In 1993 SPC created a seamless MPEG branching demonstration CD-i: Special Operations Squadron. In 1993 SPC demo’d the game ‘Special Operations Squadron’ which was one of the very first titles using seamless MPEG branching. A similar engine was later used in The Lost Ride. When the first digital video cartridge was available, it was very hard (and very expensive) to code this function. It turned out more than 50% of the processor was ‘wasted’ by seamless playing of MPEG video’s. So little power would be available to actually create a game. In the evaluation back to Philips, SPC was opting a video-maze as one of the remaining possibilities, but it had to be rendered fully digital to make that happen (Special Operations Squadron was using ‘live footage’ from stock). As this was too expensive for the budget envisioned at the time, the project was killed. When Philips was shifting this idea over to Lost Boys in 1996 to create ‘The Lost Ride’, one can imagine the disappointment of the SPC crew!
| Image may be NSFW. Clik here to view.  |
| Video shot with overlayed plane and control panel graphics. |
The sound for legs 0, 15a, 33 and 40 is ADPCM audio; the music is MPEG audio. Although the quality of ADPCM is somewhat less than MPEG, it illustrates the mixing of MPEG and ADPCM audio. The difference is also apparent when slow motion is used. ADPCM audio is played from disc and hence disabled; the MPEG music continues to play.
The Fighting Falcon project was undertaken by SPC/Vision and Philips to explore the possibilities of interactive video games on the Compact Disc Interactive platform. All three keywords are important here:
We consulted Recursive Arts, Los Angeles for creative input and together came up with the idea of a mountain maze that you could fly through, possibly fighting enemies, dropping bombs, etc. Suitable stock video footage was obtained from Patriot Film and Stargate, both in Los Angeles, and we had it encoded for CD-i by PRL (Philips Research Labs), Redhill, UK and ICDI (Interactive Creative Digital Image), Charleroi, Belgium. Audio was selected and mixed by Studiope, Utrecht
After several months of struggling with documentation, buggy cartridge prototypes and a complete lack of development tools we came up with the disc that is demonstrated here: Special Operations Squadron-DEMO 1.1. This disc shows the mountain maze that you can fly through, making decisions at key points, which is all the interactivity that was achievable in this short time period. The reasons for this are technical and explained below. There is a target that you can blast away and you are then rewarded with some video explosions.
This proves that it is possible to construct an interactive video from pre-selected parts and combine it with some interactive game aspects. It remains to be seen wether it is possible to combine the video with a real interactive game, for technical reasons as well as estethic ones.
The Digital Video extension
To understand the problems we faced, it is necessary to understand the capabilities and limits of the CD-i Digital Video extension. It can play video and audio directly from the Compact Disc as well as from memory. However, the video and audio must be specially encoded according to the international standard ISO 11172 (also called MPEG after the name of the comittee that produced it) and after this encoding the possibilities for editing are severely limited. In essence, it is only possible to cut audio and video at points that were accomodated by the encoding process.
Essentially no manipulation of the video is possible during play, apart from pause, slow motion and single step. Objects can be drawn in front of the video background, but the video itself cannot be scaled, rotated or even faded in. The only possible manipulation is selection and position of the visible rectangle from a video image that may be larger.
Audio cannot be manipulated except for controlling the volume, but this is less of a restriction. It is possible to mix the MPEG audio with regular CD-i audio and this is often good enough.
Also, the Compact Disc can be read sequentially and in that case delivers a steady data stream of about 170 kilobytes per second, which is barely enough to contain full-screen video with audio. When the read position is changed (this is called a seek) this results in a large time during which no data arrives from the disc (up to three seconds). So to keep the video playing, data must be supplied from memory. This places severe demands on CD-i player memory, which is limited to 2 megabytes when the Digital Video extension is present.
And last but certainly not least, playing Digital Video requires a significant part of the CD-i player capacity, meaning that less is available for running the game part. Since our games already use up to 98% of the available player capacity, we felt this as a severe restriction.
The abovementioned restrictions result in severe limits for any creative design, and it remains to be seen wether real interactive video games are possible.
Development problems
The single biggest hurdle we had to take to produce this disc was the total lack of Digital Video development tools. We started with the specification of the Digital Video extension, a prototype cartridge, and the tools for producing a regular CD-i. Digital Video is a significant extension, however, and is not adequately handled by the standard tools. Several months were invested in creating the necessary tools.
Another problem, to be expected with any prototype, were the bugs with the Digital Video cartridge. These were especially annoying because one of the bugs was in the provisions for seamless jumps, which is the only standard way to paste together video that was not originally mastered this way on the disc. We ended up doing this ourselves as well.
