Originally Posted by jrref
I found that there was virtually no difference between Auto, YCbCr 444 & 420. I say virtually because the OLED drifts slightly so it's almost impossible to get an exact match on every scan. What I did see was RGB Full was way off, which I knew it would be, and RGB Limited had minor but noticeable errors when compared to the YCbCr/Auto scans.
The result is on the Sony, it expects YCbCr input which is what we knew and sending it RGB Limited introduced some errors. I know Ted's disk patterns are bit perfect YCbCr and in trying a couple of patches, they matched the 6G at 1080, YCbCr, 8 bit. I wasn't able to verify the triplets as you would in a scientific experiment but the gray scale CM scan looked extremely close.
This exactly I'm talking for years, that users found easy (bombed my marketing that its the revolutionary solution) to use external 'low cost' patch generation solutions (which all are working @ RGB) and try to reduce as much as possible their errors, using RGB generated patterns (most of them, especially CalMAN users which using MobileForge which is proven here that is inaccurate to all tested devices), while their charts will look pretty with RGB signal to their HDMI input, when later they will playback movie content, errors due to different processing will be introduced.
Someone can say...why to not set your player output to RGB Video, so to match the same colorspace as your 'low cost' patch genaration solution is generating? ...In that scenario you will have some small or larger errors the player will add to convert YCbCr to RGB-Video, there player which can do that job more accurate than others.
I have posted at past some digital errors testing comparing YCbCr with RGB-Video output (see how much errors the Pioneer reference '$2500' was adding @ RGB-Video: Oppo UDP-203
or Pioneer LX-91
or Sony BDP-S5200
These are the scenarios and mistakes a lot of people are doing:
1) If you use 'inaccurate' patch generation solution with RGB output (MobileForge as data proven here) and at the end of your calibration...when you will playback movies (movies are using YCbCr colorspace for delivery) with your player with YCbCr output (because most of the known brand stand-alone players I have tested have bit-perfect YCbCr output), then during your movie playback you will have added errors to your signal chain coming from incorrect patch generation and from difference of handling RGB vs. YCbCr input of the display.
2) If you use 'accurate' patch generation solution with RGB output (LightSpace Connect as data proven here) and at the end of your calibration...when you will playback movies (movies are using YCbCr colorspace for delivery) with your player with YCbCr output (because most of the known brand stand-alone players I have tested have bit-perfect YCbCr output), then during your movie playback you will have added errors only from the difference of handling RGB vs. YCbCr input of the display.
What is the most accurate solution:
1) Use an external professional pattern generator (which can cost more money that your actual TV price...its not a solution for all pockets) which its being designed to deliver accurate YCbCr patch generation (but you can't trust anything expensive, see example
External generators you can trust (which has been verified from calibrators, not from company employers or related with marketing guys) for SDR are DVDO AVLab TPG and Accupel DVG-5000 (both have been discontinued, so you can find them at lower prices.) Both can output accurate YCbCr REC.709 8bit or real 10bit (not fake 10bit).
In that case since your player 99.9% has accurate YCbCr output, and you have an accurate YCbCr generator, you can skip doing manual cal using calibration disk from your source (to calibrate your whole video chain) but use the external generator for that job.
What I'm doing before performing an SDR calibration using internal display controls for adjustments... is to check first using DVDO AVLab TPG and by using my calibration disk as reference the digital level analysis output of the player....to see if its accurate...if its accurate...then I'm using external generator for patch generation.
If inaccurate, for example XBOX One S is always outputting RGB-Video when you playback a movie from the drive or USB input (test results here
), so in that case, I use always my calibration disk there for patch generation and not the external generator, because I want the whole video chain to be calibrated.
2) Use a calibration disk which you will playback from your actual playback stand-alone player (which is free or has small fee to get it), use a calibration disk which is correct for your software you will use, because not all calibration disks are compatible with all software, each calibration software require different patches, to Saturation Sweeps / ColorChecker etc.
While this kind of calibration can take longer time, because of manual changing the 10 or 20 grayscale patterns or 6 patterns later for CMS calibration, it will be the cheaper and most accurate solution, since you calibrate your whole video chain.