Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
12-09-2015, 05:41 AM   #46
New Member




Join Date: Sep 2012
Posts: 13
Original Poster
Ok, so, for people like Class A who are against the removal of the AA filter, found that the AA simulation system introduced by Pentax with the K-3 works fine?

12-09-2015, 05:49 AM   #47
Moderator
Loyal Site Supporter
BigMackCam's Avatar

Join Date: Mar 2010
Location: North East of England
Posts: 6,642
QuoteOriginally posted by Chosa Quote
Ok, so, for people like Class A who are against the removal of the AA filter, found that the AA simulation system introduced by Pentax with the K-3 works fine?
My own experience with the K5 and K3 is that, even when using AA filter simulation, the K3 is more susceptible to moiré. As a result, I leave the simulation switched off, and deal with any moiré in post processing. It's not a problem, and I'd rather have the extra sharpness / detail, although in most real shooting situations the difference is irrelevant.
12-09-2015, 06:45 AM   #48
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 9,193
QuoteOriginally posted by DeadJohn Quote
I think some people above might be confused between an anti-aliasing and Bayer filter.
In case you are referencing me, I'd just like to briefly comment that I'm aware of the differences between a CFA (colour-filter-array) and what's commonly called an "AA filter" but which should better be referred to as a "Bayer AA filter" because its main purpose is to achieve cross illumination between sensels in order to address the "colour blindness" of each sensel for the other two primary colours.

QuoteOriginally posted by DeadJohn Quote
An anti-aliasing (AA) filter adds a slight blur to lessen the false color from a Bayer filter.
Actually, it does not "blur", but it "splits".

I refer (again) to my old post on the subject that features a video demonstrating what (one part of) the "AA filter" really does, i.e., not "blur" as in "smear" but "split" as in "duplicate".

I'll refrain from further commenting on other posts because this thread has become beyond repair.
12-09-2015, 07:07 AM   #49
Pentaxian




Join Date: Dec 2010
Location: Ontario
Photos: Gallery
Posts: 2,554
QuoteOriginally posted by Class A Quote
I'll refrain from further commenting on other posts because this thread has become beyond repair.
I was hoping you'd comment on what makes an "AA filter" "weak" vs "strong"? The displacement of the splitting? Most explanations of the filter seem to show a 1-pixel offset, but I haven't found much that talks about the "strength" of the filter apart from photographers gabbing at each other in forums.

12-09-2015, 08:00 AM - 1 Like   #50
Pentaxian
kh1234567890's Avatar

Join Date: Sep 2010
Location: Manchester, UK
Photos: Gallery
Posts: 2,406
QuoteOriginally posted by BrianR Quote
I was hoping you'd comment on what makes an "AA filter" "weak" vs "strong"? The displacement of the splitting? Most explanations of the filter seem to show a 1-pixel offset, but I haven't found much that talks about the "strength" of the filter apart from photographers gabbing at each other in forums.
Chapter 4 of Lucas's book is still relevant (esp Fig 4.28). As are Doug Kerr's contributions in this discussion.
12-09-2015, 08:17 AM   #51
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 25,828
QuoteQuote:
The gain is quite significant. On DxO for example there more difference between K5 and K5-IIs than K5-IIs and K3 despite the K3 being 24MP. The gain seems to be 10-20% resolution.
IN photography, we deal in doubles and halves. Whether or not 10%-20% is significant is certainly debatable, and I doubt you could justify such a statement if you're talking about how much a person enjoys the image. My perspective would be , if it isn't double or half, you probably aren't going to notice.Simply stated, test chart "significance" is different than "print enjoyment " significance. Something the techno-geeks completely ignore, yet, it's the only important factor really. NO one cares what the test charts say, what they care about is "do they like the picture", it is that simple disconnect that makes sites like DxO completely irrelevant..

Another example of this would be, because I see work done by a lot of D800 style shooters, I found going to a K-3 completely made the 36 MP of a D800 un-attractive. It just doesn't give me anything more. The 33% increase in MTF is for the most part missing in action. SO it depends on how you look at these things.

They are measuring something people don't care about in the slightest. It's like measuring the pores on the hides on baseballs. Sure you can get out a micro-scope and compare the hides, or you can just go out and play baseball. Some people will try and tell you examining the hides provides valuable information you should know. To which I say pffft.. funny how in baseball, people would say pfft, but in photography people think these geeks with their benches and numbers are really important. Well they aren't, they are small minutae of passing interest. Somewhat relevant to the manufacturers, but pretty much non-sense to those playing the game.

Last edited by normhead; 12-09-2015 at 08:26 AM.
12-09-2015, 08:38 AM - 1 Like   #52
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 9,193
QuoteOriginally posted by BrianR Quote
I was hoping you'd comment on what makes an "AA filter" "weak" vs "strong"?
The ray split distance.

Only if the rays are split over the distance of a full sensel width, the Bayer AA filter is said to have 100% strength.

Often, manufactures choose a compromise, e.g. at 75% strength for the Nikon D800.
12-09-2015, 08:56 AM - 1 Like   #53
Pentaxian




Join Date: Dec 2010
Location: Ontario
Photos: Gallery
Posts: 2,554
QuoteOriginally posted by kh1234567890 Quote
Chapter 4 of Lucas's book is still relevant (esp Fig 4.28). As are Doug Kerr's contributions in this discussion.
Thanks for the references.

QuoteOriginally posted by normhead Quote
They are measuring something people don't care about in the slightest.
It's important to point this out at every opportunity and belittle people with an interest in how things work.

This might be hard to believe, but there are actually people who enjoy applying physics/math to understand the workings of their hobby, and can do this without letting it get in the way of going out and making images.

QuoteOriginally posted by Class A Quote
The ray split distance.

Only if the rays are split over the distance of a full sensel width, the Bayer AA filter is said to have 100% strength.

Often, manufactures choose a compromise, e.g. at 75% strength for the Nikon D800.
Cheers, I think I can visualize what happens with weaker filters with the textbook reference kh1234567890 gave above. There are some dandy formula to look at later.

12-09-2015, 09:20 AM   #54
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 25,828
[
QuoteOriginally posted by BrianR Quote
This might be hard to believe, but there are actually people who enjoy applying physics/math to understand the workings of their hobby, and can do this without letting it get in the way of going out and making images.
Sounds like you might be taking this a little personally. Did I hurt your feelings? Grow a pair.
Apart from feeling "belittled" did you have any comment on my suggestion that in photography , significance tends to be measured in doubles and halves, not in 10% to 20%, or did your "belittled" feelings get in the way of the technical discussion?

By the way, I took a course in Lens design, wherein I learned to apply optical physics, and did quite well in math. But... I understand their limitations when discussing manufactured lenses. IN the end, our evaluation should be the way the lens performs as a producer of pleasing images and meets your needs and budget, not a numerical evaluation.

So what I'm saying is, you shouldn't be doing that.. and providing the technical explanation of why you shouldn't be doing that.If you disagree, the logical thing to do would be to show me where I'm in error here. Not complain to me that you feel "belittled." I'm not your momma.

Last edited by normhead; 12-09-2015 at 09:29 AM.
12-09-2015, 09:39 AM   #55
Pentaxian




Join Date: Dec 2010
Location: Ontario
Photos: Gallery
Posts: 2,554
QuoteOriginally posted by normhead Quote
Sounds like you might be taking this a little personally. Did I hurt your feelings? Grow a pair.
Apart from feeling "belittled" did you have any comment on my suggestion that in photography , significance tends to be measured in doubles and halves, not in 10% to 20%, or did your "belittled" feelings get in the way of the technical discussion?
I agree with your basic premise that tiny enough details often don't matter in the final image, and that much of the minutiae can be safely ignored by a vast chunk of photographers out there.


edit I've probably overreacted, so for that I apologize. There is a constant backlash on photography forums against anything technical, largely on the grounds that Barbie would agree with (Math is hard!), and I hate seeing anyone discouraged from diving into details because it's viewed as unnecessary and with disdain by their photographic peers. Some of the minutiae may end up helping in the image creation process, other bits might just be great mental exercise. For my part, I'll try not to be so touchy and read offence where there may be none....

Last edited by BrianR; 12-09-2015 at 10:06 AM.
12-09-2015, 01:08 PM   #56
Pentaxian




Join Date: Nov 2013
Posts: 4,703
QuoteOriginally posted by normhead Quote
IN photography, we deal in doubles and halves.
For some things we consider to double for it to be miningfull like exposure time or isos. But this is not the case for apperture where the relevant factor is 1.4... This still double the apperture surface, so it let enter the double light.

Resolution, like apperture is in a single dimension, not a surface. So the relevant factor is the square root of 2, that is 1.412... Usually rounded to 1.4 (and that why we use this factor for apperture, not 2 but its square root).

Basically 33% more resolution is almost double (double is 41%) and that's relevant. This mean you can pratically print something of almost double surface (33% large in both dimensions) with same density of detailsl.

But there no 33% more pixel in both dimension between D810 and K3, The added resolution is more like 22%... If you consider it to be 33% it is that you consider that an FF sensor get better usage of its pixel than an APSC (that's a common assumption). And if it is indeed 33% it is indeed relevant by your logic.

Ironically there the same difference in resolution between 16 and 24MP there is between 24 and 36MP. For anybody that updgraded from K5 to K3 the excuse would be ironically the removal of that low pass filter because 22% resolution gain from pixel and 10-20% from low pass filter is near to the 1.4 magical factor that mean the change is relevant.
12-09-2015, 01:44 PM   #57
Pentaxian
KC0PET's Avatar

Join Date: Oct 2015
Location: Central Missouri
Photos: Gallery
Posts: 915
Interesting and entertaining debate (: I see some passions have definitely been aroused...

FWIW I enjoy the technical rants, lots of good information an I have definitely learned some new terminology here. I am an engineer and do lean that way. But also as a photographer, if I like the result I am usually not that concerned about why it came out the way it did. I have also learned as I get older that even with a "down in the weeds" technical analysis, there is usually interrelationships and complexities that are still not fully understood.

QuoteOriginally posted by normhead Quote
funny how in baseball, people would say pfft, but in photography people think these geeks with their benches and numbers are really important. Well they aren't, they are small minutae of passing interest. Somewhat relevant to the manufacturers, but pretty much non-sense to those playing the game.
I would be careful about such comments. Most likely these kind of "geeks" are the ones that designed the inner workings of your camera and without them we would still be painting on cave walls instead of shooting digital super high resolution photos that we can download to our laptop in a few minutes, tweak to our heart's content with software and send them electronically half way around the world in a fraction of a second.
12-09-2015, 03:02 PM   #58
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 25,828
QuoteQuote:
Resolution, like apperture is in a single dimension, not a surface. So the relevant factor is the square root of 2, that is 1.412... Usually rounded to 1.4 (and that why we use this factor for apperture, not 2 but its square root).
You've gone completely bonkers....
Resolution is measured in lw/ph.... line pairs per picture height..
So it' is sensor independent , it applies across the board to all sensors...

SO when we say, a K-3 has 2700 lw/ph, that is directly comparable to a D800 at 3700 lw/ph. To print the same number of line pairs you print 1/3 larger for the same lw/ph. But that doesn't take into account that if you print really large, you have to stand further away, so you still have to prove with some kind of user testing that it makes any difference at all.

If I print a 2700 lw/ph image from my K-3 at 27 inches, I'm printing 100 lines pairs per inch, each line is .01 inch.
If you print the same image taken with a D800 at 3700 lw/ph, I'm printing or otherwise displaying .0073 lines per inch. This assuming that the printer or other display option is actually capable of displaying such differences.

The difference between the widths of the smallest displayable line would be .0027 (2.7 thousandths of an inch). I know we've had people swear up and down a stack of bibles that they can see such differences, and that it makes a big difference to their images... I remain unconvinced.

The simple fact is, as lw/ph goes up. contrary to what you've stated, the differences become smaller and smaller. It takes major boosts in lw/ph, to make any difference at all.

Your 1.4 factor doesn't apply at all.

But my goal for double the resolution might, even then I'm just guessing.

The problem here is that no one has ever definitively been able to answer how many lw/ph you need to make a decent picture and some tests have shown that 100 dpi will be a sharp looking image, that's better than your computer screen.

Others have suggested that even 72 dpi is good enough.

But this is all complicated by viewing distance, pixel peeping a K-3 image on your monitor is like looking at an image more than 70 inches wide. To do that you have to back up. And when you back up you can't see small differences. That's why no one has ever been able to define exactly what lw/ph you need for a print.

It also explains why many people can't tell the difference between print at 72 DPI and one at 150 dpi, even though there should be 4 times as much information.

Last edited by normhead; 12-09-2015 at 03:47 PM.
12-09-2015, 09:38 PM   #59
Pentaxian




Join Date: Nov 2013
Posts: 4,703
QuoteOriginally posted by normhead Quote
If I print a 2700 lw/ph image from my K-3 at 27 inches, I'm printing 100 lines pairs per inch, each line is .01 inch.
If you print the same image taken with a D800 at 3700 lw/ph, I'm printing or otherwise displaying .0073 lines per inch. This assuming that the printer or other display option is actually capable of displaying such differences.
2700 lw/ph is not the max lw/ph you can get from a K3 but the max you can get from a K5. You can look at photozone for reference. The best 24MP APSC sensor with best lens on photozone already get 3700 lw/ph.

