> I bet the higher resolution makes it easier, with big*ss GPUs to clean up the raw footage.

Even if you don't do any other post-processing, downscaled video looks better than video captured at sensor resolution. This is because you can implement a steeper low-pass filter (closer to ideal brick-wall filter) in software than you can in actual optics. Hardware lowpass filters tuned to block most stuff above nyquist will also block a lot below it.

A 1080p video captured off a good 1080p sensor will have vastly lower amplitude at, say, 500 line pair vertical than a 1080p video captured on a 4k sensor and lowpassed in software with a good filter.

It was always standard to acquire at higher resolution than being delivered. Hell, they shot 35mm film for SD for decades, 16mm for non-extravagent budgets.

The first use of 4k wasn't for 4k delivery, but for single camera shoot of talking head reframing. Full frame for the wide, cropped in for the MED/CU shots. Larger image also allows for cleaner image stabilization. Allows for more detail on pulling keys in post, etc.

There are lots of benefits from larger image than delivery.

> Hell, they shot 35mm film for SD for decades

I'm not sure this is the case. Low-ISO cinema 35mm film (like Kodak VisionT 50D/250D) far exceeds SD resolution. How do you think Star Wars was ever scanned into 4k? Or 2001: A Space Odyssey? Film captures highlight detail logarithmically as well, while digital sensors still haven't really figured out how to render highlights well. This allows greater shadow detail with a skilled cinematographer with no loss of quality in brightly lit scenes. Theoretically, a single 35mm still can be scanned into 4k with no loss in detail, and higher-resolution scans won't lose quality -- there's just dimishing returns and in the end you might just be getting higher-res grain, but it is a far cry from standard definition.

Even 16mm can do HD scans just fine.

There's a reason Tarantino, Wes Anderson, PTA, Nolan, Sam Raimi, Scorsese, Spielberg, and David Lynch still swear by film (even though many of these directors have dabbled with digital). The idea that film is inherently "low-res" is a really frustrating and common misconception among younger generations unfamiliar with the technology. I find it kind of unbelievable that DPs would shoot 35mm just for an SD scan. That type of use case seems better suited for mini-DV or other contemporary video formats.

Maybe you misinterpretted what I was meaning. What you quoted from my post was meant as saying they took all of the image from a 35mm frame just to shoot a measly SD image. That's like shooting 4k to deliver SD. It's seemingly way overkill.

>That type of use case seems better suited for mini-DV or other contemporary video formats.

You realize that people were shooting film for TV production long long before DV was ever invented right?

> I find it kind of unbelievable that DPs would shoot 35mm just for an SD scan. That type of use case seems better suited for mini-DV or other contemporary video formats

Lots of shows were filmed on 35mm. For example, Star Trek: The Original Series and Star Trek: The Next Generation were filmed on 35mm, so in modern times they've been able to re-scan them and make HD Blu-ray releases. I believe shooting direct to video for anything higher budget than daytime soap operas didn't become commonplace until the mid-90's.

> Hell, they shot 35mm film for SD for decades, 16mm for non-extravagent budgets.

This probably was also to deal with period TV cameras. Video cameras of that time couldn’t deal with bright lights and had terrible dynamic range. CCDs improved this a lot. Direct to video productions looked cheap.

> Hell, they shot 35mm film for SD for decades

They shot 35mm for the big screen.

Primetime TV shows were shot 35mm. Lower budget was shot on 16mm.

They weren't doing for screen size considerations. It's just what they had. You shot 35/16 or you shot video (shudder).

What does low pass filter mean in the context of video (i understand in audio)?

The same thing: a filter that only lets through low frequencies. A blur effect is a spatial low-pass filter. A motion blur effect is a temporal low-pass filter.

What GP is talking about is the antialiasing low-pass filter used when downsampling video to smaller resolutions, or when initially sampling it at the sensor. This is the same exact concept as the antialiasing low-pass filter used when downsampling audio to lower sample rates, or when initially sampling it at the ADC.

Perfect. Thanks!

Downscaling or blurring - these are the same thing if you pretend it’s bandlimited. Which it isn’t of course.

The signal is typically bandlimited by hardware — the optics and sensor. When it's not, you get weird effects, though that's rare because cameras have optical low pass filters to avoid this.

When they don't, like the Canon 5D mark ii when used in video mode (which just threw out most of the sensor's pixels to get to 1080p), you get aliasing producing moire patterns like this:

https://youtu.be/dMHDB5fG6ys?t=218

Also, video is not actually bandlimited so the supposedly perfect filter isn’t actually perfect; you can do a better nonlinear one.

