Trinity goes to Hollywood

Which movie did the first atomic bomb test best?

The first detonation of an atomic bomb, code-named Trinity, took place at the White Sands Proving Ground, New Mexico, in the early morning of July 16, 1945. It was, by all eye-witness accounts, an intense visual experience. One of the most vivid accounts comes from Brig. General Thomas Farrell:

The effects could well be called unprecedented, magnificent, beautiful, stupendous, and terrifying. No man-made phenomenon of such tremendous power had ever occurred before. The lighting effects beggared description. The whole country was lighted by a searing light with the intensity many times that of the midday sun. It was golden, purple, violet, gray and blue. It lighted every peak, crevasse and ridge of the nearby mountain range with a clarity and beauty that cannot be described but must be seen to be imagined. It was that beauty the great poets dream about but describe most poorly and inadequately.

Poets may dream of such beauty, but filmmakers try to realize such dreams. For nearly 80 years, filmmakers have repeatedly attempted to capture the visual experience of the Trinity test. Did they capture it? Or were their attempts, in Farrell’s words, poor and inadequate? Let’s take a look.

Trinity newsreel footage (1945)

Before we get into the Hollywood footage, let’s take a look at the original footage that was taken of the actual test. Trinity was photographed by over 50 cameras, but only a few were trying to capture “general” or “newsreel” footage of the explosion. The most famous of this footage was that filmed by Berlyn Brixner. Here is a short, 12-second clip of his footage, which was re-scanned at high resolution from Los Alamos a few years back:

This is only a tiny excerpt of even this shot, but it’s wonderfully high-resolution compared to most versions of it. Peter Kuran, of Trinity and Beyond fame, has a very nice compilation of the three main “newsreel takes” of the Trinity test — the first one, which was hand-shot by Brixner, a second one that has a tighter framing, and a third one that was centered on the sky near barrage balloons, giving a very different view:

The first one is, arguably, the best footage overall. There’s also a little moment, at 0:40 in the above video, where the camera abruptly pans up. Brixner explained the moment to Robert Del Tredici in an interview with him that was later published in Del Tredici’s excellent book of photography, At Work in the Fields of the Bomb (1987):

Did you use your camera's pan-and-tilt capability?

Yes, I did. I was so amazed, though, initially that I just let the camera sit there. Then suddenly I realized that the ball of fire was going out of the field of view of the camera, so I grabbed the controls and started tilting the camera up to follow the ball of fire.

Does that mean that in this historic footage there's a kind of a jerky movement at that point?

There is. For the first twenty seconds on the standard-speed camera it’s just sitting stationary, then suddenly you will see the field of view jump as the ball of fire is going out the top of the frame.

I guess you could call that the atomic tilt technique…

Well, I guess that's it, yes. …

I think that’s pretty interesting — you’re seeing a different kind of record, of sorts: of the photographer having realized that his shock was causing him to mess up the shot.

Brixner also emphasizes, in the interview, some things not obvious in the footage. There was no audio recorded for the first atomic test (nor most atomic tests). Brixner says:

…At about thirty seconds the shock wave reached our station with a very loud bang, and then again it was silent. If you’ve ever been close to a lightning bolt and heard that, that’s the type of sound it made—just a tremendous bang and crack!

All of the above is prologue for what I really want to talk about, which is how the Trinity test has been depicted in Hollywood recreations. Below is a definitely incomplete survey of examples from across almost eight decades.

The Beginning or the End? (1947)

The Beginning or the End? was a 1947 film produced by MGM about the Manhattan Project, with official cooperation from the Department of War and the US government. The circumstances behind its creation, release, and content are recounted in great detail in Greg Mitchell’s The Beginning or the End: How Hollywood―and America―Learned to Stop Worrying and Love the Bomb (2020).

Here is their take on the Trinity test:

This footage is very interesting to me, because it is necessarily “practical” in nature. It is also pretty good, especially for its time! The cinematographer, A. Arnold Gillespie, won an Oscar for these particular special effects.

It looks to me like there are three distinct “shots”:

1. The initial fireball (0:05).

2. The rising cloud (0:20).

3. The close-up cloud (0:25).

4. The cloud base (0:51 and 1:20).

Of those three, #3 is actual footage from Trinity, and is the third “newsreel” clip in the video up above (about 30 seconds into it). As for the others, let’s also take a look at the other VFX shot from the film, which was the bombing of Hiroshima:

I did a little digging about the effects for this film and found some lore that ultimately traces back to an article from the mid-1970s which claimed:

