Engel]

8K vs 4K TVs: Double-blind study by Warner Bros. et al reveals most consumers can’t tell the difference
https://www.techhive.com/article/3529913/8k-vs-4k-tvs-most-consumers-cannot-tell-the-difference.html

Despite all this, moving to 8K makes sense for cinematographers, because of the expanded possibilities the resolution offers. Shooting 8K for the final production may be, in most cases, very distant in the future, but using the format to have more options in terms of cropping, due to the higher resolution, makes whatever is captured future-proof and a sound choice, as far as the means to shoot and edit 8K are available.
https://www.provideocoalition.com/8k-versus-4k-warner-bros-study-reveals-consumers-see-no-difference/

 

[AndreeMarkefors]

Talking about resolution as a 'hero number' seems pretty meaningless.

I have zero excitement personally for high resolutions TVs for high resolution's sake. Watching on my iPad, HD seems fine and when I'm in front my 65" LG OLED, I enjoy UHD.

I used to have a projector many years ago, and if I would get around to getting another one, I imagine watching a Masters golf (no jerky camera moves, or fast moving objects/people) in 8K on a 150" screen would be pretty immersive.

I think that if the consumer facing discussion could at least center on some common guidelines for resolution in relation to screen size*, that would be somewhat honest.

*= I wouldn't even mention viewing distance and just assume that sorts itself out based on screen size. Even if screen size vs distance might be the key value, already this is "too complex" for the average consumer.

Effectively, we would just need to communicate a PPI in order to maintain a certain IQ density when moving up in display size, while accepting that it's a sliding value that gets a bit lower with very large screens.

 

[Stacey Spears]

The nature clip was from the HDR Montage on my HDR 4K Blu-ray test disc. It was shot by Clark Dunbar on Vista Vision (I forget if the shot was Dragon or Monstro) and graded by Shane Ruggieri at Dolby. It was also the brightest shot at 3000 nits.

 

[Phil Holland]

I was a participant in this test and shockingly was able to discern the different prior to my recent eye surgery "most of the time". I do have some off the air thoughts about general testing method, but I will say from our previous test one of the key things was having 70" matched color HD, UHD 4K, and UHD 8K displays running the same content. People were able to turn to each display while the content ran in sync to do nice A/B/C comparisons. As much as this has been about the quality of the footage it's also about the quality of the display to reveal the differences.

It still was a great test. And fascinatingly one of the best shots for me was an 8K clip from Nolan's Dunkirk. That shot looked amazing in 8K in many ways to me. But it looked great in 4K as well as it was downsampled from the higher resolution source. Certainly a few things being mentioned here that underline certain concepts as to why the 4K clips also looked good.

Stacey's footage and Bug's Life stuff certainly had moment that were easy to discern. There was a scene from The Tick that also provided pretty obvious specular highlights that simply weren't visible in the 4K version as they were smoothed out in the upscale and that texture just was gone. Eyes were also a pretty nice telling target.


Personally I'm all good and many of these questions were answered for me long ago. I've been on the war path for a 4K/8K ecosystem and that is indeed exactly what's going on industry-wide. Sooner than you think you'll actually only be able to buy 8K TVs even, but we're not there yet. The concept itself is to have a display in your home that can show "everything in it's best light". Whether it's 4K upscaled through some sort of intelligence to 8K, likely an HD production upscaled similarly or via post efforts, or just showing 8K content. That is more or less the answer to all of this.

Then there is the final question that proposes, which we are far away from. Will consumer ever find the true need for a higher than 8K resolution television in the home?

That is indeed the hardest one to sell. 8K 55"-100" is pretty damn good when you are at the optimal and sub-optimal viewing distances, which reflect to really why it's a good resolution solution for most viewing situations. Even theatrically. And heck, I want 27-50" UHD 8K displays/monitors as well as DCI 8K monitors. Sort of basically what we have now for 4K industry-wide. We'll get there fast. I've been banging on doors for a few years.

The big one here though related to production, is you'll see a raising interest in monitoring in 4K on set this year onward. It's not so much that there's any real single game changer here, but rather the game itself has changed.

 

[M Most]

I've been on the war path for a 4K/8K ecosystem and that is indeed exactly what's going on industry-wide...

Not from where I'm sitting, at least not as long as the LF and the Venice (which is, admittedly, commonly shot in 6K format) are the most common choices for studio productions. That can easily change, of course. But for high end studio work, that is the current scenario.

As for on set monitoring, I'm hearing far more about HDR than anything regarding resolution, although HDR usually implies 4K these days.

 

[Clark Dunbar]

thanks Stacey - YEAH ! ...the MOOSE won!... that was shot on the 8K Dragon VistaVision (Redcode 8:1) - with the Canon 70-200 with a Canon 2X extender - so shot at 400mm) - so not a prime... which would have been even sharper with more detail.

 

[Phil Holland]

Not from where I'm sitting, at least not as long as the LF and the Venice (which is, admittedly, commonly shot in 6K format) are the most common choices for studio productions. That can easily change, of course. But for high end studio work, that is the current scenario.

As for on set monitoring, I'm hearing far more about HDR than anything regarding resolution, although HDR usually implies 4K these days.

The LF and Venice are respectively 4.5K and 6K camera sytems, which is part of a 4K/8K ecosystem. Both designed to produce a 4K delivery.

The Venice is popular, but it's pretty much ARRI -> RED -> Sony at the larger rental houses across America. Interestingly RED is more popular in some markets, but that depends on lots of factors I imagine. It's minutiae. I have not seen any instance of the Venice being more popular than RED nor Arri in any market. In fact many major markets don't even have Venice bodies. They are out here for sure though and of course in the biggest regions. And it's a decent camera.

I'll reserve some other numbers because I know how many productions are using the DXL and DXL2 bodies world-wide and you can decide personally if that is purely a Panavision camera or a Panavision/RED concept. In my mind those also count as RED, especially in how they are used in practice on set. Mostly mentioning it, because it should count too.

This year you will have HDR and 4K monitors on set. Actually in about 1 month. I've been testing a host of them nearly for the entirely of 2019. Just need more cameras with 4K out. Not really a problem for some as they have it already.


On a deeply personal note the general message and song I've been singing for 2 decades doesn't need to be sung much anymore. I'm well done with pissing contests now that every manufacturing has stepped up to the plate with 4K+ camera systems. Everybody has "made me happy" and "did the right thing" by motion picture professionals as of 2020. Each system has it's strengths, weaknesses, quirks, and warts. In the last 3-4 years my main whips have been Dragon then Monstro 8K VV for my work with a spattering of Alexa 65, Alexa LF, 2 Venice shoots, and 8 film shoots (S35, VV, and 15-per IMAX). My last big regular Alexa shoot was on Life of Pi at the end of last decade and I had one smaller shoot in the 2010s on it outside of various camera tests for studios, productions, and personal/professional reasons. I virtually filmed nothing below 4K resolution in 2010s. I am fully aware that I don't represent the entire industry, but typically in a 5 year span I am a forward trend filmmaker with an effort on 2-3 leans as trends emerge to market and grab hold of the hearts, minds, and desires of filmmakers.

Yedlin is on point to the drum I've also been banging for a long while in most ways. Film is still beautiful, be we've met and exceeded film's capabilities on technical merit in every category. And for many whose attraction to film's general workflow and common aesthetic, which is myself included, we can push our images there when needed. Or the more simple path of just working with film if you can afford it financially and temporally. It very much was our industry's primarily and only medium, but that isn't the case now.

Anyways. Long story short. No manufacture is anti-4K in this era. Every manufacturer has gotten the memo. It's now up to productions to choose their destiny on 2K or 4K deliverables and for the few of us producing content delivered in 8K, we can finish out to 4K and 2K just as well, the market much like 4K is increasing for 8K content, and will be very noticeable in the next 2-5 years. Yes and number-wise, the market presence for finished 8K content in this era is < 1% of production globally, compared to UHD 4K productions. Interestingly this mirrors relatively well to the percentage of 4K productions in a SD/HD world in the 2000s back when the industry was significantly smaller. That's delivered content and not "camera presence".

Mass market is HD (*gasp somehow still 720p being looped in there) with the heightened slope to UHD 4K due primarily to the streaming push on professional productions. Nobody in the US is doing streaming 8K outside of YouTube and Vimeo as of yet. In Japan, 8K is happening slowly but surely with now 2 streaming entities. By the end of the year and of course by summer, even more of a presence there. Lots and lots of 8K this year at the Olympics and the original content they are currently producing in 8K.

I'm happy we're here. Audiences deserve a better viewing experience. They have it now.

