Why the PC-FX sucks

[Kanon_Magic]

That’s an interesting analogy, but I think it oversimplifies both FMV and skeletal animation. You're right that sprite-based fighters like those on Neo Geo or SNES required animating full frames, while PS1/PS2-era games used skeletal rigs to animate limbs independently. But that’s not really comparable to FMV, which isn’t rendered in real time it’s pre-recorded video, not dynamic assets.
.....

Yes, I know, but it could be possible, in fact.

We'll never know if the FVM had been successful, and the PC FX 2 had been released, and, for example, they had developed an algorithm that, for example, created a 3D scene on a very expensive computer and calculated all the possible variants and put them separately in images, it would be possible, Quite possible, in fact, with quick access and everything, because you already know where each frame is and you have the most used ones stored in memory.

But that's speculation, XP. They would clearly be different from polygons. It's just a “what if.”

 

[RETRO-VETRO RETRO-VETRO]

Yes, I know, but it could be possible, in fact.

We'll never know if the FVM had been successful, and the PC FX 2 had been released, and, for example, they had developed an algorithm that, for example, created a 3D scene on a very expensive computer and calculated all the possible variants and put them separately in images, it would be possible, Quite possible, in fact, with quick access and everything, because you already know where each frame is and you have the most used ones stored in memory.

But that's speculation, XP. They would clearly be different from polygons. It's just a “what if.”

Totally fair to explore the “what if” scenario speculation is half the fun when talking about retro tech. And yeah, in theory, a system could pre-render a massive number of frames from a 3-D scene and store them for rapid access. With enough storage and clever indexing, you could simulate interactivity by jumping between frames based on input. It’s kind of like a supercharged version of Dragon’s Lair, but with more nuance.

But the challenge isn’t just technical it’s logistical. The number of permutations needed to allow fluid movement, camera control, and responsive gameplay would be astronomical. Even with DVD or Blu-ray storage, you’d hit a wall pretty quickly. And once you factor in branching paths, physics, lighting changes, and player agency, it becomes clear why real-time rendering won out: it’s scalable, flexible, and doesn’t require pre-baking every possible outcome.

So yes, it could be done in a limited way, and maybe a PC-FX with cutting-edge hardware could’ve pushed that envelope. But FMV’s core limitation is that it’s locked to what’s already been rendered. That makes it great for storytelling, but tough for dynamic gameplay.

Still, imagining what FMV could’ve become if it had evolved alongside hardware is a fascinating rabbit hole. It’s a reminder of how many paths gaming history could have taken.

 

[Jeff]

PC-FX is an interesting console, but it had a short library of games that weren't games, few of them were visual novels, software that show upcoming anime or anime that are still on in Japan, and visual novels. The only good PC-FX game I can think of is Kishin Douji Zenki FX: Vajra Fight and a few Tenchi Muyo point & click games.

 

[Kanon_Magic]

Totally fair to explore the “what if” scenario speculation is half the fun when talking about retro tech. And yeah, in theory, a system could pre-render a massive number of frames from a 3-D scene and store them for rapid access. With enough storage and clever indexing, you could simulate interactivity by jumping between frames based on input. It’s kind of like a supercharged version of Dragon’s Lair, but with more nuance.

But the challenge isn’t just technical it’s logistical. The number of permutations needed to allow fluid movement, camera control, and responsive gameplay would be astronomical. Even with DVD or Blu-ray storage, you’d hit a wall pretty quickly. And once you factor in branching paths, physics, lighting changes, and player agency, it becomes clear why real-time rendering won out: it’s scalable, flexible, and doesn’t require pre-baking every possible outcome.

So yes, it could be done in a limited way, and maybe a PC-FX with cutting-edge hardware could’ve pushed that envelope. But FMV’s core limitation is that it’s locked to what’s already been rendered. That makes it great for storytelling, but tough for dynamic gameplay.

Still, imagining what FMV could’ve become if it had evolved alongside hardware is a fascinating rabbit hole. It’s a reminder of how many paths gaming history could have taken.

You can't know that. It's like MP3s. Music used to take up a lot of space on a CD; you only had a maximum of 80 minutes. Then MP3s came along, and you could fit more than 800 mins in the same space. XP

I would think it would be something similar to Sega Saturn's Virtual Hydlide, which is an example of a game with FMV-style visuals, but which is polygonal at its core.

 

[RETRO-VETRO RETRO-VETRO]

You can't know that. It's like MP3s. Music used to take up a lot of space on a CD; you only had a maximum of 80 minutes. Then MP3s came along, and you could fit more than 800 songs in the same space. XP

I would think it would be something similar to Sega Saturn's Virtual Hydlide, which is an example of a game with FMV-style visuals, but which is polygonal at its core.

I get the comparison you're making with MP3s it’s a clever analogy. Compression absolutely revolutionized how we store and access media. But FMV isn’t quite the same beast. MP3s compress audio by removing frequencies humans don’t notice much, while FMV is video, which is far more complex and harder to compress without sacrificing quality or interactivity. You can’t just “MP3” your way into making FMV gameplay scalable.

As for Virtual Hydlide it’s actually not FMV in the traditional sense. It uses pre-rendered backgrounds and digitized sprites, but the gameplay itself is polygonal and rendered in real time. It’s more like an early attempt at blending static visuals with dynamic movement, not a true FMV engine. The player’s actions aren’t triggering video clips they’re manipulating 3D models in a 3D space.

FMV games, by contrast, are built around branching video sequences. You’re not controlling a character in a rendered world you’re selecting which clip plays next. That’s a huge limitation when it comes to responsiveness, physics, and emergent gameplay. Even with better compression or faster access, FMV is fundamentally constrained by what’s been pre-recorded.

 

[irkella irkella]

But there is one thing era-specific calculation error to mention. These businessman who are no gamer looked at FMV games that using real pictures in games and all so they thought it's better than 3D.

Even weirder: NEC decided to make an extremely odd pair of PC-FX add-in boards for IBMs and PC-9800s that had 3D capabilities, for seemingly no reason.

 

[UnDubious]

I've beaten Team Innocent, Zenki, and Zeroigar. TI is okay, and it's impressive that they were able to make a Resident Evil style game with just 2D and sprite scaling. But it's not as good as any of the good survival horror games on PS1. Actually came out before RE, meaning the inspiration had to be Alone in the Dark. The story is decent, gameplay is just alright. Zeroigar is just an average Compile-style shmup, I actually didn't like it that much. Zenki has nice combat but only about 20 minutes of gameplay plus lots of unskippable dialogue in Japanese, so it doesn't feel like a full game. It's a shame as the potential was there. Clearly the console could do some nice 2D, but I haven't seen anything on the level of the best-looking 2D Saturn games.
Apart from that, there's a couple untranslated RPG, but really not many. The one I'd like to see translated is Miraculum, it looks cool in a late 16-bit/early 32-bit way. The same people made Star Breaker and Startling Odyssey 1 & 2 on PCE-CD, none of which have a real translation either.