3 ms·
Color convergence on large CRT monitors is problematic. Back in the early 90s I spent $1200 out of my own pocket to buy a 20" CRT. At the time I was designing 3
by tasty_freeze 6y ago
Color convergence on large CRT monitors is problematic. Back in the early 90s I spent $1200 out of my own pocket to buy a 20" CRT. At the time I was designing 3D rasterization algorithms and circuits and would spend hundreds of hours tweaking constants (how much difference loss does a 5b interpolation fraction introduce vs 6b, etc) and A/B testing to get the most cost effective & high quality results.
I bought and returned two and just lived with the defects of the 3rd one when I realized it was impossible to get good convergence across the display. I would confine my A/B tests to a specific area of the screen where convergence was best.
BTW, I forget the exact details, but color convergence at the factory was set with the monitor facing some particular direction, lets say north. If you set up the monitor facing east, the convergence would be off a little bit.
When quality LCDs became available, convergence was a non-issue, but they used dithering to attempt to increase the color gradations, so it didn't really help my kind of work anyway.
- nitrogen 6y agoIn my youth I bought one of those types of high end monitor second hand. Had 5 BNC inputs and required a resolution tweaking utility to change the sync polarity. I spent hours tweaking convergence on the 9 million trim pots to get it just right. Best color gamut of any display I ever owned. Sadly the display physically broke something internal when I lost the included plastic screwdriver and tried using a metal one. The internal heads of the trim pots were apparently connected to the circuit. The display was fixed to 1024x768x75Hz exactly (though you could squeeze in a bit more with modeline tweaking), but the colors were great.