My big fights moving forward are higher encoded bit-rates for 4K and 8K streaming content, which hasn't been and won't be a well received concept by many. Though in the modern era we can support it now with the proper work done on the streaming services end to provide adaptive streams, similar to what they do now, but tied more to each users connection. HDR has a host of fun things related to panning speeds and judder which has perceptual relevance at higher contrast ratios and general brightness. A few years back before the Apple stuff was the largest global Dolby Vision release I did a bunch and I mean a bunch of tests to discern the accurate panning speeds related to format size with a maximum contrast 400, 1000, 2000, and 4000 nit finish as that was necessary for that release to understand deeply in my bones while filming. Now we'll be seeing some novel motion interpolation and temporal "remapping" to solve those problems. "Creator Mode" already seems to have the required traction to get there from here. I am still pushing for a metadata driven fps ecosystem (per content type basis) because I strongly feel 60p and even 120p for sports looks amazing, but 24p for me is very much what narrative work should look and feel like. I guess some of the bigger stuff long term will be pushing more of the 10-bit, 12-bit, and eventual 14-bit ecosystem. We're years away from panels doing that right. The real next big leap is 12-bit though. I can't do much more than suggest the path as the tech needs to catch up and become much more affordable. The multi-panel approach is one way, but not a great off-axis solution. Yet.

Since I'm online and people will likely read stuff like this, and I brought this topic up a couple years ago at an ASC MITC meeting. I like H.265 HDR at 80mbps in 4K, 160-320mbps in 8K. I would say for streaming services about 1/2 of those numbers is the next leap/goal you should look at. If you are making an adaptive and intelligent system in this gigabit and shortly 5G era, I would add to the current data rate tiers what you are doing, the 1/2 of that number, and the full target into the mix. 4 of the top 5 television manufacturers have been my clients for a while now and I've been advising for more powerful hardware to playback content internally. 80mbps is pretty easy with modern chips, but there are a few 8K sets that can indeed do "smooth" 180mbps right off an SD Card or USB based source. Not everybody on earth has a sustained 40MB/s bandwidth, but for those that do and are paying for it, it would be nice to really get the most out of everything. Some of the newer codecs, some proprietary stuff that's already in use btw, will help with this and likely a new round of data rates will need to be discerned.


Apologies for the longer than expected post. But this is important to me and should be to more discerning audiences and creative professionals. All of the above represents cost and time implications, but in my opinion, well worth the effort. For many who were truly anti-digital early on the end resulting viewing experience from the combined capture through post workflow/technology was the "main concern" about it. The cameras and display technology have drastically improved since the Viper and Genesis era thanks fully to RED, ARRI, Sony, and others with a honorable highlight mention to Panasonic (HVX200 was the keystone for all current digital cinema cameras really if you think about it and they deserve that recognition despite color being not a strong point of that camera) and Canon in there for some novel innovations berfore their proper splashes in the digital cinema scene. We have the opportunity to now button up those last buttons. We're close already. It's that last mile that's the hardest. In our case, it's the last like 3-5 miles abstractly and likely will be evident on notable technology leaps over the next few decades.

 

[Blair S. Paulsen]

A typical residential viewing setting has physical properties that, IMO, make resolution beyond 8K not worth the overhead required to deliver it. That said, once we have the equivalent of Dolby PRM contrast/DR, 12 bit color, modest compression rates and larger screens; that's when 8K will show it's inherent advantages. I'd also point out that any tests using the 8K displays of 2020 are limited by the electronics and the current state of mass market display tech at any resolution.

For me, the key point of Phil's post is that acquisition at 8K or higher is best practice for any material that might have value in the future.

For the last 25 years, 90% of my home viewing has been on a projector. Sony ES-600 4K and 108" 16:9 Stewart FireHawk currently. This has made me keenly aware of just how bad over compressed MPEG-2 can look and the delta between HD and UHD/4K. I can only imagine how dramatic the difference might be with higher contrast and more than 8 bits of color - and that's the point. In 1995 I was spreading 480x640 over a 6'x8' (120" diagonal) screen which resulted in roughly 1/3" diameter pixels - essentially 3ppi on axis, 9 ppi by area. 20 years later I'm spreading 2160x3840 over a 108" dia. screen to get 40 ppi on axis and 1600 ppi by area. What might be available to the serious home viewer 20 years from now?

