Featured Article

Why did last night’s ‘Game of Thrones’ look so bad? Here comes the science!

Hello darkness, my old friend

Comment

Image Credits: HBO

Last night’s episode of “Game of Thrones” was a wild ride and inarguably one of an epic show’s more epic moments — if you could see it through the dark and the blotchy video. It turns out even one of the most expensive and meticulously produced shows in history can fall prey to the scourge of low quality streaming and bad TV settings.

The good news is this episode is going to look amazing on Blu-ray or potentially in future, better streams and downloads. The bad news is that millions of people already had to see it in a way its creators surely lament. You deserve to know why this was the case. I’ll be simplifying a bit here because this topic is immensely complex, but here’s what you should know.

(By the way, I can’t entirely avoid spoilers, but I’ll try to stay away from anything significant in words or images.)

It was clear from the opening shots in last night’s episode, “The Longest Night,” that this was going to be a dark one. The army of the dead faces off against the allied living forces in the darkness, made darker by a bespoke storm brought in by, shall we say, a Mr. N.K., to further demoralize the good guys.

If you squint you can just make out the largest army ever assembled

Thematically and cinematographically, setting this chaotic, sprawling battle at night is a powerful creative choice and a valid one, and I don’t question the showrunners, director, and so on for it. But technically speaking, setting this battle at night, and in fog, is just about the absolute worst case scenario for the medium this show is native to: streaming home video. Here’s why.

Compression factor

Video has to be compressed in order to be sent efficiently over the internet, and although we’ve made enormous strides in video compression and the bandwidth available to most homes, there are still fundamental limits.

The master video that HBO put together from the actual footage, FX, and color work that goes into making a piece of modern media would be huge: hundreds of gigabytes if not terabytes. That’s because the master has to include all the information on every pixel in every frame, no exceptions.

Imagine if you tried to “stream” a terabyte-sized TV episode. You’d have to be able to download upwards of 200 megabytes per second for the full 80 minutes of this one. Few people in the world have that kind of connection — it would basically never stop buffering. Even 20 megabytes per second is asking too much by a long shot. 2 is doable — slightly under the 25 megabit speed (that’s bits… divide by 8 to get bytes) we use to define broadband download speeds.

So how do you turn a large file into a small one? Compression — we’ve been doing it for a long time, and video, though different from other types of data in some ways, is still just a bunch of zeroes and ones. In fact it’s especially susceptible to strong compression because of how one video frame is usually very similar to the last and the next one. There are all kinds of shortcuts you can take that reduce the file size immensely without noticeably impacting the quality of the video. These compression and decompression techniques fit into a system called a “codec.”

But there are exceptions to that, and one of them has to do with how compression handles color and brightness. Basically, when the image is very dark, it can’t display color very well.

The color of winter

Think about it like this: There are only so many ways to describe colors in a few words. If you have one word you can say red, or maybe carmine or vermilion depending on your interlocutor’s vocabulary. But if you have two words you can say dark red, darker red, reddish black, and so on. The codec has a limited vocabulary as well, though its “words” are the numbers of bits it can use to describe a pixel.

This lets it succinctly describe a huge array of colors with very little data by saying, this pixel has this bit value of color, this much brightness, and so on. (I didn’t originally want to get into this, but this is what people are talking about when they say bit depth, or even “highest quality pixels.”)

But this also means that there are only so many gradations of color and brightness it can show. Going from a very dark grey to a slightly lighter grey, it might be able to pick 5 intermediate shades. That’s perfectly fine if it’s just on the hem of a dress in the corner of the image. But what if the whole image is limited to that small selection of shades?

Then you get what we see last night. See how Jon (I think) is made up almost entirely of only a handful of different colors (brightnesses of a similar color, really) in with big obvious borders between them?

This issue is called “banding,” and it’s hard not to notice once you see how it works. Images on video can be incredibly detailed, but places where there are subtle changes in color — often a clear sky or some other large but mild gradient — will render in large stripes as the codec goes from “darkest dark blue” to “darker dark blue” to “dark blue,” with no “medium darker dark blue” in between.

