/vr/ - Retro Games

>>6324071
simple textures possibly

>>6324071
even crazier when you realize that 6mb of that are the soundtrack mp3s.

Compression. Lots of repeating textures that are easily compressed, MIDI soundtrack with as few samples as possible, heavily compressed wah-hoos and bing-bing voice samples. Not quite black magic as much as Nintendo being incredibly thrifty with the few assets that exist in the game's data.

No FMVs and no CD audio helps keep file size down a lot on a cartridge.

>>6324071
Oh it shows

>>6324071
Well they only had like 20 textures and 8 music tracks so it really wasn't all that hard. polygons themselves don't take up much data.

The skyboxes make up a majority of the .

>>6324071
In CD games most of the space is taken by FMV and music.

>>6324071
Simple meshes and animation, low texture variety. If the music was stripped from the game it would fit on a snes cart.

>>6324097
I think this is the most accurate answer.
The other games OP listed are also very limited in terms of assets, with the exception of Rayman and Lomax, which are 2D.

>>6324110
The 3D games have higher quality assets and more of them.

>>6324117
Doesn't look like the little more they might have justifies the 300 MB. It has to be audio and video data.

>>6324121
Mostly audio, which is appreciated. Mario 64's soundtrack is really really disappointing.

>>6324125
So this is an anti-Mario 64 thread? I don't get it. thought OP was praising Nintendo for the trickery they did.

>>6324127
What are you talking about? Mario 64 doesn't have a lot of music tracks and reuses them a lot. Not everything is cult of oppositions you tool.

>>6324127
op was praising it, but anti-nintendo/anti-n64 people flocked to the thread quickly.

>>6324127
OP here, I am. Even if you strip the audio from the games I mentioned. You'd still end up with over 150 MB of data. This stuff is truly black magic

>>6324130
>>6324127
I've noticed a lot of 5th gen console warring lately.
It's funny how we fluctuate between 4th and 5th gen console warring, as if they're mini-seasons.
Wonder if it's all orchestrated by the same people.

>>6324127
I don’t think anyone is bashing it. Just giving a simple explanation to a simple question. There’s no compression magic. Just good art direction and asset utilization. Nintendo aren’t technical wizards they just have really good designers that deeply care about the product

>>6324130
OP ruined his own thread as soon as he started comparing it to fucking CD games of all things.

>>6324139
yeah it almost seemed as if it's a n64 vs ps1 thread. I mean it practically is.

>>6324071
>What black magic did Nintendo use to cram all of Mario 64 content into 8 (eight) Megabytes?
Tracker music, low-res textures, vertex shading where possible. Keeping levels and asset pools small.
Overall filesize wasn't that impressive at the time, it was being constrained by the console that was the thorn in dev's sides. Outside of Crash Bandicoot, which had to delete a lot of Sony's inbuilt codestacks to use streaming to get decent levels, most of the CD-Rom may as well have been a free garbage dump for most developers, it was a bit of a running joke.

>>6324089
they're not mp3s. It doesn't take much data at all if you're using the system's chip for music.

>>6324132
They didn’t bother compressing anything or care about data management because they had more than enough storage. There’s no magic.

>>6324141
Do n64 fans need to be so insecure about the generation? I get that there's not a lot to talk about with the platform but a little bit of levity couldn't hurt. There's no winning battles lost decades ago, just play the games and be happy.

>>6324150
there is some magic in how spyro and crash worked, they used data redundancy so the laser could read more data sequentially instead of waiting for a full rotation. It helped reduce load times drastically. Spyro in particular was an important game in figuring out how data streaming worked and became such a crucial technology in the following 20 years.

>>6324158
>I get that there's not a lot to talk about with the platform
console warring is rotting your brain.
>>6324158
>there is some magic in how spyro and crash worked
Sony fanboy.

>>6324158
>Spyro in particular was an important game in figuring out how data streaming worked
Does Spyro really have that much data to stream? It looks so simple.

>>6324158
>There's no winning battles lost decades ago
Wait, you unitonically think in terms of "win" and "lose" when it comes to video game companies?
Also, I don't like 5th gen much in general. If anything, I prefer Saturn over the other 2, but that's not saying much.

>>6324163
You need to cope.

>>6324165
Of course it does, the levels are substantial in size, often make use of unique assets per stage and each one has a unique music track. It had a lot to juggle and Insomniac did a lot of heavy lifting with its streaming tech and level of detail tricks.

>>6324175
>think in terms of "win" and "lose" when it comes to video game companies?
only customers win or lose, and that generation there were a lot of winners, but also some very sore losers.

>>6324158
>there is some magic in how spyro and crash worked, they used data redundancy so the laser could read more data sequentially instead of waiting for a full rotation.
That thing is way overblown devs on 8 bit computers already knew how to optimize load times by rearranging data on fucking floppies.
CD just made thing way easier thanks to all the disc space ("data redundancy").

>>6324182
>Of course it does
Huh, well it doesn't seem like it.
As for unique music track per level, I mean Sonic CD did the same back in 1993, is that supposed to be impressive? I'm pretty sure there's many other games with unique level music on CD format.

>>6324182
>only customers win or lose, and that generation there were a lot of winners, but also some very sore losers.
It really depended on what games you had.
I grew up with a PS1 but my games were stuff like Crash, sports games and dragon ball final boy, I was jealous of my friend with N64 who had all the 1st party games.
I'm pretty sure the opposite also happened, guys with a N64 and Superman 64 being jealous of PS1 owners with stuff like Valkyrie Profile and Dragon Quest VII.
In my opinion there's no clear winners or losers depending on the console they chose. Unless they're idort.

