5D3 vs D800: Microlenses

Started Apr 22, 2012 | Discussions thread
stephenmelvin
Veteran MemberPosts: 4,547
Like?
Don't be too quick... ;)
In reply to Great Bustard, Apr 23, 2012

Great Bustard wrote:

stephenmelvin wrote:

The 5D3 sensor:

http://www.sensorgen.info/CanonEOS_5D_MkIII.html

is essentially the same as the 5D2 sensor:

http://www.sensorgen.info/CanonEOS_5D_MkII.html

I think it's more accurate to say that it's similar to the 5D2 sensor. After all, its noise characteristics are very different, from the elimination of horizontal pattern noise at low ISO and almost no pattern noise at high ISO's.

From what I've seen, the noise characteristics are quite similar for the RAW files (Canon has obviously been working on the NR for their OOC jpgs). The greater QE from the new microlenses gives a 0.57 stop advantage, and they've done some work to reduce the banding.

You clearly haven't seen any RAW files. I only work in RAW, and take it from me, the noise characteristics are very, very different. Why do you assume I'm talking about jpegs? I didn't spend $3,500 to shoot jpegs.

In other words, basically the same sensor with a tune-up.

Canon says it's "newly developed." I take them at their word.

but it has a signficantly higher QE (0.57 stops higher). This is most likely due to a more efficient gapless microlens covering.

That no doubt has helped.

Of course.

As it turns out, the efficiency of the microlens covering has much to do with how much of the light at wide apertures makes it into the photosensitive area of the pixel.

The issue is angular sensitivity, not overall coverage. Unless the individual microlenses have a wider angle of view, then they won't have any more capability to "see" the larger apertures than the existing microlenses.

It's the speed of the microlenses. The microlens needs to be faster than n x f where n is the f-number of the taking lens and f is the fill factor. So, if you've got an f/2 lens and a fill factor of 0.5, you need f/1 microlenses. (Bob, does that sound about right? )

No, it really is about the angles. The microlenses have to be able to see the edge of the f/1.2 exit pupil if it's going to affect the exposure. The speed of the microlens has nothing to do with it; it's the angle of view.

It's like our focusing screens. Modern screens are bright, because they're tuned to f/4 lenses. But they're useless for focusing fast lenses, because they're blind to the larger apertures. Ever notice that the high precision screens are darker?

The 5D2, for example, lost 0.35 stops of light at f/1.4 and 0.55 stops of light at f/1.2 (essentially making f/1.4 and f/1.2 the same effective aperture):

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/F-stop-blues

The D3s lost 0.3 stops of light at f/1.4 (essentially the same as the 5D2), and, presumably, also about the same at f/1.2.

So, the question is if the new microlens by Canon might not improve upon that significantly, and give the 5D3 a greater low light wide aperture advantage over the D800.

It would be very nice indeed. And let people who own Canon's pricey super speed lenses get what they paid for. But doing this may reduce the overall sensitivity of the microlenses.

It would have no adverse affect whatsoever.

I suspect it would reduce sensitivity, just like having a high precision focusing screen reduces the brightness of the image in your viewfinder.

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