Check out this image.

Above is a smooth gradient encoded with high color depth. Below that is the same gradient encoded with lossy JPEG encoding — different from what HBO used, obviously, but you get the idea.

Banding has plagued streaming video forever, and it’s hard to avoid even in major productions — it’s just a side effect of representing color digitally. It’s especially distracting because obviously our eyes don’t have that limitation. A high-definition screen may actually show more detail than your eyes can discern from couch distance, but color issues? Our visual systems flag them like crazy. You can minimize it in various ways, but it’s always going to be there, until the point when we have as many shades of grey as we have pixels on the screen.

So back to last night’s episode. Practically the entire show took place at night, which removes about 3/4 of the codec’s brightness-color combos right there. It also wasn’t a particularly colorful episode, a directorial or photographic choice that highlighted things like flames and blood, but further limited the ability to digitally represent what was on screen.

It wouldn’t be too bad if the background was black and people were lit well so they popped out, though. The last straw was the introduction of the cloud, fog, or blizzard, whatever you want to call it. This kept the brightness of the background just high enough that the codec had to represent it with one of its handful of dark greys, and the subtle movements of fog and smoke came out as blotchy messes (often called “compression artifacts” as well) as the compression desperately tried to pick what shade was best for a group of pixels.

Just brightening it doesn’t fix things, either — because the detail is already crushed into a narrow range of values, you just get a bandy image that never gets completely black, making it look washed out, as you see here:

(Anyway, the darkness is a stylistic choice. You may not agree with it, but that’s how it’s supposed to look and messing with it beyond making the darkest details visible could be counterproductive.)

Now, it should be said that compression doesn’t have to be this bad. For one thing, the more data it is allowed to use, the more gradations it can describe, and the less severe the banding. It’s also possible (though I’m not sure where it’s actually done) to repurpose the rest of the codec’s “vocabulary” to describe a scene where its other color options are limited. That way the full bandwidth can be used to describe a nearly monochromatic scene even though strictly speaking it should be only using a fraction of it.

But neither of these are likely an option for HBO: Increasing the bandwidth of the stream is costly, since this is being sent out to tens of millions of people — a bitrate increase big enough to change the quality would also massively swell their data costs. When you’re distributing to that many people, that also introduces the risk of hated buffering or errors in playback, which are obviously a big no-no. It’s even possible that HBO lowered the bitrate because of network limitations — “Game of Thrones” really is stretching the limits of digital distribution in some ways.

And using an exotic codec might not be possible because only commonly used commercial ones are really capable of being applied at scale. Kind of like how we try to use standard parts for cars and computers.

This episode almost certainly looked fantastic in the mastering room and FX studios, where they not only had carefully calibrated monitors with which to view it but also were working with brighter footage (it would be darkened to taste by the colorist later) and less or no compression. They might not even have seen the “final” version that fans “enjoyed.”

We’ll see the better copy eventually, but in the meantime the choice of darkness, fog, and furious action meant the episode was going to be a muddy, glitchy mess on home TVs.

And while we’re on the topic…

You mean my TV isn’t the problem?

Couple watching TV on their couch.

Well… to be honest, it might be that too. What I can tell you is that simply having a “better” TV by specs, such as 4K or a higher refresh rate or whatever, would make almost no difference in this case. Even built-in de-noising and de-banding algorithms would be hard pressed to make sense of “The Long Night.” And one of the best new display technologies, OLED, might even make it look worse! Its “true blacks” are much darker than an LCD’s backlit blacks, so the jump to the darkest grey could appear more jarring.

That said, it’s certainly possible that your TV is also set up poorly. Those of us sensitive to this kind of thing spend forever fiddling with settings and getting everything just right for exactly this kind of situation. There are dozens of us! And this is our hour.