A. Arnold Gillespie, Oscar-winner for Thirty Seconds Over Tokyo, Green Dolphin Street, Plymouth Adventure, and Ben-Hur, even created in miniature an A-bomb blast so realistic that it fooled the Manhattan Project people, and the U.S. Air Force used the footage for years in a training film. The movie was The Beginning or the End, which was supposed to show the bombing of Hiroshima, an event that was not officially photographed. Also, all technical details were classified at the time (1947). Gillespie, recalling a Tarzan film he'd seen, observed that each time the ape-man stabbed a crocodile underwater, blood would float up to the surface in the shape of a mushroom cloud. So he put his camera underwater, inside a tank, and shot his "holocaust” with exploding sacks of dye.¹

The article contains no references for its claims, and at least part of it is clearly nonsense. Hiroshima was, in fact, filmed by the physicist Harold Agnew, who was a crew member of the instrumentation plane on the mission. You can see his brief 16mm footage here for a comparison. I express extreme doubt that the 1947 footage would “fooled the Manhattan Project people” or that the Air Force used it as training footage — this sounds like Hollywood lore. It is true that some technical details were classified at the time, but the Trinity test newsreel footage was not (some of it was actually used in the movie!), and there was, by 1947, footage of Operation Crossroads that had been declassified.

Of the shots, what I’ve labeled as shot #4 of Trinity definitely looks like underwater dye packs to me. #1 might be — it has some very good Rayleigh-Taylor instability on display with regards to the stem, and that would be what a mixture of fluids in waters would give you. If that’s true, then the technique for #1 is also pretty great because of the way they are handling the lighting of it. It’s a effects shot for 1947.

#2 does not really look like dye packs to me, but I could be wrong. If it’s not a dye pack effect, I wonder if it is some kind of fog generator that is being raised. That would produce, I think, a very similar effect (and is not unlike how one can very simply simulate decent mushroom cloud effects with CGI today).²

The shots of the “base” of the Hiroshima cloud from the film look like dye packs to me, just in how they hang around and writhe. I’m less certain that the rising cloud is dye packs, although I don’t really know how dye behaves underwater. It could be. It’s a very nice effect, though.

Anyway, for 1947, this is peak. Very compelling. Better cloud dynamics than a lot of later entries. A great use of practical technique. One could imagine much cheesier effects for the time.

If I were judging it on its pure accuracy, I would say that the speed is not quite right — I would have slowed it down a bit, to give a better sense of its “mass.” And with some slightly better (less direct) lighting, one would have less of a sense that it was being filmed on some sort of movie set. It does go to show you how difficult it would be to “fake” a realistic mushroom cloud in the 1940s.

Fat Man and Little Boy (1989)

Next up is Fat Man and Little Boy, a 1989 film about Los Alamos directed by Roland Joffé and starring Dwight Schultz as J. Robert Oppenheimer and Paul Newman as General Leslie Groves. Here’s a minute of their Trinity test footage, capturing the moment of detonation:

I actually appreciate the choice of showing Oppenheimer’s face as the initial view of the test — if you can’t actually capture the experience on film, just try to capture the experience of someone experiencing it, perhaps. But in this case there are really execution issues. The sound is all wrong — instantaneous with the flash — and the arrival of the blast wave is far too quick, and choosing to represent this as Oppenheimer with a leaf blower trained on his face is, well, unintentionally comical. It robs the experience of any power it might have and looks ridiculous.

As for the detonation itself… aside from however they did it, which I don’t know, (someone needs to get Corridor Crew onto this sort of thing…), there are a lot of accuracy issues that are hard to reconcile. It looks to me like they took more inspiration from H-bomb tests in the Pacific than the actual Trinity test. It’s also such a quick SFX sequence compared to the prolonged Oppie-gasm we just witnessed, which is a strange choice. I have mixed feelings about over emphasizing the power of the weapon in this way — it’s not very accurate looking, but I can appreciate the attempt to be impressionistic about it.

Overall I appreciate the ambition and the concepts they were going for, but the execution is rather flawed and it mars the experience of it.

Day One (1989)

So this is an honorable mention, something rather obscure, and a very different and low-budget approach. Day One was a CBS mini-series that aired in 1989, starring Brian Dennehy as Groves and David Strathairn as Oppenheimer. I haven’t been able to find a copy of it, but I have found a copy of the Trinity test sequence on YouTube… but it’s the French dub. Still, amusing and interesting:

This is pretty fun. The “explosion footage” is a compilation of different actual nuclear test footage made by the government. Some of it is from the actual Trinity newsreel footage, some of it appears to be shots at the Marshall Islands.

Does it totally work? Of course not — the footage doesn’t “blend” with the rest of the film at all. Is it good-enough for a made for TV movie? Sure, why not?

