[Nicolas06]

I think that the 24MP are not necessary but a very usefull commodity to have.

As my shooting style is concerned, I'am totally for the cropping philosophy. There so many reason to crop:

- Take some wildlife... You never have enough mm... With my K3 I can take my 50-135, crop thoses 135mm shoots and get good results at equivalent 300mm. I'am no wildlife lover. But I really like the possibility to have only a modest focal lens, to not carry too heavy gear and get the shoot anyway !
- Creativity is not shoot everything at 3:2. For many subjects, 1:1, 1:2, 4:3, or 1:2.5 are also very interresting ! Each time you don't use the native ratio, you loose a few MP and go do easily to 16 or even 12MP for a pano.
- I don't like doing panorama because it is sooo boring. Post processing and taking the picture easily take 10 time the time both and you have to be carefull of first plane, moving objects, wind, change of exposure. Thoses serious about it use tripod and panoramic head... Really not fun. More res mean you take 1 shoot, crop it and you are done. Much more fun!
- Shooting buildings or anything with perspective distorsion is not always good, you might want to let some margin on the photo, correct perspective (as other optical aberations) and finally choose the final framing in post, with the final size of your subject!
- You don't always have the time to perfectly frame your shoots... Meaning some side subject can be disturbing, composition is not perfect... Judicious cropping can really enhance the shoot !

All of this mean, 24MP mean I can crop more and still get good results. There a status where one has only a 16MP sensor, maybe low pass filter, and he use a good, but maybe not perfectly sharp lens. In this position, it happened to me many time, the crop is simply not sharp enough. You see it as not sharp enough when looking at it full screen on the monitor (2MP full HD, nothing fancy). On the opposite, you go 24MP, remove the low pass filter, get a good prime. And you crop for 1:3, some wildlife or whatever... And now it is sharp enough for the screen.

24MP also improve your lenses. They produce sharper images, give more resolution. Even the low end lenses. Well it is usual that for many easy to make focal lenses, a cheapo lens on the K3 will give much more sharpness than a very high end FA limited like on K5. One think 16MP, this is a lot. But there not really 16MP. When you reduce the 24MP to 16MP and compare the sharpness with what produce a K5, both a 16MP, the K3 reduced shoot to 16MP is far sharper.

So I'am all for 24MP. I think that I could use even more. Maybe 36MP. I also think there also a limit where in too many situations on a conveniant form factor (APSC) 36MP would be limited by the lenses at most appertures. I also thing there still this law of diminushing return. And while subtle I see the impact on color deph tax you get at high iso by using K3 instead of K5.

So no more than 24MP for me until the sensors improve. But I would not want a sensor with less than 24MP now. I do not want to give up the comfort it give. For sure 24MP is the max with the technology.

The 28MP samsung might be different... if the technology improved. Anyway 28 vs 24, there is no visible difference. You'd need 36MP I think to really see a gain.

[stevebrot]

- I don't like doing panorama because it is sooo boring. Post processing and taking the picture easily take 10 time the time both and you have to be carefull of first plane, moving objects, wind, change of exposure. Thoses serious about it use tripod and panoramic head... Really not fun. More res mean you take 1 shoot, crop it and you are done. Much more fun!

I forgot to mention this use case. I regularly do severe horizontal crops to provide a pano composition.


Steve

[kh1234567890]

- I don't like doing panorama because it is sooo boring. Post processing and taking the picture easily take 10 time the time both and you have to be carefull of first plane, moving objects, wind, change of exposure. Thoses serious about it use tripod and panoramic head... Really not fun. More res mean you take 1 shoot, crop it and you are done. Much more fun!

You are doing it wrong. With good software, such as Kolor Autopano, stitching is very fast. A typical 3 shot handheld pano can be auto aligned, auto deghosted and rendered in less than a minute or so. No need to crop.

As for the rest - exactly how do higher pixel count sensors make your lenses sharper ? The linear resolution increase between for example K-5 and K-3 sensors is only about 20%.

[Nicolas06]

As for the rest - exactly how do higher pixel count sensors make your lenses sharper ? The linear resolution increase between for example K-5 and K-3 sensors is only about 20%.

It is only 20% if you don't count the low pass filter. When I compare K3 shoots reduced to 16MP and 16MP shoots from K5, the K3 shoots are much sharper. I think both resolution and lack of low pass filter combined you are more at 30-40% more resolution and this is really visible.

Some lowest end lens start to struggle with the K3 like they did struggle with K5. Many of those still take some increase (I don't know maybe 10-15%) and produce visibly sharper image with K3.