Usually “calibration” is actually a pretty simple process of making sure your TV isn’t on the absolute worst settings, which unfortunately many are out of the box. Here’s a very basic three-point guide to “calibrating” your TV:

  1. Turn off anything with a special name in the “picture” or “video” menu, like “TrueMotion,” “Dynamic motion,” “Cinema mode,” any stuff like that. Most of these make things look worse, and so-called “smart” features are often anything but. Especially anything that “smooths” motion — turn those off first and never ever turn them on again. Note: Don’t mess with brightness, gamma, color space, pretty much anything with a number you can change.
  2. Figure out light and color by putting on a good, well-shot movie the way you normally do. While it’s playing, click through any color presets your TV has. These are often things like “natural,” “game,” “cinema,” “calibrated,” and so on, and take effect right away. Some may make the image look too green, or too dark, or whatever. Play around with it and whichever makes it look best, just use that one. You can always change it again later – I myself switch between a lighter and darker scheme depending on time of day and content.
  3. Don’t worry about HDR, dynamic lighting, and all that stuff for now. There’s a lot of hype about these technologies and they are still in their infancy. Few will work out of the box and the gains may or may not be worth it. The truth is a well shot movie from the ’60s or ’70s can look just as good today as a “high dynamic range” show shot on the latest 8K digital cinema rig. Just focus on making sure the image isn’t being actively interfered with by your TV and you’ll be fine.

Unfortunately none of these things will make “The Long Night” look any better until HBO releases a new version of it. Those ugly bands and artifacts are baked right in. But if you have to blame anyone, blame the streaming infrastructure that wasn’t prepared for a show taking risks in its presentation, risks I would characterize as bold and well executed, unlike the writing in the show lately. Oops, sorry, couldn’t help myself.

If you really want to experience this show the way it was intended, the fanciest TV in the world wouldn’t have helped last night, though when the Blu-ray comes out you’ll be in for a treat. But here’s hoping the next big battle takes place in broad daylight.

More TechCrunch

Tags

The spam reached Bluesky by first crossing over two other decentralized networks: Mastodon and Nostr.

The ‘vote Trump’ spam that hit Bluesky in May came from decentralized rival Nostr

Welcome to TechCrunch Fintech! This week, we’re looking at the continued fallout from Synapse’s bankruptcy, how Layer wants to disrupt SMB accounting, and much more! To get a roundup of…

There’s a real appetite for a fintech alternative to QuickBooks

The company is hoping to produce electricity at $13 per megawatt hour, which would be more than 50% cheaper than traditional onshore wind.

Bill Gates-backed wind startup AirLoom is raising $12M, filings reveal

Generative AI makes stuff up. It can be biased. Sometimes it spits out toxic text. So can it be “safe”? Rick Caccia, the CEO of WitnessAI, believes it can. “Securing…

WitnessAI is building guardrails for generative AI models

It’s not often that you hear about a seed round above $10 million. H, a startup based in Paris and previously known as Holistic AI, has announced a $220 million…

French AI startup H raises $220M seed round

Hey there, Series A to B startups with $35 million or less in funding — we’ve got an exciting opportunity that’s tailor-made for your growth journey! If you’re looking to…

Boost your startup’s growth with a ScaleUp package at TC Disrupt 2024

TikTok is pulling out all the stops to prevent its impending ban in the United States. Aside from initiating legal challenges against the government, that means shaping up its public…

As a U.S. ban looms, TikTok announces a $1M program for socially driven creators

Microsoft wants to put its Copilot everywhere. It’s only a matter of time before Microsoft renames its annual Build developer conference to Microsoft Copilot. Hopefully, some of those upcoming events…

Microsoft’s Power Automate no-code platform adds AI flows

Build is Microsoft’s largest developer conference and of course, it’s all about AI this year. So it’s no surprise that GitHub’s Copilot, GitHub’s “AI pair programming tool,” is taking center…

GitHub Copilot gets extensions

Microsoft wants to make its brand of generative AI more useful for teams — specifically teams across corporations and large enterprise organizations. This morning at its annual Build dev conference,…

Microsoft intros a Copilot for teams

Microsoft’s big focus at this year’s Build conference is generative AI. And to that end, the tech giant announced a series of updates to its platforms for building generative AI-powered…