Usual disclaimer: there's a lot of aspects besides resolution that determine picture quality. Screen size to viewing distance ratio, color precision, micro contrast, dynamic range, ambient light, individual eyesight profiles, compression quality, and more all impact perception.

Cheers - #19

 

[M Most]

The LF and Venice are respectively 4.5K and 6K camera sytems, which is part of a 4K/8K ecosystem. Both designed to produce a 4K delivery.

The Venice is popular, but it's pretty much ARRI -> RED -> Sony at the larger rental houses across America. Interestingly RED is more popular in some markets, but that depends on lots of factors I imagine. It's minutiae. I have not seen any instance of the Venice being more popular than RED nor Arri in any market. In fact many major markets don't even have Venice bodies. They are out here for sure though and of course in the biggest regions. And it's a decent camera.

I'll reserve some other numbers because I know how many productions are using the DXL and DXL2 bodies world-wide and you can decide personally if that is purely a Panavision camera or a Panavision/RED concept. In my mind those also count as RED, especially in how they are used in practice on set. Mostly mentioning it, because it should count too..

Sorry, I misinterpreted your remark about "4K/8K" to mean 4K derived from 8K original. That is what I was basing my reply on.

I tend to deal with primarily high end (read: high budget) productions and Venice has rather quickly become a very significant player, far more than I would have anticipated.

And, yes, I also consider the DXL/DXL2 to basically be a Red camera. And while it has become a very significant player as well, Red cameras in general are often used on large productions as a specialty camera, for shots that benefit from its small footprint, overcrank work, plate work (that's one area where the resolution really matters), and mutlicamera arrays. At least that's what I'm seeing on the larger features we deal with. In the much larger general market, that might or might not be the case.

 

[Phil Holland]

No worries Mike. The main battle in my mind is really over. I swear if I'm at NAB this decade trying to convince people 4K is even a thing I'll lose my mind!

It's staggered and dynamic. Much like has been once the format options started to come into the film world in the 50s and beyond. Not everybody shoots IMAX, but there are a healthy amount of people who do and want to achieve that.

Personally if I had my lightning shooting trident and crown of fire I'd like to see digital Super 16mm become relevant again, and I think it will with some of this crazier sensors. Likely around 6K resolution there. Even though that's resolving way more than Super 16mm celluloid could do. I want Super 35mm to sort of be the backbone of the industry still and in numbers things around that format size really still are in spite of the revolutionary move to larger format digital cinema, which I personally enjoy the most for what I do.

Kodak long ago proposed a 100mm wide negative and I think due to expense and practicalities surrounding that concept, it never came to be. I don't even think digital really "needs that" either, but I am enjoying the actual choice between format sizes and resolutions now. Hell of a time to be a filmmaker even if you aren't shooting on film.

 

[John David Pope]

Is greater ability to crop in post good for cinematographers when the cropping is being done by someone other than the cinematographer?

Doesn’t it undermine the cinematographer’s vision? The frame is less the frame and instead more becomes only the outer constraint of framing.

I saw a comment on a video forum where someone said their ideal would be if everything were shot on a super-wide lens with all the framing done in post by the Director, and in that the Director or the studio would have many options. Just shoot every scene with a wide shot that is locked down. Then in post cut into the wide as desired for mediums, singles, closeups. Depth of field for days in the acquisition, shallow depth of focus added artificially in post. The lens package would be only a single lens.

 

[Blair S. Paulsen]

When a Director and DP make conscious framing, perspective, DoF, etc choices to enhance the visual storytelling - that is part of the artistry of our medium. That said, under the guidance of the DP, some post reframing can be beneficial. I would note that even with less pixels in the image, producers who lack respect for the Director and DP will reframe anyway. Then you have crappy framing AND funky changes in IQ. Happier now?

On the subject of shooting everything wide and framing in post. Even if you manufacture shallower DoF in a cropped portion of the larger frame, you are still stuck with the perspective of that optic. Not only would that look un-natural, but it could also make your on camera talent look 10-15 pounds heavier. Nothing wrong with that per se, but I'd wager it would not be popular in >90% of situations I'm familiar with. In fact, I've had creatives come to me on DIT days and ask if I can apply the "7% solution" where you scale the horizontal down to 93% or scale the vertical to 107% to make the talent look taller and thinner. Yes, I do a lot of commercials ;-)

Cheers - #19

 

[Karim D. Ghantous]

On the subject of shooting everything wide and framing in post. Even if you manufacture shallower DoF in a cropped portion of the larger frame, you are still stuck with the perspective of that optic.