>>6324194
Very different things anon, you can't forget as more RAM becomes available, there is more RAM to fill. The lessons taught on 8-bit machines loading off of floppies may have been relevant during the 32-bit days with loading off of CDs, taught valuable lessons and provided foundational concepts, but the ideas pioneered during the days of CDs became important to this very day. Maybe somewhere there's some deep memory of floppy loading on there

>>6324198
Sonic CD's soundtrack has much more seams in how it's streamed. It dumps the relevant game data into RAM and streams audio, Spyro is juggling sampling audio, streaming models, textures and level data, never fully clearing the video RAM to provide a seamless experience by always showing something on screen.

You seem eager to sell the achievements of Spyro short while we're still feeling its impact today. Can't say the same about Mario 64 or Sonic CD's technology.

>>6324219
>Can't say the same about Mario 64
Well, the same way, you seem eager to sell the achievements of Mario 64 short.
Also I'm not really against Spyro or anything, just thought streaming unique music levels isn't something new, but okay I can see where you're coming from.

>>6324226
Cartridges weren't an achievement for Mario 64, the movement was. We can talk about that if you want to talk about how its movement defined an essential way to approach 3d movement of a character we can.

>>6324219
It's not my fault the exemple you gave ("laser could read more data sequentially instead of waiting for a full rotation"), is one of the first thing devs figured out. Programmers were already doing this in the 80's by optimizing data order.

>>6324071
I honestly think it was the sound quality difference. Look at games that had versions for both consoles and the N64 version is always a smaller file size, but IMO crappier sounding music.

I always felt like the PS1 had better textures in terms of details in a lot of cases. At least, a lot of N64 games use simpler textures AFAIK.

>>6324089
>the soundtrack mp3s
10/10

>>6324302
PS1 textures are pixelated messes, N64 gouraud shading likes like shit too but at least it is somewhat organic in appearance.

>>6324071
the music and sounds on those ps1 games make up the vast majority of space. it was just devs taking advantage of the new disc medium. if ps1 used cartridges most games would probably be just as small if not smaller

>>6324071
The overwhelming majority of data in PS1 games is literally just the CD quality audio.

>>6324364
fine pixel art at a higher resolution > n64's blurry shit smears

>>6324071

>>6324479
>3D models
>pixel art

>>6324482
textures are pixel art zoom zoom

>>6324515
Textures are only pixelated on low end hardware around 1994. By 1996 even consumer hardware was past that ugly pixelated look.

>>6324516
except the n64

>>6324518
What? N64 uses texture filtering.

>>6324521
to filter super low res textures that are lower res than its contemporaries

>>6324524
Ok. It still generally looks better.

>>6324526
I disagree, pixel art looks best sharp and clean

>>6324089
truly underrated. we're dealing with next level here.

>>6324527
Being able to see the individual pixel edges takes away from whatever the texture is supposed to depict. It isn't additional detail, it's a visual artifact.

>>6324535
that's what texture filtering is, yes

>>6324527
not when that pixel art is textures in 3D games that attempt to recreate reality

>>6324540
Texture filtering works to smooth over visible pixel edges, which removes the artifacts.

>>6324071
Mario 64 had next to no texture work, and all the music is midi.

>>6324545
it smooths over texels, blurring them. Edges it does nothing for. If the texture resolution is super low there's less detail to blur the look will be extremely soft.

>>6324543
not many games from the era tried to recreate reality, and even then rough pixel art details can offer distinct benefits over blurred ones.

>>6324562
>even then rough pixel art details can offer distinct benefits over blurred ones.
Visible pixel edges are not detail.

>>6324565
Yes they are. They can be used to create areas of high contrast which has its own use for perceived detail.

>>6324568
>They can be used to create areas of high contrast which has its own use for perceived detail.
Pure nonsense, show me examples. Visible pixel edges are not detail.

>>6324570
Visible pixel edges makes the water shimmer done with a single texel more pleasant to the eye. It's not hard to imagine even without an example, that's how perceived detail works. Filter the textures and it becomes a softened mess of a texture that loses its finer details.

You need to make better posts and try thinking for yourself.

>>6324207
This, we had both and I ended up considering psx a shit console with no games, but I just didn't have the right games until I got older and bought my own games. Both consoles have a handful of great games, a bit more good enough games, and then mountains of shovelware. There's also the issue of your taste changing over time. I didn't really have any interest in the Final Fantasy games, but I did eventually open up to them and other JRPGs more.

>>6324579
>Visible pixel edges makes the water shimmer done with a single texel more pleasant to the eye
That combined with the lack of mip mapping leads to a very ugly image off in the distance. I fail to see how that pure visual noise is supposed to be pleasant to the eye.

>>6324583
If you don't find it pleasant that is personal preference, but it was not what was being discussed. If you want to discuss your shit taste you need to post something more substantial.

>>6324585
Your example would work if there weren't better, less ugly ways of achieving water shimmer effects.

>>6324579
That looks fucking disgusting.

>>6324589
There are other ways to achieve it, but they massively change the feel of the image. You keep meandering and floundering, why don't you say something more concise and to the point? If you required an example from me, I require one from you. Explain what benefits it has over my example and why sharp shimmers are undesired despite providing a clear and easy way to provide extra eye catching details at low cost and creating an appealing contrast.

>>6324596
Your shit taste is noted.

>>6324598
>Your shit taste is noted.
That skybox looks awful, I’m not sure you’re pretending that looks good. I’m not saying blurred shit-smeared textures of the N64 are anything special either but come on.

>>6324598
It's really cliche at this point to post Wave Race in a discussion about retro game water, but I'll do it anyway.

