[porphyra]

I'd have probably shot it wide open at f/2.8 rather than cranking the ISO up to 51200. Incredibly impressed at the steady hands for a sharp image at 1/4 s shutter speed though! Maybe they just let the camera float in space with the mirror up, triggering it remotely.

[throw0101d]

> I'd have probably shot it wide open at f/2.8 rather than cranking the ISO up to 51200.

One of the reasons the D5 supposedly was chosen was because of its high dynamic and good low light performance. It can go up to ISO 3,280,000:

* https://en.wikipedia.org/wiki/Nikon_D5

The D5 has been used on the ISS, including EVAs, since 2017, so is a known quantity:

* https://en.wikipedia.org/wiki/List_of_cameras_on_the_Interna...

[porphyra]

The good low light performance was amazing for its time (10 years ago), and it still holds up decently today. But let's not kid ourselves -- it has been clearly surpassed by modern backside illuminated CMOS sensors like the one on the Z9.

EDIT: sorry, it seems I'm wrong. I just checked https://www.photonstophotos.net/Charts/PDR.htm and while the Z9 has the clear edge with 2 more stops of dynamic range at low ISO, the D5 actually pulls ahead at high ISO. Perhaps the technological improvements haven't been that much for the shot-noise dominated regime.

[nelox]

Was hoping to hear from the person at NASA who made the choice and why.

[throw0101c]

Sure, but D5s were also doing EVAs for many years and much more of a know quantity in space.

[ourmandave]

You can get a D5 on amazon.com. It would be amazing if one of the astronauts did a review explaining how it performs in space.

[adamm255]

Oh man. "Rolled with the D5 on my recent trip around the moon, decent performance, very light and easy to hold in zero gravity".

[mhb]

f/8 and be there?

[jen729w]

You might misunderstand how ISO works on digital cameras. (I did.)

https://www.youtube.com/watch?v=ZWSvHBG7X0w

[deathanatos]

Good grief, that video suffers the YouTube-ism of "ramble on about how you don't understand X" for way too long.

Video alleges people think ISO makes the sensor "more sensitive or less sensitive". (I … don't think this is common? But IDK, maybe this is my feldspar.)

(The video also quibbles that it is "ISO setting" not "ISO" … while showing shots of cameras which call it "ISO", seemingly believing that some of us believe ISO is film speed, in a digital camera?)

Anyways, the video wants you to know that it is sensor gain. And, importantly, according to the video, analog gain, not digital gain.¹ I don't know that the video does a great job of saying it, but basically, I think their argument is that you want to maximize usage of the bits once the signal is digitized. Simplistically, if the image is dark & all values are [0, 127], you're just wasting a bit.

You would want to avoid clipping the signal, so not too bright, either. Turn your zebras on. (I don't think the video ever mentions zebras, and clipping only indirectly.)

The video does say "do ISO last" which I think is a good guideline. Easier said than done while shooting, though.

… also while fact checking this comment, I stumbled across Canon's KB stating to use as low an ISO as possible, which the video rails against. They should talk to Canon, I guess?

¹with the caveat that sometimes there is digital gain too; the video notes this a bit towards the end.

[harrall]

ISO changes the analog gain and in a way yes, it does make it more sensitive to a certain range of brightness.

This is because the ADC (analog to digital converter) right after can only handle so many bits of data (like 12-16ish in consumer cameras). You want to “center” the data “spread” so when the “ends” get cut off, it’s not so bad. Adjusting the ISO moves this spread around. In addition, even if you had an infinite bitrate ADC, noise gets added between the gain circuit and the ADC so you want to raise the base signal above the “noise floor” before it gets to the ADC.

Gain is not great — it amplifies noise too. You want as low ISO as possible (lowest gain), but the goal is not actually to lower gain; your goal is to change the environment so you can use a lower gain. If you have the choice between keeping the lights off and using higher ISO versus turning on the lights and using a lower ISO, the latter will always have less noise.

Most photo cameras have one gain circuit that has to cover both dark and light scenes. Some cameras like a Sony FX line actually have two gain circuits connected to each photosite and you can choose, with one gain circuit optimized for darker scenes and the other optimized for brighter scenes. ARRI digital cinema line cameras have both and both are actually running at the same time (!).

[seba_dos1]

> your goal is to change the environment

...or integration time.

[account42]

But preferably the environment since longer exposures also add more noise.

[OJFord]

> The video does say "do ISO last" which I think is a good guideline. Easier said than done while shooting, though.

> … also while fact checking this comment, I stumbled across Canon's KB stating to use as low an ISO as possible, which the video rails against. They should talk to Canon, I guess?