The perspective of that lens-to-subject distance. ;-) Even if you had huge, big league resolution, the distance of the camera to the subject is very important. As well as the perspective, which is technically where the optical axis is in 3D space, relative to the subject. I should say 4D space, actually, because camera movement is also important. You can't crop your way to a push-in, or a pull-out, etc.

Not only would that look un-natural, but it could also make your on camera talent look 10-15 pounds heavier. Nothing wrong with that per se, but I'd wager it would not be popular in >90% of situations I'm familiar with. In fact, I've had creatives come to me on DIT days and ask if I can apply the "7% solution" where you scale the horizontal down to 93% or scale the vertical to 107% to make the talent look taller and thinner. Yes, I do a lot of commercials ;-)

Cheers - #19

You learn something new every day. I've never heard this before - maybe you're giving away too many industry secrets?

 

[Phil Holland]

So. Now that this has been out for a minute and the tailing graph does reveal everybody felt 8K was "slightly better", I want to push out my quick thoughts because Monday will be real busy for me.

I'm super grateful they looped me in for this test, despite going through vision issues at the time.

Here's my impact results on each shot. Some of this is related to subtle impact due to the nature of the shot itself.

Shot 01 - Dunkirk Close - This particular shot is indeed beautiful. Richly textured cloth with bits of metal which for me were the "tell", but the subtle facial texture was a delight to see in 8K and 4K.

Shot 02 - Dunkirk Wide - The wide frame was interesting, but the specular hits on the helmets and general contrast edge detail was noticeable. Also, in my personal experience wider shots benefit a great deal from larger formats and higher resolution acquisition. I also felt this way with film. I mean it's in the name really. Interesting the film grain itself also was apparent in this shot as a tell.

Shot 03 - Brave - This was a darker scene in the woods from the film. In the comparison much of the branches and hair were observation-ally different. With it being darker and fairly low contrast down there I imagine some of the viewers had a hard time finding nuggest within the image to spot a decent tell.

Shot 04 - A Bug's Life - This one wasn't too hard to discern. The blades of grass were pretty obvious overall, but your mind certainly could play tricks on you. Interestingly I'm grossly familiar with this shot from back in the day and I seem to recall a fair bit of scaling algorithms being explored, which was common in animation and still is. It's not uncommon to soften a render to give it a more film-like appearance. Aliasing = bad, too sharp also likely feels too harsh.

Shot 05 - Tick 'Cave' - This was one of the easiest to discern. On the Tick's left pectoral muscle, the suit material had a play of specular highlights. Arthur's glasses also had some moments for sure.

Shot 06 - Tick 'Spaceship' - This was actually the hardest shot to discern mainly due to the camera operation and lens choice. However, his eyes did indeed provide enough upon close inspection. But this one was a tricky one if you were literally exploring anywhere else in the frame.

Shot 07 - Nature - Pretty obvious on this one as the hair and eye of the moose were the tells. You can also explore the frame to find other objects in focus that revealed it was 8K as well.


The one thing I'd like to find out personally is what the scanning process was on Dunkirk's frame. There's many ways to finish a frame these days as well. Subtle sharpening methods and even blurring. All in bounds in crafting your final look.


I respect that they mentioned is the general signal path here as it has been a pain in the butt getting 8K onto TVs. That gets solved with the 2020 models and the soon to be released single cable HDMI devices that support HDMI 2.1. I've mentioned elsewhere that those first cables were around $300 btw as the spec was just happening, they are now somewhere between $15-$35 bucks now.

Our previous study was done on LCD screens and not OLED and we had about an 85% "I can tell this is 8K" result. This particular test reveals that data in a different way by what appears all people feeling 8K was "slightly better". We chose to have 3 different screens of the same size with the footage matched so viewers could A/B/C compare by turning and also stepping on various marks to produce different viewing distances.

My main and only critique is likely the duration of these shots. People all scan an image differently and it's also something that changes with age as well. I think it would have been nicer to give individuals more time per shot, but I understand why they chose this route as it reflected a typical cut in the modern era. And also to keep people from having the immediate A/B comparison response by showing them in a random order. I suspect the few instances where people rated 4K better than 8K out of the 139 participants was partially due to this method as not everybody has the analytical mind nor short term visual memory to remember, even in just a few seconds or so apart.

The nod to the viewing distance is also nice, but that's also well traveled territory for many. I'd be curious if over time we see people scoot closer or away from higher resolution displays. Generally speaking, as with the work I did with THX way back when, I like content in the 40-60 degree FOV generally speaking. We do already know from testing that people with poorer vision tend to sit closer to the screen however.


The one bold thought I want to convey here though. This graph and whatever the hell this data is, is wrong:
http://www.phfx.com/temp/reduser/phfx_staricoGraph.png

Even from this test as well as my own tests in the last 20 years. It's not correct for 20/30, 20/20, 20/15, or 20/10 vision. Thus is the harm of not during research and testing on your own and just regurgitating data more or less become a game of telephone. Which alludes to why I made my Optimal Viewing Distance tool.


One other thing I'll write before leaving this one to be. The quality of modern "good" UHD 4K and UHD 8K sets is really impressive. Whether it's this LG OLED, Samsung, TCL, Sony, or whoever.

 

[Hrvoje Simic]

Film is still beautiful, be we've met and exceeded film's capabilities on technical merit in every category.

You keep repeating that but it's not true.

Technical factors still superior :


1. Highlight dynamic range

2. Motion rendition



Also, who's "we" ?

 

[Hrvoje Simic]

Is greater ability to crop in post good for cinematographers when the cropping is being done by someone other than the cinematographer?

No.

Doesn’t it undermine the cinematographer’s vision?

Yes.
And cinematography.

Unless previously decided by the cinematographer on set for specific case and framing and not used as side wheels post festum.

 

[Phil Holland]

You keep repeating that but it's not true.

Technical factors still superior:


1. Highlight dynamic range

2. Motion rendition


Also, who's "we" ?

1. Within a single exposure you are incorrect. Kodak and Fuji both disagree with you. Film is more forgiving and how it captures over exposed highlights and renders them, i.e. more forgiving. Especially if we're talking about still photography and exploring uncommon shutter speeds that aren't used in cinematography typically. But within the negative for a single scanned exposure, 13 stops-15 stops. Kodak states 13 stops, I'm stating 15 because you "can do it" if you know how to. Not to mention digital is significantly cleaner and more sensitive, but I digress.

2. You can make this argument, but most won't follow you off this cliff.

We very literally did likely the last true tests in 2018 on this subject and I just don't think it's very interesting to discuss anymore. The fire was much hotter when Dragon 6K hit the scene.


But I'll throw you a real nugget on topics that matter.

The absolute thing you can go after though pertains to OLPFs and various optical artifacts each digital cinema camera has that film does not. And in stressing the digital image you can certainly make it look real bad in ways film won't look bad (vice versa for film actually). Most of us aren't paid to shoot that poorly though. Similarly we can focus on the deficits film has inherently as well, but that gets back to pros and cons. I happily haven't dust busted much in the last decade. Everything else can be emulated really accurately. I mean you haven't noticed any of my digital shots on a few films. In some more recent larger blockbusters too despite the IMDB only listing film as the medium.

Much like I guess 4K not being a thing, this is also not a thing for me 2020 onward. There's stuff digital can get better at, and that I will give you.

But in short, choose whatever medium you want. Shoot film if you want to shoot film. I'm just stating there's little reason to. This wasn't true until last decade, 2010s. And nobody is even attempting to make the argument anymore because of how far things have come.

 

[Hrvoje Simic]

1. Within a single exposure you are incorrect. Kodak and Fuji both disagree with you.

Please show me Kodak and Fuji statements on highlight capture range being superior on digital and if it's well argumented I will consider re-examining my logic with clipped tail lights, street lights, candles, etc. with a toneless hole, continuosly manifesting in digital acquistion.
Unlike film.

But within the negative for a single scanned exposure, 13 stops-15 stops. Kodak states 13 stops, I'm stating 15 because you "can do it" if you know how to. Not to mention digital is significantly cleaner and more sensitive, but I digress.

Not to mention you are talking about scanning DR and not film medium recorded DR.
And scanner is.... ?

2. You can make this argument, but most won't follow you off this cliff.

Number of followers has no effect on basis of my statements.

