A Question about megapixels and image size.

I'm not sure if this has been answered, taking good composition into play is there anything gained by using a 12.8 MP CMOS vs. a 6.5 MP CMOS and using software to convert it to a 12.8 MP equivalent? Software I have I can increase image size by 300% without readily noticeable errors.

1. What are limitations of software?

2. What are advantages of larger sensors excluding camera features.

3. If you want a 16 x 20 or larger print from a 6.5 MP or lower camera is software going to resolve a non-pixilated print? If so what is "sweet spot" for the number of MP a CMOS should have?

--Chris

...not as excited about getting 5D invest for full frame and less noise?

I've had 16x20 prints from the DRebel made by Pictopia, there's no noticible pixilation. I followed their recommendation of NOT upsizing and let them do the interpolation.

I'll probably upgrade to the 5D when the housings come out. The main reason is not more pixels but better signal-to-noise and dynamic range, that and the 15mm FE I have will be a 180 degree lens. The advantage of a large sensor is that it captures more light for a given exposure and improves the signal-to-noise over the smaller sensor by about the area ratio if the number of pixels are the same.

I cant say that I completly under stand your question. 12mp will allways have 2x the data as 6mp.

Yes Software may be able to smooth out the lines, and avoid pixelation, but that does not increase the ammount of detail in the shot. So the interpolated image will still only have half as much detail as an image that started off at 12mp.

An example of detail. Take a picture where a fish shows up as 100 pixels. You will be able to see the shape of the body and where the fins are, but the details of the fish such as scales will not be visable. You can interpolate that fish to 200 or 400 pixels but no ammout of software processing can make scales show up on that fish. But if your camera had higher resolution, the fish would take up more pixels from the start, and the scales might be visable.

Clearly for small prints, the extra detail will not show up so there is no point to haveing more mp. Prints are usually done at 300DPI so up to a 10"x6" there will be no difference between a 6 and 12mp camera. But any print bigger than that, the 12mp image will contain more detail, at least if you look close enough.

The point where a person picks up on the extra detail depends on the person, and the viewing distance. If you never print bigger than an 8x10 or you dont look at your 16x20 from 6 inches away, you may never see the difference.

But eventually as increas print size (or decrease viewing distance) you will notice the lack of detail and at what ever size that is, you would be able print a 12mp one twice as big.(in terms of area) before you noticed it.

Is the extra deatil worth spending $3000? That one is up to you.

is there anything gained by using a 12.8 MP CMOS vs. a 6.5 MP CMOS and using software to convert it to a 12.8 MP equivalent?

 

Yes.

 

1. What are limitations of software?

 

Software creates pixels where there were none. It's not real detail. How and when this difference shows up varies greatly.

 

While I would love to have 300 pixels for every inch of print, much lower densities may be fine depending on subject and viewing distances. I was always amused when testers whipped out there loupes to examine prints of different resolutions. If you need a loupe to see a difference, does it really matter?

 

If you printed a 6mp, 12mp and a 6mp upsized to 12mp at 8x10, you wouldn't see a difference among them.

 

Interpolation programs have gotten better. If you start with 6mp, you can go quite large without much problem.

 

2. What are advantages of larger sensors excluding camera features.

 

a) Pixel density/size, which may equate to increased dynamic range. B) Better angle of view utilitly with certain lenses. Example: a 28-70mm FF lens may be a great range for shooting, but when it becomes the equivalent of say, 42-105mm, it's not as handy in the field.

 

3. If you want a 16 x 20 or larger print from a 6.5 MP or lower camera is software going to resolve a non-pixilated print? If so what is "sweet spot" for the number of MP a CMOS should have?

 

I'm not clear on the question, but I will take a stab. A 6.5mp camera can easily do a 16X20 print without visual pixelation. As mentioned, most high-end printers use sophisticated RIP programs that will prevent any printed pixelation, but a higher pixel count camera may provide finer detail.

 

There is no sweet spot. Using the term CMOS is also not applicable. It doesn't matter if it CMOS or CCD. For Light Jet prints, having print files that are 304.7dpi for the print size I am doing would be ideal, but lesser resolutions work well all the way down to 150dpi.

Great!

I wasn't very clear but you guys hit it on the head. With my 6.3 MP (where did I get 6.5 ) Rebel I am able to get images and prints that have satisfactory detail and resolution. I cannot see a difference between 35mm film and digital prints in size 4 x 6.

I need to make a 36 x 72 vertical print for my hallway. Wall/Open sea.

Plan: I'm going to use Magic Filter, wide angle, RAW, use software to edit and increase, obtain print. I think this will work as a temporary solution.