What if you took a selfie from a billion kilometres away?

From a billion kilometres, all of Earth fills about one pixel of a Voyager-class camera. To see yourself you would need a lens over 1,300 km wide. Or you could cheat with a laser and become the brightest point on the planet.

The 2020 reprocessed Pale Blue Dot image from Voyager 1, with Earth a tiny pale point inside a band of scattered sunlight
Photo: NASA/JPL-Caltech, Public domain, via Wikimedia Commons

If you took a selfie from a billion kilometres away, you would not be in it in any visible way: with a Voyager-class camera, the whole Earth covers about 1.4 pixels, and each pixel covers about 9,250 km, some 18 million times your width. To actually resolve a 0.5 m patch of you in visible light, the diffraction limit demands a telescope aperture of about 1,342 km. The trick that works is to stop trying to resolve yourself and become a light source: a 7 kW laser aimed at the camera would outshine the sunlit Earth around you.

Key takeaways
  • 55.6 minutes is how long light takes to cover a billion kilometres, so the selfie shows you as you were an hour ago.
  • 1.4 pixels is Earth's width in a Voyager narrow-angle camera at that distance.
  • 1,342 km of aperture is needed to resolve 0.5 m at 550 nm, from the Rayleigh criterion.
  • About 100 m is what an Earth-sized telescope like the Event Horizon Telescope would resolve there.

    Calculated in this post from the Event Horizon Telescope's 1.3 mm observing wavelength and a baseline equal to Earth's 12,756 km diameter, at the billion-kilometre distance this what-if assumes.

  • 7 kW of laser, the power NASA used to reach Psyche, would make you a bright point inside the pale blue dot.

We already have real "selfies" of Earth from deep space. Voyager 1 took the Pale Blue Dot from about 6 billion km in 1990, Cassini photographed Earth from Saturn at 1.45 billion km in 2013, and NASA's Psyche caught Earth and Moon from 290 million km in July 2025. A billion kilometres sits between Psyche and Cassini. So the camera question is real, and the answer can be worked out exactly.

What would Earth look like in a selfie from a billion kilometres?

Earth would be a single bright dot, about 1.4 pixels across in the best camera Voyager carried. That follows from simple angles: size divided by distance.

The story

On 19 July 2013, Cassini slipped into Saturn's shadow and turned its cameras back toward the inner Solar System, about 1.45 billion km from Earth. For the first time, people on Earth were told in advance that their planet was being photographed, and the imaging team led by Carolyn Porco asked everyone to go outside and wave. More than 1,400 photos of waving people from over 40 countries and 30 US states were sent in and stitched into a collage. In Cassini's own frame, Earth was a blue speck below the rings, and NASA's project scientist said plainly that no individual human could be picked out.

NASA JPL: Cassini releases image of Earth waving at Saturn (Aug 2013)

That was the largest group selfie ever attempted from the outer Solar System, and nobody was in it. Let me work out why, and how far off we are.

Saturn's rings with Earth and the Moon as tiny points of light below them, seen by Cassini on 19 July 2013
Earth and the Moon from Saturn, 19 July 2013. At about 86,620 km per pixel in the wide-angle frame, the whole planet is smaller than one pixel.Photo: NASA/JPL-Caltech/Space Science Institute, Public domain, via Wikimedia Commons

The camera we'll use

Voyager's narrow-angle camera had a 1,500 mm focal length at f/8.5, a 0.42° field of view and an 800 × 800 vidicon frame, which NASA's Planetary Data System catalogue for the camera gives as 0.00053° per pixel, about 9.25 microradians. The aperture is therefore 1,500/8.5, or about 176 mm. It is the camera that took the Pale Blue Dot, so it is a fair benchmark.

  1. Assumptions. Distance \( d = 10^{9} \) km \( = 10^{12} \) m. Earth's equatorial diameter \( D_\oplus = 12{,}756 \) km (NASA). Pixel scale \( p = 9.25\ \mu\text{rad} \).
  2. Earth's angular size:
    \[ \theta_\oplus = \frac{D_\oplus}{d} = \frac{1.2756\times10^{7}}{10^{12}} = 12.8\ \mu\text{rad} \]

    Calculated in this post from Earth's equatorial diameter (NASA) and the billion-kilometre distance this what-if assumes.

  3. Earth in pixels:
    \[ n = \frac{\theta_\oplus}{p} = \frac{12.8}{9.25} \approx 1.4\ \text{pixels} \]
  4. Check against the real Pale Blue Dot. At 6 billion km, the same sum gives 0.23 pixels for the full disc. NASA reports a crescent of 0.12 pixel, consistent because only part of the disc was lit.
  5. Light delay:
    \[ t = \frac{d}{c} = \frac{10^{12}}{2.998\times10^{8}} = 3{,}336\ \text{s} \approx 55.6\ \text{min} \]
    A shutter command from Earth takes 55.6 minutes to arrive and the picture 55.6 minutes to return, so the round trip is about 1 hour 51 minutes.

    Calculated in this post from the speed of light, 299,792,458 m/s (NIST) and the billion-kilometre distance this what-if assumes.

How far away real cameras were when they photographed Earth
See the data as a table
Distance from Earth ( million km)
Psyche, Jul 2025290
This what-if1000
Cassini, Jul 20131446
Voyager 1, Feb 19906000

Source: NASA JPL (Psyche 2025), The Planetary Society (Cassini 2013), NASA Science (Voyager 1 1990) · Voyager figure is NASA's rounded 6 billion km; some sources give about 6.4 billion km.

How big a lens would you need to see yourself?

You would need an aperture about 1,342 km wide to resolve a 0.5 m patch of yourself in green light. The limit is not pixels; it is diffraction, the spreading of light through any finite opening.

The Rayleigh criterion, in plain words

Light passing through a circular aperture of diameter \( D \) makes a blurred spot, the Airy disc, instead of a point. Two points are just separable when one sits on the first dark ring of the other, at an angle of

\[ \theta_{\min} = 1.22\,\frac{\lambda}{D} \]

No sensor, software or AI upscaling beats this for a single aperture; it is a property of waves. To see something twice as small from the same distance, you need a lens twice as wide.

  1. Target (this what-if's premise). One pixel covering a 0.5 m patch of you, \( s = 0.5 \) m, so \( \theta = s/d = 5\times10^{-13} \) rad. Wavelength \( \lambda = 550 \) nm.
  2. Aperture:
    \[ D = \frac{1.22\,\lambda}{\theta} = \frac{1.22 \times 5.5\times10^{-7}}{5\times10^{-13}} \approx 1.34\times10^{6}\ \text{m} = 1{,}342\ \text{km} \]
  3. Focal length to put that patch on a 1 µm pixel:
    \[ f = \frac{\text{pixel}}{\theta} = \frac{10^{-6}}{5\times10^{-13}} = 2\times10^{6}\ \text{m} = 2{,}000\ \text{km} \]
  4. Compare Voyager. Its 176 mm aperture has a diffraction limit of \( 1.22 \times 5.5\times10^{-7}/0.176 = 3.8\ \mu\text{rad} \), which at a billion km is a 3,800 km blur. You are 7.6 million times too small for it.

    Calculated in this post from the Rayleigh formula 1.22λ/D (Wikipedia, Angular resolution), Voyager's 0.1765 m aperture (NASA PDS) and this what-if's 550 nm wavelength and billion-kilometre distance.

What you want to seeSizeAngle at 10⁹ kmAperture needed at 550 nm
A city10 km10 nrad67 m
A cricket stadium200 m0.2 nrad3.4 km
A car4 m4 prad168 km
You, as one pixel0.5 m0.5 prad1,342 km
Your face, recognisably1 cm0.01 prad67,100 km

The last row is the punchline. A recognisable face needs a mirror about five Earths wide.

camera, aperture D d = 10¹² m (55.6 light-minutes) Earth, 12,756 km angle θ = size ÷ distance Earth: 12.8 µrad (1.4 Voyager pixels) you, 0.5 m: 0.0000005 µrad, needs D = 1.22 λ ÷ θ ≈ 1,342 km Airy disc: the smallest spot any lens makes
The geometry of the impossible selfie. Distance shrinks the angle; diffraction sets the smallest angle a given aperture can split.