The absolute thing you can go after though pertains to OLPFs and various optical artifacts each digital cinema camera has that film does not.

OLPF has nothing to do with motion.
Shutter and sensor properties do.


Regarding OLPF negative factors, they could be avoided by:

a) ditching the OLPF and oversampling beyond the lens resolving ability - in which case the aesthetic becomes less appealing on anything than soft glass resolving barely 4K and highlight range gets reduced due to smaller pixels.

b) ditching the OLPF with somewhat higher acqusition resolution and alterantive pixel arrangement (Fuji X trans) - in which case the same potenital limitations under a)

c) same as b) but "Quad Bayer" concept recently implemented by Sony, with specific AA processing - yet to be seen

d) no OLPF, standard Bayer 8K acquisition with pixel and AA processing for 4K master (softer than typical 50% scale) - unlikely due to 8K sexyness

e) 4K+ CCD - unlikley

f) 3x 4K+ CMOS - highly unlikely for S35 and up

Therefore > larger format digital > few more technical limitations compared to film.

 

[Graeme Nattress]

OLPF does pertain to motion though, in that edge sharpness/contrast alters our perception of judder. And that comes even more to bite us as we move to more HDR displays. HDR can look wonderful as it allows us to do away with some of the highlight rolloff and its unwanted effects on the image, but it also further exacerbates motion judder perception.

Ditching the OLPF is not really an option. Unless you only want to shoot very shallow DOF objects with no repeating patterns. Alternative arrangements such as the Fuji are just as susceptible to aliasing, if not more so, and it's only bonkers marketing that make people think they're not. Sony quad bayer is for HDR capture, and it's bound to screw up fine details in the scene, but as cell-phones images go through so much NR and look so bad pixel for pixel, it's probably a reasonable trade off in that situation. I doubt it has enough finesse to go beyond that.

3 chip + beam splitter / Foveon style solutions don't eliminate the need for OLPF, just ensuring that all aliasing is luma aliasing rather than a mix of luma and chroma aliasing. A small amount of chroma aliasing can be well handled as our chroma resolution perception is much lower than our luma resolution perception (which is why bayer in general and 4:2:2 and 4:2:0 and old colour TV with reduced chroma bandwidth on transmission actually work). But they still leave luma aliasing alongside all the other drawbacks of those arrangements.

So what's the answer - it's the same one it always has been - higher resolutions with a matched reduction in OLPF pitch.

Graeme

 

[Hrvoje Simic]

Thanks for adding to discussion, Graeme.

OLPF does pertain to motion though, in that edge sharpness/contrast alters our perception of judder.

Yes, the perception part.
Not affecting motion capture specifically. Temporal recording performance i.e. motion blur fallof/smoothness and rolling shutter pixel offset is unaffected.
I'd say OLPF has lower effect on judder interpretation than display tech.

Ditching the OLPF is not really an option. Unless you only want to shoot very shallow DOF objects with no repeating patterns.

Alternative arrangements such as the Fuji are just as susceptible to aliasing, if not more so, and it's only bonkers marketing that make people think they're not.

I wrote that in the context of oversampling where adequate scaling of 6K > 4K should have minimal issues for normal viewing. I need to do more tests with Fuji.
To be clear, I'm not advocating ditching the OLPF for motion imaging.

Sony quad bayer is for HDR capture, and it's bound to screw up fine details in the scene, but as cell-phones images go through so much NR and look so bad pixel for pixel, it's probably a reasonable trade off in that situation. I doubt it has enough finesse to go beyond that.

Sony has different readout modes for quad bayer, HDR with varying gain is one. Different pixel group readout, full resolution. Makes little to no sense for tiny sensors, but at 36mm size it might bring a legitimate choice between higher acuity and higher DR.

3 chip + beam splitter / Foveon style solutions don't eliminate the need for OLPF, just ensuring that all aliasing is luma aliasing rather than a mix of luma and chroma aliasing.

Understood, I oversimplified.

Some cameras deal with AA in processing.

So what's the answer - it's the same one it always has been - higher resolutions with a matched reduction in OLPF pitch.

It's the answer which comes at a different cost.

Higher resolution = smaller pixel pitch.
Smaller pixel pitch negatively affects saturation performance, stability over exposure range and falloff to extremes.

Higher resolution also brings file size/compression limitations. Higher compression reduces image detail/consistency...and there is the processing requirement part.