[y0chang]

I went from K-5 to K-3 to K-1. Image quality wise the K-1 is unparalleled. But what I am finding is that rather than it being the beast that I use all the time, the K-1 is the best when I am doing portraits, while for landscapes I fluctuate between crop and full frame. Part of the reason is that my lens collection is very eclectic. But I now spend way more time than I want picking before an outing, and now I need more space in the bag for which lens body combination that i want.

[Ian Stuart Forsyth]

One was produced with a fraction of the light, but they look the same on my monitor. There is no reduction in light in the output. The output the light values are equal, and will be even if you were to print to 100 inches.

I can clearly see what that decrease in light had on the image, one of the biggest is the increase in noise and other artifacts


Take for example this image was taken with an exposure using iso 2000 that is 1/4 the size of your Raven image exposure but it shows less of the artifacts that you can see in your image, this is what happens when you crop you decrease how much light the image is made up of

Even though heavily cropped, there is exactly the same light intensity used on each pixel, there has to be, it's the same image.

No its not the same image, here's a hint it has a narrower FOV and in doing so you increase the magnification of the artifacts found within your image, noise being one of them

FF sensor - Focal Length 300, ƒ8 2 meters DoF, 10meters. DoF .52 meters shutter speed 1/60s (actual measurement)
APS-s sensor 'Focal length 200mm, F5.6. 10 meters, DoF .55 meters. shutter speed 1/125s ((actual measurement)
The APS-c sensor uses Double the light for half the time. So the exposure is the same.

How can the APSC sensor use double the light ,I take it that you mean that crop body allowed twice the amount of light because of the F/5.6 is being used over that of the FF at F/8. The major problem with this thinking is that they both lenses let the same amount of light for the final image. We know this because the hole letting the light in from the lens was both are almost the same size 200/5.6= 35mm opening while the 300/8=38mm size of opening. In order to keep the same output image brightness the cropped body needed to decrease the shutter speed to maintain the output image. If one body had to decrease the shutter this should be a clear indication to you that it is made up of less light.

The APS-c sensor uses Double the light for half the time. So the exposure is the same. The only place the FF has an advantage is in total surface area. More light gathering capacity. But in actual practice the reduction in size ou need to view an image means you throw away a lot of the light the APS_c image cpatured, and even more of that captured by the FF. You add pixels during enlargement to compensate for less total light. And in most instances like the one above, you are actually throwing away. If you take an APS_c image and crop to reduce the image size to 3650 x 2100 for your 4k TV, you throw away. 20 MP of your 4k image.

And why do you have to reduce your image by the means of cropping to make it 4k I would tend to think it would be better to reduce the image size by scaling and use how you have decided to frame your image at the time of capture, very strange way to use your camera ?

My images look fine on a 110 dpi computer screen but I'm sure 100 dpi would be the same. . So the 6000x 4000 24 MP image of my K-3 will lbe optimal up to 60 inches and probably well beyond, because upscaling software creates the effect of adding extra light you get with the larger sensor for you, while the software that down scales your image throws away light.

Upscale only magnifies the unwanted artifact that you will see in the image and in now way give you the effect of adding more light, while scaling will change the artifacts like noise and place those artifacts into a smaller area that occupies less of your final output image. To think so differently leads me to believe that you don't know the difference between the 2. again all you have to do is look at the 100% crop from your raven and the uncropped image of the raven at 4mp output and you will see exactly that a decrease in noise in the final image.

I have a 42 inch print on my wall, taken with a K-3. And it looks great. I suggest the only time the additional light surface used on the larger sensor would be for images over 60 inches. No one has ever established the point at which people will notice a difference between a large print printed with an FF and one printed with APS-c. ( I actually don't believe there is such a point and people have taken to comparing even 4/3 images to FF with no noticeable difference. And honestly when you look at the numbers, 6000 pixels wide compared to 7300 pixels wide, that's not surprising.

