SCI/TODO

From ScummVM :: Wiki
< SCI
Revision as of 16:25, 13 June 2010 by M kiewitz (talk | contribs) (pq3 driver face seems to have been fixed)
Jump to navigation Jump to search

Missing features

SCI0-SCI1.1 missing features

Graphics

  • SCI1.1 Palette morphing (kPalVary and kAssertPalette)
    • kAssertPalette is used to make sure that the palette stays fixed to a custom palette number and is not affected by sprite palette. An example where this is utilized is QFG1VGA (for the day/night effects)
    • kPalVary is used for palette variation effects. An example is the first room of SQ5, when the simulator pod is lowered and Roger exits it. There is a palette variation (kPalVary) effect in that scene, where the initial red room palette slowly turns into the final blue palette. Currently, as a workaround, the final palette is set immediately
  • Scaling inside kAnimate - scaleSignal bit 1 triggers setting of scaleX/scaleY (animate.cpp)

Sound

  • Instrument changes in SCI0 must be sound resource specific - in lsl3 door slams will change an instrument in mt32 emulation and when the actual music is resumed, it will be using that instrument instead of the one that was meant to get used (see bug #3007449)
  • Channel remapping. Songs that are played simultaneously might be using the same MIDI channels. E.g. in LSL1VGA, in the copy protection screen, there's the Larry music and a "pling" sound when a question is answered. Currently, the pling sound affects the instruments of the Larry song. Let's say that the Larry song is playing and it's using channel 1 to 5. Then the 'pling' effect is played, it uses channel 4. The instrument change in that song will change channel 4, and this will also affect the Larry song. In order to avoid this, the pling could be remapped to channel 6. We need some sort of lookup map, so that we redirect notes directed to channel X to go to channel Y, and we would need one such map per song. Ideally, we should find out what algorithm Sierra uses for this. We could try out own channel remapping implementation first, and then if we need it we could See what sierra does sometime in the future
  • Voice mapping. This is used by all SCI games to assign hardware voices to midi channels in the Adlib and FB-01 drivers. So, for example, if a song tries to play more than 9 notes at the same time, we need to stop an earlier note. We currently do this round-robin, while we should be respecting voice assignments. If a game sets 3 voices for piano, we should not take one away unless 4 piano notes are played, then it's OK to stop one piano note. Currently if a song were to use 5 notes at once we would just play them all, even if that midi channel was set to use 2 voices, and those 3 extra notes might stop more important notes, e.g. in the melody. This has been implemented for SCI0, but it's not hooked up to Adlib at the moment. Also we're not sure if the initial voice assignments are in the MIDI stream in SCI1 (they're not in SCI0). Of course, channel remapping has an effect on this too
  • Amiga (SCI01/1) and Mac (SCI1/1.1) sound support (same format)
  • Proper song volume support, currently volume is reset to max whenever a new song is loaded
  • kDoAudio subops 12 and 13 (used in Freddy Pharkas: subop 12 is called while people are talking and subop 13 is called while the narrator's voice is heard)
  • Implement a CMS driver

Parser

  • Support for non-English text input

Mac-specific

  • Support for the SCI1 Mac version of kGetEvent
  • Support for the SCI1 Mac version of kNewWindow
  • Support for modified SCI1.1+ scripts (all but KQ6)
    • Looks like some parts are LE and some parts BE (as opposed to KQ6 having all BE)
  • Fix QuickTime videos - assigned to clone2727
    • The videos now play in KQ6 Mac, but require multiple edit list support (see here)
  • Support for SCI1.1 partially hardcoded icon bar
    • Images stored in PICT resources (tags: 'IBIN' (not-selected), 'IBIS' (selected), 'PICT' (inventory image))
    • kPlatform subop 4 is called various times for the icon bar. One call sets up the scripts for the icons, etc.


SCI2-SCI2.1 missing features

This list is VERY incomplete, SCI2-SCI2.1 support is still in very early stages

Graphics

  • Proper support for text views
  • Support for Robot (RBT) videos

(TODO)

Sound

  • Sound fading (there are differences from SCI1.1, e.g. music in the the Sierra logo screen in GK1CD)

(TODO)


SCI3 missing features

Support for SCI3 is non-existent

Graphics

  • Support for Duck TrueMotion 1 videos (only used in Phantasmagoria 2)

(TODO)

Sound

(TODO)