Medium quality lenses (like tamron/sigma 17-50, sigma/pentax 17-70 or maybe even 18-135) are able to get the full burst of resolution gain at least in ideal condition. Condition I typically have when shooting landscapes. So before they wheren't mayber exactly as good as the best lenses (FA31, FA77 or DFA100 macro come to mind) but if you shoot thoses best lens with K5, and the medium quality lenses on K3, the medium quality lenses are visibly sharper.

I know that many think a lot of sharpness. They have high opinion on the best/newest sigma f/1.4 that is razor sharp at f/1.4 (or at least f/2 or f/2.8). I mean this is money thrown away if you don't pair them with a body that can really use this sharpness. This is K3, not a K5 or even K5-IIs. If you don't really care much of sharpness, you should not care... But you should not need expensive lenses neither excepted maybe from bokeh/rendering but surely not what MTF they give in tests.

Myself I don't care much as soon it is good enough. Meaning it can get good results at some usable apperture. But I know many want more than that, so K3 or equivalent is critical for them.

I still think the biggest argument to K3 is improved AF... but I instantly noticed my image where much more sharp, very visibly and that I had much more cropping capabilities. This is very conveniant as I do not fear to crop.

---------- Post added 10-05-14 at 01:16 PM ----------

You are doing it wrong. With good software, such as Kolor Autopano, stitching is very fast. A typical 3 shot handheld pano can be auto aligned, auto deghosted and rendered in less than a minute or so. No need to crop..

This is not exactly that. Maybe the software itself or so take only 1 minute (maybe in practice 2-3 minutes in many cases). This also add hassle in the processing flow, become much longer if you do HDR too and add complexity also in the shooting process: change to a longer lens so you can make the pano, go manual mode to keep shooting parameters constant, take a fake image before and after to quicky figure out it is a pano...

I mean I do not want to say that you should not do it, or that you should think it is boring too. Everybody is different. But I don't like the "you are doing it wrong" because I don't like stiching panoramas and I prefer when possible to use the possibility given by my gear. The process is much faster to crop 1 photo. It was a little risky with K5 as you could have only a 1200-1500 pixel high picture (1500 + maybe some little reframing) in the end and the low pass filter would make it effective of 800-1200 pixels. With the K3 you get 1500-2000 effective pixels (2000 + same eventual reframing) high without the low pass filter tax.

It happen to be enough while it was not with K5. You are much more comfortable with cropping with K3 than K5. You can see it otherwise: the K3 resolution increase is rougly the resolution you would get natively with K5 by doing a pano of 2 pictures (maybe a little less, but there still the low pass filter thing). But it come for free, without any stiching software, without deghosting, without needing to bother with the pano in the first place.

I still make a few panos, but this is when I want more angle of vision than what my lens provide (like I want 120 or 180°). Thoses are much more than 2-3 shoots and take quite some time to process. I really think that this should be handled directly in the camera like you have on smartphones or with Fuji. I least directly when shooting I can see the result, see it was a success or not and I can forget entirely the software in the post process. Much better.

---

Last edited by Nicolas06; 10-05-2014 at 04:26 AM.

[kh1234567890]

It is only 20% if you don't count the low pass filter. When I compare K3 shoots reduced to 16MP and 16MP shoots from K5, the K3 shoots are much sharper. I think both resolution and lack of low pass filter combined you are more at 30-40% more resolution and this is really visible

I am still waiting for anyone to show me that the low pass filter (or its absence) makes any difference in real life.

And even 360 degree panos only require 6 shots with a fisheye lens

[Nicolas06]

I am still waiting for anyone to show me that the low pass filter (or its absence) makes any difference in real life.

And even 360 degree panos only require 6 shots with a fisheye lens

You asked how K3 improved sharpness I explained my own experience with K5 and K3. I'd give you an example but honestly I don't want to spend lot of time trying to convince. The evidences are here, easy to grab and refine. This thread is not for everybody to share opinions, not to argue and explain others fail to see how removing the low pass filter change nothing or how one should always go out with is fish eye to make the ultimates pano !

So here the old example I already given. I did that before I brought the K3 to check the real high iso performance and sharpness of K3 vs K5. The example are in fact from the raw of dpreview, at 6400isos both taken by DFA 50mm macro at f/5.6

The K50 crop:
K50.jpg_crop
by Nicolas Bousquet, on Flickr

The K3 shoot:
K3-16MP_crop
by Nicolas Bousquet, on Flickr

You can help yourself and compare at 100isos (do so with raws, that's important) and you can check also many more pictures on the net. You can also waste your time between K5 & K5-IIs shoot if you want to argue on how much come from the low pass filter removal and how much come from the added pixel count.