If this is your thinking and that you need to crop the most from a format to display an image on a 4 k monitor then many of the new bridge cameras will even give you better results over what you are using if we use your train of thought

It has more to do with the light and the amount that you can store in adition the additional pixels allow you to more aggressively use NR and sacrifice detail for less noise. We want more light because of how sensors capture light and anything that falls in the mid tones and lower are only capturing 1/8- 1/16 of the light in an exposure at base iso. When we factor in how the colors are captured you will find that most of the time the red channel is only using 1/32 of the sensors capacity to store light in anything that falls where our subjects fall mid tones, we also want more resolution to help combat how we capture red details and how they are limited by how the sensors work.

So there you have it. The theoretical, and the practical. My guess is if you aren't printing more than 60 inches wide there you are getting nothing for the extra light collected by the FF sensor. It's of no practical value. And its quite possible that even up to 100 inches wide, you get nothing of value.

As shown by your crow crop you can see a difference at 4mp

Most of the time you throw it away a lot of the collected light downsizing an FF image. And even on over 60-100 inch prints it provides no demonstrable practical advantage. Especially if you look at the print from a practical viewing distance. You may be able to see increased detail, but no one has ever shown that, that level of increased detail makes viewing the print more enjoyable.

Again with downsizing you are not throwing away the light you have collected you are shifting those artifacts to a resolution that is hidden to the outgoing image display resolution.

---------- Post added 11-11-2018 at 12:41 AM ----------

So two identical fields of view, one with f8 and one with f5.6. So one aperture stop difference. Well guess what, the camera will want to use half the shutter speed for the f5.6 shot compared to the f8 shot. I learned this 40 years ago when I was practising with my MX !

The f stop is 1 stop different but the aperture letting light into the image is the same, we know this by what is written on the lens
take the 300mm F/8 that is a formula for the size of the aperture 300mm/8= 38mm with the 200mm/5.6=36mm so both lenses for their given format let roughly the same amount of light but because one of the images used the shorter shutter speed and used less exposure time this should tell you that less light was captured in the final image.

[normhead]

Take for example this image was taken with an exposure using iso 2000 that is 1/4 the size of your Raven image exposure but it shows less of the artifacts that you can see in your image, this is what happens when you crop you decrease how much light the image is made up of

Sorry Ian, despite your effort to tell me what to see, I don't see it. As per usual, when challenged on anything, you just make up stuff.

If this is your thinking and that you need to crop the most from a format to display an image on a 4 k monitor then many of the new bridge cameras will even give you better results over what you are using if we use your train of thought

That depends entirely on the light. We have sold images take with 1:1.2 cameras. Right tool for the job and all that.

If this is your thinking and that you need to crop the most from a format to display an image on a 4 k monitor

Ya well stop right there cause I don't think that.

Upscale only magnifies the unwanted artifact that you will see in the image and in now way give you the effect of adding more light, while scaling will change the artifacts like noise and place those artifacts into a smaller area that occupies less of your final output image. To think so differently leads me to believe that you don't know the difference between the 2. again all you have to do is look at the 100% crop from your raven and the uncropped image of the raven at 4mp output and you will see exactly that a decrease in noise in the final image.

You just see what you want to see. Once again, I don't see what you see. You act as if it's obvious but if it was, you wouldn't have to spend so much time explaining what we're supposed to be seeing.

This has always been your method. I post images to demonstrate a point, you nit pick the images insult everyone else's work post some of your own you claim are better, then go in in your own little world of invisible artifacts, noise anything subjective. Some little thing only Ian can see that means you are right and everyone else is wrong.

In any case, I have no more interest in carrying this on. If you don't understand the points i laid out, fine, no sweat off my back. You really aren't competent to discuss other people's images. You don't seem to appreciate anyone's work but your own. and you make up all kinds of nonsense reason why yours are better.

BY the way, the cropped Raven has lots of like here, on facebook and on flickr, your complaints don't matter to anyone but you as far as i can tell. And your biggest complaint from my perspective is you didn't take it. At least that's how it appears from here.

