mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Astronomy (https://www.mersenneforum.org/forumdisplay.php?f=142)
-   -   James Webb Space Telescope (https://www.mersenneforum.org/showthread.php?t=27092)

xilman 2022-07-15 17:23

[QUOTE=retina;609572]You necessarily have to throw away information to "fix" the images.

If you only care about the images being used for artistic purposes then go ahead and mangle them to your hearts content. But if you care about the science then you won't retouch them just to make them pretty.[/QUOTE]Your first two statements are undoubtedly true.

The third is open to interpretation. By re-arranging the information you have you can make parts of that more prominent at the expense of the other parts. The now more prominent information can be of significant scientific interest. For an example of useful deconvolution, see [url]https://britastro.org/observations/observation.php?id=20181109_015500_a5089659d4038718[/url] and [url]https://britastro.org/observations/observation.php?id=20181111_015500_44cd9ab1092010e3[/url]

kriesel 2022-07-15 18:06

[QUOTE=retina;609572]You necessarily have to throw away information to "fix" the images.[/QUOTE]No. Always preserve the raw data, and with good backups. I would argue that ignoring the diffraction transform is throwing away data and resolution.
[quote]if you care about the science then you won't retouch them just to make them pretty.[/quote]It's not subjectively retouching them that I proposed, it's generating a scientifically justified corrected image, more accurate than the raw observation data, compensating for as much of the diffraction transform error in a copy of the observed data as practical. One could use observations of certain well known objects to generate reference corrections across the JWST field of view. See [url]https://en.wikipedia.org/wiki/Point_spread_function#Astronomy[/url]
Where I used to work, we would commonly calibrate certain custom-ordered instruments we produced, as well as measure reproducibility, to small portions of a microradian, over a range of a significant portion of a radian. (Resolution/repeatability to as good as ~1 in 2 million.) And we accomplished this with room air in the path of the calibration optics. That takes some care. Including knowing the errors in the calibration hardware.

xilman 2022-07-15 20:14

[QUOTE=kriesel;609583]No. Always preserve the raw data, and with good backups. I would argue that ignoring the diffraction transform is throwing away data and resolution.[/QUOTE]
We are in violent agreement! That is what I do. It is also consistent with what I stated. If you check out the posted links you will see that they are two different ways of retouching the original data. If you wish, I can send you the raw data for you to reprocess and you may well do better than I. My PSF was a first-look and the deconvolution algorithms chosen because they are easily implemented rather than optimal by some other metric. The metadata lives in a PostgreSQL database and the raw bits on the backup-server here.

In principle, and I have no evidence to back up this claim, there might be scientifically valuable information contained within the diffraction spikes which is not currently evident but may be obvious after deconvolution.

kriesel 2022-07-15 22:15

Re Hubble image processing: [url]https://esahubble.org/static/projects/seminars/presentations/schirmer_astroprocessing.pdf[/url]
Since JWST images 0.6 microns to ~28 microns, anything not red or red-orange is false color for certain.
[url]https://jwst.nasa.gov/content/about/faqs/facts.html[/url]

retina 2022-07-15 22:24

[QUOTE=kriesel;609592]Since JWST images 0.6 microns to ~28 microns, anything not red or red-orange is false color for certain.[/QUOTE]Of course. The images would be quite boring presented in real colour, since humans can't see the longer wavelengths.

Here is the real colour version of the image from the EHT.
[size=9]██[/size]

tServo 2022-07-19 20:02

Webb's first deep field image
 
Among the images released in the first batch was one that showed an absolute riot of galaxies.
This space.com's article explains what we are looking at.
[URL]https://www.space.com/james-webb-space-telescope-deep-field-explained?utm_source=notification[/URL]

Uncwilly 2022-07-19 21:39

Technically it wasn't even a deep field. It was only ~12 hours. A deep field would need to be about 4 x 24 = 96 hours. Once we start seeing those (and I don't expect one until after the new year) the astronomers are going to lose their minds. Also, once JWST does a collaboration with scopes working in other parts of the spectrum, papers will be flying out like the output of a line printer.

chalsall 2022-07-19 22:55

[QUOTE=Uncwilly;609870]Technically it wasn't even a deep field. It was only ~12 hours. A deep field would need to be about 4 x 24 = 96 hours. Once we start seeing those (and I don't expect one until after the new year) the astronomers are going to lose their minds.[/QUOTE]

Completely agree. Wow! What a wonderful time to be an observer!

[QUOTE=Uncwilly;609870]Also, once JWST does a collaboration with scopes working in other parts of the spectrum, papers will be flying out like the output of a line printer.[/QUOTE]

Also completely agree. Including the engineering stuff. One mirror has already taken a signifgant "hit".

Everything is a learning experience.

retina 2022-07-28 01:02

[QUOTE=retina;609392]Why do the images have 8 spikes?

I can understand the 6 larger spikes. What are the other 2 smaller spikes caused by?[/QUOTE]Found it.

[youtube]cWXTy_GeCis[/youtube]
[url]https://www.youtube.com/watch?v=cWXTy_GeCis[/url]


All times are UTC. The time now is 16:55.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.