Resolution vs Pixel Count (by Steve Yedlin, ASC)

Indeed, I definitely don't need that resolution, but that Alexa 65 test was SWEET. And to think an a9 storing RAW video could be just as good in terms of resolution, and not far behind in terms of dynamic range...
 
Yep, insanely good. Although not really new to anybody who is not drinking the 4K coo-laid. But it's nice seeing somebody actually putting together such a test to show it. I have not watched part 2 yet though. But I will when I get some time.

I'm already wondering what arguments will the 4k crowd use to dismiss Yedlin's findings.
 
Yep, insanely good. Although not really new to anybody who is not drinking the 4K coo-laid. But it's nice seeing somebody actually putting together such a test to show it. I have not watched part 2 yet though. But I will when I get some time.

I'm already wondering what arguments will the 4k crowd use to dismiss Yedlin's findings.

All in all a great video series by Yedlin. However, there are some stuff that is missing (from part 2), like that film has more of a temporal resolution (almost like pixel-shifting cameras) due to the fact that the film wiggles (and every frame is a bit different with grain), so it will look better resolution wise when rolling (but that is also harder to "define" then). I also prefer boring resolution charts instead of pixel peeping an actors eye (that can be just a tad out of focus because of difference in acting, motion blur, etc.). It is great for "real world stuff" with actors so definitely useful, but not great for resolution comparison. Also, I don't take a film being scanned at 6K or 11K as it will have that resolution (and most people I worked with don't take the K-value as the "true resolution" of the scan, but just simply as safety for over-scanning since film simply isn't "pixels" and thus will not have the built in "perfect average" that digital camera sensors have (if they hit a 4K sensor and records in 4K), and also as I said earlier that film has resolution of many frames (temporal) so setting highest K according to a still frame is misguided. Scanning film is almost like the same thing as when a digital camera over-scans 8K sensor to store in 4K...

Also, don't get me wrong. I personally have stayed away from 4K hype on TV's cause it has not been worth it (I've chosen more expensive OLEDs Full HD instead before), but for my projector that is gonna be 135" and I will view 10-12-ish Feet away from I will have a semi-4K (Full HD but with pixel shifting which at least makes it a little bit better than Full HD, but it was HDR that was more important). I want 4K if I have a large angle of viewing (which I try to have), but resolution isn't as important to me as deep blacks and color...
 
Last edited:
Biggest advantage to 4k and higher resolutions is not the increase in apparent detail, but having enough pixels to push common digital artifacts like aliasing and moire further out of visual range. The limiting factor should not be the pixel resolution of the display or camera, but the optical resolution of lenses and the human eye. Images look smoother with more organic edge transitions. If the comparison is to 35mm film, the best resolution I ever saw out of a projected 35mm SMPTE test film was about 850 lpph, or about the same as the best alias free 1080p.

Pixel resolution conforms to Nyquist sampling limits for spatial resolution. You want twice the pixel resolution of the highest spatial resolution you want to resolve to avoid artifacts.

If you miss critical focus on a 4k camera by just a half pixel radius of COC, you are in fact looking at 2k.
 
Last edited:
My take from the video is that it reinforces two things I already knew:
* In terms of camera resolution, 4k is plenty enough. I want higher frame rates at 4k, but definitely not 6k or more, that would just make my workflow heavier with no visible gain.
* In terms of viewing screens, unless I'm getting a really HUGE one, 1080p is enough. Last year I bought a 75" 1080p TV that I see from a 3.3m distance, and it has more resolution than I can see. For 100" I'd like to go 4k, but that will take at least 5 years to become affordable.
 
Where were you guys when I was getting downvoted into oblivion on the Netflix thread?

You were not the only one getting the childish downvotes in that thread. But the whole thing became a joke once it came out that Netflix didn't stop Scorcese using an Alexa in his Netflix project.

But if you think about it, most people with an investment in 4K won't admit that it doesn't really make that much of a difference. That attitude is to be expected.
 
Wow! That was super informative. I've always said I would prefer better pixels not more of them. We have seen this race for more pixels for a long time now starting with digital stills cameras. At least Sony did the unmentionable when they released the A7s with less pixels and the ability to have larger photo sites to gather more light. Hopefully this type of innovation will make other manufactures focus on better use not more.
 
Let me go against the grain here.

Pretty much everyone knew that the returns in terms of resolution were decreasing as the world moved from 480 to 1080 to 2160 to ... whatever. His statement is hardly revolutionary. In still photography, a higher pixel count allows for the various crops and magnifications that is usually, but not always, absent from the moving pictures.

An 11K scan might be quite useful for the 8K home theater whenever that 8K home theater is technically and financially feasible. At this point, comparing the 11K scan to a 3.2K image is not different than comparing GH2 to Alexa during that Zacuto shootout on a 1080p/2K timeline and screen. No current display can handle 11K (aside of maybe some in the military) and no one knows when that might become available for the public at large.

