>> Friday, July 27, 2007
The multi-layer background of Lucky Luke was also something I never thought it would be possible on CD-i. How did SPC Vision do this rain and snow effect on CD-i? Which parts were not possible without the DVC extra memory? Lucky Luke doesn't show any full motion video, but it doesn't play with the emulator. "Lucky Luke was done in 1994/1995/1996. It uses the DV cartridge for its sound effects, which is the reason why it won't play without. The multilayer thing is relatively simple: use both CD-i planes to contain graphics, and overlay them with sprites. This requires background restoration on each frame but that was made possible by the extra memory in the DV cartridge. SPC used background erase in most other titles, which is faster. Also, I think that Lucky Luke as a 25 Hz frame rate, not 50 Hz, but I might be wrong here."
Apparently Philips used to have different plans with Lucky Luke. According to the crew it was planned to look more like Zelda than your traditional horizontally scrolling game. However, these first developments never got any far as both changed plans radically. Work on Lucky Luke started in 1994.
"It was a painful release. I will go into some detail below because it was quite typical (Dimo's Quest, Steel Machine and The Apprentice had similarly painful releases but for different reasons). I did the release handling, but PixelHazard did all the programming (Arjen and Luke had already left SPC at this point [to form PixelHazard]).
I have bugfix lists for v1.01 and v1.02 dated January 23, 1996 and March 10, 1996, respectively. These were "normal" bugfix releases, as was typical for game titles: a screen flash here, a crash there, an unplayability elsewhere, and pretty soon you have an unreleasable disc Version 1.03 had most of these fixed.
However, for version 1.03 weird, unrepeatable freezes and crashes were reported on several players, most notably Mono-II, at random points in the game. After some deliberation Philips sent us a set of driver updates to include in the disc, and this became version 1.04. This did not fix all the problems, so we switched from PCM (CDDA) music to ADPCM music for version 1.05 (at some loss in music quality), but left the driver updates in there. This turned out to be a mistake (the drives fixed some bugs but apparently introduced others), so we took them out again as requested by Philips, making no other changes; the result was version 1.06.
Version 1.06 appeared to fix the random freezes, but now we got random dirty disc messages on some players, especially the Roboco (450) with ROM version 1.1. These were very hard to reproduce; at SPC we didn't succeed at all and PIMC reported that the problem went away when they attached the NIRD (Non-Intrusive Realtime Debugger) to the player. Later, however, they could provide us with a few crash traces. We worked around the problem (some calls returned seek error codes; we just retried them in this case) and this became version 1.07 which was released to manufacturing (i.e., pressed).
The saga doesn't end there, however. Version 1.07 still has occasional crashes on some players; I won't bother you with the details. I ultimately diagnosed the problem, but a version 1.08 was never produced.
Note that after version 1.04 no changes were made to the game itself; only initialisation and music/sound effect playing. This is also typical. As an extreme case, midway during testing The Apprentice was restructured to run as two separate processes to fix a problem with the interrupt scheduling of the Mono-II players; this restructuring left the game itself largely alone.
The problems were usually related to the "extreme performance" approach that SPC took to games; the game programmers sometimes took shortcuts that appeared to be legal from the documentation but then turned out not to work on some player models due to hardware differences. After the Dimo/Steel/Apprentice releases we instigated procedures and safeguards to avoid such problems, but they were apparently not foolproof.
Anyway, we learned how to effectively debug CD-i titles on a consumer player; We were, I think, unique in this respect; on at least two occasions SPC programmers were baffled when Philips engineers had problems downloading into a consumer player, which we did quite regularly at SPC.
I can also remember the discussion about Lucky Luke being allowed to smoke or not, but I think this was quite late in the development..."
Thanks to an ex-member of SPC Vision