What does a better filter look like?

Selective Gaussian blur or hq2x are examples of image filters.

Font hinting is an example of a nonlinear downscaling though most people don’t think of it that way. It preserves line sharpness, which is a non-bandlimited feature since lines are infinitely sharp = made of infinite frequencies.

Overly sharp linear filters produce artifacts like blockiness and ringing; nonlinear ones would try to avoid that but you could still get more difficult ones like moiré.

This is mostly obsolete with High-DPI screens though; we're at the point where such hacks are no longer needed and we can use close to ideal filters (e.g. lanczos, which is a windowed sinc). macOS no longer does font subpixel AA or strong hinting by default for this reason.

This is good, because it means we can finally treat computer graphics as properly scalable images without introducing "blurriness".

Note that lines were never infinitely sharp; pixels aren't infinitely sharp, they have a shape and the screen construction determines how sharp they look. They are, however, sharper than what you'd get in a nyquist-limited system.

> Also it looks like 4K is their minimum standard now. I bet the higher resolution makes it easier, with big*ss GPUs to clean up the raw footage.

Resolution is far less important than bit color depth. We barely need 4K(Alexa used to be 2K) much less 8K footage for something to look good on a tv.

I would think that capturing at higher resolutions makes it feasible to crop and zoom? In digital production, you always want to work at high resolution for all intermediate stages, then reduce bit depth only during the final output stage.

When I worked in music production a while back, we often had the problem of source material recorded at low amplitude to 16-bit media (e.g. DAT, CDDA, ADAT). While 16-bit is theoretically sufficient, in the real world you need headroom while recording in order to avoid clipping. So we'd often wind up with grungy sound from inadequate bit-depth during capture and intermediate production.

I know recording live music can get hairy, but yikes! How was 16bits not enough? Were all the preamps set to max? Who records at -12db - -6db? Record at a way lower level and boost at the amp stage?

Imagine you're recording highly dynamic music, like Jazz. Clipping is absolutely never OK. You have to leave tons of headroom, there's no other way. So 16 bits is not enough for live capture — especially when the source is dynamic or the recording conditions are chaotic.

The studio I used to work at was one of the first ADAT studios back in the 90s, and they did a lot of rock. One of the ways they got past the limitations of the medium was to compress and limit to tape. You generally don't want to do that unless you have tons of time to set up and get it right (a la Rudy Van Gelder), which was not in most budgets. But it worked well enough and the ADAT recordings out of this studio sounded way better than most.

Mercifully storage has gotten cheaper in the decades since and you really shouldn't capture to 16 bits any more.

How much post-production dynamic range do you really need/want for live jazz recordings? Surely 50 dB would be plenty, and even that would be pretty impractical to listen to except in very quiet environments. Wouldn’t 16 bit digital audio easily give you 50 dB of usable dynamic range and still have like another 50 dB of headroom?

You need 8 bits to do 48dB, so if you want your music to have 96dB of dynamic range and you want your editor to be able to slide things around 48dB without perceptible loss of quality, you need to record at 24 bits of depth.

If you record at 16 bits of depth you are pretty much automatically forcing your engineer to choose between quality and range when they are normalizing the recording.

Listening to music with 96 dB of dynamic range is extreme though, right?

You seem quite determined to leave inadequate margin for error. :(

The most important rule of audio recording is "always be ready to roll tape". What the engineer does is of little consequence compared to the talent. Don't ever fuck up a keeper take.

Recording with some extra bits means that even if your gain staging was bad, you just get some extra uncorrelated noise, not correlated quantization distortion.

As for whether this stuff is perceptible, there are two issues at play. First, differences which are not perceptible during production can result in a perceptible difference in the final product. This is easiest to understand with a visual analogy: if you have inadequate resolution of an image during processing, the end result may be slightly defocused and smeared.

Second, as I described elsethread, quantization distortion is a particularly pernicious form of degradation, poisonous in small amounts. https://www.youtube.com/watch?v=8KkS4NSv-0M It is hard to listen for it during production, so we defend against it by giving ourselves lots and lots of headroom.

I get it. I’m asking what dynamic range you need for the completed production, and how much additional headroom you need when recording. I don’t think you normally would need 96 dB for the completed production, because that would be very impractical to listen to.

16 bit certainly suffices as a delivery format. Many listening environments provide far less dynamic range (e.g. cars) and environments where the lowest level details aren't buried in the ambient noise floor are going to be few and far between. Even though it is possible to encode information below the noise floor in practice information that far down isn't going to be perceptible. (A practical definition of "perceptible" might be "able to be distinguished in an ABX test by a trained listener".)

As for how much safety margin is "enough" during initial capture... We're looking at a stark choice between either 2 or 3 bytes per sample. 2 bytes seems to be borderline, while 3 bytes seems to be plenty. It would be hard to quantify a safety margin more precisely than that.

Say that I need a 300x400 image for a webpage. Does it make sense to set my camera to take 300x400 pictures? What happens when we crop and postprocess?

If we want the best quality end product, free from low-res artifacting, we need to preserve an excess of resolution all the way through production until the final export. Exactly how headroom much is enough? Enough that artifacts arising from insufficient resolution are not perceivable in the end product.

Yes, I understand the reasons to have enough dynamic range and enough headroom. My question is about what specifically is enough.

I think 16-bit quantization noise level is usually lower than the noise of all the analog hardware before the ADC, but 24-bit quantization noise seems to be lower than even the internal noise of audio ADCs so there seems to be no reason to go higher.

I had the privilege of visiting the van Gelder studio in high school. I asked him about analog vs digital and surprisingly he said he was all in on digital and spending his days remastering his old work with modern tools.

My understanding is that 24-bit source material (the standard when I was last recording which was several years ago) provides enough dynamic range that you can bring up the desired volume without overly muddying quiet sections with noise, especially when mixing multiple sources where that noise adds up quickly. My own personal experience recording at 16 bits was… “I must play louder and overstate my song’s understatements” because I never could get a 16 bit output that didn’t sound like garbage. Granted I’ve never been more than a novice at recording and I’m certain there’s more I could learn to work with the lower bit rate but… why?

Also: this got me wondering about how much the “loudness wars” and “brick wall” music products overlap (in time) with the almost total transition to digital recording. It also got me wondering how much the more gradual digital transition affected musical style and taste. If analog recording captured a much more flexible dynamic range, it’s easy to imagine a lot of other recording artists similarly deciding “fuck it play it louder”. And it’s easy to imagine that culturally influencing louder recording targets generally.

The resolution bottlenecks of early digital as described in this thread are not related to the loudness wars of the aughts. If you were doing big budget recordings back then you were either doing 2" analog or you had enough setup time to set gain levels carefully and "fill up the meters" for your digital recorder.

Recording close-mic'd drums to analog tape is actually a useful technique when maxing out levels because analog tape saturation can be aesthetically superior to digital peak limiting for shaving drum peaks — analog tape saturates high frequencies first. It wasn't uncommon to either record basic tracks to analog then transfer to Pro Tools, or even to track to Pro Tools then bounce the drums out to analog tape and back.

For the record, there were loudness wars back in the vinyl days, since it was seen as important to have your 45 rpm single "compete" with others in the jukebox. It wasn't something new in the aughts.

> If analog recording captured a much more flexible dynamic range

It doesn't, especially not in comparison to modern digital. Dynamic range is not a strong point of analog because of tape hiss.

The issue with low res digital is that truncation distortion is non-harmonic (not an integer multiple of the input frequency) and so is disproportionately aesthetically damaging even in small quantities — in other words, digital grunge sounds worse than tape hiss.

It is trivial for modern hardware to work at 24bit. 16bit hardware is probably old, overly value engineered, or deliberatly operating below its capacity. These are all more plausable reasons for it to sound bad.

> capturing at higher resolutions makes it feasible to crop and zoom?

The more you capture, the more you have to play with at later production stages.

And pay for. Larger data size across the chain isn’t costless.

Overall, data storage is not a large fraction of the cost of a TV series or movie production. If your assets grow too big and you don't plan on accessing them, you can store them in cold cloud, or archive-grade BD and tapes.

I wish we would move to higher bit color for playback as well. Dark scenes very easily result in banding (especially with streaming compression). With video shot with cameras there is usually noise and grain (real or artificial) to help break it up some, but dark scenes in video games almost always look like garbage to me. Some games, like Playdead’s Inside get around this by, again, add a ton of noise to the image, but that doesn’t always work stylistically.

Also compression. ARRI 2.5k cameras made decent images as they were lot less compressed as some 4K cameras at the time.

Netflix streams 4K but often doesn’t look so great over internet compared to any cinema in my area which projects only 2K on a huge screen but from bigger files.

> Resolution is far less important than bit color depth.

I discovered this playing with my scanner, looking for the best settings. It surprised me.

Sure is nice to crop in though.