I can clearly see what that decrease in light had on the image, one of the biggest is the increase in noise and other artifacts

Ya well, that doesn't mean they are there. It's well documented that some people's vision is affected by what they think they should see. They expect to see something, and they see what they expect. You might be suffering from that.

One thing that should color the analysis is the fact that resolution isn't a matter of total number of dots. It's how many dots there are in a given area. I haven't actually calculated the numeric difference, but I suspect my K-50's resolution is roughly the same as the K-1. It's just that there's half again as many dots, total. But when you go to print from the two onto the same 5*7 paper, there's a lot more dots per inch from the K-1's picture because you're shrinking the image more to get it to fit into that space. Conversely, enlarging two such images to the same size will result in more rapid pixellation with the K-50's image, just because there aren't as many dots to expand into the resultant space. But I don't think there's really any signficant difference in resolution at the sensor's level.

[normhead]

For a 5x7 printed at the capacity of a Canon printer, you need 300 DPI. that would be 1500 x 2100. The K-1 is roughly a 3.5x oversample. The K-3 is a 3x oversample. The difference being on the K-1 you've reduced the image form 7300 to 2100 , with the K-3 you're reducing 6000 to 2100 and with a K-5 you're reducing 5000 to 2100.

Once reduced each of those 2100 x 1500 pixels is a value for three channels containing numbers between 0- and 255. You can't tell looking at the numbers what size sensor it was taken with, or how much total light created the original image providing it was reduced to an 8 bit jpeg for printing.

I've seen no evidence that there is any benefit at all to the k-1 image on a 5x7 print. So you print at the same resolution with all three images. The only issue is how many pixels you combined to create that image. But the number of pixels in the end is the same. And that will be true right up to the K-5s 4900 x 3200 image size (or 18 x 11). You definitely won't see any difference on a 5x7 print.

One of the practical things no one seems to ever mention is that upscaling in digital can actually clean up your image. I once did some comparisons with a telephoto lens and an upscaled image from a shorter lens. The upscaled version looks better. It cleans up some barely visible ink smudging on my test chart, and some people actually preferred the upscaled version to the higher res telephoto version. It was just more aesthetically pleasing.

I once printed a canoe in a barn at 72 dpi, because I no longer had the original. It looked great, and was everything I wanted from the print.

If you get out there and test these theories you find there is a lot of truth people who get wrapped up in theory miss. It's fine to put forward, these speculations based on how you perceive the world, but sooner or later, surely you should be testing them. It's fine to spout theory , but sooner or later you have to get out of you own head. Of course that won't help you unless you can be unbiased enough to correctly interpret the results. Blind testing can be a real eye opener. People who aren't looking for the things the photographer might object to, often have no issue with perceived technical flaws of a photograph.

This is way more complex than iti's being made out to be. The simple fact is, the highest resolution is not always the best for displaying a photograph. Resolution has to match the subject to present the subject in it's best light. Photgraphy is the art of making the right choices for what is being portrayed. And those choices involve, lens choice, distortion, field of view, distance form the camera the subject will be, the amount of resolution required, ,how much noise will be acceptable for that image, and many other factors. Many factors can affect affect an image and even with distortion, I've had images taken with fisheyes, that look better than if i'd taken them with a rectilinear lens. These factors are never always negative or always positive. The photographer who sees the image that is going to be more pleasing taken with a fisheye than a rectilinear lens is way ahead of the game. I've even seen photos where noise was part of the effect of the photo.

Thinking total light makes much difference is just wrong, it's way down near the bottom of the list. which is why Ian has to claim I have noise and artifact issues no one but him can see (or if they can see them, they don't care about them) to make his point.

It must really annoy photographers who are focused on technical abilities of a camera rather producing images people like looking at, the amount of technically mediocre images that people rave about because of their artistic merit. You now the type, you're in the gallery looking at some fantastic prints, and some photographer is buzzing away like a little fly telling the world what he sees that's wrong with it. A lot of amateurs (and even some pros) need to spend more time appreciating other's work and less time glorifying their own.