Microsoft upgrades its AI app-building platforms

The U.K.’s data protection watchdog has closed an almost year-long investigation of Snap’s AI chatbot, My AI — saying it’s satisfied the social media firm has addressed concerns about risks…

UK data protection watchdog ends privacy probe of Snap’s GenAI chatbot, but warns industry

U.S. cell carrier Patriot Mobile experienced a data breach that included subscribers’ personal information, including full names, email addresses, home ZIP codes and account PINs, TechCrunch has learned. Patriot Mobile,…

Conservative cell carrier Patriot Mobile hit by data breach

It’s been three years since Spotify acquired live audio startup Betty Labs, and yet the music streaming service isn’t leveraging the technology to its fullest potential — at least not…

Spotify’s ‘Listening Party’ feature falls short of expectations

Alchemist Accelerator has a new pile of AI-forward companies demoing their wares today, if you care to watch, and the program itself is making some international moves into Tokyo and…

Alchemist’s latest batch puts AI to work as accelerator expands to Tokyo, Doha

“Late Pledge” allows campaign creators to continue collecting money even after the campaign has closed.

Kickstarter now lets you pledge after a campaign closes

Stack AI’s co-founders, Antoni Rosinol and Bernardo Aceituno, were PhD students at MIT wrapping up their degrees in 2022 just as large language models were becoming more mainstream. ChatGPT would…

Stack AI wants to make it easier to build AI-fueled workflows

Pinecone, the vector database startup founded by Edo Liberty, the former head of Amazon’s AI Labs, has long been at the forefront of helping businesses augment large language models (LLMs)…

Pinecone launches its serverless vector database out of preview

Young geothermal energy wells can be like budding prodigies, each brimming with potential to outshine their peers. But like people, most decline with age. In California, for example, the amount…

Special mud helps XGS Energy get more power out of geothermal wells

Featured Article

Sonos finally made some headphones

The market play is clear from the outset: The $449 headphones are firmly targeted at an audience that would otherwise be purchasing the Bose QC Ultra or Apple AirPods Max.

5 hours ago
Sonos finally made some headphones

Adobe says the feature is up to the task, regardless of how complex of a background the object is set against.

Adobe brings Firefly AI-powered Generative Remove to Lightroom

All cars suffer when the mercury drops, but electric vehicles suffer more than most as heaters draw more power and batteries charge more slowly as the liquid electrolyte inside thickens.…

Porsche Ventures invests in battery startup South 8 to boost cold-weather EV performance

Scale AI has raised a $1 billion Series F round from a slew of big-name institutional and corporate investors including Amazon and Meta.

Data-labeling startup Scale AI raises $1B as valuation doubles to $13.8B

The new coalition, Tech Against Scams, will work together to find ways to fight back against the tools used by scammers and to better educate the public against financial scams.

Meta, Match, Coinbase and others team up to fight online fraud and crypto scams

It’s a wrap: European Union lawmakers have given the final approval to set up the bloc’s flagship, risk-based regulations for artificial intelligence.

EU Council gives final nod to set up risk-based regulations for AI

London-based fintech Vitesse has closed a $93 million Series C round of funding led by investment giant KKR.

Vitesse, a payments and treasury management platform for insurers, raises $93M to fuel US expansion

Zen Educate, an online marketplace that connects schools with teachers, has raised $37 million in a Series B round of funding. The raise comes amid a growing teacher shortage crisis…

Zen Educate raises $37M and acquires Aquinas Education as it tries to address the teacher shortage

“When I heard the released demo, I was shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine.”

Scarlett Johansson says that OpenAI approached her to use her voice

A new self-driving truck — manufactured by Volvo and loaded with autonomous vehicle tech developed by Aurora Innovation — could be on public highways as early as this summer.  The…

Aurora and Volvo unveil self-driving truck designed for a driverless future

The European venture capital firm raised its fourth fund as fund as climate tech “comes of age.”

ETF Partners raises €285M for climate startups that will be effective quickly — not 20 years down the road