Rivergate Software Inc. - makers of DVDAfterEdit

This thread will contain a list of common quirks in player's implementation of the spec which can cause problems. The idea is to help out users migrating from abstraction-layer applications especially. This first post will be updated periodically so that it isn't necessary to read the whole thread.

1. When setting subtitle streams, audio streams etc. it is advisable to do so at the last possible moment - ie. in the pre-commands of the PGC in question. Some players manage to "forget" these settings, especially when moving between the VMG and VTSM, for example.
2. It is a good idea to reset all registers ( GPRMs ) in the first-play PGC. Some software players do not reset these variables when inserting a new disc.
3. It is safer to Link to a PG than a cell - for example, "Link C2" is not always honoured correctly by some players.
4. The behavior of "Set Button r4" is inconsistent if r4 ( or whatever GPRM you use ) is equal to zero. Some players will perform a "Set button B1", others will simply not change the button.
5. Care should be taken with cell-commands - if a user winds or chapters through a cell-coundary, the originally intended command may not be executed when the cell boundary is reached.
6. SPRM 14 cannot be used reliably in all players. Some, notably Sony, return incorrect values.
7. It has been alleged that SPRM 4 ( TT ) is also incorrectly implemented in many players, sometimes giving the VTS TT instead, ie. SPRM 4 and 5 will return the same value. It would be interested to hear confirmation of this this...
8. Force Activated Buttons - A button set to forcibly activate in an HLI (FOAC_BTTN in the spec or "Force action button number" in the button highlight info in dvdAE ) won't be activated by many (if not most) players.
9. Randomness - Some players random numbers generator allegedly have a better randomness when using the maximum (65535) value, then doing a modulo to get the correct range.
10. Button commands should never be NOPs. That will stop some players when they are activated.
11. PGCs containing non-zero PGC Still Time values cannot be one-sequential. Using Cell still times will be fine.
12. If a Title PGC is one-sequential, it should only have a value of "0" in the Next and Previous PGC settings to be spec compliant (Menu domain PGCs can only have a value of "0" in the next and previous settings, as well) It's not clear whether ignoring this rule would have serious consequences, though.
13. If a Title PGC is not One Sequential, timecode values and chapter info ( PTT number, ie. SPRM 7 ) cannot be relied on. Many players will still store the PTT in SPRM 7 anyway, but up to 50% may return the last PG instead, or an unpredictable result.

Refinements, corrections or insights to the above are welcome !

Ian