compa
Well-known member
Recurvirostra said:If 5% of your theoretical JPEG is blown-out you will have far less blown-out pixels in the RAW because it has many more brightness levels below pure white in which to store the detail, before you reach the next level down in the JPEG file (one gradation below pure white).
In real life, even if the photo is not perfectly exposed, whatever area is blown-out in the JPEG will be much smaller in the RAW because the RAW file has a whole bunch of levels below pure white, but still brighter than the JPEGs second brightness level, in which to store detail.
Ahhhh ... now I understand where you are making your error. You are assuming that the 8-bit value of 256 represents 30 values (4066-4096) in a 12-bit file. This is not correct. This top value is used to represent blown out pixels and nothing else. 256 = 4096
255 may be used to represent 4065-4095 or something similar. In reality the exact numbers assigned does vary from one RAW converter to another (and from one model camera to another) but usable data (above 0 and below the maximum) is never discarded, just compressed.
Levels detail is lost in this average 30 to 1 compression without question. But pixels don't drop to 0 or become blown out because of it. The people who write the RAW converters are smarter than that!
I came into this knowledge from quite a number of sources over time. A couple of websites that give a little insight into this subject are:
http://www.dpreview.com/learn/?/key=ad_converter
http://www.luminous-landscape.com/tutorials/understanding-series/u-raw-files.shtml