[Nicolas06]

I actually haven't seen any threads about lenses not being able to keep up with a K-3.... I'd be interested in seeing those. My understanding was your image resolution would be increased using a K-3 no matter what lens you use... so, I'm kind of surprised that there are threads stating definitely the contrary. Anyone got links?

Are we talking real world examples or geeky speculation?

I think this a problem with the 18-55, when the first tests with K3 where done some reviewer did it with the 18-55 kit lense and there was no real improvement compared to a K5.

I would suppose other basic lenses like some 70-300 would maybe show no improvement.

I know for sure that 18-135 or DA21, FA50, DA50-135, DA35 all show improvement with K3 through.

That's absolute hogwash, for it to be true the 24MP sensor would have to destroy resolution... it might produce the same lw/ph, it couldn't be worse. There's a difference between a video and a reliable source. The only qualification to make a video is... you have a video camera. You don't have to know a thing.

I would expect Canon's 24-70 to perform better on a 24 Mp APS_c sensor than on an 18 Mp APS-c sensor, oh wait, does Canon even have a 24 Mp APS-c to test it on... how would they even know how a 24-70L performs on a K-3 type camera?

So I think I get it now. The 24-70 L performs poorly on cameras Canon doesn't make. That seems like a reasonable argument. In Nikon terms, a D7100 with the same lens will produce maybe 100w/ph, or less than a 5% advantage for the full frame. So yes there is a slight disadvantage to shooting on APS-c in terms of resolution. But it's not something to buy or not buy a camera for, and you need a test facility to measure it.

if both FF and APSC sensor have 24MP, there only 10MP worth of detail on the FF shoot meaning the sensor is far less demanding in term of fine detail than the APSC sensor.

[osv]

you can mount the canon 24-70L onto any sony mirrorless camera, so it could be tested at 24mp aps-c... i am skeptical that it could actually be worse.

not unheard of tho, the Tamron SP 150-600mm is supposed to be worse on aps-c than it is on ff, tested by dxo? Tamron SP 150-600mm F/5-6.3 Di VC USD Lab Test Review: Digital Photography Review

[normhead]

if both FF and APSC sensor have 24MP, there only 10MP worth of detail on the FF shoot meaning the sensor is far less demanding in term of fine detail than the APSC sensor.

huh?

I think this a problem with the 18-55, when the first tests with K3 where done some reviewer did it with the 18-55 kit lense and there was no real improvement compared to a K5.

I rarely take what "some reviewer" says as gospel.
Imaging resources, using actual MTF measurements said some cheaper lenses improved their resolution as much as 60%, when they went to a D7000 to a D7100 (16MP APS-c to 24MP APS-c). The better lenses were limited to about a 20% increase, they were already pretty good on the D7000. So, I'd like to see what "some reviewer" is talking about, just so I can point out how he messed up.

[formercanuck]

Well, I have some issues with Tony Northrup's video in general... not that I have the years of photography experience that he has... but in general - if I'm looking at DxO mark, for Pentax K mount, it pretty much doesn't matter what lens (FF or APS-C) is shown, the P-MPix is fairly similar.
Also for going up to high end glass for FF, I'm not sure why he's picking FF zooms on a crop sensor - makes for a large/heavy zoom, IMHO - especially if you use ones with OS in the lens.
I don't expect any 'real' difference perceived or otherwise on a prime... possibly sharper (hey wait ... he actually showed that at the end ).

Interesting piece to note - with DxO mark. Take a look at my favorite cheap FF lens... Samyang 85mm f/1.4
HUGE difference in Nikon between a D810/D800E and say a D3300/D7100, 23 P-MP to 13P-MP. Go to say a Canon, and there's not much difference between it with 5D Mark III, 7D, 700D, 1Ds Mark III (relatively speaking). 13P-MP to ~10P-MP.

---

Last edited by formercanuck; 02-23-2015 at 11:38 AM.

[audiobomber]

Imaging resources, using actual MTF measurements said some cheaper lenses improved their resolution as much as 60%, when they went to a D7000 to a D7100 (16MP APS-c to 24MP APS-c). The better lenses were limited to about a 20% increase, they were already pretty good on the D7000. So, I'd like to see what "some reviewer" is talking about, just so I can point out how he messed up.

The big jump for Pentax occurs when the blur filter is dropped:

Sharpness comparison between Pentax K-5 IIs and K-5 II - DxOMark

The increased resolution from K-5 IIs to K-3 is not as impressive for most lenses:

Sharpness comparison between Pentax K-5 IIs and Pentax K-3 - DxOMark

The Sigma 35mm 1.4 showed a 23% increase in resolution going from the K-5 IIs to the K-3, much higher than others tested, which means that most Pentax lenses are nearing their max resolution at 24mp with no blur filter on an APS-C sensor. Resolution on the upcoming Pentax FF body will increase for FF compatible lenses.

[Nicolas06]

huh?

The APSC framing is smaller than the FF framing. When the 2 sensor are said to have the same MP (eg K3 & A7 for example) if you take a crop of the FF to match the APSC area you'd get 10MP. 10MP is less demanding for the lense than 24MP and thus easier to max out.

A lense on FF need only 2/3 (1/1.5) on the lp/mm to cap the sensor resolution in the center because the pixels are much bigger.

So the question finally is how much quality your lense can deliver. Your hypothesis normhead is that any lense is so good that it cap any sensor, and so it is mostly the sensor that will change the performance.

The reality of measurement is different some lenses have very high resolution and, as an example macro lenses tend to perform better than some other. High end lenses tend to also perform better than some others. This mean their resolution in corners might be better, the overall contrast might be better and the resolution again might be better.

As said here, Sigma 35 Art gain more resolution from K5-IIS => K3 than many more lenses. This mean this lense has more margin than many.

If we look at DxO what do we get by going from K5-IIs to K3, counting in theory we should get 50% more MP on their score if the lense was perfect:

- DA20-40: 9=>9MP 0%
- 18-270: 6=>7MP 16%
- DA35 macro: 12 => 13 8%
- DA55-300: 7 => 7 0%
- DA70 ltd: 10 => 11MP 10%

And if we compare K5 => K3:
- DA20-40: 7=>9MP 28%
- 18-270: 5=>7MP 40%
- DA35 macro: 9 => 13 44%
- DA55-300: 7 => 7 0%
- DA70 ltd: 8 => 11MP. 37%

Most of the difference in fact as explained comes from the low pass filter removal. The best lenses like DA35 macro and DA70 ltd still see more improvement by going to K3 but low end lenses like 18-270 or DA55-300 do not see that much improvement.

For me I have no issue with that it really match what I observe:
- I saw most of my lenses being better from K5 => K3 but of course I don't know what is more responsible for it: added resolution or low pass filter removal.
- I see that my best lenses are sharper/better than the worst one, very noticably. We can really see it DA18-270 is much worse than DA35 macro for example.

If we go back to APSC vs FF, FF need less lw/mm to achieve the same picture resolution in the end so at least in the center it is less demanding than APSC lenses.

[Rondec]

huh?



I rarely take what "some reviewer" says as gospel.
Imaging resources, using actual MTF measurements said some cheaper lenses improved their resolution as much as 60%, when they went to a D7000 to a D7100 (16MP APS-c to 24MP APS-c). The better lenses were limited to about a 20% increase, they were already pretty good on the D7000. So, I'd like to see what "some reviewer" is talking about, just so I can point out how he messed up.

It doesn't make sense to me that cheaper lenses would benefit more from a resolution bump than expensive lenses. Maybe something like a 50mm f1.8 or the 35 f2.4, but consumer zooms are probably pretty maxed out on 16 megapixels and just won't see much of a bump going to 24 megapixels.

I have nice lenses and honestly I have to make sure everything is just right in order to see a bump in resolution. It is there, but it just doesn't always show up in real world shooting.

[Sagitta]

It doesn't make sense to me that cheaper lenses would benefit more from a resolution bump than expensive lenses. Maybe something like a 50mm f1.8 or the 35 f2.4, but consumer zooms are probably pretty maxed out on 16 megapixels and just won't see much of a bump going to 24 megapixels.

I have nice lenses and honestly I have to make sure everything is just right in order to see a bump in resolution. It is there, but it just doesn't always show up in real world shooting.

I have cheap lenses. I'll have to employ some unscientific testing to see if they fare better than they did with the K-x/K-30.

If you think about it, the reverse could be true as well - that high end lens you bought for the FF would be worse for the cropped sensor, which somehow doesn't make any sense for me.

