PIX 2015

Quality of a pixel

Started Oct 6, 2012 | Discussions thread
Shop cameras & lenses ▾
Theodoros Fotometria
Senior MemberPosts: 2,090
Re: Please, reply to his argument... OK
In reply to Jack Hogan, Oct 10, 2012

Jack Hogan wrote:

aggressor wrote:

Total noise of a sensor is determined by the amount of useful information collected divided by the amount of noise introduced in the process. If the amount of useful information collected becomes smaller (because more surface area is wasted, for example) while amount of noise introduced in the process remains the same, total noise of the sensor will increase.

...using lower resolution on otherwise identical production process would allow better noise performance

...the way sensors are built today, there is a surface area overhead per pixel, so given specific level of production technology, when you put more pixels on the sensor, total wasted area will be larger

so back to the original thesis:

The overhead of higher resolution eats up SNR and DR. More space is wasted in between the pixels, if you will.

If the two sensors (with the same physical size) are produced on the same line using the same technology, the sensor with lower resolution will have better noise level and dynamic range.

Seems like a reasonable induction to make.  So what you are asking is whether the naive inference holds true in practice, in the context of 2012 DSLR sensors.

That's an easy question to answer: in practice, given equal output size and the state of the art of sensors circa 2012, sensel size makes very little difference in terms of SNR and in fact smaller sensels are often made to produce slightly better overall DR.  Many contributors to this thread can explain why, if asked nicely.

So no, if two sensors of the same physical size are produced on the same line using the same technology, the sensor with the lower resolution will not necessarily have better noise level and DR in their spec'd working range.



It is really difficult to comment on the subject.... we can only guess..., in theory the depth of the pixel is also important... the shallower it is, the lesser the noise, OTOH deep pixels are favoring DR... another factor is that the center pixels receive direct light while the edge ones under angle, yet another factor is the shape of the microlenses... My opinion is (not expert but I have an engineering degree and have had some discussions with people in P1 and ex-Imacon as well as Sinar) that larger pixels is the safer way to go... As I said its not an expert opinion but it is based on current results I had from looking at D4, D600 and D800 sensors and compare them in equal size prints... I did insignificantly prefer D800 or D600 up to Iso 400, but it was the other way around there after and I didn't mind the difference at low Iso... My other impression in life with both D800 and D800E is that Nikon chose the 36mpx sensor for the "E" only and that D800 was probably made to provide a step and educate the market... I don't think that Nikon would have chosen the 36mpx sensor if it wasn't for the "E" to hit the market...

-- 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