Personally, I can tell 4K from 1080p when right in front of the screen. I also prefer a higher frame rate when watching sporting events. Most film goers, on the other hand, were not pleased with Peter Jackson's 48p use for a fiction/fantasy project nor were they enthralled with the visuals from "Long Halftime Walk". Technically, the F65 120p visuals were "better" but they weren't better for the project at hand. Nonetheless, when the future development of, at least, the broadcast and streaming is discussed, 8K at 120p is the goal.

Overall, this seems - and, to me, this is the most negative part - is that he is mostly dismissing technology and propping up the cinematographer. Which is what he is. In this sense, he is defending his own turf. But others, from different disciplines, may counter that by saying that content is still the king.
 
he is mostly dismissing technology

He absolutely is not dismissing technology. He wrote his own image-processing program. He talks far more about the tools in the colorist's suite than on the cinematographer's set. However, both of their tool sets are, by definition, "technology."

He's not arguing against technology. He's arguing against technology that makes no practical difference. "Once something has completely disappeared, it can't disappear MORE as you move ever farther away. 'Less than zero visibility' might be evocative rhetoric, but in the literal sense, there is no such thing as negative visibility; zero is the minimum." --- http://www.yedlin.net/141014.html
 
But that's also a straw argument because no one is claiming that they can see a 4K vs. 2K difference on a small screen from far away. To me, his points are very hodge-podge. He's comparing a 3.2 and a slew of 4K cameras with an 11K camera/film scan during the era when not many can even see a "true" 4K image on an upclose 4K screen (owners of the Blu Ray 4K players excepted). And that's tilting at the windmills. Why do you need to film in color when all everyone has is a 20" B&W set with rabbit ears? Then, two decades later, when a 27" TV getting a clean cable signal becomes the norm, no one wants to watch B&W. In this instance, he is much like many DP's of the 00's who were arguing against the digital for a multitude of reasons, half of them being quite cogent and quite obvious, but still losing the main point that the digital was here to stay and soon outperform film on the criteria beside the workflow.

Anyway, I actually I think it's great that a working DP of note is engaging the unwashed masses for free. I only wish his videos didn't buffer so heavily.
 
a small screen from far away

This comment makes me wonder if you watched the video. He's not talking about watching on a small screen, from far way. In fact, he said just the opposite.

His demo simulates watching 4K twice as close as you are sitting from your computer or TV screen while watching his video.

I only wish his videos didn't buffer so heavily.

Direct links, so you can download the videos and then watch:
https://theasc-assets.s3.amazonaws.com/YedlinResolutionDemoPt1.m4v
https://theasc-assets.s3.amazonaws.com/YedlinResolutionDemoPt2.m4v
 
I agree with a lot of his points, but he seems to keep trying to make a point that 2K is more than anyone needs while showing us 4k images that obviously have more detail. It is an important note that the 2k-4k increase isn't nearly as big as SD-HD was, but in my opinion that doesn't make it irrelevant. I think he may have a point for a small theater, but that hardly seems like the experience to be aspiring to with people sitting inches away from 5K monitors all day long these days. I think partly his point gets blurred with his surprise at how soft the 6K film scan is compared to 3K digital.
 
Last edited:
^ Well, I think he has two points that get somehow mixed at times:
* 2k is all you need
* "6k or more" does not necessarily have more detail than good 4k

This EPIC-W vs a7R II stills comparison here shows that RED cameras use a super strong OLPF, which makes their mpix count translate pretty badly with actual resolution (they think it's the OLPF, and I think so too, but it could also be something else, like a sub-par debayering algorithm). And a film scan can't generate information that's not already in the film, so going for more k there hits diminishing returns relatively early.

Remember those resolution tests with 1080p cameras where we saw that some were a lot sharper than others, even though they were all 1080p? People tend to forget that this also happens at higher resolutions.
 
HD is all i need,mm i guess my eyesight went when i moved to 4k because i saw a BIG difference.As i see it a lot of people with 4k kit love to lower the sharpness as low as possible making films look HD.
 
Last edited:
HD is all i need,mm i guess my eyesight went when i moved to 4k because i saw a BIG difference.

Confirmation bias is a wonderful thing. I assure you that if I spent the money and the time to upgrade my HT to 4k I would certainly see the difference too. At least, I'd be adamant about it with my wife. :cheesy:

Sony says that in a theater, all you really need is 2k. Unless you're sitting close. The argument is really about how close is close enough. This is why IMAX works -- all the seats are close enough that you can actually take advantage of the higher resolution.

That said, I just got back from seeing Dunkirk at my local 70x52 foot IMAX screen. Apologies to Nolan and Hoytema, but I thought Dunkirk's visual quality just a little short of what I saw from Mendes and Deakins' Skyfall, shot at 2.8K with Arri cameras and projected at 2k at my local theater, where I tend to sit about the middle of the screen height, so fairly far back. Skyfull was sharp, clear, visually stunning. I couldn't ask for more.

From Dunkirk I could and do ask for more -- more things in focus. It doesn't matter what kind of resolution it can carry if 90% of the screen in any given frame is out of focus.

There's a lot more to visual quality than simple resolution. Mendes and Deakins know this. Nolan is more into film than he is into visual quality IMHO. Sacrilege, I know. Yet there it is on screen for all to see.
 
Back
Top