What was the Jaguar truly capable of?

[philipj]

1 minute ago, phoenixdownita said:

I'll try.

 

The N64 had no BIOS to speak of  https://retrocomputing.stackexchange.com/questions/1364/why-is-a-bios-dump-not-required-to-emulate-nintendo-64-games-in-most-modern-emul

 

The RSP was microporgrammable to begin with 

https://retrocomputing.stackexchange.com/questions/6261/nintendo-64-microcode-format

link to the gory details

http://ultra64.ca/files/documentation/silicon-graphics/SGI_Nintendo_64_RSP_Programmers_Guide.pdf

 

so CJ comments are smack onto the face of the fact that your statements are no more than technobabble in the way they are written.

 

Mind you I understand the general sentiment but you need to be factually correct first, this is technology, you can't make this stuff up as you go.

 

Gotcha...

[A Black Falcon]

52 minutes ago, phoenixdownita said:

I'll try.

 

The N64 had no BIOS to speak of  https://retrocomputing.stackexchange.com/questions/1364/why-is-a-bios-dump-not-required-to-emulate-nintendo-64-games-in-most-modern-emul

 

The RSP was microprogrammable to begin with 

https://retrocomputing.stackexchange.com/questions/6261/nintendo-64-microcode-format

link to the gory details

http://ultra64.ca/files/documentation/silicon-graphics/SGI_Nintendo_64_RSP_Programmers_Guide.pdf

 

so CJ comments are smack onto the face of the fact that your statements are no more than technobabble in the way they are written.

 

Mind you I understand the general sentiment but you need to be factually correct first, this is technology, you can't make this stuff up as you go.

Most N64 games use the default microcode, but a couple of developers convinced Nintendo to allow them access to the custom microcode, to make their own changes to it.  The two external developers who are best known for custom microcode are Boss Games (as seen in World Driver Championship and Stunt Racer 64) and Factor 5, as seen in Indiana Jones and the Infernal Machine and Star Wars: Battle of Naboo.  I'm not sure if Rogue Squadron had custom microcode, probably phillipj was thinking of Battle for Naboo would be my guess.  Battle for Naboo is a pretty amazing looking game and one of the best-looking games of the generation.

[CPUWIZ]

3 minutes ago, A Black Falcon said:

Most N64 games use the default microcode, but a couple of developers convinced Nintendo to allow them access to the custom microcode, to make their own changes to it.  The two external developers who are best known for custom microcode are Boss Games (as seen in World Driver Championship and Stunt Racer 64) and Factor 5, as seen in Indiana Jones and the Infernal Machine and Star Wars: Battle of Naboo.  I'm not sure if Rogue Squadron had custom microcode, probably phillipj was thinking of Battle for Naboo would be my guess.  Battle for Naboo is a pretty amazing looking game and one of the best-looking games of the generation.

 

LOL

[VladR]

On ‎8‎/‎29‎/‎2019 at 1:44 AM, InMain said:

What if.. he was right tho.

I just checked N64 Rouge Squadron on YT. 

 

I actually programmed exact same kind of terrain engine, back in the day. It was on ~300 MHz Athlon with Geforce2 card.

 

The gameplay itself is surely simple enough for jag to handle. But the visual quality of terrain would have to be reduced way too drastically to hop over 10 fps mark. Of course, no textures on terrain. Even flatshaded it couldn't run at 10 fps at that view distance and terrain vertex grid spacing. If we reduced the terrain detail by a factor of 16x (4x less vertices in each dimension), then perhaps, 10 fps...

 

I have tried some high poly terrain on jag. I know exactly how much it can handle at 60, 30, 20, 15, 12 and 10 fps. Hint:It's nowhere near density of Rouge Squadron. There's a hard limit on how many scanlines can Blitter execute per frame, even if they are just 1 px short. There's easily over 20,000 scanlines in such terrain. Just go and do a GPU loop that counts to 20,000, let alone handles edge tracing 20,000x . For best multitasking, go make coffee in the meantime  

 

And remember , 10 fps is just 6 frames worth of each processor time. Not much given the details we are talking about here.

 

Object Processor is completely useless here. Blitter too, given how short the spans would be (just couple pixels, not even worth spinning Blitter up, forget about 64-bit blitting POWAH here, totally useless at those short blits). 

 

The only advantage Blitter gives us here in this scenario is Z-Buffer, but I seriously doubt that its internal precision is so high that it would handle such complex 3D scene without any of the typical nasty Z-buffer visual glitches (even if we ran it at one fps). Hell, those Z-Buffer glitches were pretty nasty even on Geforce and it took a lot of tweaking to reduce them to acceptable levels. 

And GeForce2 is a serious rendering HW compared to Blitter...

 

 

 

Now, if the question was "Is Jaguar capable of rendering one frame of such 3D scene" ? Then the answer is : "Yes. Eventually "

[agradeneu]

3 hours ago, CPUWIZ said:

The Jaguar is capable of making people hate each other, over some solder and plastic, more than any other system.  Hmm, maybe NeoGeo, who knows?

Haters gonna hate, it's just a matter of projection.

[philipj]

3 hours ago, A Black Falcon said:

Most N64 games use the default microcode, but a couple of developers convinced Nintendo to allow them access to the custom microcode, to make their own changes to it.  The two external developers who are best known for custom microcode are Boss Games (as seen in World Driver Championship and Stunt Racer 64) and Factor 5, as seen in Indiana Jones and the Infernal Machine and Star Wars: Battle of Naboo.  I'm not sure if Rogue Squadron had custom microcode, probably phillipj was thinking of Battle for Naboo would be my guess.  Battle for Naboo is a pretty amazing looking game and one of the best-looking games of the generation.