It's also interesting that when I post an image to illustrate a point, the point is often ignore and time is wasted critiquing the image.

But back to the topic at hand. 36 MP FF gives you more flexibility in cropping, more dynamic range (though you won't often use it) more resolution, (though you won't often need it) and more control of DoF, one extra stop on the wide open end, which you also won't need very often.) Concentrate on those things in your camera selection and FF or APS-c becomes a straight choice, those are the benefits, what are those things worth to me? Are those improvements that would help me out or that i want to experiment with? At a given point in your photographic journey, one of those things may be of interest, and a full frame may be for you. But total light? My opinion would be don't waste your time on that.

[Merv-O]

K-1ii with the DA lens crop function is the best camera I've ever owned from any manufacturer (including my stellar Leica typ 240/M9 kit). Using DA lens offers little fallout in clarity and color palette. The FF lenses produce the full clarity and power of 36,4 mp. Even smaller reproductions of photos have denser clarity in smaller print/detail (i.e., 8.5 x 11). At 13 x 19, the clarity is professional and surprises me. Using fast 50mm 1.4 prime lens yields excellent results. Love it. I bought 6 FF lenses (3 primes 15mm, 35mm & 50mm, along with 3 teles: Tamron 25-25mm mini, 28-105mm kit lens & 70-300mm SIgma). But the real joy is using my several DA lenses. It's a total kit. just wonderful all around.
I've attached a few samples so you can check the exif data for comparison.
The first two were taken with the K-1ii in crop mode using (intentionally) the inexpensive and much maligned 50-200mm DA lens; Note the 2nd photo of the dog was cropped by 60% (still excellent even with a cheap DA lens) & the 3rd was taken with the K-3 and the acclaimed 18-135mm DA lens. I still think the K-1ii delivers a bit more on the palette and overall.
CONCLUSION: Buy the K-1ii, you won't miss the K-3/ii one bit. If you're "jonesing" for an APS-c camera, buy a leftover/demo K-S2 to tide you over, but I assure you, many K-1/ii fans aren't switching back once the new rumored K-3/ii replacement hits the market.

[Merv-O]

The 36mp makes a big difference in cropping. It also has extreme clarity when viewing printed materials within a medium size photograph. The sensor of the new K-1ii needs less correction for me as it has excellent balance and color palette. The weight of the K-1ii is difficult at first, but the weight at times controls shake by forcing you to stiffen your forearm. (I personally hold my breath to stiffen any movement as well as having the Image stabilization activated). Cropped photos with the better quality DA lenses is also astounding. the full frame along with the large pixels has great depth of field and the viewfinder cropped shows other areas allowing one to change the frame of the photo and improve your image.


The longer I spend time with the K-1ii, the more I'm pleased with Ricoh/Pentax and the system itself. I am so bullish on the K-1ii, I am sellng my K-3 and purchased a new K-S2 leftover with the funky 18-50mm collapsible lens (for $380) and, I have to tell you it is a great back-up for the K-1ii. Those of us that maligned the K-S2 probably never spent time with it. Its excellent APS-c sensor (20.4 mp) is excellent. It also has most of the K-3 features and the small difference in mp power is not that big of a problem.

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[normhead]

The 36mp makes a big difference in cropping.

Many time that issue is more efficiently addressed by using a longer lens, or moving closer. The simple fact here is, if you crop your FF image to less than 15 MP, you would have 30% more resolution with a K-3. Someone should graph out exactly where the mesh point is, where 36 MP cropped is the same as 24 MP cropped in resolution. My guess is its somewhere around 28MP, but my ability to do that kind of math vanished long ago.

If I'm right, if you crop your image to around 28 MP, or less, you would have equal or better by pre-cropping with a crop sensor and taking advantage of the higher pixel density. (if you didn't need the DR or high ISO performance.)