>>6324602
Cum on ya face, it looks gorgeous.

>>6324604
Wave Race is just the example I was waiting for. Notice how solid it looks in its colors, the water itself flows smoothly with blooming white gradients in an obvious radius around the player. The water has no visible shimmers or foaming, it all blends into the same mess of detail.

It's both cliche and a very different feel for the water. There is no texture. Water shouldn't always look like gelatin.

>>6324613
I mean water also isn't an ugly flat pixelated mess but you didn't hear me say that.

>>6324619
Look up the game in action, it looks very nice.

>>6324620
Honestly the water looks even worse in motion with how clearly tiled it is and it's lack of any sort of waves.

>>6324634
And despite all of that, it's very detailed. That was the entire point of the discussion.

>>6324654
>And despite all of that, it's very detailed
No it isn't. You seem to think that pixels = detail but that just isn't the case.

>>6324658
Actually yeah, more pixels is more detail. You can also do more things with high contrasts filtering doesn't allow you to do.

>>6324660
>Actually yeah, more pixels is more detail
Not necessarily especially when the pixels are completely unfiltered resulting in an artifacted image.

>>6324668
>more detail isn't more detail
uhhhhh wow
>artifacts
you don't really know what this means, do you?

>>6324071
>comparing CD based games with a cartridge game

>>6324675
Visible pixel edges are not detail and are a visual artifact as a result of 1994 3D hardware.

>>6324705
It gives room to create the perception of detail, my illiterate friend. That's how sharp contrasts work. That's also not an artifact.

>>6324712
>It gives room to create the perception of detail
Pure nonsense as your attempt at an "example" already demonstrated.
>That's also not an artifact.
Yes it is.

>>6324719
>Pure nonsense
Not quite, it's how the human eye works, as demonstrated before. Something like that would be impossible with an aggressive texture filter. Don't tell me you're not human as your next argument.

>>6324724
>Something like that would be impossible with an aggressive texture filter
Yes creating an aggressively ugly pixelated image is kind of what filtering is supposed to avoid.

>>6324729
Unfortunately a lot more is lost in an aggressive filter. It succumbs to artifacting, the real kind not the kind you think texel edges are. Combine that with ultra low res textures and you have a substantial loss of detail. A lose/lose situation.

>>6324735
>Unfortunately a lot more is lost in an aggressive filter.
God forbid we lose the game looking like a pixelated mess.

>>6324739
Yes, fine pixel art going to waste is a real shame.

>>6324746
Textures are not pixel art.

>>6324749
Sure they are. The artists placed every dot intentionally to create fine details in the texture. Coincidentally you can observe what happens when they're compressed to fit into a tinier texture budget. You might not be very familiar with how these 5th gen games were made, but simply put the skills honed during the 3rd and 4th generations nice pixel art became applied very easily to 3D environments with the diffuse texture.

You're a living contradiction, do you love being retarded and wrong all the time? A glutton for abuse.

>>6324759
>The artists placed every dot intentionally to create fine details in the texture
Which also applies to textures that were designed to be filtered, said textures benefit from detail without visible pixel edge artifacts effecting the creator's vision.

>>6324771
>Which also applies to textures that were designed to be filtered
Now you're getting it! Good job. I knew you weren't a lost cause.

>>6324779
Ok, now explain how pixel edge artifacts are beneficial. Clearly you must acknowledge they are preventing the texture artist from more accurately translating the idea in their head to the game itself. I don't think people's minds imagine pixel edges.

>>6324782
>Clearly you must acknowledge they are preventing the texture artist from more accurately translating the idea in their head to the game itself
Sure they do, they account for the fact they are present, use gradient like colors to blend them to soften the transition to them, all the consoles at the time had options for dithering filters to soften the image (actually I think the saturn didn't?) and would also rely on old techniques like composite blending.

If they didn't want you to see them, they would do something about it. If it added to what they were trying to do, such as how Quake uses fullbrights for some textures, they would be tagged as such. The more control you have over this the better, don't you think? If everything you do is forcibly filtered it really limits your options as an artist.

>>6324796
>Sure they do, they account for the fact they are present
So they are a limitation on an artists creative input. Sounds like a negative outright.

does the trolling ever just stop or is this board condemned also.

>>6324806
Try reading the rest of the post my poor little retard.

>>6324071
>PS1 sizes
Misleading.
Most of the disc space for those was taken up by videos and sound. There are some exceptions, like games that made heavy use of prerendered backgrounds, or streamed chunks of levels.

For example, Rondo of Blood is actually around 25 MBs of actual game data.

>cram all of Mario 64 content into 8 Megabytes
1- barely any textures used at all, everything is untextured polygons. Level geometry is full of places that would have had signs, all made invisible. When textures are used they're repeated, or stretched and blurred a lot (Zelda OoT cranked this up)
2- heavy asset reuse, for polygons, sound samples, animations, etc
3- everything is compressed with the MIO0 compression (as is the tradition for launch title Marios, SMW is half the file size of what it has any right to be, because it's heavily compressed compared to most of the later snes library)

>>all that shit
>>6324782
>>6324796
Mods

>>6324814
I dunno, the N64 fans really are unhinged lately.

>>6324815
"arguments" like
>If they didn't want you to see them, they would do something about it
Make no sense when we're talking about a hardware limitation. In a way they did do something about it in fact. Since by 1996 all gaming hardware going forward would support texture filtering.

>>6324826
Yes, higher resolution textures mitigated the desires for unfiltered textures. It wasn't until fairly recently Doom and Quake communities realized there was an art to the unfiltered textures and people started to embrace them more. There's been an uptick in developers embracing unfiltered textures for aesthetic reasons, realizing the unfiltered look lets you translated 2d pixel art to 3d surfaces with appealing results.

