Pages: [1]   Go Down

Author Topic: Question about image quality  (Read 2738 times)

wmchauncey

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 793
Question about image quality
« on: March 08, 2016, 10:45:28 am »

I opined on another forum that by using various PS photomerging techniques, one could obtain 50 MP images by using a 22 MP camera, Canon 1Ds3.

The thread progressed as expected until the term "Nyquist limit resolution" was introduced. 
It was indicated that my gear produces about 4.7 LP/mm, whereas the new 5Ds coughs out 7.1 LP/mm.

My question...does the term "Nyquist limit resolution" have any relationship to IQ?
Logged
The things you do for yourself die with

Telecaster

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3686
Re: Question about image quality
« Reply #1 on: March 08, 2016, 02:26:22 pm »

My question...does the term "Nyquist limit resolution" have any relationship to IQ?

Yes, it sets a hard limit on how much genuine spatial detail your camera/lens can resolve. Of course spatial detail isn't the be-all and end-all of image quality. In your example the Nyquist limit isn't likely a practical concern: you can build your composite from more pics taken with longer focal lengths and end up with however many pixels you want—up to your long lens limit—while avoiding spatial clipping (aliasing).

-Dave-
Logged

BrianVS

  • Full Member
  • ***
  • Offline Offline
  • Posts: 163
Re: Question about image quality
« Reply #2 on: March 08, 2016, 04:12:43 pm »

> It was indicated that my gear produces about 4.7 LP/mm, whereas the new 5Ds coughs out 7.1 LP/mm.

What was this calculation based on? It seems off by a factor of 10. I can understand 47 LP/mm and 71LP/mm, figure you need two pixels minimum to resolve a line.
« Last Edit: March 08, 2016, 07:08:58 pm by BrianVS »
Logged

wmchauncey

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 793
Re: Question about image quality
« Reply #3 on: March 08, 2016, 04:58:46 pm »

Quote
What was this calculation based on
Don't have a clue...is that Nyquist thing even important in terms of IQ?
Logged
The things you do for yourself die with

BrianVS

  • Full Member
  • ***
  • Offline Offline
  • Posts: 163
Re: Question about image quality
« Reply #4 on: March 08, 2016, 07:13:49 pm »

Nyquist means you need at least 2 samples to resolve a line. You really want more than that- at least 10% more samples to really see anything. I figure a 1930s Sonnar is a perfect match for a Leica M Monochrom.

Sonnar_Test_Modern_Photo by fiftyonepointsix, on Flickr
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13723
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Question about image quality
« Reply #5 on: March 08, 2016, 07:25:37 pm »

The answer is that there is no direct relationship in your case.

When you stitch, you virtually increase the size of your sensor, simulating what you can get with a medium format or even with a - yet to exist - large format sensor.

I have 2 meters large stitches made from a 24mp d3x + Zeiss 100mm f2.0 that are in the 300~400 mp and leave very far behind any print I have seen from single captures done with top notch medium format backs.

Needless to say, much better can be done nowadays with the latest sensors from 5Ds/a7rII/D810 with the best glass such as the Otus lenses, but your 1dsIII is still a very good camera and stitching can bring it ahead in image quality compared to any of the latest cameras in the hands of a lazy person not interested in stitching. ;)

Cheers,
Bernard

BrianVS

  • Full Member
  • ***
  • Offline Offline
  • Posts: 163
Re: Question about image quality
« Reply #6 on: March 09, 2016, 05:47:58 am »

If you take an image and then shift the sensor by half a pixel, acquire an image, then merge them- you are effectively doubling the sampling in the direction of the shift.
Logged

wmchauncey

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 793
Re: Question about image quality
« Reply #7 on: March 09, 2016, 10:08:14 am »

It sounds as if you guys are saying that if I take a series of images using longer glass and photomerge them to the same field of view of an
 image captured using a high MP camera, assuming my merging skills are up to task...my merged image will be of equal or greater quality.

Is that right?
Logged
The things you do for yourself die with

kers

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3537
    • Pieter Kers
Re: Question about image quality
« Reply #8 on: March 09, 2016, 11:16:47 am »

If you take an image and then shift the sensor by half a pixel, acquire an image, then merge them- you are effectively doubling the sampling in the direction of the shift.

Do you mean, just make a picture - shift 0,0024mm and take another one- merge them together.... :) ?
my hands shake too much
-
It sounds as if you guys are saying that if I take a series of images using longer glass and photomerge them to the same field of view of an
 image captured using a high MP camera, assuming my merging skills are up to task...my merged image will be of equal or greater quality.
Is that right?

yes, but taking 1 image only has some unique advantages, for instance shooting moving subjects... etc


That said, i must say i am happy where i am now- with a Nikon d810 and some good lenses...
To have 36MP sharpness on a moving subject is often impossible and usually not necessary.
« Last Edit: March 09, 2016, 11:22:44 am by kers »
Logged

BradSmith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 735
Re: Question about image quality
« Reply #9 on: March 10, 2016, 02:04:08 pm »

It sounds as if you guys are saying that if I take a series of images using longer glass and photomerge them to the same field of view of an
 image captured using a high MP camera, assuming my merging skills are up to task...my merged image will be of equal or greater quality.

Is that right?

The stitched image will be of higher resolution.  That's one of the main reasons why people stitch images.    Doesn't a 42 megapixel sensor yield a "greater quality (resolution)" image than a 12 megapixel sensor?  Similarly, a stitched image that ends up with 4 or 5 times as many pixels as a single exposure yields a "greater quality (resolution)". 
Brad
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11303
    • Echophoto
Re: Question about image quality
« Reply #10 on: March 12, 2016, 03:50:03 pm »

Hi,

The Nyquist limit defines the boundary between resolution or fake detail. If your lens resolves beyond Nyquist (*) fake detail will result. Most cameras limit resolution using an OLP filter. The sole purpose of that filter is to limit resolution of the lens to match the resolution of the sensor.

Stitching will generally increase resolution. It is essentially the same as enlarging the image area. Let's assume that you shoot with a 20 MP 24x36 camera. Now, put it in the vertical position and shoot three images with the sensor shifted -15, 0, and +15 mm. The image area is now (15+15+24) x 36 mm. That is 54x36 mm. So your 20 MP camera now takes 54/24 * 20 -> 45 MP images.

Best regards
Erik

I opined on another forum that by using various PS photomerging techniques, one could obtain 50 MP images by using a 22 MP camera, Canon 1Ds3.

The thread progressed as expected until the term "Nyquist limit resolution" was introduced. 
It was indicated that my gear produces about 4.7 LP/mm, whereas the new 5Ds coughs out 7.1 LP/mm.

My question...does the term "Nyquist limit resolution" have any relationship to IQ?
Logged
Erik Kaffehr
 
Pages: [1]   Go Up