If your crop is a minor crop to 30MP or more. I would expect the K-1 image to be better. But the fact remains, if you are able to frame with your lens as you wish the image to look, through the viewfinder an APS-c image will never be better than the equivalent K-1 image, but it might be worse. But, that's a pretty big "if" with the weight and lack of portability in FF zooms, especially long zooms and lenses. For being able to frame without lens changes, or in the long end or macro, APS-c has the advantage.

I bought the FF knowing it wouldn't help with every image, but that for some images it would be noticeably better. And that's been my experience. And when I was studying photo-arts, sometimes I shot an 8x10 view camera, sometimes I shot a 4x5 view camera, sometimes i shot 645 or 6x6, sometimes I shot 35mm. Usually the strengths of one system, in no way negates the strengths of another. In the big picture, there is no "best format". There is however a format which might best suit your own personal shooting style, if you can only have one. Having grown up with no fewer than 3 formats just from the cameras in my house, I'd find that very restrictive. Others are up to the challenge of making do with one format. I'm not sure there's a right or wrong here. Just individual inclinations. There is simply no irrefutable argument that can prove any one format can do everything the others do. That is why we have so many formats available. If IQ is the game and you shoot all your images between 28-105mm, don't shoot macro, wildlife or action, a K-1 will be a step up. Include macro, action and wildlife and now you're looking at compromises.

Bottom line, at least for me, even if I have APS-c I want my Q and XG-1 to go with my K-1. I f my choice is just one it's APS-c, as my sig points out. If you have your other bases covered, portable, higher pixel density, etc. etc. a K-1 is great addition. Others seem to be happy with just a K-1.

This would all be simplified by a poll, with some carefully selected questions. Unfortunately the polls i've seen on line were put together by amateurs.

They said I should buy a Canon 5d mk III. If it was a person in the room with me, I would have explained to him what a piece of junk his poll was and why, right after I gave him the finger.

You pays your money and you takes your pick.

[Merv-O]

Many time that issue is more efficiently addressed by using a longer lens, or moving closer. The simple fact here is, if you crop your FF image to less than 15 MP, you would have 30% more resolution with a K-3. Someone should graph out exactly where the mesh point is, where 36 MP cropped is the same as 24 MP cropped in resolution. My guess is its somewhere around 28MP, but my ability to do that kind of math vanished long ago.

If I'm right, if you crop your image to around 28 MP, or less, you would have equal or better by pre-cropping with a crop sensor and taking advantage of the higher pixel density. (if you didn't need the DR or high ISO performance.)

If your crop is a minor crop to 30MP or more. I would expect the K-1 image to be better. But the fact remains, if you are able to frame with your lens as you wish the image to look, through the viewfinder an APS-c image will never be better than the equivalent K-1 image, but it might be worse. But, that's a pretty big "if" with the weight and lack of portability in FF zooms, especially long zooms and lenses. For being able to frame without lens changes, or in the long end or macro, APS-c has the advantage.

I bought the FF knowing it wouldn't help with every image, but that for some images it would be noticeably better. And that's been my experience. And when I was studying photo-arts, sometimes I shot an 8x10 view camera, sometimes I shot a 4x5 view camera, sometimes i shot 645 or 6x6, sometimes I shot 35mm. Usually the strengths of one system, in no way negates the strengths of another. In the big picture, there is no "best format". There is however a format which might best suit your own personal shooting style, if you can only have one. Having grown up with no fewer than 3 formats just from the cameras in my house, I'd find that very restrictive. Others are up to the challenge of making do with one format. I'm not sure there's a right or wrong here. Just individual inclinations. There is simply no irrefutable argument that can prove any one format can do everything the others do. That is why we have so many formats available. If IQ is the game and you shoot all your images between 28-105mm, don't shoot macro, wildlife or action, a K-1 will be a step up. Include macro, action and wildlife and now you're looking at compromises.

