Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Thread Tools Search this Thread
Old Jan 6, 2007, 4:24 PM   #1
Junior Member
Join Date: Jan 2007
Posts: 4

Astro photographers have long been using webcams etc. in conjunction with programs like Registax to capture highly detailed images of planets etc. For those of you not familiar with Registax, look at this link, and observe the difference in detail in the first (real) image and the last (aggregated / multi sampled) image: http://www.threebuttes.com/RegistaxTutorial.htm

The basic principle of multi sampling (or stacking or whatever the correct term might be) is that multiple images of the same exact subject taken at slightly different times or from slightly different viewpoints actually are not precisely identical. If you have a large number of such "near identical" images (say 100) you in fact have more detail available than any single one of these images contain by itself. By combining these images in Registax (or similar software) you can obtain a significantly more detailed image. (And as a bonus you get heavy lossless noise reduction.)

Then to my actual question:

Can this multi sampling technique be applied to regular (non astronomical) high resolution telephoto to enhance detail in photos?

Obviously the subject would need to be completely motionless (i.e. no animal photography). To simplify debate let's stick to one single example: You the photographer are standing on the ocean shore. A mile out in the ocean is a solid rock with an old lighthouse. You want to produce as detailed a photo of that lighthouse as is possible.

... To make the scenario more interesting lets say you've entered a competition where a crazy billionaire has invited the worlds foremost photographers to his mansion overlooking the shore. He is offering 100 million dollars to the first photographer able to read the text written on a sign posted on the lighthouse. The photographers are not allowed to step outside the mansion, and can not bring their own equipment. Each photographer is given an identical setup with a super stable tripod, a state of the art DSLR camera, and a cheap (!) 300mm telephoto lens (so obviously the lens is the limiting factor not the sensor), and a computer with any desired software installed. You each have 24 hours to your disposal. You step onto the balcony and mount the camera on the tripod. Zooming in on the lighthouse you easily spot the sign. It is a modern white sign with black lettering. You take lots of RAW pictures using various apertures, load them onto the computer, select the sharpest ones, and run various kinds of channel splitting and sharpening using your favourite post processing tool. After post processing you can almost read the blurred (certainly not jagged) letters, but only almost. The rules are that each photographer will get one single chance to recite the sign, and it must be correct by the letter, otherwice they are escorted out of the mansion straight away.
Can you win the 100 million dollars using multi sampling (or some other fancy trick)?

Any experience/testing you (or others) might have using multi sampling for terrestial purposes would be very interesting to hear about. As would explanations or theories concerning the inevitable limiting physical factors at play.

What other techniques might be available for enhancing (midtone) details beyond what is available through normal post processing?

alvevind is offline   Reply With Quote
Sponsored Links
Old Jan 6, 2007, 6:34 PM   #2
Senior Member
Join Date: Jun 2005
Posts: 804

Dude......interesting concept.

I've never heard of this sort of thing, but just from your description and the pix on the linked page, this seems something like Neat Image on steroids.

I can see where multiple images of a stationary object could allow a program to sort the noise from the actual detail. The noise is more random where the image detail is fairly stable. Multiple pix could well provide areas of consistent data (detail) that a program could map against inconsistent data (noise) to allow it to reinforce the consistent data and eliminate the inconsistent -- the result being image sharpening.

I don't really know if that's the way this program works, but you say that you use webcams to capture the data? What would happen if you hooked a really good digital video camera to the telescope? Or even a good-quality DSLR and shot multiple pix at high resolution?

You might do better at getting people with whom you can carry on an informed discussion in the Digiscoping section of these forums. Unless someone is conversant with the theory behind the techniques you mention, it's hard to speculate about your questions.

I hope that someone less ignorant than myself comes along, here!

granthagen is offline   Reply With Quote
Old Jan 6, 2007, 7:46 PM   #3
Senior Member
BillDrew's Avatar
Join Date: Jun 2002
Location: Hay River Township, WI
Posts: 2,512

Max Lyons (http://www.tawbaware.com/maxlyons/) has a program to do that. He does point out that it is possible to stack images without his program, but at the price he charges, I don't think it is worth the effort to figure out how. His site is well worth a visit - stacked image gallery at http://www.tawbaware.com/maxlyons/cg...e.pl?gallery=8
BillDrew is offline   Reply With Quote
Old Jan 7, 2007, 6:02 AM   #4
Junior Member
Join Date: Jan 2007
Posts: 4

I recently discovered another program that seems to do somewhat of the same that Registax does, but is less geared towards astronomy. I have not tested it yet but it looks very interesting.

Another thing I've come to suspect is that to apply this technique the optics must be of superior resolution compared to the sensor. So using a low quality lens on a high quality sensor might yield no real benefit (apart from generally reducing electronic noise inherent in the sensor).

My question then is: How do I determine wether the lens is the inferior component? Can I take a photo of a high contrast subject (like overlapping pieces of black and white paper) and simply visually examine the edge in high magnification; if the edges are jagged the lens is superior; if the edges are blurred the lens is inferior? Any practical rule-of-thumb guidelines for this?
alvevind is offline   Reply With Quote

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off

All times are GMT -5. The time now is 4:06 AM.