[falconeye]

Keeper rate, not ratio

Equally bad, doesn't help.
All keepers could be from a fast burst at the beginning of a sequence with little change of focus. Moreover, AF performance should no just measure a camera's burst rate.
The solution is like I wrote above.

[Nicolas06]

Equally bad, doesn't help.
All keepers could be from a fast burst at the beginning of a sequence with little change of focus. Moreover, AF performance should no just measure a camera's burst rate.
The solution is like I wrote above.

We can criticize test protocol but this doesn't change the point 1 bit...

we are still waiting to see the theoretical superiority of Sony on sensor PDAF... Until then this is just wishfull thinking.

[IchabodCrane]

We can criticize test protocol but this doesn't change the point 1 bit...

we are still waiting to see the theoretical superiority of Sony on sensor PDAF... Until then this is just wishfull thinking.

On-senosr PDAF is not just "theoretical". Olympus legacy 4/3 lenses can focus at a snap with the E-M1 and PDAF where they take a long time to focus when using CDAF on the other bodies.

[Nicolas06]

On-senosr PDAF is not just "theoretical". Olympus legacy 4/3 lenses can focus at a snap with the E-M1 and PDAF where they take a long time to focus when using CDAF on the other bodies.

I do not say they don't exist... it is more that they are better for action/sport that I challenge...

And well you know ANY DSLR focus at a snap, it is not like it is extraordinary.

[kadajawi]

You may speculate so.
But people now shooting FF have no intention to go APSC ever again in their life. Nobody of those *I* know, at least.

If anything, they add smaller than APSC gear to their arsenal, like L/RX100.

BTW, a pro-level 24-70/2.8 may be heavy and expensive. But I found it worth every penny. Between an RX100 24-70 and FF 24-70, there is no room for APSC. The RX100 is (in equivalent terms) an F/4.9-7.6 lens, the FF F/2.8 of course. An APSC kit zoom is worse, a DA* 16-50/2.8 is an F/4, just a bit better than the RX100.

So, the advantage of an APSC dSLR with zoom is that it is cheaper. Otherwise, it hase been eaten up from below and above. Already.

The only room for APSC I can see in the foreseeable future are pro-level APSC mirrorless cameras like NX1 or A7000. When they beat SLRs in something. But here again, cameras like the A7 may eat that cake too.

For me FF doesn't make much sense. SR loses effectiveness with larger sensors too (just look at the A7R II). APS-C is a wonderful format to me, not too big, not too small, close to Super 35, ... the only thing I'd like FF for is the larger viewfinder, but an EVF would solve that, or a variable translucency mirror.

[Pål Jensen]

What format you prefer is totally subjective. Anyone who tries to make their choices into some objective truths are really just displaying their psychology...
I agree, APS is pretty much an ideal format but that's just my subjective opinion.

[Mikesul]

For me FF doesn't make much sense. SR loses effectiveness with larger sensors too (just look at the A7R II). APS-C is a wonderful format to me, not too big, not too small, close to Super 35, ... the only thing I'd like FF for is the larger viewfinder, but an EVF would solve that, or a variable translucency mirror.

The A7RII won't be available for us to look at for 3 weeks. Perhaps you meant the A7II which has a very effective shake reductions system. Shake reduction seems to work very well on FF.

---------- Post added 07-19-15 at 06:51 AM ----------

What format you prefer is totally subjective. Anyone who tries to make their choices into some objective truths are really just displaying their psychology...
I agree, APS is pretty much an ideal format but that's just my subjective opinion.

Well put. There are too many variables for such judgments to ever be definitive.

[kadajawi]

The A7RII won't be available for us to look at for 3 weeks. Perhaps you meant the A7II which has a very effective shake reductions system. Shake reduction seems to work very well on FF.

---------- Post added 07-19-15 at 06:51 AM ----------



Well put. There are too many variables for such judgments to ever be definitive.

From what I recall from video samples (maybe the A7 II, can't remember) the video was quite stuttery with longer lenses. On APS-C (Pentax) 200mm is stuttery, for the Sony it was earlier, while Olympus has no problems at 200mm. Something like that.

For stills SR has little meaning for me. I don't really need it... there is little difference. Video is the thing where I notice SR a lot.

[Winder]

SR loses effectiveness with larger sensors too

Shake reduction seems to work very well on FF.

Larger sensors don't need as much compensation. The old shutter speed rule of 1/focal length applies to FF, but with APS-C you really need 1/focal length x 1.5. For equalized output you are magnifying the APS-C image more which means you are also making the blur from camera movement more visible. Glass for a Pentax 67 doesn't need to be as sharp as glass for an APS-C for the same reason. For APS-C to really compete with FF at the same output size, APS-C has to have better SR and better glass.

The larger the sensor, the more depth you get in the image at a given output size. If I'm printing at A3+ and I take a picture with both a D810 and a K-3.

A3+ is 158,907mm^2
D800E sensor size is 864mm^2
K-3 sensor size is 370mm^2

To print at A3+ the image from a D800E has to be enlarged 183 times.
To print at A3+ the image from a K-3 has to be enlarged 430 times.