Could a telescope array the size of Earth do it?

An Earth-sized array gets you to about 100 m resolution at a billion kilometres in radio, which is still 200 times too coarse. Linking dishes across the planet is called very long baseline interferometry, and it is how the Event Horizon Telescope photographed a black hole.

Calculated in this post from the Event Horizon Telescope's 1.3 mm wavelength, a baseline equal to Earth's 12,756 km diameter, the billion-kilometre distance and the 0.5 m patch of this what-if.

Interferometry replaces one giant mirror with many small ones spread over a baseline \( B \); the resolution scales roughly as \( \lambda/B \). It does not collect as much light as a filled mirror, but it splits angles as finely.

  1. EHT's numbers. Wavelength 1.3 mm, baseline up to about Earth's diameter, \( B \approx 1.28\times10^{7} \) m:
    \[ \theta \approx \frac{\lambda}{B} = \frac{1.3\times10^{-3}}{1.28\times10^{7}} \approx 1.0\times10^{-10}\ \text{rad} \approx 21\ \mu\text{as} \]
  2. At a billion km: \( 1.0\times10^{-10} \times 10^{12} \approx 100 \) m. Enough to see a stadium, not a person.
  3. An optical array at 550 nm needs \( B = \lambda/\theta = 5.5\times10^{-7}/5\times10^{-13} = 1{,}100 \) km of baseline, with the light from every telescope combined to a fraction of a wavelength. That has been done over hundreds of metres on the ground, not over a thousand kilometres in space.
First M87 Event Horizon Telescope Results. IV. Imaging the Central Supermassive Black HoleThe Event Horizon Telescope Collaboration · The Astrophysical Journal Letters 875, L4 · 2019

The collaboration combined radio dishes across the planet, observing at 1.3 mm in April 2017, into one virtual instrument with baselines up to nearly Earth's diameter. The resulting image showed a ring about 40 microarcseconds across around the black hole in galaxy M87. It is the best demonstration that an Earth-sized aperture can split angles of tens of microarcseconds. The limit for our purposes: it works in radio with sparse coverage, and the same trick at visible wavelengths would need optical paths matched to nanometres.

Read the paper (free)
A line of large white radio dish antennas of the Very Large Array in New Mexico
Dishes of the Very Large Array in New Mexico. Arrays like this trade one giant mirror for many small ones spread out.Photo: The Modern Polymath, CC BY-SA 4.0, via Wikimedia Commons

Is there enough light to make the picture?

Surprisingly, yes. Light is not the bottleneck; resolution is. A 1,342 km mirror would collect about a hundred million photons a second from a sunlit 0.5 m patch of you.

  1. Sunlight on you. NASA's Earth fact sheet gives solar irradiance \( S = 1{,}361 \) W/m². Take a patch \( A = 0.25 \) m² with albedo \( a = 0.3 \) (my estimate for skin and clothes): reflected power \( P = S A a \approx 102 \) W.
  2. Into the telescope, treating you as a Lambertian reflector seen from overhead:
    \[ P_{\text{tel}} = \frac{P}{\pi}\cdot\frac{\pi (D/2)^2}{d^2} = \frac{102}{\pi}\times\frac{1.41\times10^{12}}{10^{24}} \approx 4.6\times10^{-11}\ \text{W} \]
  3. Photons at 550 nm carry \( hc/\lambda = 3.6\times10^{-19} \) J each, so about \( 1.3\times10^{8} \) photons per second. A short exposure would do.
From a billion kilometres, light is not the problem. The width of the glass is.

The catch is that only a giant aperture collects that much. Shrink the telescope to something buildable and you lose both the light and the resolution.