I didn't see D800 on photozone, the latest FF lenses tests have be done with a D3X that is 24MP FF and the Sigma 50mm Art achieve a bit more than 4000 lw/ph on it, 4066 to be exact. Simple math give us that with 36MP assuming you keep the same kind of low pass filter (or lack of it), if you can find a lens that is good enough, you could expect then up to sqrt(36/24)*4066 = 4979 lw/ph, not 3700 on a D800. On latest A7R-II that would be a bit more sqrt(42/24)*4066 = 5378 lw/ph.

but that doesn't change anything about the fact that if you keep your "good enough" printing quality of Xdpi, whatever that X might be, if you have 3700 lw/ph in one case and 2700 lw/ph in the other case, you can print the 3700 lw/ph picture 37% large both horizontally and vertically and it mean you print surface is 1.87 time greater. If you round it because we now that double that count you can print twice larger while keepîng same resolution or same pixel density or same number of lines per inch.

Another way to see it is a TC is providing you an enlargment of 1.4, here we are at 1.37 so that like a free TC... TC that typically cost you 1 stop of light and degrate a bit picture quality.

By your notion that we speak in half and doubles, 3700 vs 2700 is at this level, rougly you double you printable surface whatever you min acceptable dpi might be.

As to what dpi is the industry standard for print, this is not 72, neither 100 or 150. The standard is 300. This is because it is assumed that if you print higher than 300dpi would not notice the difference.

This is actually not true. Studies has been made to evaluate the max resolution eyesight can take benefit from. This is higher than the standard of 300dpi. If you show people print with 600dpi worth of detail instead of 300 and ask them to tell wich print is better they tend to choose the print with 600 dpi level of details even you didn't tell them anything on the print quality.

I agree with you that there diminishing return of course because you might expect that most people will not start at your print from near distance or because you are not going to print 60" large and get "only" 100 dpi on your 24MP sensor uncropped or "only" 50dpi if you cropped quite heavily. But that on the level of another double or half we have in photography. In

For most situation 1/4000s max is enough and 1/8000s shutter speed isn't that necessary. You could have a 70-200 f/2.8 but many people are just happy with 70-200 f/4... Some people want to have an iso 50 setting while other don't care and are happy with iso 100 as the minimum. Some people better high iso is relevant and some don't care much. So that depend who you are.

But in term of double the D810 is offering quite a few indeed compared to a K3. There 2EV more of dynamic range. There a 50 iso mode, you can bump the iso 2.3 time (or 1.1EV)... You can print 33% larger (4978/3700) so more double your printing surface area...

Do we have to care, I'am the first to think that's maybe a No, but for sure there many "doubles" involved here and depending of what you do, you may very well care.
12-09-2015, 09:56 PM   #60
Pentaxian




Join Date: Nov 2013
Posts: 4,703
QuoteOriginally posted by normhead Quote
But this is all complicated by viewing distance, pixel peeping a K-3 image on your monitor is like looking at an image more than 70 inches wide. To do that you have to back up. And when you back up you can't see small differences. That's why no one has ever been able to define exactly what lw/ph you need for a print.
This depend a lot on the monitor. on a typical monitor that has arround 100dpi, true. And then most will notice that a pixel shift picture look MUCH better at 100% crop on that monitor. This mean that more resolution than 24MP without pixel shift would indeed look better to most at 60-70" printing size.

Now if you add some margin for post processing (crops, perspective correction, optical corrections) and that this is achieved only on the center, and you'll see that this D800 is going to provide noticable improvement at 30", that you don't really need 60" to see it. And then that's not so big anymore. You don't need to be that far from a 30" print to see it entirely and you can resolve quite some level of details already.

While you or me may not care, and I'am the first to say one doesn't need an FF if he isn't making a living out of photography, that far from being useless improvement.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, dslr, photography, sensor
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Is ther a K-x or other AA Pentax for sale on the Forum kamicota Pentax DSLR Discussion 21 08-17-2015 07:33 AM
Is a K-3 with a full frame sensor enough for you? jake14mw Pentax Full Frame 58 03-28-2015 11:39 AM
Pentax staying small. An advantage? stillshot2 Photographic Industry and Professionals 27 12-22-2013 02:34 AM
Bayer filtered camera sensor brian101 Troubleshooting and Beginner Help 4 05-08-2013 01:10 PM
What is the advantage of 21/35/40 prime vs a f2.8 zoom lens raider Troubleshooting and Beginner Help 23 06-21-2010 10:57 AM



All times are GMT -7. The time now is 10:52 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top