Bottom line, at least for me, even if I have APS-c I want my Q and XG-1 to go with my K-1. I f my choice is just one it's APS-c, as my sig points out. If you have your other bases covered, portable, higher pixel density, etc. etc. a K-1 is great addition. Others seem to be happy with just a K-1.

This would all be simplified by a poll, with some carefully selected questions. Unfortunately the polls i've seen on line were put together by amateurs.

They said I should buy a Canon 5d mk III. If it was a person in the room with me, I would have explained to him what a piece of junk his poll was and why, right after I gave him the finger.

You pays your money and you takes your pick.

Your point is well made and I intend to stay with APS-c. I bought a new leftover K-S2 recently with the retracting (in my opinion crappy 18-50mm lens) so I could stay in the Pentax APS-c game as my venerable, but aging K-3 was sold. When (or "if" say the naysayers) Ricoh decides to spruce up the K-3/ii line, I'll be there with my $$$. To me the KP is a sprucing up of the K-S2 in many ways, though it has some obvious strengths and advantages (as it is newer than both K-3/ii & K-S2).


That being said, I recently did the same thing with my Leica kit. I kept the M9 (had the sensor replaced: CCD is like Kodachrome) FF, and recently bought the new CL (Leica's APS-c camera) to use all that M-glass (w/adapter). Both kits have an FF and an APS-c option.


The K-1ii is really becoming my favorite camera as it is also strengthening my wrist and forearm muscles due to its girth.

[Rondec]

I'm not sure why all the focus seems to be on resolution. APS-C and full frame both give plenty of resolution for standard printing sizes. The question is what can you get out of your images and do you push the limits of dynamic range, high iso performance much. If you do, then you either need lenses with faster apertures or cameras with larger sensors or both. Shooting at iso 100, both formats are fine. I still can push a K-1 image a lot more than a K3 image, but you can usually get what you need out of both. Go up much in iso and it isn't the same. Whether or not you need that performance is hard for anyone else to answer, but it is nice to have in situations where you don't have a tripod with you.

I would definitely say that for landscapes, full frame is a treat -- particularly if you choose to do much pixel shifting.

New Orleans at Night (iso 12,800)



New Orleans at Night (iso 10,000)



The other thing that does get mentioned is shallow depth of field. I'm not big into this, because I like stuff to be in focus, but if you do a lot of portraits then it can be handy. Also working distance with lenses like the FA 77 is better with full frame than with APS-C. It is a nice lens on both formats, but you have to stand farther back with APS-C.

DA *55 at f1.5 on K-1 II



If it sounds as though I'm saying APS-C is terrible, I'm definitely not. I'm just saying that full frame is a little better in certain areas. It is up to the individual photographer to decide if they need that performance boost for what they shoot.

[Tranzistors]

Many time that issue is more efficiently addressed by using a longer lens, or moving closer.

Technically true. But at least in my photography I often crop half of the image away, for example, if I see that a background is too distracting, I will crop a landscape portrait into a portrait portrait. This can easily make a 36mp image into a 16mp image. You might argue that a 16mp “should be enough for everybody”, but lenses tend not to be super sharp, so those 16mp images can easely turn into effective 8mp images. And 8mp essentially are 4K screen size.

Having a 24mp APS-C will not help in any way. I suspect that effective megapixels will even be lower for APS-C for the same lenses.

[normhead]

Technically true. But at least in my photography I often crop half of the image away, for example, if I see that a background is too distracting, I will crop a landscape portrait into a portrait portrait. This can easily make a 36mp image into a 16mp image. You might argue that a 16mp “should be enough for everybody”, but lenses tend not to be super sharp, so those 16mp images can easely turn into effective 8mp images. And 8mp essentially are 4K screen size.

Having a 24mp APS-C will not help in any way. I suspect that effective megapixels will even be lower for APS-C for the same lenses.

You really need to go and read a few lens charts. Or go to Imagine Resources and they camera reviews.
The K-5 and K-5ii were about 1900 distinct line captured per inch, the K-3 is 2700-2750. For from less resolution it's about a 50% increase in resolution, (with a 50% increase in pixels.) There's absolutely no sign of a resolution drop off.