>>6324830
>There's been an uptick in developers embracing unfiltered textures for aesthetic reasons
You mean talentless hipsters cashing in on nostalgia for system limitations.

>>6324837
Who hurt you?

>>6324097
>>6324108
>>6324479

>CD games are all CD audio

Why is this still going around? You guys do realize that a good chunk of Saturn and PS1 games use Midi-based music or compressed audio Streams like ADX right?

None of these examples are CD Audio:
https://www.youtube.com/watch?v=BGDs5NkJiiU
https://www.youtube.com/watch?v=CZMWszd5SRk
https://www.youtube.com/watch?v=nkQAkLBjwuc
https://www.youtube.com/watch?v=i36tXq7yK6c
https://www.youtube.com/watch?v=df5JD0-P3gM
https://www.youtube.com/watch?v=exdt3YEXVOc
https://www.youtube.com/watch?v=exdt3YEXVOc

unga bunga n64 bad

>>6324863
Last link didn't copy and paste properly:
https://www.youtube.com/watch?v=J6GrZYE2il0

>>6324613
Are you belittling the best water effects of that era in defence of water that looks like carpeting in a convention hall without a semblance of undulation or movement? Well I suppose you at least have muh foam.

>>6324117
>>6324097
Once you do 7zip, IF the emulator supports it?
Some gamecube game are even 100-150MB, since the compression for video isn´t aggressive.
Most PS1 games land in 10-15mb range with some exceptions once you remove the FVMs or compress them to death with modern algorythms.

The biggest advantege of the playstation model is 700mb of disc space + 2 hours of Redbook full audio. Possibly a dual track.
But as most people who dabble knows, its rare.
Quake did, but most Source ports do not have a real standard for how to OST. Meanwhile you need to record fucking bin + complimentary data files for what is in 95% one single audio track.

>>6324894
That still looks better than the water in Wave Race 64

>>6325236

>>6325236
the game looks more fun too

>>6324613
>just the example I was waiting for
You seem very invested in this console comparison game.
I love both GITS and WR64, but WR64 does have the most impressive water effects.
The GITS water looks good on screenshots, but when you play the game it just feels like you're strafing through a carpet, it doesn't have water physics, transparencies, etc.
There's better water effects on Panzer Dragoon II when it comes to 32-bit.

>>6324109
I’d like to hear that, actually. Someone should figure out a way to feed N64 MIDI data into the SNES soundchip, like using a soundfont but with hardware/emulation.

>>6324863
>midi-based music
Using what? I always had the impression that PSX didn’t do MIDI. Also, I don’t think many people distinguish “CD audio” from compressed audio streams. They’re both recordings and not being played live via instructions interpreted by the sound chip ala SNES and N64.

>>6324071
triangles son, triangles

>>6324147
>vertex shading
You mean using vertex colours instead of textures?

>>6325285
>Using what?
It's internal sound chip. The Saturn and PS1 both have some pretty advanced internal sound chips. The PS1's can do up to 24 PCM channels and can play back ADPCM samples as well. It's pretty much the successor to the SNES Sound chip. These are all examples of the PS1 playing back MIDI:
https://www.youtube.com/watch?v=J6GrZYE2il0
https://www.youtube.com/watch?v=QY1Vetd7OCs
https://www.youtube.com/watch?v=qDD-iYkHBhc
https://www.youtube.com/watch?v=lgb6peaKCSc
https://www.youtube.com/watch?v=CZMWszd5SRk
https://www.youtube.com/watch?v=3v6BtJaBQmo
https://www.youtube.com/watch?v=3jho-peCAKs
https://www.youtube.com/watch?v=RVkcmx2l3WA
https://www.youtube.com/watch?v=fPoOtOcuRqw
https://www.youtube.com/watch?v=BPOSBDrqoDY

The Saturn has it's own custom Yamaha chip that can do 32 PCM channels, though it can't play back compressed audio samples, they have to be decompressed in RAM first. However the Saturn's chip can also do FM Synthesis. Each channel can act as a 1op FM synth channel, or they can be chained together. So you could do 8 4op FM Synth channels, or mix and match for say 6 4op FM Synth + 8 PCM. These are all examples of the Saturn's internal sound chip playing back MIDI/FM:

https://www.youtube.com/watch?v=JwdiACmTQfs
https://www.youtube.com/watch?v=rTmrdYyRbpY
https://www.youtube.com/watch?v=5z-crDa7etg
https://www.youtube.com/watch?v=_X3u39tIbD8
https://www.youtube.com/watch?v=VESzdVBIqhc
https://www.youtube.com/watch?v=JxyZgtSdmbY
https://www.youtube.com/watch?v=xhf3dpaEyIc
https://www.youtube.com/watch?v=tESPxv7VHOQ
https://www.youtube.com/watch?v=fLloK0eGNZU

>Also, I don’t think many people distinguish “CD audio” from compressed audio streams

They should because things like ADX can significantly reduce the footprint on the disc that the audio takes up giving more space for the game.

>>6325285
>>6325532
Going a little bit further, the funny thing is the N64 is actually the system that doesn't have a sound chip, at least not in the traditional sense.

The N64's audio is pretty much CPU brute force by the RCP. As a result developers had to play the balancing act of "Do I want nice music or nice graphics?" Late in it's life some developers like Rare were able to kind of achieve both, but even those aren't on par with the best you get out of the PS1 and Saturn.

Sure on paper it says it can do up to 100 PCM channels, but that's only if all the system's resources are dedicated to doing that. In reality you tend to have a lot less, and in most cases less than even the PS1 and Saturn. That's why a lot of N64 chiptunes tend to sound so empty or shallow by comparison. They're not using the same amount of channels as PS1 and Saturn stuff to create more depth.

>>6325597
The N64 had the worst audio from that generation in general but it was also the only one with surround sound support in certain games.

>>6325608
Yeah about that:
https://segaretro.org/Dolby_Surround#Saturn
https://www.mobygames.com/attribute/sheet/attributeId,136/p,6/

>>6325597
I hate how the N64 only has two games with good music, it's such a stupid fucking hardware design.

>>6325608
love how you've been running around with that assumption for a while now, stop talking out of your ass any day now.

>>6325638
What two games anon?

>>6325642
mystical ninja starring goemon and goemon's great adventure

>>6324579
Bilinear filtered textures produces the exact same result. Mip-mapping is what reduces that long distance shimmering you describe (which looks shit anyway.)

>>6325628
That is a seriously under advertised feature on some of those games, it's not even on the back of box in all cases. Still I was totally wrong.

>>6325643
>he doesn't like Bomberman Hero's soundtrack
https://www.youtube.com/watch?v=Fv-2XIhmxPE

>>6325654
meh, put it back in the youtube analysis video soundtrack with the nier ost

>>6325660
>he watches youtube analysis videos

>>6324071
It's a big, empty world with nothing to do but jump around shouting Italian curse words and collect everything

that's what the n64 was designed for, big game no content collect a lot of things

>>6324182
>It had a lot to juggle and Insomniac did a lot of heavy lifting with its streaming tech and level of detail tricks.
Mario 64 also uses level of detail. Two versions of it actually. Low detail versions of character models (only when models are very far from the camera) and mipmaps for textures.

>>6324219
>Spyro is juggling sampling audio
Only for sound effects. Spyro uses redbook audio for music.

>>6324579
>Visible pixel edges makes the water shimmer done with a single texel more pleasant to the eye
Tends to look alright in still shots, but aliasing looks bad in motion because the sampling error is uncontrolled based on camera movement, not a controllable effect like an animation. That's essentially what textures without proper filtering do - produce sampling errors.

To be fair, even trilinear filtering at steep angles produces sampling errors. That's why you need anisotropic filtering to make it perfect. That being said, certain art styles are designed to look good even with sampling errors, and can look worse when hardware attempts to correct those errors.

>>6324613
>Water shouldn't always look like gelatin.
If you want to see gelatin water, look up Crash 3. At least WR64 water is fully alpha blended.

>>6324819
>everything is untextured polygons
Not sure why people keep trying to suggest the N64 is special for using gouraud shading when it is much more prominent on PS1. Look at Final Fantasy VII, Spyro, Crash, etc. I've noticed N64 games tend to shun use of gouraud shading for low-res textures if anything. It's true for Mario 64 as well, what people keep implying is untextured is actually textured.

>>6325256
>There's better water effects on Panzer Dragoon II when it comes to 32-bit.
Problem is that the scrolling of the water is completely perspective incorrect (the actual mapping is perspective correct though) because it's a VDP2 layer. It's also just a projected single flat plane with distortion effects.

>>6325628
>https://segaretro.org/Dolby_Surround#Saturn
These Saturn games don't use real-time Dolby Surround encoding, but baked encoding. You can bake Dolby Surround into any console that uses PCM channels. It would be possible to bake Dolby Surround into NES games. Hell, there's even a handful of SNES games like Super Turrican that use baked Dolby Surround.

What the N64 can do that the PS1 and Saturn can't is generate a real-time Dolby Surround matrix on the fly and combine it with a PCM stream. The PS1's sound chip is wholly incapable of this (purely fixed function setup), while the Motorola 68000 audio supervisor on the Saturn doesn't have the matrix performance to do it.

>>6325597
>The N64's audio is pretty much CPU brute force by the RCP
No, the RCP doesn't produce audio through brute force like a CPU does. It has a vector unit (a programmable and re-purposeful one) . What do you think a sound chip has inside of it?

>>6326132
Actually, NES games can't support Dolby Surround since it's only mono without mods/emulation, and even then, the mods are fake stereo.

>>6326154
Point taken. Should have said, any console capable of true stereo PCM.

>>6325597
>Sure on paper it says it can do up to 100 PCM channels, but that's only if all the system's resources are dedicated to doing that. In reality you tend to have a lot less, and in most cases less than even the PS1 and Saturn. That's why a lot of N64 chiptunes tend to sound so empty or shallow by comparison. They're not using the same amount of channels as PS1 and Saturn stuff to create more depth.
The 100 PCM channels = 100% RSP usage is based on launch day audio microcode. It's badly outdated. Later audio microcodes like MusyX are more efficient than that and use 0.7% of RSP per channel. To replicate channel-heavy usage of the PS1 sound chip, you'd be looking at 20% of RSP used up. Good thing RSP is about 4 times more powerful than the PS1's geometry processor GTE.

In any case, most N64 aren't even bottlenecked on RSP performance, but RDP pixel <-> RDRAM bandwidth, so the console usually has plenty of RSP performance to spare anyway.

If I would guess why many N64 tunes didn't sound so good it would be a combination of low amount of cartridge space, and the "default" audio microcode from Nintendo not being feature complete (IIRC it didn't support reverb! you wanted reverb you either needed a better audio microcode or do it on the CPU).

>>6326132
>while the Motorola 68000 audio supervisor on the Saturn doesn't have the matrix performance to do it.

So have one of the SH-2s or one of the DSPs do it.

>>6326192
>To replicate channel-heavy usage of the PS1 sound chip, you'd be looking at 20% of RSP used up. Good thing RSP is about 4 times more powerful than the PS1's geometry processor GTE.

