Does a "PAL/NTSC" switchable Agnus have any idea about its input clock value? (I wouldn't think it does)
In PAL (28.37516 MHz input clock, 25 frames/sec) each frame would be 1135000 ticks long.
In NTSC (28.63636 MHz input clock, 29.97 frames/sec) each frame would be 955500 ticks long.
So if the Agnus has no idea of the input clock value it must rely on counting to provide sync.
However these counting values are correct only if the matching oscillators are used.
If the Agnus NTSC-counts with a PAL oscillator, it'll provide 28375160 / 955500 =~ 29.70 fps, resulting in a somewhat "slower" NTSC timing.
And if the Agnus PAL-counts with an NTSC oscillator, it'll provide 28636360 / 1135000 =~ 25.23 fps, resulting in a somewhat "faster" PAL timing.
Granted, they're not far off, but they're not exact timings either. Good approximations at best.
Has this been documented/confirmed/disproved anywhere?
In PAL (28.37516 MHz input clock, 25 frames/sec) each frame would be 1135000 ticks long.
In NTSC (28.63636 MHz input clock, 29.97 frames/sec) each frame would be 955500 ticks long.
So if the Agnus has no idea of the input clock value it must rely on counting to provide sync.
However these counting values are correct only if the matching oscillators are used.
If the Agnus NTSC-counts with a PAL oscillator, it'll provide 28375160 / 955500 =~ 29.70 fps, resulting in a somewhat "slower" NTSC timing.
And if the Agnus PAL-counts with an NTSC oscillator, it'll provide 28636360 / 1135000 =~ 25.23 fps, resulting in a somewhat "faster" PAL timing.
Granted, they're not far off, but they're not exact timings either. Good approximations at best.
Has this been documented/confirmed/disproved anywhere?