What is the version that would actually work?

Stop trying to see yourself and shine instead. A powerful laser aimed at the spacecraft turns you into a point source that can be brighter than the sunlit patch of Earth around you, even when the camera can't resolve anything smaller than a city.

This isn't science fiction. NASA's Deep Space Optical Communications experiment used a 7 kW uplink laser at Table Mountain to send light to the Psyche spacecraft, reaching about 460 million km in July 2024.

  1. Assumptions (this what-if's premises). Spacecraft camera with a 1 m aperture (\( A_t = 0.785 \) m²) at \( d = 10^{12} \) m. Ground laser at 532 nm, 7 kW, with the beam spread to \( \phi = 5\ \mu\text{rad} \) by the atmosphere (my estimate).
  2. Laser power received: spot radius \( \phi d/2 = 2.5\times10^{6} \) m, so
    \[ P_L = \frac{7000}{\pi (2.5\times10^{6})^2} \times 0.785 \approx 2.8\times10^{-10}\ \text{W} \]

    Calculated in this post from the 7 kW uplink power NASA JPL gives for the Table Mountain laser and this what-if's premises: a 1 m camera, a 5 µrad beam and a billion-kilometre distance.

  3. Earth's glare in the same resolution element. The 1 m camera resolves \( 1.22\lambda/D = 0.65\ \mu\text{rad} \), a 650 km patch of ground. Sunlit at albedo 0.3:
    \[ P_{bg} = S\,a\,A_p\,\frac{A_t}{\pi d^2} = 1361 \times 0.3 \times 4.2\times10^{11} \times \frac{0.785}{\pi\times10^{24}} \approx 4.3\times10^{-11}\ \text{W} \]

    Calculated in this post from the Rayleigh formula 1.22λ/D (Wikipedia), the solar irradiance of 1,361 W/m² and Earth's Bond albedo of 0.294 (NASA Earth Fact Sheet), rounded to 0.3, and this what-if's 1 m camera and billion-kilometre distance.

  4. Contrast. \( P_L/P_{bg} \approx 6.5 \) with no filter at all. A narrow green filter would cut the broadband sunlight by a further factor of hundreds while passing the laser.
  5. Sensitivity. Drop to a 1 W laser and the beacon is 7,000 times fainter, below the glare without a filter. Double the distance and both the laser and the background fall by four, so the contrast holds; only the photon count drops.

    Calculated in this post as the ratio of the 7 kW uplink power (NASA JPL) to a 1 W laser, which is this what-if's premise.

each square = one 650 kmresolution element (1 m camera) dark square: your 7 kW laser,about 6.5 times brighter thanthe sunlit ground around it you are not resolved;you are detected
Resolution versus detection. The camera can't see you, but it can see a light you control, exactly as stars are detected without being resolved.
Alien Maps of an Ocean-Bearing WorldNicolas B. Cowan, Eric Agol, Victoria S. Meadows and others · The Astrophysical Journal 700, 915–923 · 2009

The team pointed NASA's Deep Impact spacecraft (the EPOXI mission) at Earth for two full days and recorded the planet as a single unresolved point in seven wavebands from 300 to 1,000 nm. As Earth rotated, its brightness changed by 15–30%, and two "eigencolours" explained 98% of the colour change; mapped back, they matched oceans and cloud-free continents. It shows how much an astronomer can learn from one pixel of Earth, which is what a billion-kilometre selfie really is. The limit: two days of data at gibbous phase, and a much nearer, brighter Earth than an exoplanet.

Read the paper (free)
A Voyager spacecraft with its high-gain antenna dish undergoing testing in a clean room at JPL in November 1976
A Voyager spacecraft under test at JPL in 1976. Its narrow-angle camera, 176 mm across, took the Pale Blue Dot 14 years later.Photo: JOHN GREGOIRE, NASA/JPL, Public domain, via Wikimedia Commons

Calculate your own deep-space selfie

Use this to check what you just read. Set the distance and the smallest thing you want to see.

Selfie from deep space

So, can you take a selfie from a billion kilometres?

A selfie in which you appear as a face is ruled out by diffraction for any instrument we could build. A selfie in which you appear as a point of light is within reach of hardware that already exists.

Verdict: possible with changes

Resolving you is out: the Rayleigh criterion needs a 1,342 km aperture for a 0.5 m pixel and about 67,000 km for a recognisable face, while Voyager's 176 mm camera blurs to 3,800 km. Change the goal from being seen to being detected: point a 7 kW ground laser, the same class NASA used for Psyche, at a 1 m camera a billion km out, and your spot is about 6.5 times brighter than the 650 km of sunlit Earth around it, before any filter. You become the brightest pixel on the pale blue dot.

Calculated in this post from the Rayleigh formula 1.22λ/D (Wikipedia), Voyager's 0.1765 m aperture (NASA PDS), the 7 kW uplink power (NASA JPL) and the premises of this what-if (a 0.5 m pixel, a 1 m camera, a billion kilometres).

For the lens side of this, read how focal length and compression work, whether megapixels matter once diffraction kicks in, and how phone cameras compute detail they can't optically resolve. More from the studio at Kala.

What if next?

Questions people ask

How far away was the Pale Blue Dot photo taken?

Voyager 1 took it on 14 February 1990 from about 6 billion km (NASA's figure), using its narrow-angle camera and violet, blue and green filters. Earth was a crescent just 0.12 pixel in size.

Can you see people in the Pale Blue Dot?

No. Earth is less than one pixel, so every person alive in 1990 is inside a fraction of a single pixel. Even in Cassini's 2013 "wave at Saturn" image, NASA said no individual could be distinguished.

What is the most distant photo of Earth?

The Pale Blue Dot, at about 6 billion km, is still the most distant photograph of Earth. Cassini's 2013 image from 1.45 billion km and Psyche's 2025 image from 290 million km are much closer.

Why can't a better sensor fix it?

Because the blur comes from the aperture, not the sensor. The Rayleigh criterion, θ = 1.22λ/D, sets the smallest angle a lens can split; more pixels only sample the same blur more finely.

Sources

  1. NASA Science: Solar System Portrait, Earth as Pale Blue Dot (1990) (0.12 pixel crescent, narrow-angle camera, three filters, 32° above the ecliptic)
  2. NASA Science: Voyager 1's Pale Blue Dot (14 Feb 1990, about 6 billion km)
  3. USGS Astrogeology: Voyager wide and narrow angle cameras (camera overview; the per-pixel figures in this post come from the NASA PDS Voyager 1 catalogue linked in the text)
  4. The Planetary Society: The Day the Earth Smiled (2013) (1.445858 billion km, 86,620 km per pixel)
  5. NASA JPL: Cassini releases image of Earth waving at Saturn (2013) (1,400+ images, 40+ countries, 30 states)
  6. NASA JPL: NASA's Psyche captures images of Earth, Moon (2025) (290 million km, July 2025)
  7. NASA NSSDCA: Earth Fact Sheet (equatorial radius 6,378 km, solar irradiance 1,361 W/m²)
  8. EHT Collaboration: First M87 Event Horizon Telescope Results IV (2019) (1.3 mm, Earth-sized baseline, 40 µas ring)
  9. Cowan et al.: Alien Maps of an Ocean-Bearing World (2009) (Earth as one pixel, 15–30% diurnal variation)
  10. NASA JPL: Laser comms demo makes deep space record (2024) (7 kW uplink, 460 million km)
  11. EarthSky: Voyager 1 one light-day from Earth (2026) (25.9 billion km, 18 Nov 2026)
  12. The Planetary Society: The pale blue dot 35 years later (2025)
Pale Blue DotVoyager 1Cassinidiffraction limitRayleigh criterionEvent Horizon Telescopeinterferometrylaser communication
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Artblog.

Keep reading in Artblog

From the other blogs