E-M5 has DR 13!

Started Apr 13, 2012 | Discussions thread
Detail Man
Forum ProPosts: 14,950
Like?
Re: Output Dynamic Range Specifications are limited by ADC Bit Depth
In reply to dotborg, Apr 15, 2012

dotborg wrote:

cmorse wrote:

I think what hes saying is that if you're only allowed ten numbers you could write

1, 2, 3, 4, 5, 6, 7, 8, 9, 10
or
1, 3, 5, 7, 9, 11, 13, 15, 17, 19

You still have the same number of pieces of data but have spread them over a greater range.

Of course in an image if you did something as simple as that you'd end up with some pretty ugly posterization. With a longer sequence of numbers though they might come up with something clever that extends the range a bit but is subtle enough that the gaps aren't overly noticeable. No idea or if they've done any such thing though.

Precisely.

From what I understand, the photo-sites of a CMOS sensor (yes, I know that the sensors used in m4/3 cameras are not strictly CMOS)

They are PIN photodiodes, semiconductor devices that are different than the Complimentary Meta-Oxide Semiconductor Field Effect Transistors (MOSFETs) that interface with the photodiodes.

... are charged with +3.3v.

Voltage is not charge. It is an electrical potential that has units of Joules/Coloumb of charge.

The photons bombarding the photo-site then replenish the electrons, reducing the charge to zero. Therefore, the pre-amplification voltage range is +3.3v - 0v. The dynamic range of a sensor is determined by the capacitance of the photo-site ...

Which is meaningless unless the MOSFET amplifier circuitry and ADC that follows that transducer can provide an output Dynamic Range that can be measured as the RAW-level data output ...

... and the noise of the circuitry rather than the number of bits it's encoded to at the AD converter.

A statement that is significant only in the situation where the DR (or SNR) of the transducer is significantly lower than the DR (or SNR) of the ADC. This is true when the image-sensor is fully illuminated - but not at all true when the image-sensor is very dimly illuminated (or not illuminated at all), which is the situation that exists when Dynamic Range is tested ... which is what I explained in my original post (the title of which this post of yours is threaded onto). How about reading my first post on the subject-title again ?

Depending on the capacitance, a photo-site could record 10 stops with that 3.3v range or 15 stops.

Please explain how 10 stops (EV) magically transform into 15 stops (EV) in your statement above ?

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