Aligning the JWST - why?

[DaveC426913]

TL;DR Summary

Why align them physically? Why not softwarily as needed?

When I'm Photoshopping or film editing or 3D modeling, , I wait until the last possible moment to combine anything. Once combined, the individual data would be lost.

Why bother physically combining the images from JWSTs segments? Why not receive the data as 18 channels, store them separately and stack them digitally at our leisure in the air conditioned comfort of our offices? Then we can pick and choose - and even replace - our stacking algorithms.

It seems to me that it's the equivalent of flattening all my .PSDs into GIFs before saving. Or throwing away my .BLEND files and only keeping the exported .STLs.

I can see an 18-fold reduction in bandwidth, but otherwise...

 

[Drakkith]

The segments operate together as a single larger mirror to create a single image. For starters, this increases resolution because a larger mirrors allows for a smaller diffraction pattern (airy disk) at the image plane which generates a higher quality image. It also increases SNR of the images compared to simply stacking 18 separate exposures, as the camera sensor itself adds noise to each image. With a single exposure you only have a single round of noise generation, not 18.

 

[Ibix]

Because the imaging sensor only records intensity, not phase information. That means that you cannot post-combine the eighteen images with the same level of detail you get from optically combining them (you can only add $|A(x,y)|^2$, not $A(x,y)\sin(\phi(x,y))$). As Drakkith says, you effectively get eighteen small telescopes instead of one big one.

In radio frequency our electronics are fast enough to record phase, so we can combine telescope images from separate telescopes. That's what Very Long Baseline Interferometry is.

 

[DaveC426913]

I don't doubt you, I just don't grok it.

Surely:

• Anything that can be optimised physically at the start can be better optimized digitally downstream.
• Noise in 18 separate images is the way to know what's noise and what isn't. That's why we stack separate images in astrophotog.

No?

 

[DaveC426913]

Because the imaging sensor only records intensity, not phase information. That means that you cannot post-combine the eighteen images with the same level of detail you get from optically combining them...

Ah! The phase.

But why can't you sync the signals like we do with Very Lon...

In radio frequency our electronics are fast enough to record phase, so we can combine telescope images from separate telescopes.

Ah! Because radio waves have way low freqs. Got it!

 

[Drakkith]

Anything that can be optimised physically at the start can be better optimized digitally downstream.

Absolutely not. In fact the reverse is true. Anything you can optimize physically is better than digitally. Unless it costs too much. Hence why we don't have giant radio antennas that are kilometers across. They'd be better than moving smaller telescopes during aperture synthesis, but they are far, far too large to be practical.

Noise in 18 separate images is the way to know what's noise and what isn't. That's why we stack separate images in astrophotog.

Unfortunately stacking doesn't solve the noise issue. Your SNR will increase roughly proportional to the square root of the number of images stacked. So with 18 images your SNR will be about 4x better than the single images. But a single image with 18x the light has an SNR that's only affected by one round of sensor noise and isn't subject to the square root rule (in regards to sensor noise only). If there were no other sources of noise in the image then it would be up to 18x better in SNR than a single exposure from an individual mirror segment. In practice the gains aren't as great because of other noise sources of course.

I hope all that makes sense. I'm very tired and about to go to bed.

 

[BillTre]

Absolutely not. In fact the reverse is true. Anything you can optimize physically is better than digitally. Unless it costs too much.

This is why in the movies, physical effects (like Ridley Scott often uses) often look better than the computer generated effects.