How Peter Jackson Made WWI Footage Seem Astonishingly New | Hacker News Reader

How Peter Jackson Made WWI Footage Seem Astonishingly New

nytimes.com

192 pointsIfOnlyYouKnew50 comments

About 10 years ago Peter Jackson made a "short" movie called Crossing the Line. It was the first real footage shot with a Red camera. It's kinda like a trailer for a movie that was never made. It clearly shows Jackson fascination with WW1.

https://www.youtube.com/watch?v=QleIfcA8mRQ

https://en.wikipedia.org/wiki/Crossing_the_Line_(2008_film)

Oh man you don’t know the half of it. He’s been working on The Dambusters for years [1]. Also, he owns this incredible collection of old WWI and II planes, most of which have been restored from scratch! [2]

[1] https://www.stuff.co.nz/entertainment/celebrities/108520139/...

[2] https://www.omaka.org.nz/

He also owns a company that builds and sells brand new, airworthy WWI replica planes.

http://thevintageaviator.co.nz/

As I recall its production was subsidized by Red with prototype cameras, to sell their next generation of 4K/5K cameras to filmmakers. It worked, I see Red in use all over Vancouver, all the time.

This is great film, I highly recommend it. It was released here about a month ago (NZ).

Though I didn't know that there was a 3D version as well.

What is the name of the film?

From the second paragraph of the article:

>”With “They Shall Not Grow Old,” Jackson has applied new technology to century-old World War I footage to create a vivid, you-are-there feeling that puts real faces front and center and allows us to hear their stories in their own words.”

> The footage had a herky-jerky feel because it had been shot on hand-cranked cameras [...]. Jackson’s team retimed the footage, speeding up the frame rate, adding extra frames digitally and smoothing out the movement.

I don't know how they did this. Sure, I can understand filling in frames at regular intervals, to bring one speed up to another speed, say from 18 to 24 frames per second. But if it was hand-cranked, then the old speed wasn't exactly 18 frames per second, and I don't know how you can tell precisely what speed the cameraman was cranking at any given time.

Someone might suggest watching for when the movement of the subject or camera changed speed, but how do you know that it wasn't the person himself who slowed down or sped up, in real life, or that the cameraman's handheld pan wasn't really slowing down or speeding up?

Either this was a lot of manual trial and error, or software for this kind of thing is far better than I can imagine.

I'd say it's a little bit like automatic stabilisation. If everything in your frame moves in the same direction at the same speed, then it's probably the camera moving, so you correct for it and the video becomes stable.

Similarly you could see multiple things in the frame accelerating and decelerating at the same time and deduce it's not the things, but the camera that's doing that. That's the retiming part. Filling in frames is probably a whole lot more magic though..

Just out of interest, could a GAN be used to clean up video, filling in all the blanks and creating a higher resolution, cleaner image and of course colouring too? From what I have seen of moving GAN images, they can jump around and be a bit jittery.

It’s been actively worked on, you can see plenty of samples on YouTube. Like many ML applications, it sort of works, but the quality leaves plenty to be desired.

Yes, sort of: You can use a GAN to do almost everything described in the article. That's really interesting in a way, because they mention several distinct steps, and all of them seem readily amendable to machine learning, including the totally-unrelated-to-coloring lip reading they did to add sound, etc.

BUT: reading the article with ML in mind already, I noticed the frequent references to their work with historians and other experts to get it right.

In other words: ML could possibly get you results that are entirely plausible, yet wrong. They mention the color of buttons on the uniforms and, even less likely to be amendable to technology, commands and other speech shouted off-screen.

Probably the best solution would be a combination of the two: the ML algorithm would also learn from human examples and corrections.

I think this will be the solution in a lot of problems and industries: ML does the boring and repetitive work, and a human sanity-checks it and fixes the edge cases where ML is 'dumb'

I wonder if the before and after could be used to train a computer to do this automatically, well, for the future. Obviously not the sound, but the picture.

There was a popular post a little over a month ago [2] about this very topic. That particular example used deep learning in a similar way to what you described.

It looks like there are previous attempts [2] that have even been turned into APIs [3]

[1] https://news.ycombinator.com/item?id=18363870

[2] https://github.com/richzhang/colorization

[3] https://demos.algorithmia.com/colorize-photos/

There are many excellent hires copies of silent movies run on TV now and then. With colorization, and adding a soundtrack with dubbed dialog, they'd make fine modern movies.

Why the request for colorization and dubbed dialog? Metropolis did just fine for a restore and without any of those changes.

For the same reason Jackson has.

Upvoted, though remember, much of silent film is acted in a particular, stylized manner because of the lack of sound. But in documentary footage, people actually speak in a normal manner.

That's remarkable. I'll be in an audience somewhere on the 27th.

Pity this article has been released after the limited run has seemingly ended

The showings are tonight, the 17th, and on the 27th. It's not too late.

I think those are the US showings, which for some reason are a month after the armistice? Not sure why. The UK ones were in November so no chance to catch it in the UK now, is that’s what the person meant.

Thanks for clarifying, not sure why I didn't think of that!

it's available for download all over internet, just watching 720P X265 BRrip

Yep I'm going to buy it, but unfortunately there's no 3D version available

I'm ambivalent about WW1 since the US didn't do much fighting (too late into the war), and it was mostly trench warfare that wiped out European royalty, not necessarily a bad thing IMO.