VM

  • Support for the new SCI3 scripts


Known bugs

SCI0-SCI1.1 known bugs

Graphics

  • KQ5 Mac: The game uses a screen size of 320x190 and the offTop needs to be set to 0. However, setting the offTop to 0, messed up the priority screen. Without changing offTop, the menu bar will flicker when hovering over it.
  • KQ6 CD: Flickering right in the first room when discovering the chest - caused by 2 lists that are being used on kAnimate. One rect overlaps with the ring animation, that's why it flickers.
  • KQ6/CD: several sync resources aren't available in the resource files. The windows version does only use the additional sort of "raw-data" appended to those sync resources. We are currently using the sync resources inside kPortrait instead of the raw data, that's why lip sync isn't working for those cases. We need to find out how this raw data is interpreted, probably we could even create virtual sync data out of it for the DOS scripts to use
  • LB2: The credits windows (about screen from the options menu) are incorrectly drawn above the character sprites
  • LSL1 (new graphics code regression): doors at the casino wont open. A workaround is in place till the actual issue is found. Issue seems to be caused by signal of the door set too late or kAnimate being call too few times. If you move the .doit calling code to the end of kAnimate door will open. It may be caused by VM actually. Old graphics code worked fine, because kAnimate was differently implemented
  • PQ2: about window is drawn onto status bar (sierra sci seems to move window accordingly)
  • PQ3: Highway (day 1) - speech window incorrectly placed
  • PQ3: Hospital (day 1), Computer (day 1), Court (day 4), Oak Tree Mall (day 4) and others - speech window isnt fully removed
  • SQ1: Ulence Flats - when entering the ship, another ship appears. We see this ship shortly completely and then it starts appearing - kAssertPalette exists in SQ1VGA, possibly related
  • SQ1: Some effects missing (example: when destroying lasers on kerona) - kAssertPalette exists in SQ1VGA, possibly related
  • XMAS88: View position off by one (plus +1 to vertical coordinate) in fireplace scene (maybe also goes for other views?) - seems to be also the case in kq4 old (for all views) - needs to get investigated further

Sound

  • The Tandy/PCjr driver needs more work
  • LSL1: when getting killed, the mixer sound isn't fully played, but only the first part
  • LSL1: when getting killed, the sample (?) when larry is getting up with the platform isn't played

VM

  • Castle of Dr. Brain: Loading a saved game made inside the puzzle room (room 360) shows an empty grey screen. That particular room does not have a background
  • Iceman: displays the warning "NodeValue() on a NULL pointer attempted" at numerous points during the map table scene - isSaneNodePointer: Node at 0000:0000 wasn't found! is displayed when starting to plot
  • Iceman: acknowledging "green board" isn't possible (game breaking) - issue is script 383 calling kSaid with spec "[!*] / board < green" - needs to get checked - workaround: use "confirm green board"
  • Iceman: acknowledging depth isn't possible - issue is script 384 (see green board)
  • Iceman: repairing the diver should work with "put washer on shaft" - doesn't work here - workaround: enter "install washer"
  • Iceman: "climb down" on the main ladder in the submarine doesn't work (room 41), 2nd kSaid is returning true and shouldn't - workaround: enter "decend"
  • Island of Dr. Brain: right at the first puzzle clicking on the bricks on the bottom of the screen makes the engine freeze (endless loop)
  • Jones/CD: walking animations and portraits are turned off by the scripts on some computers caused by speed throttler only occurs on Mac OS X, does not occur on Windows 32-bit, Wii
    • Occurs when the system slows down when receiving too many update screen calls. See here.
  • Longbow: After leaving the cave and the conversation that follows, a list is not sane anymore (perhaps it's related to pathfinding? There's a relevant warning just before this occurs)
  • LSL1: bartender animation when ordering a whiskey and the rat animation in the room with the drunken man are much too fast
  • LSL2: Restoring a game in the first screen moves the cels up by some pixels (e.g. the road sign). Cel positions are fixed when exiting and reentering the scene
  • LSL5: one screen nearly at the end with text disappears far too quick
  • PQ2: when restoring a game in room 25 (753 third street) our car moves up some pixels, so trunk animations dont match
  • PQ2: room 26 isn't exitable - happens if you warp from inside the car at the start into room 26 - happened to me during the game although i'm not unable to reproduce it anymore. I guess I did something differently last time.
  • PQ2: when getting game over later in the game restoring results in kAnimate called with argv somehow invalid (valgrind/app verifier)
  • PQ3: VM goes mad when going in computer to tools / city map / plot crimes (warnings, heap issue (only valgrind/app verifier))
  • QFG1 EGA: displays the warning "List node at x is not sane anymore" when the dinosaur snaps at you during the intro
  • QFG1 VGA: when winning any random monster fight, the game uses the unimplemented kMergePoly function. No apparent side-effects are visible

SCI2-SCI2.1 known bugs

This list is VERY incomplete, SCI2-SCI2.1 support is still in very early stages

(TODO)


SCI3 known bugs

Support for SCI3 is non-existent

(TODO)