The more magnification you apply to an image the flatter the final image. You magnify all of the flaws in the lens, all the distortion, camera movement, & all of the noise. This is one of the reasons DxO scores are of limited real world value.

[Pål Jensen]

Larger sensors don't need as much compensation.

..but there's significantly more mass that needs to be moved (and stopped). Probably the reason we haven't seen it implemented in the 645...

[D1N0]

I think the array would only be marginally larger when you look at the amount of free space around the sensor.

[eyeball]

Actual lenses being sold show that both WA AND tele are bigger on FF than on APSC and also noticably more expensive. The WA thing might be actually true for mirrorless FF of Sony, but in exchange their tele are even bigger.

Just see what a DA15 or DA70 provide in term of size/weight and we can discuss what FF 22mm and 105mm lenses on FF allows for the same weight/size/quality as what is available on APSC.

Examples of what I had in mind : The Pentax-m 20mm F4.0 on FF is as compact as the 15mm limited on APS; a 135mm F2.5 on APS serves as well as 200mm 2.8 on FF.

[Fogel70]

Larger sensors don't need as much compensation. The old shutter speed rule of 1/focal length applies to FF, but with APS-C you really need 1/focal length x 1.5. For equalized output you are magnifying the APS-C image more which means you are also making the blur from camera movement more visible. Glass for a Pentax 67 doesn't need to be as sharp as glass for an APS-C for the same reason. For APS-C to really compete with FF at the same output size, APS-C has to have better SR and better glass.

The larger the sensor, the more depth you get in the image at a given output size. If I'm printing at A3+ and I take a picture with both a D810 and a K-3.

A3+ is 158,907mm^2
D800E sensor size is 864mm^2
K-3 sensor size is 370mm^2

To print at A3+ the image from a D800E has to be enlarged 183 times.
To print at A3+ the image from a K-3 has to be enlarged 430 times.

The more magnification you apply to an image the flatter the final image. You magnify all of the flaws in the lens, all the distortion, camera movement, & all of the noise. This is one of the reasons DxO scores are of limited real world value.

As long as FOV is the same, the requirements of the SR will be "equal" on all sensor sizes. The resolution of the SR system has be the same (over the whole sensor) as long as the pixel count is the same, and movement isrelated to the size of the sensor, so the movement of a larger sensor has to cover a larger area at the same duration. The smaller sensor need to make finer movement as the pixels are smaller.

Not sure which is most difficult to do, regulate a larger and heavier sensor in larger movements, or smaller and lighter sensor in smaller movements.
Pentax had no problem designing the SR system for Q, so the change in sensor size between APS-C and FF might not make much difference in SR system.

[ogl]

The more magnification you apply to an image the flatter the final image. You magnify all of the flaws in the lens, all the distortion, camera movement, & all of the noise. This is one of the reasons DxO scores are of limited real world value.
[/COLOR][/LEFT]

A3 printed photo doesn't depend on such calculation.
It depends on quality of printing (200, 240 or 300 DPI, you can even find places where you could print 360 or 400 DPI), 300 DPI is basic now.

11,7×16,5 x 300 x 300 = 17.3 MP is enough for A3 printing.
24 MP is good even for SUPER HQ photos (360 DPI, for example).

I've printed A3 photos from K200D with 240 DPI in photo lab and it looks not worse than the same from Canon 5D.

[Nicolas06]

Larger sensors don't need as much compensation. The old shutter speed rule of 1/focal length applies to FF, but with APS-C you really need 1/focal length x 1.5. For equalized output you are magnifying the APS-C image more which means you are also making the blur from camera movement more visible. Glass for a Pentax 67 doesn't need to be as sharp as glass for an APS-C for the same reason. For APS-C to really compete with FF at the same output size, APS-C has to have better SR and better glass.

The larger the sensor, the more depth you get in the image at a given output size. If I'm printing at A3+ and I take a picture with both a D810 and a K-3.

A3+ is 158,907mm^2
D800E sensor size is 864mm^2
K-3 sensor size is 370mm^2

To print at A3+ the image from a D800E has to be enlarged 183 times.
To print at A3+ the image from a K-3 has to be enlarged 430 times.

The more magnification you apply to an image the flatter the final image. You magnify all of the flaws in the lens, all the distortion, camera movement, & all of the noise. This is one of the reasons DxO scores are of limited real world value.

You mix 3 things: noise, distorsion and camera movement.

SR fix only the movement of the hand. If the hand move by 2mm, you have to move 2mm in the revert, this doesn't change on APSC or FF (3 degre of correction work like this. But if you change the angle, the wider the sensor, the more this angle mean in term of absolute movement and the more effort you need.

For noise, this is true if you have same sensor technology but the key is to know when the signal is strong enough... You aren't going to see the difference from iso 100 to iso 400 if you don't look at your picture at 100%. That's like printing twice the size of A0 if you display your pixels 100% on a full HD screen on 23"...

For the flaws of the lenses, many lenses have greater problem on borders so that's not all shiny for the FF format neither... If somebody goes with its 300 f/2.8 on APSC, you may need you 450mm f/4 on FF to get the same framing... And if you have you 600mm f/4 on APSC, you'll have to look for a 900mm f/5.6 on FF...