Then why does even the best sounding N64 audio still sound worse than average PS1 and Saturn stuff? Not talking about CD Audio obviously.

>>6326624
>So have one of the SH-2s or one of the DSPs do it.
Doubt the DSP is flexible enough to do it, but sure the SH-2 could. Problem is the Saturn is already starved for T&L performance as it is (unless you find a good use for the DSP, which is not always the case) so crunching real-time Dolby matrices on the CPU isn't going to really be practical.

>Then why does even the best sounding N64 audio still sound worse than average PS1 and Saturn stuff?
Compression perhaps, and some of it down to very few Japanese devs. But it's nonsense to suggest that the best N64 sequenced tracks sound worse than "average" PS1 and Saturn sequenced tracks.

https://www.youtube.com/watch?v=1YRGbfufHG8
https://www.youtube.com/watch?v=0Y0RwyI8j8k
https://www.youtube.com/watch?v=cu1RI9rzKoE
https://www.youtube.com/watch?v=VuQSKUTe89g
https://www.youtube.com/watch?v=6rcpLWo-GeU
https://www.youtube.com/watch?v=vVPHmok3hxA
https://www.youtube.com/watch?v=5L2-ISU87rU
https://www.youtube.com/watch?v=8GQe5Jye7ik (0:41)

I mean you could argue Chrono Cross (a game with sequenced audio) sounds better than any of these, but that's not an average game is it?

>>6324071
- very low resolution textures, if any at all (most things in sm64 are gouraud-shaded)
- low quality sample-sequenced music

playstation games were mostly not small, not because they couldn't be (check out some old playstation rips, they can be very small), but because they didn't have to be, they had tons of free storage to blow on video, streaming cdda/xa music, large unique textures in every level, etc, etc

the thing about cartridges is that the bigger the game is, the more expensive the cartridges, so there's a huge incentive to make smaller games, while on the playstation, you had a fixed 650MB, it didn't cost any more to make a 500MB game than it did a 16MB game

don't get me wrong, it's still impressive they fit it all in 8MB, but it would have been far better if they didn't need to be so conscious about keeping the ROM as small as possible

>this thread
It's been so long and sonyjabronies are still seething about sm64.

>>6326678
>most things in sm64 are gouraud-shaded
Wrong

>>6326624
re2 is one of PS1's best soundtracks and the N64 version's music sounds literally identical or better

>>6324071
>Mario 64's size was 8 MB
>Look at ROM dump
>Is 32MB

>>6326678
The true horror is that N64 can do both phong and gouraud.
But unlike on PS1, there isn´t radically different how the sampling works for the end result.

>>6326698
LOL

>>6326824
The N64 doesn't support true phong shading. You can do a lot of interesting things with its pixel shader, but it's too fixed function to support a genuine phong shading algorithm.

You can kind of fake phong shading with clever use of environment mapping though, which the N64 supports.

>>6324604
Wave Race 64 is the only good game on the N64

>>6324148
>the system's chip for music

>>6326836
Digital Foundry already proved it. Voices more compressed on N64 though, but they couldn't find a difference with music quality.

>>6324071
The majority of space on old CD-based games is the audio. Look at the filesize of Shining Force CD, despite having graphics the same as SF2 and being a smaller game it weighs in quite heavily because of its high quality music.

In addition, when you have a CD full of free space there is no need to optimize or compress files, which improves loading times.

>>6326857
>In addition, when you have a CD full of free space there is no need to optimize or compress files, which improves loading times.
Not to mention it improves load times on CD to just make copies of game data over and over throughout the disk to help the CD head get there faster.

>>6326852
>but they couldn't find a difference with music quality.
holy shit, this changes everything. This proves the music is absolutely 100% better in every way imaginable. (or the same.)

>>6324089
Nice troll

>>6324819
SMW is compressed?
From my ROM corruptions and viewing of the ROM Map it seems pretty uncompressed.

>>6326861
Never knew that, makes sense.

>>6324071
You compare N64 with PS1. You should learn how both systems operate and how they differ and how they create characters and graphics. It is two different systems. You should also learn about the sound systems that were used. It is all really so different.

>>6326710
8 Mariobytes

>>6327071
Yea, in the end all that matters is that the N64 is better.

>>6326652
>Doubt the DSP is flexible enough to do it

The Saturn has 2 DSPs in it, one in the SCSP, and one in the SCU. The One in the SCU is pretty powerful actually:

https://www.youtube.com/watch?v=n8plen8cLro

As for the SH-2s being starved, Burning Rangers which is probably one of the most CPU intensive games on the system with all the graphical effects it's doing still has enough CPU time to handle the entire ADX audio decompression routine that plays multiple streams on top of each other and are streamed off the disc in real time. Hell that system even allows the game to keep playing ADX Audio streams for the disc while the game is loading other data off the disc.

As for the examples, some are impressive like the Perfect Dark one and the Goemon one, but a lot of them while having nice instruments and effects, they're still pretty flat when it comes to the number of instruments playing. There's more Polyphony going on in this track from NiGHTS than most of those N64 samples you posted:

https://www.youtube.com/watch?v=5z-crDa7etg

>I mean you could argue Chrono Cross (a game with sequenced audio) sounds better than any of these, but that's not an average game is it?

As far as MIDI playback goes I'd say so. Just about every major JRPG on the system uses sequenced MIDI to save disc space and just about all of them sound incredible. Considering how there's a massive plethora of those kind of games I'd definitely say it's representative of your typical PS1 game that uses sequenced audio.

>re2 is one of PS1's best soundtracks

From a composition stand point sure, but from a technical stand point not as much. It's nice that the N64 was able to match it, but there's definitely better examples of Sequenced PS1 music out there.

>>6327494
>The Saturn has 2 DSPs in it, one in the SCSP, and one in the SCU.
The SCSP DSP is very fixed function. It basically allows the layering of different preset audio effects.

>The One in the SCU is pretty powerful actually
Relatively speaking, yes. It’s a good helping hand for the SH2s, but it’s not really a chip that can do much T&L stuff in isolation. Compared to RSP, of course, it’s far behind. SCU DSP can do 1 add and 1 multiply per cycle at around 14 MHz, while the RSP can do 1 add and 8 multiplies per cycle at around 62 MHz.

The one of the bigger problems is that the SCU DSP has a very limited instruction set. It doesn’t really have any matrix instructions, those functions have to be cobbled together in assembly. That’s fine for certain tasks (as in the Travellers Tales example), but inefficient for others.

For processing Dolby there would also be as massive memory pipeline issue. You would need to get the 3D direction of a particular object causing a sound, then the Dolby Matrix would have to be calculated with that direction and combined with a PCM stream. The SCU DSP was already criticised for its location so this scenario would be particularly bad.

>it's doing still has enough CPU time to handle the entire ADX audio decompression routine that plays multiple streams
ADX isn’t intensive. Even the SNES had implemented an audio decompression routine in hardware without it being too intensive in transistors.

>There's more Polyphony going on in this track from NiGHTS than most of those N64 samples you posted:
Not really, if you carefully try to seperate the different instruments by ear. You just like the track.

That also said, Nights isn’t an “average” Saturn game.

>Just about every major JRPG on the system uses sequenced MIDI to save disc space and just about all of them sound incredible
Small reference pools. Try playing something that wasn’t made by Square

>>6327494
>, but a lot of them while having nice instruments and effects, they're still pretty flat when it comes to the number of instruments playing.
I still Banjo Kazooie/Tooie encapsulates the benefit: A area can have one soundtrack, but each subarea can use different instruments and reverb.
A Teddybear´s Picnic with different instruments, reverb and tempo trough out the main area is fantastic.

But as is said: When the N64 got as a few games as it did, the early devkits having a worse implementation is a big detractor.
You also see the same on the system library: There is a big number of "early" games with primitive graphics, which stands taller due fewer total games. Body Harvest and Herakles stand a lot taller than they would on the PS1, where the massive amount of games lead to improved devkits.

>>6324071
You should be asking that for Star Fox 64, with all the voice acting in the game that had to have pushed the cartridge to its limits.

>>6324071
You tell me.

>>6331535
>glitchy romhack mario 64 (1996) vs spyro year of the dragon (2000)

>>6331542
https://youtu.be/CduvEG1ItsM Mario is just an ugly game, that's why it only takes 8MB.

>>6331542
not him, and that's obviously not an accurate comparison, though the romhack screen isn't really worse than what's in sm64, and spyro 1/2/3 are all pretty close in terms of graphics
while it's not strightforward to compare an obviously cartoony style to an obviously realistic (though exaggerated) style, i don't think anyone really believes that sm64 looks as good as spyro 1-3

>>6324071
>Spyro 1 300MB
>Croc 308.35 MB
>Rayman 429.06 MB
>Lomax 547.24 MB
all of these games have CD quality music which take up majority of disc space
>Crash Bandicoot 439.42 MB
the devs intentionally tried to take up as much disc space as possible so it would load certain files faster and therefore reduce load times.

>>6332836
>the devs intentionally tried to take up as much disc space as possible
this is very common on disc-based systems
discs read from inside to outside, and disc are also fastest on the outside of the disc
this is why hdd's start at the outside of the platter(s) and work their way in
presumably cd's were designed to read from the inside to accommodate alternate disc sizes, since unlike a record, the read mechanism is entirely automatic
so you can get performance benefits from keeping your data at the end of the disk, so many games considerably smaller than a CD have several hundred megabyte dummy files just to "push" the game near the end of the disc
for crash specifically, it wasn't hard to find a 56MB rip in a minute

>Crash Bandicoot games on PS1 was so small someone managed to fit it on one disc

>>6324071
>see thread
>do not believe
>remember other n64 games being like 50 mb
>check filesystem
>8.1ish mb

Why compare psx games though? ISO/IMG is always huge just becuase of what it is.

Other n64 are like a 32 mb. Four times larger is still a hmmmm OP without being faggy and using ISO images.

>>6324596
Ask me how I know you're a zoomer.

>>6333380
the biggest n64 games were 64MiB, most were 16MiB (i believe), only a couple were 64 (like re2), with most "big" games being 32 (like oot)

>ISO/IMG is always huge just becuase of what it is.
the overhead of "being" an iso/img is very small, probably only a handful of 2KiB sectors, storing things like subchannel data is another issue which can make an image larger, but that's mostly irrelevant (only used in later games for copy protection), and can be mitigated with ECM, a program which removes predictable subchannel data, leaving only the copy protection behind (a lossless transformation)

Anything for me on the Niggy69 if I dont like collectathon platformers? 5th generation is a huge blind spot for me in gaming history, I really only played the Playstation survival horror games and the Saturn arcade stuff. But I am very well versed in 4 and 6.

>>6333796
Unironically, the Crusin' series. Also, Snowboard Kids.

>>6324071
Simple, low-resolution textures. It's why it still looks nice too, because they made it cartoonish. Grass is a simple texture with a single flower sprite repeated.

>>6333806
Lol looks fun I'll check them out

>>6333796
007 goldeneye
super smash bros.
mario kart 64

>>6324071
Iwata wizardry?

>>6333796
F-Zero X with the expansion kit translated to English
Star Fox 64
Sin and Punishment
Wave Race 64
Winback
Goemon's Great Adventure

>>6324091
it actually wasn't MIDI but a sequencer that was developed for the N64. Must have been programmed really efficiently to work with that 90 MHz processor. Some soundtracks are really amazing like link related

https://www.youtube.com/watch?v=OopYRt_-4Ls

>>6336471
>it actually wasn't MIDI but a sequencer that was developed for the N64.
sample-based sequenced music was not really that dissimilar to MIDI + a custom soundfont, and it was hardly a new thing by the time N64 came out, similar techniques were used on systems like the amiga, or the playstation, as well
and by no means is a 90MHz cpu a problem here, in fact, the cpu isn't really doing the music at all, the N64 (and basically anything else for that matter) has a chip dedicated to processing and outputting sound (the spu / sound chip)

since we're adding examples;
https://www.youtube.com/watch?v=qLAScOKONaw
a 4 channel sample-based track on the amiga, from 1992
i don't know what kind of exact specification the game required of an amiga, since i had the DOS version (same music, but using a software MOD player). needless to say, it did not require a 90MHz cpu, even the dos version ran great on my 66MHz 486

>>6336545
thanks for the info. I once tried to figure out how the N64 sequencer works, but it seems to be really weird. In snes emulators you can just toggle the sound channels 1-8 on and off (I used this to sample single instruments froms SNES osts), but that's not possible in N64 emulators. You can extract the samples with software tools, but all you get are weird sound files that seemingly have nothing to do with the actual soundtrack.

>>6332774
>though the romhack screen isn't really worse than what's in sm64
The real game isn’t quite as abstract as the image shows. It looks more like a caricature of SM64 than the real game. Also the sky isn’t filtered.

>spyro 1/2/3 are all pretty close in terms of graphics
Spyro 2 and 3 look similar, but their graphics are much better than Spyro 1.

>i don't think anyone really believes that sm64 looks as good as spyro 1-3
Generally true because Spyro has a better art style, but unlike Spyro SM64 doesn’t have the aggressive LOD which strips distant objects of textures. The LOD is particularly aggressive in Spyro 1. Most of the images I see of Spyro don’t show the LOD in action, They are either cherry picked or from emulators which do not have LOD working.

>>6336471
>>6336545
>>6336570
N64 doesn’t have a dedicated sound chip at all. But that doesn’t mean it processes audio on the CPU. It has contains a hardware vector unit which can accelerate audio processing in the same way that a sound chip does. But that chip also has to do vertex shading.

Rayman is only 4 MB.
Mario 64 is bloat.

>>6336620
i don't think the LOD is an issue in spyro, and this is from someone who 100%'d all three games on original hardware
the textures being missing from far objects is noticable, however
- they fade in/out, no popping
- the simple shading underneath mimics the basic colours of the texture, so you barely notice it if you aren't looking directly at it
- this occurs at a pretty decent distance, nothing you'd consider 'close' has missing textures
- low-LOD models and scenery exist in the game, and i actually don't recall them even existing, it wasn't until i looked at sites like TCRF that i become aware this was happening, so clearly they did they job well there

i'm not saying either game is objectively better, and it is a little unfair comparing games two years apart (which was a very significant amount of time for 3D games during that period), just that it doesn't surprise me that spyro is bigger

>>6324071
I think these other games could be just as small if it weren't for their soundtracks and probably a few better textures

>>6336639
>they fade in/out, no popping
They do pop for two reasons. First is that the polygon LOD is not done through tesselation, where the polygons gradually increase/decrease based on detail, but a swap from one set of models to another. The game tries to hid the transition with a little semi-translucency blending but it’s not fully effective.

The second is that texture transitions (including the transition to gouraud shading) is applied both on a small range of presets (instead of a large range of mip layers) and the swap-over between the different levels is per-vertex instead of per-pixel. Not really Spyro’s fault, they did the best they could with the hardware. The only way to do seamless LOD transitions is through trilinear mipmap filtering where the transitions are pixel perfect.

>the simple shading underneath mimics the basic colours of the texture, so you barely notice it if you aren't looking directly at it
Well yes, you’d hope they colors would match up.

>this occurs at a pretty decent distance, nothing you'd consider 'close' has missing textures
The distance is pretty close in Spyro 1. The sequels pushed it back a bit, but it’s still pretty obvious.

>low-LOD models and scenery exist in the game, and i actually don't recall them even existing
Well of course they wouldn’t be so noticeable. You’d more readily see the stripped out textures to Gouraud shading. What I think is more obvious in terms of geometry LOD is the enemies more obviously look less detailed.

>>6336662
I should clarify that Spyro does actually do a limited amount of tesselation, but not in a way that affects visible detail. It just draws landscapes with more/smaller triangle strips as you get closer to prevent warping. There’s a seperate LOD landscape as well with a different mesh structure.

>>6336662
>The game tries to hid the transition with a little semi-translucency blending but it’s not fully effective.
i concede i do mean it's not very noticable during gameplay, that is, textures don't instantly go from none to full contrast, i meant more "no noticable/annoying popping", not that they don't technically have a point where they switch between states
they did a pretty good job at minimizing it, and it's not something you can really notice during normal gameplay, you can tell exactly what it's doing if you run around while looking at distant objects.. but i mean...
>>6336662
>What I think is more obvious in terms of geometry LOD is the enemies more obviously look less detailed.
i don't recall seeing this, though i suppose it was quite a while ago now. even if i did see it, it evidently wasn't so bad that i remembered it