Still this is not on the point that the pictures are sharper and that this sharpness show also on medium end lenses and make them look better. This is just how it is. Like it or not

After is it usefull? Well for everyone to check depending on his usage and way to take photos !

[kh1234567890]

You asked how K3 improved sharpness I explained my own experience with K5 and K3. I'd give you an example but honestly I don't want to spend lot of time trying to convince. The evidences are here, easy to grab and refine.

Unfortunately they are not. Not real life, hand-held shots taken with the same lens and under the same light conditions.

[stevebrot]

I am still waiting for anyone to show me that the low pass filter (or its absence) makes any difference in real life.

I am really alive and I know that in at least one area, absence of a low pass filter does make a difference.

Not that it makes any difference in the real world except for continuation of a thread hijack. To be honest, I am not sure I understand your point or what you are trying to accomplish. A new thread perhaps?


Steve

---------- Post added 10-05-14 at 12:45 PM ----------

Not real life, hand-held shots

Comparisons are typically done on tripod to remove the variability introduced by hand-held shooting. In a real world comparison of 100 shots hand-held between the K-5 and K-3 of a suitably detailed subject, with a blind comparative ranking between shots (softest to sharpest), the K-5 shots will cluster on the soft end. Come to Vancouver and we can do the comparison side-by-side.


Steve

[kh1234567890]

As they say, 'this thread is worthless without examples'. Let's see some comparison shots (taken with the same lens, hand-held, under the same light etc. conditions) where the extra megapixels or the absence of a low pass filter make an appreciable difference.

[RonHendriks1966]

As they say, 'this thread is worthless without examples'. Let's see some comparison shots (taken with the same lens, hand-held, under the same light etc. conditions) where the extra megapixels or the absence of a low pass filter make an appreciable difference.

I never spend any time on such things. I use high quality lenses so all my camera's can make sharp images.

[carrrlangas]

I don`t know about resolution*, but without doubt the K-3 has more resolving power because it`s pixels are half the size as those from the K-5IIs. So, it should be able to see smaller details, as much as half the size the smallest detail the K-5 can see, assuming the K-3 is able to keep noise as low as the K-5 at any (equal) given settings.

Sensor / pixel dimensions comparison: Pentax K-3 vs. Pentax K-5 II - Sensor Comparison

*I`d like to see a K-3 vs K-5IIs hand held comparison too. Using the same lens, same ambient light conditions

[Vitalii]

my k-x still has smoother colors. =) so 16MP would be just fine...
---------------------------------------------------------------------------------------
people who crop, will burn in hell...

[stevebrot]

As they say, 'this thread is worthless without examples'. Let's see some comparison shots (taken with the same lens, hand-held, under the same light etc. conditions) where the extra megapixels or the absence of a low pass filter make an appreciable difference.

Ok...you first. You post a link to your best with the K-5 and one of us will do better. Start a new thread though.

• Hand held
• JPEG with link to TIFF
• Same pixel dimensions as a K-3 image (yes, you will have to upsample 1.23x to get there)

Points subtracted for PP artifacts.

Steve

[Nicolas06]

Unfortunately they are not. Not real life, hand-held shots taken with the same lens and under the same light conditions.

I'am really sorry for anybody that has hand shake problem that prevent the images they take being sharp as I guess there much more problem involved with that symptom than just blured pictures.

On my case the K3 give me sharper images. It does so even with medium/low end lens like 18-135 at f/8... But it does it continuously with lens like my DA 35 plastic wonder, my FA50, my DA50-135 or my FA77. This allow me to crop more and to not avoid stiching panoramas. That may be of no use to some, I completly agree and those should not buy the K3 for the added resolution. But the difference is there.

I don't know what you want to conclude from this. I'am supposed to be a liar? Or victim of some placebo effect? Or maybe I have better hands than average people?

Anyway, it work for me, and I'am sure it would work for other. Too bad it didn't work for you. I'am mean, if you tried at least.

[kh1234567890]

Ok...you first. You post a link to your best with the K-5 and one of us will do better. Start a new thread though.

• Hand held
• JPEG with link to TIFF
• Same pixel dimensions as a K-3 image (yes, you will have to upsample 1.2x to get there)

How would that show anything ? If I had a K-3 body handy I'd be happy to oblige. Surely there must be someone with both a K-3 and a K-5ii/iis body to hand who could take a quick shot, handheld, outside in daylight, swap the lens, take another one at the same settings and post the two ? It does not have to be a work of art.