The Camera Sensor - Impact and Utilization

[Pica]

Hello Everyone,
I'm now reading through the inputs received on the OP I created, according to my thanks in:
The Camera Sensor - Impact and Utilization


I believe this to be a fruitful teamwork, and my first supplementary question concerns the following which I found interesting:

As for Pica's concerns, as far as I know, the D5300 produces it's lower image sizes by re-sampling - unlike my Sigma which uses true binning on-sensor.

Ted and Everyone,
It would be valuable to learn more about the second part, i.e. “uses true binning on-sensor”. I have as always made an effort to research the to me new concept of “true binning on-sensor”, however haven't found anything informative yet. Please, Ted (everyone) would you like to elaborate?

Thanks.
-

[xpatUSA]

Ted and Everyone,
It would be valuable to learn more about the second part, i.e. “uses true binning on-sensor”. I have as always made an effort to research the to me new concept of “true binning on-sensor”, however haven't found anything informative yet. Please, Ted (everyone) would you like to elaborate?

Pica, as far as I know, true on-sensor binning is only done with Sigma/Foveon sensors:

http://www.foveon.com/article.php?a=71

Other manufacturers do binning but in software, not in hardware on-sensor. It's not quite the same, IMHO.

Foveon's binning method is to literally short photodiode outputs together. In the case of 2x2 you get 4x the area but also 4x the capacitance. Therefore, the output of a Foveon binned pixel is the arithmetic average of the pixels it's made up from. For 2x2 you get a half-width image, a much smaller file size and about 0.7 of the noise that would come from a full-size image, all other things being equal. The low resolution image is sharper (on a per pixel basis). Moire is more likely if the scene has high frequency details, especially with a steady hand and a good lens.

[Manfred M]

Pica - "binning" is when you combine the data from one or more photodiodes that are adjacent to one another and effectively making them act as if they are a single photocollector.

The Foveon sensor design (the three colour layers stacked on top of one another, much like traditional colour film) makes doing this a bit simpler than with a traditional Bayer array that is used in all other cameras.

My tests with my Nikon and Panasonic cameras suggest they use software to downsize the image, not the binning technique described by Ted.

[dem]

The Fuji EXR technology uses binning - the raw file coming from a 16Mpx sensor is essentially an 8Mpx file of dual pixels. You have three choices of how to use the sensor:

1) increase the level of detail, this means L-sized JPEGs as every pixel is processed individually

2) decrease the noise by averaging two neighboring pixels as Ted described above

3) increase dynamic range by applying different signal amplification (different ISO) to the two neighbouring pixels and then reconstructing the image as in HDR technique.

If you want to access options 2) and 3), you will have to do with the M-sized JPEGs, i.e. 8MPx.

[Pica]

Hello Ted, Manfred and Dem,
This is to say thank you for taking part during my supplementary topic on: “True binning on-sensor”. I appreciate it.

With a background from different roles in engineering as well as teaching, I appreciate too your link, Ted. In it, I do like the graphics, etc on the concept of binning. (And I agree with you concerning your comment that doing an action in HW versus SW, are different things.)

Thank you Manfred for your second informative post to support this camera sensor quest of mine! I read the first one with great interest as well.

Dem I recently learned about the concept of Fuji EXR; thus I appreciate your post which describes more about it. Thanks for conveying the information to me.

I’ll be back if needed. All the best from Scandinavia!
-

[Pica]

Hello Ted, Dem and Everyone at this forum,
I “seasoned” researcher who often finds what I’m searching for. However this time I don’t know where to begin.

As an example, say that I am wishing to learn more about the binning capabilities of a specific camera and model. If it is not in the specification, where do you go to find the information (e.g. below) and if it is performed via HW or SW? (The vendor’s homepage or some other place?)

Thanks, Pica

Pica, as far as I know, true on-sensor binning is only done with Sigma/Foveon sensors ...

The Fuji EXR technology uses binning …

[dem]

In my case you buy a Fuji EXR camera (I've got the X-S1) and then start wonder why raw files have half the number of Mpx of JPEG files. There are lots of blogs and forum discussions about the X10, X20 and X30 (most of them 3-4 years old) that are dedicated to the EXR sensors.

As a side note, Fuji have discontinued all their small sensor X cameras (recently X30 and XQ2, X-S1 about 2 years ago) and now seem to concentrate on the X series with APS-C-sized Trans sensors.

[Pica]

Hello Dem,
Thank you for your reply on the last item. I appreciate it.