New Canon on 22nd March?

Started Mar 7, 2013 | Discussions thread
bobn2
Forum ProPosts: 31,491
Like?
Re: New Canon on 22nd March?
In reply to qianp2k, Mar 11, 2013

qianp2k wrote:

bobn2 wrote:

jonas ar wrote:

I meant that I have seen picutures from Sony cams using the two sensors with the same exposure (I believe posted by dpr member RicksAstro). They were very similar. The quantitative measurements at DxOmark from a55 and a77 also seem to indicate that they would perform essentially the same and it does seem that the read noise is somewhat smaller in the a77 from your own sensorgen figures? At least in Sony's implementation of the sensor.

Actually, the Sensorgen figures show that the read noise of the 24MP is lower than the 16MP (about 2.5 e- per pixel as opposed to about 3 e- per pixel), which would make the 24MP about equal integrated read noise, as you suggest. It looks like my memory is playing tricks here. So, even that exception to the rule goes.

It only means Sony 24mp sensors are designed later with better read noise but not because it has more MP. Sony could update its 16mp APS-C sensor to have even better read noise. I personally believe 24mp is marketing gimmick - a) it doesn't gain much resolution than 16mp on diminishing return; b) its pixel level IQ is pretty bad even at base ISO that largely makes irrelevant.

-- hide signature --

Hi Peter,

The two things actually go together. As the sensor processes are enhanced, the optimum size for a pixel goes down. Low read noise depends on a small read transistor, but if you have a small read transistor the 'full well capacity' of a pixel reduces (it's limited by the read transistor swing, and small transistors produce a higher voltage for the same electron count than big ones). So designers, as they make the transistor smaller to increase fill factor, or reduce read noise face having a small FWC, which means raising the base ISO (so reducing the amount of light each pixel receives) or making the pixels smaller (which means reducing the light per pixel but maintaining it overall). In the end, the second is probably the better solution, and as Canon always managed to achieve a balance of all three, base ISO, read noise and pixel size. The first generation Nikon CMOS in the D3 wasn't quite as good as the contemporaneous Canons, but went for raising the base ISO as above to achieve seemingly better 'high ISO' performance at the cost of low ISOs. Canon still leads the field in absolute read noise. They do with 18MP what the others do with 24, but if they do make more progress, they too will have to reduce pixel size or raise the base ISO.

-- hide signature --

Bob

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