Twin Peaks: The Return (2017)

Jumping forward, we have a clip of the Trinity test from David Lynch’s Twin Peaks: The Return, the 2017 season 3 of his famed surrealistic noir television show. I have not (yet) actually watched Twin Peaks (I am sorry!), so while I was made aware that it involved the Trinity test, I know very little about its significance in the show other than it is significant somehow. So I am looking at this as on purely aesthetic terms:

So this one is clearly CGI. I appreciate the choice to do it in black and white, which aligns well with the original footage. The audio design is appropriately intense. The long zooming is an interesting choice as it misses most of the “early cloud” of the original footage and focuses instead on the “late cloud” that looks more mushroomy. It is not entirely clear to me whether this is meant to showing the evolution of the cloud in realtime. The “base” surge of the cloud seems like it is expanding very slowly, and is pretty large in proportion to the rest of the cloud.

The actual modeling of the mushroom cloud effect is… not that accurate. My guess is that whomever did this did not really understand the underlying physics that they were meant to be emulating very deeply and sort of “winged it” with some inspiration from footage of Trinity and other nuclear tests. This may be why it looks more like a videogame mushroom cloud than a real mushroom cloud.

It is possible to do pretty good mushroom cloud physics with modeling tools like Blender these days (for example, this approach is a pretty good place to start, and is not hard to set up or tweak), and certainly there are other films and shows that have managed to make more plausible CGI mushroom clouds. (I will not be surveying all of those, right now, because I am focused exclusively on the Trinity test — but at some point I’ll do another of these that looks at other films/shows/etc., like The Day After, The Sum of All Fears, and Fallout.)

But while there is room for improvement, the aesthetic choices are great and I suspect most people aren’t going to be nitpicking it to quite the same level that I would, and would not be bugged that it is not really displaying much understanding of Rayleigh–Taylor instability. The starkness of it, the length of the bright flash, the intense audio, the strange flying into it — it all works aesthetically.

Oppenheimer (2023)

And, of course, we have our most recent entry — Christopher Nolan’s 2023 film Oppenheimer, which has the Trinity test about midpoint through the 3 hour film:

Nolan’s approach is a combination of practical effects — he famously eschewed doing this with CGI for artistic reasons — and editing, with lots of cuts that jump back and forth between different perspectives, the explosion, the fireball, etc. It’s largely silent, which is both uncharacteristic of him (I kept waiting for the THRONGGGG noises from Inception) and allows him to center it on the breathing of Oppenheimer.

I have very mixed feelings about how it came out. I think most of the footage that lets one see the “scale” of it makes it look small. It does not look very “nuclear” to me. It’s not hot enough, for one thing — the colors are too orange and too red, all too early. It’s not very mushroomy. It doesn’t move at the right speed for it to be as large as it was. Large things appear to more slowly than smaller things, because there’s much more to move. It’s why plausible kaiju are pondering and lumbering and men-in-suits kaiju are so obviously men in suits.

Also regarding the scale, there are way too many “sparks” — each of which would be the size of an automobile or larger if the explosion was properly scaled to kiloton range. Nuclear explosions don’t send out a shower of visible sparks.

If I were presumptuous enough to suggest what kind of editing might salvage this, I might suggest: color-grade it so that it felt hotter (more white-yellow, less orange-red) for longer, focus more on the zoomed-in shots, and slow it down considerably. That would get you something more like the third of the three original Trinity newsreel clips, done practically. For any other shots I would simply not show it in a way that would make it clear what its actual scale was, because it’s really hard to fake kiloton scales with what was probably only a few tons of TNT equivalent.

I think the final mushroom part of the fireball looks good — starting around 1:29 in the above clip — and I would linger on that a lot longer.

This is not my main critique of the Nolan film, but it’s bugged me ever since I saw it, because I feel like even with the commitment for purely practical effects, which I am totally fine with, there is a lot of room for improvement here. It doesn’t convey the awe to me.

But, again, I’m not the target audience, so maybe my sensations of it weren’t universal, and other people were more impressed by it.

Am I being too nit-picky? It’s entirely possible! But for me, the goal of a Hollywood version of the Trinity test should prioritize, above almost anything else, trying to convey that experience of it that someone there might have had. Something that goes beyond just watching the original footage. Awe, horror, “now we’re all sons of bitches,” the whole shebang.³

That’s the holy grail of a Hollywood Trinity test. I’m not sure any of the above fully pulled it off, but they each have their glimpses of it.

1. James M. Martin, “The Best of All Impossible Worlds,”  American Film 1, no. 5 (March 1, 1976): 28-32, 49

2. For example, here is an example of how to render a mushroom cloud in Blender which involves making an invisible, smoke-emitting sphere which moves upwards through the scene. With a decent smoke simulation, this produces an effect that is not so visually unlike how actual mushroom clouds work.

3. To be fair, the Able detonation of Operation Crossroads (1946) was also viewed by observers as a disappointment. There were some good reasons for this — the observers were very far away from it, and it missed its target, etc. — but a lot of it was also probably because of their high expectations for what it ought to look like, based on the descriptions and stills from Trinity.