Camera Megapixel Rating versus Actual CMOS Sensor size
August 12, 2017
This year as in every year the trend is an ever escalating MP war by manufacturers to have a higher and higher megapixel rating on their cameras. I fear that buyers are confused into thinking that a higher megapixel means the imaging is better or that they use it as part of their selection process.
Do not confuse Quantity with Quality
For me, I totally ignore the MP rating because it does not mean a thing. Internally most cameras on the market use only a handfull of different CMOS sensors in total. The MP ratings on these sensors are 2, 3, and 5. Now if you are a manufacturer and you actually use a 30MP sensor please do not get upset.
My thoughts are this: “you can’t get blood out of a turnip”.
If a CMOS sensor is a two dimensional array of pixels, 2 million in total, where each pixel contains an RGB value among other things, how do you save a JPG onto the SD card with a 30MP rating?
If you stored the entire CMOS array of 2 million pixels as a file to the SD card, would it not be a 2MP image file?
Well, the term “interpolation” comes to mind. Interpolation means that software creates pixels in between existing pixels as we stretch or blow up an image from 2MP to 30MP. It uses math, which we won’t get into but the process is what we need to understand.
Now if I was to study the original photo in 2MP and zoom in, then study the interpolated 30MP image and zoom in, will I actually see more details in the 30MP?
What might confuse the issue is that the camera does not save a native 2MP image uncompressed because it is huge. It compresses to a JPG file first then saves this to the SD card. Now when you compress an image into a jpg, there are factors used in the compression to get a smaller or larger file, with a corresponding poorer or better resulting image quality.
Could it be that if the camera simply saved the 2MP image as a 2MP jpg file but increased the compression quality factor, the resulting image would be all we need? I believe so. I just do not feel that interpolating gives you anything more. It does create larger and larger JPG files and the illusion you can zoom way in but do you really see anything more there? I say no.
Do I want a 30MP camera with a minimum setting of 14MP? No. The files are too large and require more CPU and disk space to handle.
So, we could compare two cameras that both use the very same 2MP CMOS image sensor, one with a 30MP rating on the retail box and one with a 2MP rating on the box. Will consumers buy the 30MP camera? Probably ….
Should they? Well the answer is complicated. Much of the image quality is a factor of firmware engineers’ image processing code and the jpg image compression settings. Two identical cameras can produce very different images because of a different firmware (software).
So, this year we have 30MP, what next year? 40Mp, then… 50, 60, 80, 100MP ?????
When will it stop?
This is just stupid.
Please do not buy a camera based on its MP rating. This is just nonsense….
ok, back to camera testing now…. 🙂