How big is the X10 sensor size?

[Steen Bay]

You can see that the size difference between the 1/2/3 and the 1/1.7" which is 1/1.7" (7.49 x 5.52 mm) on the Canon S100, looks to be minimal but still would affect pixel density and likely IQ at least at higher ISOs to some degree I would think.

I am not sure why it is thought the difference is minimal, one third (or even one quarter) could not be considered minimal no matter how one looks at it.

Well, one third, that's like the difference between iso640 and iso800, and isn't that pretty minimal?

As for pixel density, even at 12mp full resolution the X10 has the lowest pixel density of all current premium compact cameras, when you consider the EXR modes (for low light) the pixel density is less than half that of the closest competition.

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

When you consider the X 10's bright lens, (and the very slow lens at tele of the S100) there will likely be quite a gulf between the two.

From my experience with compact cameras, the sensor size and pixel density is of primary importance and one of the most reliable ways to determine the likely image quality.

Yes, as a rule of thumb, but as the iso1600 RAW comparison below shows, then the 1/2.3" BSI-CMOS sensor in SX40 is actually able to match the larger 1/1.7" CCD sensor in G12 at pixel level, even though the SX40 has 2mp more than the G12.

http://forums.dpreview.com/forums/read.asp?forum=1010&message=39793941

 

[Daniel Lowe]

I am not sure why it is thought the difference is minimal, one third (or even one quarter) could not be considered minimal no matter how one looks at it.

Well, one third, that's like the difference between iso640 and iso800, and isn't that pretty minimal?

No, actually, it's quite significant.

As for pixel density, even at 12mp full resolution the X10 has the lowest pixel density of all current premium compact cameras, when you consider the EXR modes (for low light) the pixel density is less than half that of the closest competition.

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

When you consider the X 10's bright lens, (and the very slow lens at tele of the S100) there will likely be quite a gulf between the two.

It's almost F6!!

From my experience with compact cameras, the sensor size and pixel density is of primary importance and one of the most reliable ways to determine the likely image quality.

Yes, as a rule of thumb, but as the iso1600 RAW comparison below shows, then the 1/2.3" BSI-CMOS sensor in SX40 is actually able to match the larger 1/1.7" CCD sensor in G12 at pixel level, even though the SX40 has 2mp more than the G12.

Oh well, I guess the premium compact concept has just been proven all wrong and we would all be stupid to buy one.
Glad that's all resolved.

 

[Steen Bay]

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

I understand that binning and downsampling both will reduce the amount of noise (and detail) in the image. It really doesn't matter much which method is used.

 

[Daniel Lowe]

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

I understand that binning and downsampling both will reduce the amount of noise (and detail) in the image. It really doesn't matter much which method is used.

The EXR sensor doesn't "bin" pixels.

I stand by my comment.

 

[silps]

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

I understand that binning and downsampling both will reduce the amount of noise (and detail) in the image. It really doesn't matter much which method is used.

Pixel Binning does not refer to throwing away pixels. It doubles up pixels to capture light at high iso with less noise. Read this for a full explanation http://www.dpreview.com/previews/fujifilmx10/page3.asp

 

[Steen Bay]

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

I understand that binning and downsampling both will reduce the amount of noise (and detail) in the image. It really doesn't matter much which method is used.

Pixel Binning does not refer to throwing away pixels. It doubles up pixels to capture light at high iso with less noise. Read this for a full explanation http://www.dpreview.com/previews/fujifilmx10/page3.asp

Binning/combining two pixels in hardware, and downsampling to half the pixel count, both methods will result in a 41% increase of the SNR (if looking at the shot/photon noise alone).

 

[Sir Monty Golfear]

This is the email I sent to AP on Tuesday last.......I have not had a reply yet !!......perhaps there are some red faces at AP.

Hi ,

I read with interest your first look preview of the Fujifilm Finepix X10 camera ( 12th November issue ).

In the build up to the release of this camera I have been looking at the Specifications on the Fujifilm website. One of the big things it is promoting is the size and performance of the EXR sensor . It states that it is 2/3 inch sensor. In the review Matt Gallagher says that it is indeed a 1/2.3 inch sensor and that it is the same size as the sensor employed in the Pentax Q camera. My understanding is that the 1/2.3 inch sensor is considerably smaller than that of the 2/3 inch sensor. I have also been told by an employee of my local branch of Jessops that the sensor size is the same as the one used in the Pentax Q . Which one of us is correct ?.

Again, surely this is a good case to give actual physical dimensions of the sensor or perhaps publications could try and work with manufacturers to implement a size code to make it easier for everyone concerned .

In the meantime I will ponder many other of life's mysteries , and carry on reading your great mag.

Cheers,

 

[min32]

Not sure can the any full size resolution photo do the pixel binning but think it is possible. Is it easy to combine the information of the two side by side pixels after photo had take? How the information of pixel 1 and pixel 2 combined if the format not arrange probably before the photo taken. Is it more easy to manage pixel binning with same format before taken the photo? If noise more depend on the single size of individual sensor rather than average total sensor area.Than the sensitive to noise of every single pixel sensor should be same whether using the full resolution or half resolution. By combine two pixels inform, the chance of getting data will be increase also. But, how effective? will the optimal condition also depend on the brightness of light level? Just curious how the EXR SN work and how effective?Any one familiar with it?

 

[Midwest]

A 12mp S100 image can easily be downsampled to 6mp, which will have pretty much the same effect on noise and detail as a lower native MP count.

Utter nonsense, you still clearly don't understand how the EXR 'SN' mode works.

I am not 'up' on the EXR 'SN' mode but I do know that you can't change the characteristics of a sensor by 'downsampling' the images it captures, or else all the 12 and 16mp cameras with 'noise' problems could be resolved simply by shooting at a lower pixel count - and that doesn't work.

We're on the same side for once, how about that!

 

[Sir Monty Golfear]

I think Fuji state somewhere that they interpolate the pixels. See their website

 

[Steen Bay]

I am not 'up' on the EXR 'SN' mode but I do know that you can't change the characteristics of a sensor by 'downsampling' the images it captures, or else all the 12 and 16mp cameras with 'noise' problems could be resolved simply by shooting at a lower pixel count - and that doesn't work.

DxOMark seem to think it works. All their 'print' figures are for images downsampled (or 'normalized') to 8mp, and you can see how the downsampling has increased the SNR and DR, compared to the values at the cameras native resolution (the 'screen' figures).

 

[Daniel Lowe]

Down sampling is not the same as the EXR sensor 'sn' mode.

The EXR sensor is using the full area of the sensor in the first instance, ie., to replicate the results you would have to down sample a standard sensor twice the size of the EXR sensor for your position to be correct.

With the EXR sensor, the physical size of the photo sites and the pixel density at the time the shot is taken is already at those values, there is no way down sampling a standard sensor of the same physical area is going to match the results of the EXR sensor in 'sn' mode (or 'M' size for manual shooting).

I am not 'up' on the EXR 'SN' mode but I do know that you can't change the characteristics of a sensor by 'downsampling' the images it captures, or else all the 12 and 16mp cameras with 'noise' problems could be resolved simply by shooting at a lower pixel count - and that doesn't work.

DxOMark seem to think it works. All their 'print' figures are for images downsampled (or 'normalized') to 8mp, and you can see how the downsampling has increased the SNR and DR, compared to the values at the cameras native resolution (the 'screen' figures).