The seamless branching was a problem. It was clearly needed for interactive titles such as the CDi games you have and is also used in the "multiangle" feature in DVD. Theb problem is that MPEG compression does not produce the same amount of data for each video frame. Some are easier to compress than others (eg a frame in an active football match uses a lot of bits, a black cat at night not so many). However a CD (at least in those days) has a strict limit on how fast the data can be read from it. So to match the fixed rate of delivery to the decoder in the CDi player to the varying number of bits the decoder "eats" for each frame, MPEG uses memory to pre-read some data and smooth out the peaks and troughs. This is fine when you go through the frames in the correct order beacuse the encoder monitors the state of the buffer that the decoder will use to make sure that the decoder will never get in difficulties. But when you jump or branch things can go badly wrong (buffer over-run/under-run etc) because the encoder can no longer do this. So what we did was to control the video encoders carefully so we could mark out "splice" points in the streams where it would be allowed to jump to with a guarantee that the buffering would be properly managed. But that's about all i know.
- GAME SPC/Vision has proved in the past that they can program CD-i games that utilize the capabilities of the CD-i player to the limits.
- VIDEO The capabilities of the new Digital Video extension for CD-i had to be explored in this area.
- INTERACTIVE The interactive aspect had to be stressed; we did not want to use video as a background or reward only.
We consulted Recursive Arts, Los Angeles for creative input and together came up with the idea of a mountain maze that you could fly through, possibly fighting enemies, dropping bombs, etc. Suitable stock video footage was obtained from Patriot Film and Stargate, both in Los Angeles, and we had it encoded for CD-i by PRL (Philips Research Labs), Redhill, UK and ICDI (Interactive Creative Digital Image), Charleroi, Belgium. Audio was selected and mixed by Studiope, Utrecht
After several months of struggling with documentation, buggy cartridge prototypes and a complete lack of development tools we came up with the disc that is demonstrated here: Special Operations Squadron-DEMO 1.1. This disc shows the mountain maze that you can fly through, making decisions at key points, which is all the interactivity that was achievable in this short time period. The reasons for this are technical and explained below. There is a target that you can blast away and you are then rewarded with some video explosions.
This proves that it is possible to construct an interactive video from pre-selected parts and combine it with some interactive game aspects. It remains to be seen wether it is possible to combine the video with a real interactive game, for technical reasons as well as estethic ones.
The Digital Video extension
To understand the problems we faced, it is necessary to understand the capabilities and limits of the CD-i Digital Video extension. It can play video and audio directly from the Compact Disc as well as from memory. However, the video and audio must be specially encoded according to the international standard ISO 11172 (also called MPEG after the name of the comittee that produced it) and after this encoding the possibilities for editing are severely limited. In essence, it is only possible to cut audio and video at points that were accomodated by the encoding process.
Essentially no manipulation of the video is possible during play, apart from pause, slow motion and single step. Objects can be drawn in front of the video background, but the video itself cannot be scaled, rotated or even faded in. The only possible manipulation is selection and position of the visible rectangle from a video image that may be larger.
Audio cannot be manipulated except for controlling the volume, but this is less of a restriction. It is possible to mix the MPEG audio with regular CD-i audio and this is often good enough.
Also, the Compact Disc can be read sequentially and in that case delivers a steady data stream of about 170 kilobytes per second, which is barely enough to contain full-screen video with audio. When the read position is changed (this is called a seek) this results in a large time during which no data arrives from the disc (up to three seconds). So to keep the video playing, data must be supplied from memory. This places severe demands on CD-i player memory, which is limited to 2 megabytes when the Digital Video extension is present.
And last but certainly not least, playing Digital Video requires a significant part of the CD-i player capacity, meaning that less is available for running the game part. Since our games already use up to 98% of the available player capacity, we felt this as a severe restriction.
The abovementioned restrictions result in severe limits for any creative design, and it remains to be seen wether real interactive video games are possible.
Development problems
The single biggest hurdle we had to take to produce this disc was the total lack of Digital Video development tools. We started with the specification of the Digital Video extension, a prototype cartridge, and the tools for producing a regular CD-i. Digital Video is a significant extension, however, and is not adequately handled by the standard tools. Several months were invested in creating the necessary tools.
Another problem, to be expected with any prototype, were the bugs with the Digital Video cartridge. These were especially annoying because one of the bugs was in the provisions for seamless jumps, which is the only standard way to paste together video that was not originally mastered this way on the disc. We ended up doing this ourselves as well.
With some old SPC Vision crew members I saw the tech demo of Special Operation Squadron. Here the same technique had been applied. I hope you can shed a light on the material, about how it was developed, which demo's has been presented, and also how it got to be that the whole system was cancelled.
Well in Redhill we developed an encoder with both the best picture quality we could achive and a very high degree of controllability (see above). The price we paid was that it worked in slower than real time. Nevertheless it was widely used both for the games titles and for the movies (where good - well the best possible until DVD came along - quality was needed). That system was mainly used for these titles.
Well in Redhill we developed an encoder with both the best picture quality we could achive and a very high degree of controllability (see above). The price we paid was that it worked in slower than real time. Nevertheless it was widely used both for the games titles and for the movies (where good - well the best possible until DVD came along - quality was needed). That system was mainly used for these titles.
