Found a good PDF to prove my point.
http://www.compression.ru/download/articles/color_space/ch03.pdf
Remember why even today there is the YUjv limitation of 16-235 from RGB. Most people get confused and in practice there is a loss of alrowed codewords. Then more banding effects because the performance is more like 7.5bits.
The reason is when NTSC was created analog transmitters had modulated subcarriers excursions of 20%; then it was included noise and a little reserve to allow true black point calibration on bigger chains of equipment.
Going above 8bit has considerable hardware wiring cost so a better codeword selection has to optimize the signal removing some legacy codewords. Some quantization will always have to be made because human perception change in Weber law. For sound or luminance Barten has better equations.
0-3 :: 1020-1023 are reserved for control of CRT tube positions.
3-63 reserved
64 is black
512 is gray or achromatic
940 nominal luma peak
960 chroma peak
941-1019 reserved
Effectively a 9.5 bits system, maybe a 8bit full range with temporal dither would surpass the narrow range PQ2084.
Now the M4K blu-rays used the xvYCC. This 8bit system used the reserves 4-15 236-252 to extend the gamut. There are many samples that show how much better the colors become.
xvYCC is found in 99% 1080p TVs. Some papers show that has wider gamut than DCI-P3. So a 1080p display with xvYCC can surpass a 4k PQ2084 in terms of final image quality!
The problem with tools and standardization killed xvYCC in content production.
I am not saying to abandon everything just to pressure for a even better HDR today.
Sent from my Moto G using Tapatalk
http://www.compression.ru/download/articles/color_space/ch03.pdf
Remember why even today there is the YUjv limitation of 16-235 from RGB. Most people get confused and in practice there is a loss of alrowed codewords. Then more banding effects because the performance is more like 7.5bits.
The reason is when NTSC was created analog transmitters had modulated subcarriers excursions of 20%; then it was included noise and a little reserve to allow true black point calibration on bigger chains of equipment.
Going above 8bit has considerable hardware wiring cost so a better codeword selection has to optimize the signal removing some legacy codewords. Some quantization will always have to be made because human perception change in Weber law. For sound or luminance Barten has better equations.
0-3 :: 1020-1023 are reserved for control of CRT tube positions.
3-63 reserved
64 is black
512 is gray or achromatic
940 nominal luma peak
960 chroma peak
941-1019 reserved
Effectively a 9.5 bits system, maybe a 8bit full range with temporal dither would surpass the narrow range PQ2084.
Now the M4K blu-rays used the xvYCC. This 8bit system used the reserves 4-15 236-252 to extend the gamut. There are many samples that show how much better the colors become.
xvYCC is found in 99% 1080p TVs. Some papers show that has wider gamut than DCI-P3. So a 1080p display with xvYCC can surpass a 4k PQ2084 in terms of final image quality!
The problem with tools and standardization killed xvYCC in content production.
I am not saying to abandon everything just to pressure for a even better HDR today.
Sent from my Moto G using Tapatalk
.