If you’re shooting 4K, you have to be calculated with your lens configuration to actually get that resolution at the sensor. After that kind of effort, cropping your shot in post is going to be a silly step to take.

You don’t have to crop in significantly, but just trim edges/etc.

When you used actual film it was long lengths of the stuff, hence footage. Lots of words have weird no longer relevant sources. Like “logs”.

Not what I expected at all:

log (n.2) "record of observations, readings, etc.," 1842, sailor's shortening of log-book "daily record of a ship's speed, progress, etc." (1670s), from log (n.1). The book so called because a wooden float at the end of a line was cast out to measure a ship's speed. General sense by 1913.*

So it does have a relationship to the wooden log - it was originally a series of measurements from a floating bit of wood.

https://www.etymonline.com/word/log

On set, depending on the size of crew, part of the job for an AC or 1st AC is to make a camera report. Each shot is logged to note things like which lens was used, what filters were used, the height of the camera from the floor, the distance of the film plane to the focus point, the frame rate, aperture, ISO, etc. All of this info can then be used if they needed to do a pick up later to put the camera in the same setup. Also, it can be used in post production so that the camera can be recreated in 3D to have the special f/x match.

Also, log can refer to the gamma curve which affects how roll-offs in highlights/shadows are saved (rough description).

"Log" is still used daily in production work.

This may not be universal, but when I was a 2nd AC/Loader - the 1st AC didn’t log anything. Always 2nd AC along with keeping the slate.

"depending on size of crew"

To get to 2nd AC, you had a good sized crew

I suppose, but we’re talking about Netflix here. DP, Operator, 1st AC, 2nd AC, and a loader is pretty standard on even sub-10 mil budgets (which is nothing).

Debating the size of the crew is fun and all, and yes, traditionally for a full crew, this is the purview of the 2nd AC. As the crew gets smaller, those duties roll up to the next person.

The fact that a "log" is still a thing in today's productions is not dependant on which crew member is doing it. In fact, with all of the digital f/x, they are crucial.

Absolutely. No disagreement here.

See also the unit of speed for ships and aircraft, knots, after the knots tied in the line attached to the log.

"mark twain" means 12 ft. Samuel Clemens adopted it as a pen name.

https://www.mvd.usace.army.mil/Portals/52/docs/regional_floo...

Yeah and knots from the literal knots in that rope attached to that literal log. Not sure what you expected though, look at the etymology of any word and you get back to something concrete and tangible eventually (with rare exceptions for a few feelings and such.) We haven't been abstract that long, so most of it is metaphor or analogy.

It's becoming less and less of a thing as film is used less and less, but some of the older NLEs would allow you to view your time as feet+frames and allowing to specify between 35mm and 16mm.

Film was also sold by the foot. You didn't by film by 15 minutes. You bought film by 1000' rolls (35mm).

Or a fraction of a second at 10,000fps, where 400’ would be spooled through in under 3s (16mm)

As ingenuous as our mechanical past was, I am so happy to have lived through the transition to 100% solid state everything.

> Film was also sold by the foot. You didn't by film by 15 minutes. You bought film by 1000' rolls (35mm).

But that just makes sense doesn't it? It's only 15 minutes' worth if you run it at 1000'/15m.

Yes, I was just mentioning that for those that might not be aware.

Developing your film was also charged by the foot. The soup didn't care about your framerate. Just how much was need to process.

Thank you. Now I understand why it’s "métrage” in French, and also why a “long métrage” means a long movie!

> Also it looks like 4K is their minimum standard now. I bet the higher resolution makes it easier, with big*ss GPUs to clean up the raw footage.

I assume it's their minimum standard for Netflix Originals as they want more content in 4K rather than making it easier for them, particularly as 4K is an up-charge subscription.

From what I've read on their documentation [1], it's mostly for future-proofing their own content.

[1] https://partnerhelp.netflixstudios.com/hc/en-us/articles/150...

--- StartQuote ---

Why does Netflix require UHD on Netflix Originals?

Answer: In 2014, Netflix made the decision to begin shooting and delivering all Originals in UHD. This decision was made for several reasons, the most of important of which is to future-proof our content. UHD is here, and adoption of UHD in the home is increasing. In just a few years, it will be harder to find an HD television than a UHD television. For this reason, we feel it only makes sense to shoot natively in the format that most of our customers will see for years to come. The experience for customers viewing HD is still fantastic. Our encoding pipeline takes a UHD master and produces beautiful HD (and SD) streams for all those customers who are viewing on smaller or older displays.