But the inter-war years and WW2 are a fascinating epic.

The WW2 European conflict resulted from Germany losing WW1, but not being occupied. The WW2 Pacific conflict resulted from the USA embargoing Japan after they "colonized" Manchuria.

Royals start wars; commoners fight them and bear their hardships

(A point exemplified by World War One)

The first world war is a bad example. While the war was absolutely started by royalty, they did send their sons to the front to die along with commoners.

> Although the great majority of casualties in WW1 were from the working class, the social and political elite were hit disproportionately hard by WW1. Their sons provided the junior officers whose job it was to lead the way over the top and expose themselves to the greatest danger as an example to their men.

> Some 12% of the British army's ordinary soldiers were killed during the war, compared with 17% of its officers. Eton alone lost more than 1,000 former pupils - 20% of those who served. UK wartime Prime Minister Herbert Asquith lost a son, while future Prime Minister Andrew Bonar Law lost two. Anthony Eden lost two brothers, another brother of his was terribly wounded, and an uncle was captured.

https://www.bbc.com/news/magazine-25776836

Also, Princes Oskar and Eitel Friedrich of Prussia both served in front-line combat roles and both were wounded.

The hubris of the British at the time, where they fully believed that the war would only last a few weeks and they'd be 'Home for Christmas', is mainly what led to many sons of the Upper Class willingly signing up to fight in droves.

Alas, as we all know, the war lasted a lot longer than 'a few weeks'.

WW1 directly impacted the ability of many large estates across the UK to survive beyond the 1940s. It changed the cultural and class landscape of the UK significantly.

This has probably been thought, at least publicly, about almost every offensive war waged by a democratic nation. Few civilizations (there are exceptions) would stomach a war that they knew from the outset would last for years.

Cousins King George V and Tsar Nicholas II actually made a concerted effort to mediate with another cousin, Kaiser Wilhelm II. However he seemed to be more influenced by his war-mongering generals, especially Von Moltke. If the blame for starting the war can be laid at any door, it is his.

In some sense the first shot fired was at royalty, by the student Gavrilo Princep.

WWI is super fascinating. It sort of marks the final transition period from stand in a polite line and shoot at one another style warfare to modern warfare.

I would put that transition at perhaps the Austro-Prussian or Franco-Prussian wars, at least for Europe.

WW1 was the first war with tanks. Pretty fascinating all the technology developed in WW1, and the war itself is really cool also.

The documentary has plenty of shots of tanks entering the battlefield. Probably my favorite part of the whole thing to be honest. It goes from men on horses to seeing tanks roll through.

Highly recommend this documentary.

And now we are back in a situation where the British Army has more horses than tanks!

Seriously? That's fascinating.

Fascinating isn't the word I would use... :-|

https://www.bbc.co.uk/news/uk-scotland-highlands-islands-229...

“and the war itself is really cool” that is such a weird statement to read. Maybe “interesting“ would fit, but “cool“? I wonder if any of the people who died in the Ypres/Passchendaele region, or Somme, would call the war “cool“.

Please mind that you may not be talking to a native or fluent English speaker, and your comment just comes across as rude and unwarranted then.

One also could interpret WW2 in the Pacific to have been caused by the US forcing Japan to open up to foreign trade by sending Commodore Perry and his fleet. This ultimately resulted in the fall of the Tokugawa shogunate and lest to the nationalistic Meiji restoration. After that some elements in Japanese politics tried to get respect from Western powers by making nice nice and imitating Western culture. That didn't bring the hoped respect and better trade deals though. Japan ended up learning from Western empirialism and got it's respect by showing some military might on its own.

You could only interpret it that way if you believed that every step in the process was a necessary progression from the previous one. That is a very big stretch over the time scale of the 1850s to the 1940s. At every stage there were possible alternative choices that would have resulted in different outcomes. Don't make the mistake of believing that what actually happened was set in stone at the beginning like a row of falling dominoes.

This is a difficult criticism since it applies to nearly every human interpretation of history. Cause and effect are so wired into our perception of the world. It's like saying, don't be fooled by your eyes, you're not seeing the truth, just a reflection of photons on your retina. Technically true, but totally useless.

No, it's not like that at all. The outcome was not determined in advance because there were multiple possible outcomes. People could have chosen to take different actions than they actually did. The events of the 1940s were not set in stone in the 1850s.

I don't disagree with that. I'm saying that no events are ever set in stone, so you can't talk about history without tying together events that are only partly related.

The same is true for the US response to Japan's invasion of Mongolia and Japan's response to the US response. If any step along the way had been different the outcome would have likely been as well.

An important point here is that it's much harder to find a clear starting point of WWII from a Japanese perspective. The whole thing pretty much was a mess starting with Commodore Perry's arrival.

Check out the movie "The Blue Max". Initially, WW1 pilots were selected from the nobility. The death rate was so high that they ran low on nobility and were forced to recruit pilots from commoners. TBM is about the transition.

The soundtrack is also a favorite of mine. I don't think any other piece of music captures so well what it's like to fly.

I thought somewhat similarly about World War I not being as interesting as later 20th century history until I listened to Dan Carlin's excellent "Blueprint to Armageddon". It's a multi-part Podcast going through WWI. I loved his style. Each episode is a few hours long, but it goes quick.

Here's part 1:

https://www.dancarlin.com/product/hardcore-history-50-bluepr...

Reply on news.ycombinator.com