[audiobomber]

If you think about it, the reverse could be true as well - that high end lens you bought for the FF would be worse for the cropped sensor, which somehow doesn't make any sense for me.

I like that DXOMark tests the lens and camera as a system. Here are my conclusions from viewing their sharpness results with various combinations:

1. A larger sensor increases sharpness, all else being equal
2. More pixels increases sharpness, if the lens is capable of higher resolution
3. A blur filter decreases sharpness, all else being equal

APS-C can show greater sharpness than FF, depending on items 2 and 3. A D7100 (24mp, no blur filter) can usually out-resolve a Df (16mp, with blur filter). Given a capable lens, an a6000 (24mp, APS-C, with AA filter) beats an A7S (12mp, FF, with AA filter).

Matched with the best lens in the world, a perfect 12mp sensor can only resolve 12mp, regardless of the size of the sensor.

[normhead]

I'm still waiting to see results from a lens, that in some circumstance doesn't benefit from a higher resolution sensor, maybe you guys who look at these numbers could highlight one for us. People keep saying this " lens out resolved by the sensor thing" but the only guy I know who calculated that came up with 51 MP as the number where APS-c sensors start out resolving lenses, and we area a long way from that with current technology.

I mean really... has anyone got anything? Anything at all?

[ZeljkoS]

I saw a test for 40mm F2.8 XS lens on dxomark.com and it had a higher sharpness score when used on K-5IIs then for K-5 II or original K-5
all three cameras have the same sensor, sans AA filter on S version. It would be interesting to see this same lens tested on the future FF.

[formercanuck]

Tony's video also doesn't take into account of 'why' people (Pentaxians) purchase FF lenses vs. say Canon/Nikon/Sony users.

1. Almost all legacy glass for Pentax is FF (I'm sure the same can be said for Canon/Nikon/Sony)
2. There's a lot more cheap legacy prime glass than current. This includes old A, M42, etc.
3. There's a lot of fast primes out there.

Can anyone find say a 135mm f/3.5 for < $100 ?
How about a 28mm f/2.8 ?
50mm and 35mm plastic fantastic are the closest you'll find.

My best overall reference from FF to APS-C with this is Samyang 85mm f/1.4
Nikon D610 - 24MP vs. Nikon D5300 24MP.
D610 = 17P-MPIX, D5300 = 15P-MPIX.

Not much of a difference, when comparing a $2k body with a $800 body for the same FF lens.

[audiobomber]

I'm still waiting to see results from a lens, that in some circumstance doesn't benefit from a higher resolution sensor, maybe you guys who look at these numbers could highlight one for us. People keep saying this " lens out resolved by the sensor thing" but the only guy I know who calculated that came up with 51 MP as the number where APS-c sensors start out resolving lenses, and we area a long way from that with current technology.

I mean really... has anyone got anything? Anything at all?

The D FA 100mm macro does not show a resolution improvement used on the K-3 vs. the K-5 IIs.

Sharpness comparison between Pentax K-5 IIs and Pentax K-3 - DxOMark

[normhead]

You do know that DxO tests in less than Ideal lighting conditions? I have no idea how their data relates to normal shooting conditions, and as far as I can tell either do they. I know that DxO rates lenses, they don't offer any images to let you see what they are actually talking about so I find them pretty useless. You know what their numbers are, you don't know how they translate to actual IQ.

Here's a couple shots I just ran off.....




I'd be amazed if anyone can't tell the difference between the two cameras. These taken with my weakest lens, the A-50 ƒ2.

Admittedly the difference are small, but in my experience, the difference between 16 and 24 isn't that much.

[FreeSpirit9]

You do know that DxO tests in less than Ideal lighting conditions? I have no idea how their data relates to normal shooting conditions, and as far as I can tell either do they. I know that DxO rates lenses, they don't offer any images to let you see what they are actually talking about so I find them pretty useless. You know what their numbers are, you don't know how they translate to actual IQ.

Here's a couple shots I just ran off.....




I'd be amazed if anyone can't tell the difference between the two cameras. These taken with my weakest lens, the A-50 ƒ2.

Admittedly the difference are small, but in my experience, the difference between 16 and 24 isn't that much.

K3 first image
K5 second image

So which is which Norm?