tape timing improvements
This commit is contained in:
1 parent
527ad0a23a
commit
1031e44f77
4 files changed
+35
-23
No files matched your search
@@ -7,21 +7,23 @@
|
||||
// We have the following possible CPU frequencies:
|
||||
// - C64 PAL: CPU frequency 0.985248 MHz
|
||||
// - C64 NTSC: CPU frequency 1.022727 MHz
|
||||
// Average is 1.0039875 MHz and we will take this value for calculation (but it really does not change much).
|
||||
// The clock difference is about 4% - it is unlikely to cause problems.
|
||||
//
|
||||
// We have to distinguish the following pulse lengths (http://sidpreservation.6581.org/tape-format/):
|
||||
// - S (short) 352 us
|
||||
// - M (medium) 512 us
|
||||
// - L (long) 672 us
|
||||
// - S (short) tap value $30 (PAL: 352 us)
|
||||
// - M (medium) tap value $42 (PAL: 512 us)
|
||||
// - L (long) tap value $56 (PAL: 672 us)
|
||||
//
|
||||
// This gives us the following thresholds
|
||||
// - S vs M 432 us (108 or $6C 4-tick periods)
|
||||
// - M vs L 592 us (149 or $95 4-tick periods)
|
||||
// The TAP format (http://unusedino.de/ec64/technical/formats/tap.html) uses a resolution of 8 PAL ticks;
|
||||
// this implementation sets the timer for twice the precision (timer B is triggered every 4 CPU cycles),
|
||||
// probably an overkill, but in realityy it does not cost us anything.
|
||||
//
|
||||
// So -= we have the following thresholds:
|
||||
// - S vs M $72 (in our units of 4 ticks)
|
||||
// - M vs L $98 (in our units of 4 ticks)
|
||||
//
|
||||
// We need to recognize the following pulse sequences:
|
||||
//
|
||||
// 4-tick periods should be enough - TAP file format description (http://unusedino.de/ec64/technical/formats/tap.html)
|
||||
// tells it uses exactly 8 PAL ticks as a resolution
|
||||
|
||||
// The following encoding is used:
|
||||
// (S,M) = 0 bit
|
||||
// (M,S) = 1 bit
|
||||
// (L,M) = new-data marker
|
||||
|
||||
Reference in new issue
Block a user