So you can suspect whatever you want. There are real tests that will give you real numbers, done by people doing their best to make sure the tests are reliable and comparable.

Our laboratory resolution chart reveals sharp, distinct line patterns down to about 2,700 lines per picture height in the horizontal direction, and to about 2,750 lph in the vertical direction in best quality JPEGs, though some aliasing can be seen at around 2,300 lph. Complete extinction of the pattern didn't occur until just past 3,000 lines in both directions. An Adobe Camera Raw conversion yielded the same numbers, though individual lines were more crisp up to the same resolution, and complete extinction of the pattern was extended past 3,800 lines. The ACR processed RAW images do however show a lot more color moiré, as they often do. (Note that AA Simulation was set to its default of off, for maximum resolution.) Use these numbers to compare with other cameras of similar resolution, or use them to see just what higher resolution can mean in terms of potential detail.

Pixel Shift Resolution Mode
~2,850 to 2,900 lines of strong detail from JPEGs, about the same from converted RAW files.

Our laboratory resolution chart revealed sharp, distinct line patterns down to about 1,900 lines per picture height in both the horizontal and vertical direction in JPEGs. (Some might argue for over 2,000 lines, but aliasing artifacts begin to appear earlier.) Complete extinction didn't occur until around 2,800 lines in both directions. We were able to extract more resolution (to about 2,100 lines) with RAW files processed through Adobe Camera RAW, with complete extinction extended to around 3,200 lines. The ACR processed RAW images were also crisper (thanks to strong/tight unsharp masking in Photoshop), and didn't show as many jagged edges on fine diagonal lines as the camera JPEGs did, though color moire was a bit higher. Use these numbers to compare with other cameras of similar resolution, or use them to see just what higher resolution can mean in terms of potential detail.

You might want to examine the process the lead to your suspicion, and discard it from your "how I form my suspicions" repertoire.

Then look at something like the Panasonic FZ2500



1 inch sensor 12mm x 9mm, 20 MP
An in-camera JPEG of our laboratory resolution chart revealed sharp, distinct line patterns up to about 2,550 lines per picture height in the horizontal direction, and to about 2,500 lines per picture height in the vertical direction. Some may argue for more, but aliasing artifacts begin to interfere and lines begin to merge at that point. Complete extinction of the pattern didn't occur until about 3,000 to 3,100 lines. An Adobe Camera Raw conversion produced similar results, but complete extinction of the pattern was extended to about 3,100 to 3,200 lines and color moiré is more visible. Use these numbers to compare with other cameras of similar resolution, or use them to see just what higher resolution can mean in terms of potential detail.


That's 283 distinct lines per mm, the K-3 is 183 distinct lines . IN other words, the parsonic packs the same resolution into a 9x12 sensor the K-3 does in 23x15 sensor. Lets not start worrying about higher MP gains not achieving that until we get to like 50 MP or something closer to the current limit. The issue to date is not resolution in these high density cameras, it's the way the smaller sensors render, and their pathetic low light performance. Putting Panasonic sensor tech into an APS_c camera should produce about 40 MP. SO we have empiracle proof that's possible When we get over 40 MP, maybe then we can start worrying about the effect of adding more MP possibly not adding any resolution. We know from this it may be possible to get even more resolution out of this type of sensor, but we know for sure we can get that much.

[Tranzistors]

. For from less resolution it's about a 50% increase in resolution, (with a 50% increase in pixels.) There's absolutely no sign of a resolution drop off.

I think you have misinterpreted me. When I talk about resolution drop off, I don't mean that the sensor is to blame, but rather the sensor-lens combo. If you have a razor sharp D-FA 50mm at f/2, moving from 16 to 24 MP on APS-C will probably give you about 33% more detail. However, if you have a pinhole camera, those extra megapixels on the same surface won't give you much more detail, but increased surface will.

