It's the same issue you encounter with audio mixing. You have to clamp out-of-range values, even though they don't occur a lot. If you don't you get awful artifacts, and have to lower everything so that it can never overflow your range.
Gladly this isn't that much of an issue anymore since most audio mixing nowadays is done using floating point numbers (either 32 or 64 bits).
Last week I made a test in Bitwig where I boosted a 0dBFS sine wave by 64dB into "clipping", exported it to 32 bit float wav, imported it again, reduced gain by 64dB and there the old sine was in all it's glory. Theoretically now with rounding errors and a loss of precision, but nothing audible.
Of course as soon as you go into 24 or 16 bit fixed point representations (and you will have to eventually) that clipping becomes a problem.
There should be little loss of precision exactly because floating point is by definition precision-invariant over most of its domain – ie. 1.23456*10^-30 and 1.23456*10^30 are essentially the "same" number. Precision loss normally only occurs when combining numbers of different magnitudes.
To be pedantic (apologies), I’m certain you know this, but not everyone thinks about floats - precision loss is almost always occurring in practice with every floating point operation you do, and happens often even when combining numbers of exactly the same magnitude (exponent). They’re just very small losses on the order of 0.5 ULP on average, and it would take many many of those to add up to audible artifacts. It is sometimes possible, but quite rare to do any float math and have no precision loss, that only works if you can change the exponent without touching the mantissa, or in the narrow range of add & subtract operations where the exponents of the operands and the result are all the same. Multiplication is always rounding even when all input & output exponents are identical.
This is of course why we have audio compressors that lower everything while it’s loud and bring it back up when it’s quiet and try to do it quickly but without anyone noticing too much. It’s theoretically possible the same trick could work on the N64, except it would require & cost extra math ops that weren’t there.
> Did you ever wonder why explosions and other effects looked so much cooler on the original PlayStation than they did on the Nintendo 64?
Begging the question, aren't we?! Of the examples displayed, I much prefer Star Fox's fx to Silent Bomber's. They fit the game's style well, and the explosions when killing an enemy are just the right amount of rewarding, while not being so ostentatious as to be distracting. SF64 nailed the game feel of destroying enemies, those small little intangibles that make the game satisfying on a visceral level, as Nintendo is so good at doing.
The Starfox explosions lack a realistic brightness level. They don’t appear to be emitting light the way the Playstation’s explosions do. Even the afterburner effect on Starfox’s ship looks muted.
> Of the examples displayed, I much prefer Star Fox's fx to Silent Bomber's. They fit the game's style well
...I feel like the "style" you're reacting to is the "N64 style", which is to say that you've become habituated to seeing games built to the N64's specific limitations. If the N64 had launched with better alpha blending, then "N64 style" would include PS1-esque explosions, but it didn't, so you don't.
It's like how people prefer 24 fps, because movies have long been recorded at 24 fps.
Obviously, this is just my intuition and we'll never the know the alternate history. But objectively, I don't see why explosions which resemble physical spheres better fits the aesthetics of Star Fox. But, those sorts of explosions absolutely do evoke the N64. If I was creating an N64-esque indie game, I would want to avoid alpha blending.
The point is that good art direction within technical constraints matters more than the constraints. For example, Wind Waker holds up significantly better than any realistic PS2 or X360 game from that era. Is the GameCube a less powerful console than its competitors, yes. But Nintendo built a beautifully stylistic game within those limitations, while 'realistic' games without stylised art direction just ended up looking both shoddy and generic as their platforms became obsolete.
You missed the point, it is not about art direction. The effects in Star Fox have normal alpha blending and look "solid", you can see them as a floating poster board, while the PS ones actually look like light effects.
You know what else didnt support additive blending? S3 ViRGE decelerator, Matrox Mystique mistake, nor NEC PowerVR PCX 1/2, afaik not even Permedia 1/2.
Another pain point was multiplicative blending missing in even more early 3d accelerators.
Then came 3dfx Voodoo doing all the blending modes for free.
Wasnt December 1995 Comdex premiere still a closed door dont look behind the curtain $80K SGI box showcasing what the chip will do? with actual boards shipping a year later starting October 1996. Flood began after 1997 ram price collapse dropping initial Voodoo1 $299 price down to $199.
What I'm getting from the article is that to work around the lack of clamping in the RDP, it is set up for additive rendering with low intensity onto a 32-bit framebuffer to make use of its higher dynamic range. But that's way too dark for display, so the RSP is used to scale and clamp this down to a 16-bit framebuffer that is then displayed.
Using the RSP makes this possible at a reasonable framerate, but it requires writing custom microcode for it and Nintendo didn't provide the necessary tools or documentation back in the day.
My read is that this makes pseudo-HDR rendering techniques possible on the N64.
> Nintendo 64 games running on custom microcode benefit from much higher polygon counts and more advanced lighting, animation, physics, and AI routines than its competition. Conker's Bad Fur Day is arguably the pinnacle of its generation combining multicolored real-time lighting that illuminates each area to real-time shadowing, and detailed texturing replete with a full in-game facial animation system.
Probably more accurate to say the custom microcode tools were not given to third party developers. From what I remember, you had to list the microcode used on your submission form, and Nintendo checked for it in your ROM. Did some third parties make custom microcode? Highly probable, but I do not know how many were able to ship with it.
Factor 5 famously customized microcode in essentially all of their N64 games. According to some emulator devs, the F3DEX source was available in the SDK and intended to be customized [1], so it's interesting to hear a different account .
Interesting. I was on PlayStation so I only heard about N64 dev at lunch. Factor5 were much closer to Nintendo than we were, and also known to be a bit good. Plus SDKs change over time. Take PS1 for example. We didn’t get GTE register details and inline functions initially, just libraries, and the more open it became the more we wrote in assembly.
> However, it is possible to implement saturating addition and subtraction in software without branches, using only modular arithmetic and bitwise logical operations that are available on all modern CPUs and their predecessors, including all x86 CPUs (back to the original Intel 8086) and some popular 8-bit CPUs (some of which implement the Z80 instruction set) are still in production.
Does anyone know what the implementation of this is, without conditional moves?
move.b #200,d0
move.b #100,d1
add.b d0,d1 ; 8th bit goes to X flag
subx.b d0,d0 ; d0 = $00 or $ff depending on X flag
or.b d0,d1 ; d1 = 255 if there was saturation
You can do the same with all others that somehow store
the 8th bit carry somewhere then allow using it for substraction.