--- EndQuote ---

Well that and 4K60 is the new standard definition. It’s been 4 years since flagship phones that could record 4K60 were released and most of those devices are now discontinued and obsolete. It’d be hilarious if studios used their fancy optics with cameras that have a lower res than something you can get for $80 on eBay

Eh, just because it has the pixels doesn't mean it has the resolution. Often 1080p content from high quality systems still looks more detailed than 4k content from smartphones, even fairly modern smartphones, as the pixels can only record detail that gets through the optics and isn't marred by noise. With phones having relatively low quality lenses (for the size of the pixels they must resolve) and smaller sensors, this can quickly become an issue. With that said, it is great to see 4k becoming standard for productions!

Regarding "footage", notice that you also refer to them as "film-makers." There really isn't film involved. And Netflix is named after the term "flick" which refers to the flickery effect you get with movies (especially in the old days), which hopefully isn't there anymore.

Netflix farms out all this post work. That's why they have these delivery specs. You follow the white papers and then they run it through their automated systems that do QC (followed by people verifying). Then they run it through an auto pipeline to create final streaming files and archives but they don't do any further manipulation of the image. Once you start doing one Netflix job they return to your post facility consistently.

> (Footage? WTF? Video measured by distance? Maybe we should say "frameage" or "pix.")

I'll start saying one of those when my phone dialer icon stops looking like a Western Electric Model 500 handset.

https://en.wikipedia.org/wiki/Model_500_telephone

> The stuff with the red N on the poster -- some of it is REALLY good

An overwhelming majority of it is total garbage. It's such a bad ratio that I treat the red N as a negative signal when I'm choosing something.

The McDonald's of TV show production. With very, very few exceptions (Midnight Mass, maybe?) I at least know what I'm getting into: designed-by-committee, very very "safe", out of place (most likely contracted) references to other Netflix shows to try to cross-pollinate viewers... not exactly a recipe for success.

Similar to how Amazon Originals now seem synonymous with "we cut the writer's outline from 12 episodes to 6, enjoy the rushed character and plot development!"

> designed-by-committee, very very "safe"

This is why I was so surprised by some of the background references Inside Job was able to drop in its first season. (Do they still call them seasons when they release all at once? Either way it's really good)

Agreed, except your last example. Most shows are much too long now, and extend way past after they have exhausted what little they had to say.

Yea, I can agree to Midnight Mass. Best quality was the plot of “X discovered by Y who thought it was Z”. But now that I think about it, the cinematography was pretty good. I know some people were put off by the long monologues, so it wasn’t without some odd choices.

I thought Dark was excellent and not at all like you described.

DAЯK was excellent, but as the saying goes, it's the exception that proves the rule.

Agreed. Netflix has so much pulp now it's not even funny. It used to produce much higher quality content. After other copyright holders started pulling out they seem to have started focusing on quantity.

The streaming services are in strong competition for original content. Sometimes buying exclusive rights to completed work off the shelf. This content seems to get labeled as an "XYZ original," be it Apple, Netflix or something else.

Do you differentiate between content greenlit and produced by the streaming services from those purchased? I'm curious because they are two different capabilities and it seems each service is attempting some amount of both.

Yes, there is a clear quality difference between the stuff they acquire/license (some of which is quite good) and the utter dreck they produce themselves.

I wasn’t aware that „dreck“ is being used in English. TIL

What other language(s) is that word used in?

It's a moderately rare word in everyday spoken English.

“Dreck” is a German word.

I didn't know that! I got it from my mum who presumably got it from her British mum.

It's traveled fairly broadly on account of having been adopted into Yiddish - I picked it up as a very young child from an issue of MAD Magazine, and was somewhat surprised to find the New Englander side of my family very familiar with it and only bemused at the happenstance of hearing it from a five-year-old.

There is also some original content which is quite good.

Yeah they greenlight almost anything. It means they take a lot of risks which can produce some really creative and unexpected treasures but also a ton of crap. It's basically the VC model applied to content creation. One blockbuster is worth ten flops.

Most stuff on Github is garbage as well, but all the best projects are there.

This analogy doesn’t hold since the best projects aren’t on Netflix. HBO’s TV catalog destroys Netflix from a quality and critical acclaim perspective. I can’t even think of a Netflix show I’d consider to be HBO-level quality.

> total garbage

Which historically correlates with commercial success.

It's why Netflix is more than happy to green-light awful shows such as Emily in Paris.

Definitely.

McDonalds grosses a helluva lot more than Wolfgang Puck -- $19.21 billion vs $57.50 million.

Yeah, footage.