That is to say, if a lens is not super sharp, but is about as sharp in the centre as it is in the corners, you can get more details out of a picture by moving from APS-C to FF and keeping pixel density, rather than increasing megapixels on the APS-C. However, that if is a big if. Lenses tend to be noticeably sharper in the centre. I have not seen any studies of FF Pentax lenses in this regard. It is perfectly possible that for a same FF lens K-3 will give more detail than K-1, I just suspect that 35mm f/2.4 and 100mm f/2.8 macro will give more detail on K-1.

[normhead]

I think you have misinterpreted me. When I talk about resolution drop off, I don't mean that the sensor is to blame, but rather the sensor-lens combo. If you have a razor sharp D-FA 50mm at f/2, moving from 16 to 24 MP on APS-C will probably give you about 33% more detail. However, if you have a pinhole camera, those extra megapixels on the same surface won't give you much more detail, but increased surface will.

That is to say, if a lens is not super sharp, but is about as sharp in the centre as it is in the corners, you can get more details out of a picture by moving from APS-C to FF and keeping pixel density, rather than increasing megapixels on the APS-C. However, that if is a big if. Lenses tend to be noticeably sharper in the centre. I have not seen any studies of FF Pentax lenses in this regard. It is perfectly possible that for a same FF lens K-3 will give more detail than K-1, I just suspect that 35mm f/2.4 and 100mm f/2.8 macro will give more detail on K-1.

K-1 and K-3 for the 100 macro, same aperture and minimum focusing distance from the subject, if you set up for the K-3, the K-1 will have less subject detail but more of the area around the subject. If you set up for the K-1, then the K-3 will not show the whole subject, but what it shows will have more detail.

All this of course assumes you are photographing something with detail one can resolve. Of course in real life you back the K-3 up. Once you reach minimum focusing distance for true macro, you can't move the K-1 in more, except with extension tubes or bellows. The ultimate resolution with some time consuming work arounds can help you exceed your K-3 macro-capability with a K-1. But you need extra equipment and time.

With telephoto work, there's pretty much nothing you can do. Once you reach the point where you have to crop your K-1 image to 28 MP or less, you get better with a K-3 as long as you can keep your ISO under 640 ISO. The K-1 has more pixels for more resolution, but it is are harder to effectively use them.

Used to it's strength the K-1 is better. But for those occasions when a K-1 can't be used to it's strengths, and there are many reasons that could happen, you're probably better off with a K-3.

Blanket statements are almost always misleading oversimplifications that ignore less common but important shooting situations. I often start close to a small object and then back out until I hear focus lock to get my image at minimum focusing distance. If you don't do that, maybe the difference in macro isn't apparent to you. I also shoot birds at distances where I have to crop my K-1 images, and in fact also have to crop my K-3 images. In those instances the K-3 is the better camera to use. The K-1 provides more resolution, until it doesn't. It's not absolute.

However, if you have a pinhole camera, those extra megapixels on the same surface won't give you much more detail, but increased surface will.

This is kind of a distraction. There simply are no modern lenses (like made in the last 15 years) that don't benefit from smaller pixels. I'm all for illustrating a concept, but if you have to go as extreme as a pinhole camera, it's probably irrelevant to modern photography. So, why go there? It's the practical vs. the never to be encountered hypothetical.

To people who don't shoot macro or telephoto (and I suspect the majority don't) , my observations are meaningless.

Think how much money I would save if I was happy shooting all my images between 28-105mm on a K-1 and can always carry a camera bag big enough to hold them both. If that's the limits of my photographic expectations, the K-1 is king.

And when that's the case, I only take the K-1.
Yet half the time I take the K-3, and because of the advantages of burst mode at least 2/3s of my exposures are still taken with K-3.

[Tranzistors]

Yes, macro photography is a special case, where smaller sensor is a direct benefit for extra close up. I think this conversation can be summarised to “Ask your doctor, if full frame is right for you”.