M51 Through Three Telescopes: DWARF 3, Hubble and JWST

M51 was the first target I pointed the DWARF 3 at, on August 12, 2025. Since then it has collected 8 hours and 14 minutes of my backyard sky across three sessions. This project puts that stack next to two telescopes I will never own: Hubble and the James Webb Space Telescope.

All three datasets went through Siril with the same two stretch commands. One DWARF 3 pixel covers the same sky as about 89 x 89 JWST pixels. The three pages below show what that difference looks like, step by step.

The three datasets at a glance

DWARF 3HubbleJWST
Aperture35 mm2.4 m6.5 m
CameraTelephoto, Sony IMX678ACS Wide Field ChannelNIRCam
Light recordedVisibleBlue to near infraredInfrared
FiltersAstro filter (color sensor)F435W, F555W, F814W, F658NF115W, F200W, F335M
WhenAug 2025 to Jun 2026, three sessionsJanuary 2005June 14, 2023
Exposure8 h 14 min2720 s (F435W), 1360 s (others) per pointing2920 s per filter
Pixel scale2.79″0.05″0.0312″ (F335M: 0.0629″)
Final image3564 x 1960 px7613 x 11521 px11249 x 4109 px
Sky coveredabout 166 x 91 arcminabout 6.3 x 9.6 arcminabout 5.9 x 2.1 arcmin
Where fromMy own sessionsMAST, program 10452MAST, program 1783 (FEAST)

The DWARF 3 frame covers about 1200 times more sky than JWST in this image, and JWST sees about 89 times finer detail in each direction. Hubble sits between them on both counts.

Side by side

M51 inner disk from the DWARF 3, Hubble and JWST, top to bottom, all placed on the same JWST pixel grid
Top to bottom: DWARF 3, Hubble, JWST. All three are on the JWST pixel grid, so each panel shows the same patch of sky, about 5.9 x 2.1 arcminutes.
Data: DWARF 3, AK; Hubble: NASA, ESA, Hubble Heritage Team (STScI/AURA); JWST: NASA, ESA, CSA, STScI. Processing: AK.

To compare the three directly, each final image was placed onto the JWST pixel grid using the sky coordinates stored in the files. After that, the same pixel in all three shows the same point on the sky; a check after the alignment measured an offset of 0 pixels.

Two things to keep in mind when you look at it:

  1. The DWARF 3 image in the comparison is the comparison version from its page: stretched with the same commands as the other two, with no denoise, no sharpening and no AI tools.
  2. Putting the DWARF 3 image onto the JWST grid enlarges it about 89 times. That makes it the same size on screen, but enlarging adds no detail. Each DWARF 3 pixel turns into a patch of about 89 x 89 JWST pixels.
The same spiral arm region of M51 from the DWARF 3, Hubble and JWST, left to right
One section of a spiral arm, 2000 x 2000 JWST pixels or about 62 x 62 arcseconds. Left to right: DWARF 3, Hubble, JWST. The DWARF 3 covers this area with about 22 x 22 of its own pixels.
Data: DWARF 3, AK; Hubble: NASA, ESA, Hubble Heritage Team (STScI/AURA); JWST: NASA, ESA, CSA, STScI. Processing: AK.

One recipe for all three

A linear astronomy file looks almost black, because the faint parts of the galaxy are only slightly brighter than the sky. Every page uses the same two commands to fix that:

  1. asinh, a curve that lifts faint values strongly and bright values only a little, so the faint disk comes up while the core keeps its structure.
  2. autostretch -2.8 0.08, which sets the black point 2.8 noise levels below the sky and brings the sky to 8% brightness.

The asinh strength comes from one rule: strength = 0.025 / (faint disk level minus sky level).

The gap between faint disk and sky is very different for each telescope, so the strength is too:

DWARF 3HubbleJWST
Asinhasinh 10 0.1104asinh 28000 (H alpha: 225000)asinh 1000
Black point in asinh0.1104, just under the skynonenone
Autostretch-linked -2.8 0.08-2.8 0.08, unlinked-2.8 0.08, unlinked
Why linked or notOne color file, already color calibrated with SPCCOne file per filter, no color calibrationOne file per filter, no color calibration

The target of 0.08 is the same on all three pages, so the three skies end up at the same brightness and the images can sit next to each other.

The DWARF 3 file needs a black point in the asinh command because its sky carries about 7200 counts of light pollution and camera offset. The space telescopes have no light pollution, and autostretch alone sets their black point.

What each telescope adds

The DWARF 3 is the only one of the three that shows M51 in its surroundings: the whole galaxy, its companion NGC 5195 and the field of foreground stars, in one frame. It is limited to detail a few arcseconds across, set by its 2.79″ pixels.

Hubble covers both galaxies at 0.05″ per pixel, in visible light and near infrared. Its H alpha filter records the glowing hydrogen clouds along the arms, where new stars are forming.

JWST covers a band across the inner disk at 0.0312″ per pixel, all in infrared. Its F335M filter picks up the 3.3 micrometer glow of PAH molecules in dust, which shows where the dust and the star forming regions are.

The three pages

Read them in this order. The DWARF 3 page has the most steps and explains the stretch in the most detail; the other two build on it.

  1. DWARF 3 page: an 8 h 14 min stack, cropped, gradient removed, color calibrated with SPCC and stretched. Two versions: a comparison version with no AI, and a showcase version with deconvolution, StarNet, GHS and Cosmic Clarity.
  2. Hubble page: the Hubble Heritage ACS mosaic, four filters cropped to one box, stretched and combined, with H alpha blended into red.
  3. JWST page: three NIRCam filters aligned from their sky coordinates, stretched and combined, with SCNR.

Each page ends with the full Siril command list.

Frequently Asked Questions

Can a DWARF 3 image be compared fairly with Hubble and JWST?

On processing, yes: the same commands and the same sky target were used on all three. On detail, no: the pixel sizes differ by up to 89 times, and that limit is in the data.

Is the Hubble or JWST image here the official release?

No. Both are my own processing of the public archive files, using only the steps on their pages.

Do I need my own telescope to try this?

No. The Hubble and JWST files are free to download from MAST, and Siril is free. Only the DWARF 3 page needs your own data.

Why is the JWST image so narrow?

The NIRCam mosaic in this observation covers a band across the inner disk, about 5.9 by 2.1 arcminutes, not the whole galaxy.

Credits

Hubble: based on observations made with the NASA/ESA Hubble Space Telescope, obtained from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. These observations are associated with program 10452. Data: Hubble Space Telescope multi-color ACS mosaic of M51, a High Level Science Product by the Hubble Heritage Team (PI S. V. W. Beckwith), retrieved from MAST, DOI 10.17909/T9GP4C. Image credit: NASA, ESA, Hubble Heritage Team (STScI/AURA).

JWST: based on observations made with the NASA/ESA/CSA James Webb Space Telescope. The data were obtained from the Mikulski Archive for Space Telescopes at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-03127 for JWST. These observations are associated with program 1783 (FEAST, PI A. Adamo). Image credit: NASA, ESA, CSA, STScI.

The Hubble and JWST images here are my own processing of the archive files, not official releases, and are not endorsed by NASA, ESA, CSA or STScI.

DWARF 3 data, all processing and text: AK.