Isn't ISO last the same as setting it as low as possible? Obviously it's always set to something, so I thought 'doing it last' means start with it low, set exposure & shutter, increase as necessary?

(Shutter speed being dictated by subject and availability of tripod, essentially it's just exposure & ISO which becomes about how much light there is and how it's distributed, I suppose.)

I'm not really into photography though, so perhaps that's all nonsense/misunderstanding.

[jen729w]

Sounds like you'll be spending your day making a better video! :-)

[account42]

Why when he already made a comment that has a much better content density than any video would. Not everything needs someone rambling in front of a camera.

[bingkaa]

i think we need to differentiate between raw or derivative format. canon KB might cater to wider audience thus the latter

[throwaway290]

If you think they only took one shot, you're not a digital photographer)

In this special situation you get as many as you can a few dozen at least. Then only publish the one that looks the best. If it's f4 then f4 worked best.

[js2]

Wide open generally sacrifices lens sharpness.

[porphyra]

Sure, but less grain is often worth it. There's a reason why fast lenses exist. The high quality lens being used here can probably still resolve 20 MP adequately even wide open.

[dgxyz]

I had that lens. It’s soft as fuck around the edges open.

Peak sharpness is about f/8. They should have had the D5 on aperture priority auto iso, pushed the exposure comp either way and then just fired at f/8 and let the camera make the decisions.

But they are astronauts not photographers :)

The modern Z lenses are far better and sharper open but much larger generally.

[kccqzy]

> Incredibly impressed at the steady hands

I’m an amateur photographer and I’m actually incredibly impressed at how good the optical image stabilization tech is these days even in entry-level cameras. I wouldn’t hesitate to use 1/4 s shutter speed when necessary (such as in many indoor environments) and I’m quite amazed at the sharpness.

[narmiouh]

I would imagine since they are not circling the earth, that there will be pull of gravity and the camera would start to move relative to the spacecraft. But may not fast enough for a short exposure

[gus_massa]

Once you are out of the atmosphere and turn off your thrusters, you are on "fee fall" and the gravity on the camera, you and the spaceship produce the same acceleration and they cancel and it looks exactly like "zero gravity". It doesn't matter if you are in orbit around the Earth, going to the Moon.

[echoangle]

Actually not quite correct. The camera and spaceship will generally have different starting positions of their center of gravity but the same starting velocity, leading them to drift apart.

The only real relevant thing for the photograph is rotation though as long as the camera doesn’t float in front of the window frame, and airflow is probably much more relevant for both points than gravity.

[gus_massa]

> will generally have different starting positions of their center of gravity but the same starting velocity

That are the tidal forces, they are quite small for a system of the size of Orion.

[echoangle]

Right, I was talking technically. That’s what I tried to say in the second paragraph.

[dotancohen]

The gravity of the Earth (and moon, and everything else) is uniform (i.e. no gradient) on the scale of things the size of that capsule at the distance that capsule is from them, on the order of time of the exposure of that photograph. So the gravity (from any source) will pull on the spacecraft and on the camera in the same fashion.

To fully answer the question, the moon's gravitational gradient does pull on the Earth, the ocean closest to the moon, and the ocean furthest from the moon differently. But those are objects separated by thousands of kilometers, having hours of gravitational influence acting upon them.

[mr_toad]

They’re not circling the Earth, but they’re still orbiting it. Their orbit is highly eccentric, and will be near the Moon at apogee.

[narmiouh]

Yeah, makes sense. I was assuming the vehicle was also deliberately accelerating which would make a difference on a floating camera, but if it isn't, then gravity is the only force and no relative difference.

[chrisbarr]

It was shot at f/4, so opening up 1 stop to f/2.8 would only reduce the ISO to 25600

[treis]

They're in space so they only sort of need to hold the camera.

[sdenton4]

Or maybe press the timer and let it float...

[embedding-shape]

Bad idea, shutter speed was 1/4 apparently (https://news.ycombinator.com/item?id=47632457), even the small rotational inertia everything in zero gravity gets from a human "dropping" it would probably be enough to be annoying, you'd get a better shot holding it.

[plaguuuuuu]

or sticky-tape it to the window.

d5 has an actual shutter yeah? not mirrorless? I think the shutter moving will spin the camera.

[cloudfudge]

I would love to see the effect of the mirror's effect on the motion of the camera in a weightless environment. I bet it's enough to measurably affect the picture, especially on a long exposure. Net torque of it opening and then closing should be near (but probably not exactly) zero, but while it's open the camera should spin a tiny amount.

[hackerdood]

I haven’t looked at the manual but it likely has the ability to flip and keep the mirror up for direct capture on the sensor without the mirror flipping up and down between exposures.