Luminous Landscape Forum

Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: AaronA on September 02, 2005, 07:12:57 pm

Title: CMOS Sensor size vs pixel size
Post by: AaronA on September 02, 2005, 07:12:57 pm
a photosite is simple, the bigger the area, the less the noise(initially, DNR after the image is taken and actual context of the image will also have the same effect).

It makes me super happy to see the size on the 5D so large, I really beleve that they took the design of the 1Ds and 1Dmark 2 and peformed a hybrid on them.....something like that, since I see for the sensor a lot of common attributed of both sensors.
Title: CMOS Sensor size vs pixel size
Post by: sedunaway on September 03, 2005, 09:43:46 am
I followed the thread and this helped. Based on the thread you would say the 20D is a better image, not getting into the discussion of image size (1.6x).

I use the 20D for astrophotography and when you look at SBIG cameras (pro $$$$$) they always talk about pixel size, typically decreasing, "finer detail", yet when you throw in "noise" increasing with smaller pixels then you rob peter to pay paul, so to speak.

I assume the only way to really judge this is to take exact pictures in the environment you want to compare the two cameras and get into PS.
Title: CMOS Sensor size vs pixel size
Post by: BernardLanguillier on September 04, 2005, 10:11:04 am
To add a small comment to this, common sense does indeed claim that smaller photosites do result in more noise, but it is extremely difficult for us, end users, to test the validity of this theory.

The reason is that, even if the theory were true, it would only be true all other things being equal.

The fact is, there are no 2 sensors on the market using the exact same technology. The 5D might be a bit cleaner than the 20D, but we have no way to know if this is the result of the larger photosites, or whether it is just the result of the 5D being one year newer, and using therefore a better technology.

Common sense would lead us to think that a thicker steel part is less likely to rupture, but all mechanical engineers know that the very opposite can be true in some cases.

The technology involved in sensors is just too complex to be modelled by linear equations linking one output (noise) to a single input (photosite size). Testing of the system (hardware and software) is probably the only way to confirm if such a model holds, and this is something we cannot do because of the reason explained above.

Based on this, to my eyes, the "larger photosite means less noise" theory is not proven.

Title: CMOS Sensor size vs pixel size
Post by: AJSJones on September 04, 2005, 10:44:54 pm
It won't be too long before production models are out there and someone does the comparison you are in need of - MR or Phil Askey will probably compare it with other Canon sensors with respect to noise.  The 5D has the DigicII chip and bigger pixels than the 20D so, a priori, it should not be worse than the 20D.

Title: CMOS Sensor size vs pixel size
Post by: sedunaway on September 02, 2005, 07:06:42 pm
When I read dpreview of the Canon 5D versus Canon 20D there was a diagram that showed the CMOS sensor (graphics) and out came a surprise for me....pixel size. The 5D has a 8.2um pixel, and the 20D has a 6.2um pixel and the 1DsMkII has a 7.2um pixel. Now the question comes, independent of the frame size what does this pixel size have to do with the ultimate picture/print when we get to the printer for a 13x19 print? Or is there an issue when you crop and enlarge? Is there ever an issue that is PRO small pixel so that you would select the 20D just for this????

Size matter?
Title: CMOS Sensor size vs pixel size
Post by: AJSJones on September 02, 2005, 07:58:27 pm
This thread (;f=4;t=1152) is one you might check out.  In some situations, the closer spacing is advantageous and even though the sensor elements are smaller and have increased noise, the images do not suffer significantly except at extreme ISOs.  A summary would be, if you have the glass to compose the image you want, use as many highest-"quality" pixels as you can...
Title: CMOS Sensor size vs pixel size
Post by: AJSJones on September 03, 2005, 02:32:10 pm
I'm not sure who can lay claim to the original but "photography is a just a series of compromises".

Noise vs resolution, based on pixel size is a perfect example.  Decrease pixel size as long as the noise isn't noticeable under the conditions you want to shoot, to increase the ability to distinguish details.  For maximal resolution (e.g. astro as you suggest, presumably limited by wavelength range < pixel dimension) decrease noise as much as possible, by cooling the sensor.  Capturing as much of the sky as possible, for a survey for example, needs sheer numbers of pixels coupled with the "right" glass - the LSST has a 2.8 gigapixel sensor array but, I think, only one "lens"  

To your questions, I would say that the 20D produces a more detailed, but smaller image than e.g. 1Ds or 5D, based on its pixel spacing and superb noise characteristics.  Some of the issues you need to balance can be addressed without resorting to detailed side-by-side testing.  However, your plan is the right one because it addresses the whole process and you'll see how you weight the factors : weather sealing vs not, gloved operation in the cold  , total pixels vs spatial resolution; weight /size /balance vs features , $$,   etc etc.  Then to extremes : if you will always desire high ISO, what's your balance between noise and resolution at the temperature you'll be working at.  Even though it's a multifactorial issue,  it may end up as a no-brainer after you really order your priorities.  

Did you check out the 20Da, which was modified for astro work?
Title: CMOS Sensor size vs pixel size
Post by: sedunaway on September 04, 2005, 08:21:54 pm
Yes noise vs temp is well known in the CCD arena, more towards "hot pixel" but the CMOS doesnt have this exact anomoly. When I shoot in the 40 deg f to 20 deg f noise isnt an issue for 1600 ISO for 3-4min. Noise Nija SW plug-in has been a blessing in astro work, removing a fair amount of noise. The 20da has more alterations with sensor filters, which is the prime dislike for certain wave lengths of light, the IR filter is a problem. Yet when I show up with a 13x19 photo via Epson 2200 they all gasp.

I see more discussions over the pixel size (area) etc. versus noise on the 5D and makes me nervous for ASTRO work, then terrestrial work. I can shoot 5-6min at ISO 800 with minimal noise for faint objects. ISO 3200 is just not the ISO to use, even at 20 deg! I really appreciate the comments on this topic, as +$3000 makes me want to understand this better.

I would love to shoot a full 35mm frame, but to inject more noise in my (astro) environment over the 20D would be a big $3000 mistake.