I've circled them in this edit:

Source: https://commons.m.wikimedia.org/wiki/File:New_Horizons_Approach_to_Arrokoth.ogv

Note, this has generated frames, see the source. The twinkling stars are regular though, almost taking turns to twinkle.

I know it is sped up, are those variable stars, that vary that much regularly over the course of hours and days? If so wow, how many stars are like that?

Are there any good videos that show the sky in time lapse demonstrating what it would look like to see the stars twinkle like that?

  • $\begingroup$ Very cool! Different but slightly related (and not the answer) Which are stars and which are noise in this comet photo? When it approached the exposure times were much shorter, so the stars didn't streak or creep. If one chooses playback speed of 0.25 in YouTube they do interesting things! I'm wondering if some frames of the video are interpolated and zoomed to make the growth of the size of the asteroid smooth in the final video. $\endgroup$
    – uhoh
    Mar 24, 2023 at 12:40
  • 1
    $\begingroup$ I hear that. And very interested thank you. I will try to make a GIF of the stars later. They flash very regularly, and are changing position with everything else, seems real. Would like a direct analysis. $\endgroup$
    – Rabbi Kaii
    Mar 24, 2023 at 12:56
  • $\begingroup$ Side question, why did they "zoom" in? Just to create the feeling of movement? $\endgroup$
    – Rabbi Kaii
    Apr 10 at 19:31

2 Answers 2


I got a still from the video and uploaded it to nova.astrometry.net. This successfully identified the field and provided an astrometrically calibrated image.

The two stars are:

The fainter, large amplitude variable object (on the right hand side) is at 18h18m48.2s -20d50d05s (J2000) and appears to coincide within an arcsecond with 2MASS J18184819-2050058, which has been identified as a Mira variable in the OGLE survey and in Gaia DR3. It has a pulsation amplitude of 1.11 mag (G-magnitude, quite low for a Mira variable) and a pulsation period of about 290 days (quite typical for a Mira variable). However, I am puzzled by the fact that in the video you show, which is taken over the course of around 4 weeks, this source appears to pulse 7 times, which would make the period more like 3-4 days...? Thus I am not convinced that this star is the source of this rapid variability and indeed, high amplitude periodic variables with periods of a few days are quite unusual.

The brighter, but lower amplitude variable object (on the left hand side) is at 18h18m49.8s -20d50m08s and appears to coincide with 2MASS J18184976-2050077, which has not been noted as anything special, though probably is the optical counterpart to a Chandra X-ray source (CXOPTM J181849.7-205007 - see Pravdo et al. 2004). Given that this was found in an X-ray survey of a star forming region, it is quite likely that this is a variable young protostar. The timescale for variability can be anything from hours to decades for such objects and is associated with changes in accretion and/or the rotational modulation of bright and dark spots.

As to "how many stars are like that"? Well, the empirical data are in front of you! The circle you have drawn contains about 20 stars as bright or brighter than the varying objects. The circle occupies a few per cent of the area of the image. The two objects you've noticed appear to be the only ones that I can discern varying. That is not to say that none of the other stars are variable; just that if they vary it must be at lower amplitude or much longer timescales. Thus the answer to the question would appear to be about 0.3%, with a large error bar.

Caveats to that:

  1. It is actually more complicated than that because the stars are at different distances in the image. Both protostars and Mira variables (which are red giants) are much more luminous than the vast majority of stars, so can be seen to greater distances. They are thus over-represented in any image of the sky. Thus the 0.3% figure is a big over-estimate of the true proportion but could be a realistic figure for the proportion of stars that vary in this way at a given apparent brightness.

  2. If both of these variable objects turn out to be protostars, then because star formation tends to be concentrated in certain regions of the galaxy (including this one), then you are much more likely to see variable stars when looking towards those regions.

  3. Essentially all stars are variable at some level. The number above applies to stars that vary in that way.

  • 1
    $\begingroup$ Kudos for successful use of astronomy.net! $\endgroup$
    – uhoh
    Apr 11 at 2:00
  • $\begingroup$ You rock! Thanks for spending so much time, it is an honour to pass some bounty points and a green tick to you ProfRob $\endgroup$
    – Rabbi Kaii
    Apr 15 at 8:05

I followed the link and viewed the video there expanding the window size. The video is quite low resolution and the stars in that area were ‘twinkling’ because they were passing from one pixel to another in the CCD camera. Others throughout the view but they weren’t as noticeable as these, not sure why that was, but it was all due to the large pixels in relation to the star points.

  • $\begingroup$ There's no way the camera for this mission would have a pixel spacing (or for that matter, dead area between pixels) so large compared to the diffraction-limited images that they could "fall through the cracks" between pixels and cause twinkling. This is uninformed speculation. *Welcome to Stack Exchange! SE is unlike most other sites, answers need to support their assertions with authoritative citations and/or references and/or calculations. These images are taken by LORRI $\endgroup$
    – uhoh
    Jan 21 at 4:07
  • 2
    $\begingroup$ which uses the Teledyne e2V CCD47-20 Back Illuminated NIMO Frame-Transfer High Performance CCD Sensor with 13 micron pixels and 100% active area because it's backside-illuminated. With a 20.8 cm aperture and 263 cm effective focal length, Rayleigh criteria diffraction limit at 550 nm is a little over half-pixel at 8.5 microns, so sometimes a star is mostly on a single pixel, and sometimes it's split across two, but with 100% active area via backside illumination there be no twinkling! $\endgroup$
    – uhoh
    Jan 21 at 4:14
  • 1
    $\begingroup$ Maybe whoever made this page took out a very small area of the original screen size expanding the size of that view and the pixels along with it, or something in the processing. The video didn't go full screen, but, it did get bigger than the embedded one on the page. I could see the pixels on the video and the blinking stars were just moving from one pixel to another. But, I don't think it indicated any type of variability in the stars themselves. $\endgroup$
    – Pogo
    Jan 21 at 4:47
  • 1
    $\begingroup$ I see two "twinklers" only, and they are almost the same distance from the center of the zoom but have very different "twinkling frequencies". There are plenty of variable stars out there, in fact it's hard to find a field of view that doesn't have some. The time lapse nature of this image sequence makes periods of hours or days easy to see. I don't think there's any need to speculate about exotic processing phenomenon when variable stars are so likely to be the explanation. $\endgroup$
    – uhoh
    Jan 21 at 5:03
  • 1
    $\begingroup$ See Table 2 in Variable stars across the observational HR diagram Something like 2-3% of stars are periodic variable. $\endgroup$
    – uhoh
    Jan 21 at 5:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .