Is the Panasonic GX7 a Sony NEX7 killer?

Started Aug 2, 2013 | Discussions thread
EEmu Forum Member • Posts: 69
Re: Is the Panasonic GX7 a Sony NEX7 killer?

jim stirling wrote:

EEmu wrote:

kcamacho11 wrote:

Bigger sensor >>>> better image quality, better depth of field, better detail

End of story.


Maybe if you're comparing FF to M43, but Sony's APS-C? It's not all that different.

Sony's APS-C is only 2/3 stops bigger than m43 by area, with most of that in the width because of the the different aspect ratio. Cut down to 4:3 it's only 1/2 stop larger. At that point differences will be more sensor dependent than format dependent.

Actually the Sony APS sized sensor has a .75 stop advantage regarding DOF control which may not sound a lot but it boils down to an F1.8 on Sony/Nikon/ Pentax APS having the same DOF as an F1.2 Len on mFT. Though the size difference should dictate a 0.75 difference the best APS bodies ( D5200/D7100) have a 1.4 stop advantage in DR and colour depth over the best mFT options .|0/(brand)/Nikon/(appareil2)/793|0/(brand2)/Olympus

(Was on vacation, but since this thread is still active I figured I'd reply.)

First off, if we're going to be picky then it's worth pointing out that the size difference is .705 = log2(23.5*15.6/17.3*13.0) or somewhere in between 2/3 (5.4% error) and 3/4 (6.4% error) stops. That's just the size, however. The aspect ratios are different so it's not really a meaningful comparison. In reality, an APS-C sensor is 1.2x taller and 1.36x wider.

Without going to far into the physics, pixel size trades off resolution with noise and, to a very limited degree, dynamic range. If an APS-C sensor has the same pixel size and sensor quality as a m43 sensor than we can expect that, per pixel, they will have the same noise and DR (and the APS-C will have more resolution). If you instead keep the same vertical resolution then the APS-C will have only a log2(1.2*1.2) = .53 = 1/2 stop advantage in noise and DR along with a few pixels on the sides. In the real world, APS-C comes somewhere in between (somewhat larger pixels, somewhat more resolution).

In terms of dynamic noise, the comparison you posed is unhelpful. The D7100's sensor has more dynamic range not because it's bigger but because it's better. If you compare the parameters of the D7100 and the E-M5 you'll see the D7100 has exceptionally low sensor/read noise at low ISO while the E-M5's is just okay.  This gives it really excellent dynamic range because read noise (which dictates the black level) is one of the biggest factors in calculating that. Because DR depends so strongly on individual sensor performance it's a lousy point of comparison between formats. Or are you trying to say that you think a Canon S100 is almost as good as an E-M5 (DR 11.5 vs 11.9)?

If you compare the SNRs, however, you'll see that they're basically the same. That's because senor quality doesn't really factor in there. The D7100 has 16% more vertical resolution so as predicted it has similar noise to the E-M5 per pixel (see DxO's Screen SNR). If you scale the images down to the same resolution (DxO's print resolution) you get 1/2 stop better noise performance and a bit more width, just as predicted by theory.

Depth of field is well approximated by the image magnification (skipping the math; it's long). Because an APS-C sensor is 1.2x taller, you can magnify the image 1.2x more and thus use 1/1.2x the f number. This equates to an advantage of 2*log2(1.2) = .53 = 1/2 stop.

Obviously that applies to images where the vertical is the limiting factor. APS-C is 1.36x wider and thus has a DoF advantage of .88 or basically 1 stop for images frames along the wider direction. This is probably APS-C's single biggest advantage, but really is only applicable for portrait oriented, well, portraits, IMO. For completeness, I'll mention that diagonally the DoF advantage is 3/4 stop which is the reason for the often cited 3/4 stop advantage. I, however, never match my image magnification to the diagonal, so I find that to be of little practical value.

There's the long story. APS-C does have some nice sensors now (beating full frame) if you care about DR at low ISO. Beyond that, however, they generally only have about a 1/2 stop advantage in terms of DoF and IQ (where I weight noise higher than 13 vs 12 stops of DR). If you like framing against the long side of the sensor, however, you can get almost a full stop improvement in DoF and IQ to an extent.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow