4 ms·
I have never understood how we can point to video that has gone through an unknown converter into a television screen recorded by an unknown camera and say defi
by mmcgaha 6y ago
I have never understood how we can point to video that has gone through an unknown converter into a television screen recorded by an unknown camera and say definitively what the expected result will be. I feel like the timing on multiple devices is guaranteed to give us some strange artifacts.
- toast0 6y agoI don't think there's an expectation for an exact recorded output, but in this case, it's known how the arcade system draws the levels in ram, how the video output chip works from the ram, and if things show up on screen inconsistent with that, but consistent with how an emulator outputs video, that's pretty fishy for contest that's supposed to run on real hardware. It's also an indication that the emulator isn't totally accurate at all times, even if it might be during gameplay.
- bhickey 6y agoMitchell also described his recording set up as using a 2-bit brand converter. This device is not double buffered, making double buffering artifacts implausible. On top of this, DK hardware generates a badly timed sync pulse. The 2-bit converter does not correct for this and typically generates a black and white image when connected to a DK board.
- toast0 6y agoEven if the converter was doing some buffering or interpolation (which ugh, who wants?), it couldn't reorder things to show out of order. At most, it would make things more simultaneous, or fade in.
- gentleman11 6y agoFor one thing, the pixels in the original are huge by today’s standards. With or without noise, you could roughly analyze a good percent of them fairly accurately