Atari Emulation of CRT Effects On LCDs
An anonymous reader writes "A group at Georgia Institute of Technology has developed a fun little open source program to emulate the CRT effects to make old Atari games look like they originally did when played on modern LCD's and digital displays. Things like color bleed, ghosting, noise, etc. are reproduced to give a more realistic appearance."
Well, in the case of the Atari games, it is fairly obvious that the programmers used the effects to create nicer-looking graphics without going to extra work. Sort of a reverse anti-aliasing effect.
Take a look at some of the comparison images in the article. The 'Enduro' image is particularly interesting: The skyline looks extremely fake on an LCD, but with the CRT emulation it looks almost realistic. The effect basically gives a continuous-color blend which would be impossible using just the colors available to the program.
So really, you can argue that this is how the games were meant to be seen like this, and this is actually how it should look.
'Sensible' is a curse word.
One of the most widely used Blargg's NTSC libraries. Many console emulators make use of them. This new one just looks to be more advanced than most of the preceding ones.
This page has screenshots showing the difference. Many games were designed with NTSC artifacts in mind.
The tech to get the response times so low does tend to jack colors and produce some odd artifacts, but none of those relate to how quickly it displays the data on the screen, not as far as I've ever heard anyway.
The 2ms 'response time' is just about the pixels response to the electrical signal. In other words, those 2ms means: this LCD can change a pixel from black to white in 2ms. It doesn't mean: this LCD will change the pixel 2ms after the computer or console tells the screen to change the pixel.
Since a couple of years, LCDs have a 'image enhancement' mode that adds some lag, from 40 to 105 ms. This is precisely to have a buffer that lets the chip preprocess some stuff and reduce ghosting or other things. I think that the 15ms or less to change a pixel is also possible only because of this processing.
You can't say that 105 ms is not noticeable, and this is probably what the GP is talking about. And DLP HDTVs seems to have up to 250ms of lag.
However, modern LCDs have also a 'gaming mode' with (virtually) no processing lag, but with the usual ghosting and other LCD classic issues.
We are Turing O-Machines. The Oracle is out there.
LCD response time, latency and motion quality has nothing to do with human reaction time. Humans can distinguish differences in time interval much shorter than their reaction time. Look at graphs of beat length variance of skilled drummers.