 

Yes I was... Hi res butter smooth 3D graphics; Battle for Naboo looked incredible... Rouge Squadron was a great looking game as well; I had that game and played it often. Another game "Phantom Menace" was a pretty good looking game that shine with the graphics. Great game at the arcade and the N64. Being that both the Jaguar and N64 are two very different systems the Jag using general purpose JRISC instead of the tried and true MIPS the way the PS1 and N64 did, butter smooth 3D graphics isn't possible in the same way for the console to come after it except for when it came to the 2.5D stuff (Doom, Phase Zero, Super Burnout i.e. OP). Not even "Open GL" would work on the Jaguar because it was most likely designed to work on the MIPS even though it was created around the same time the Jaguar was. It became my conclusion 2.5D or fake 3D would be something worth looking into and with the Jag having so many processors there was really no reason why making a smarter 2.5D engine wasn't couldn't be done. I still believe that today as I did in yester-years.

[Lost Dragon]

Good lord

 

We are now entering the stage of how the Jaguar would of handled polygon 3D titles seen on the N64?.

 

It had enough issues running it's own polygon 3D titles.:-)) 

 

 There is a danger  this thread could head  back into ye oldie era of RetroGamer Magazine Jaguar coverage, by suggesting only the Jaguar allowed development teams to code to the metal in order to get best performance from the hardware, where as likes of Sony insisted everyone used their 3D libraries. 

 

Talk to the likes of NEON who did Tunnel B1 and Viper (Aka Hornet)  on PlayStation for starters.

 

They developed and used their own 3D libraries as they apparently  were not keen on using the ones Sony provided.

[philipj]

Well... back in the day, for me, no idea was too big, too small, or too crazy for a very deeply flawed Jaguar system. Like I mentioned earlier I'm a little more sober minded today as I was in yester-years. Not the hardcore mindset I use to have that someone might consider to be idiocy; I didn't know how to program so I had absolutely nothing to loose therefore all bets were off. But that was back in the day, I'm a bit more realistic today about all of that kind of stuff now then I was in my youth... I think my only regret is not being able to prove my theories; I still have my old ideas written down from those old days so I'm glad I still have them... I get to see where my head was when I came up with them versus what's more feasible. But if people don't always get what I post, I perfectly understand and appreciate the feedback and corrections, I wouldn't have gotten this far without them.

[Lost Dragon]

 

During the time i had my Jaguar, i wanted both Freelancer 2120 and Legions Of the Undead so badly.

 

 

Both my type of games, both had promise of being flagship Jaguar titles.

 

 

Both failed to appear on Jaguar, PC or Playstation. 

 

The former never going further than a demo created for purposes of something to show the press, the latter seemingly similar, running on PC but badly. 

 

My point is basically that the industry gets your hopes up, whets your appetite with often fake screens and fake annoucements. 

 

 

All it does is make you a little more cautious and appreciate the flagship games you did get, that little bit more.

 

 

But at times with the Jaguar community and even with games like Zoop ? it's headlong into the fantasy zone and without the welcome to the...and get ready!  that goes with the Sega classic. 

 

Can  we not  just appreciate the work Minter, Eclipse, Rebellion, ATD, I.D, HVS etc did then and Songbird/Cyrano Jones/Piko Interactive/A.A do now, that has meant the Jaguar has the devoted following it has now?. 

[VladR]

2 hours ago, philipj said:

Well... back in the day, for me, no idea was too big, too small, or too crazy for a very deeply flawed Jaguar system. Like I mentioned earlier I'm a little more sober minded today as I was in yester-years. Not the hardcore mindset I use to have that someone might consider to be idiocy; I didn't know how to program so I had absolutely nothing to loose therefore all bets were off. But that was back in the day, I'm a bit more realistic today about all of that kind of stuff now then I was in my youth... I think my only regret is not being able to prove my theories; I still have my old ideas written down from those old days so I'm glad I still have them... I get to see where my head was when I came up with them versus what's more feasible. But if people don't always get what I post, I perfectly understand and appreciate the feedback and corrections, I wouldn't have gotten this far without them.

The only way you will prove the theories is by implementing them.

 

And you can't really do that using some high level language that produces a horrendous ASM code. Let alone if the produced code is for 68000.

 

You need to get down to the metal - Risc ASM - and as rewarding as it is, it's a tedious and painful process. 

[Clint Thompson]

5 hours ago, VladR said:

The only way you will prove the theories is by implementing them.

 

And you can't really do that using some high level language that produces a horrendous ASM code. Let alone if the produced code is for 68000.

 

You need to get down to the metal - Risc ASM - and as rewarding as it is, it's a tedious and painful process. 

The ability to do something in assembly doesn't translate to something great or anything at all in most cases since it apparently takes a decade of learning just to implement and another decade to produce anything worthwhile from it. Most people don't have the spare time or willingness to dive that deeply just to try and get theoretical results from a 25-year old console but more power to you or anyone else that does I guess. I've only been around 25-years, what's another 25. (I mean, I'll be dead before then but you know, whatever lol)

 

@philipj - Ideas mean nothing if not acted upon and you can't prove anything if you don't act. You're the only one stopping you from creating something cool at this point. Even if it sucks at least you tried. You've either got to shit and get off pot or die trying ? I hear there's this cool software package that lets you create you're own games on the Jaguar that doesn't require 10+ years of experience, but I've probably mentioned it x times before:

 

https://github.com/ggnkua/bcx-basic-Jaguar

 

                                 

 

Go on, embrace realism. I dare you...

 

[philipj]

7 hours ago, VladR said:

The only way you will prove the theories is by implementing them.

 

And you can't really do that using some high level language that produces a horrendous ASM code. Let alone if the produced code is for 68000.

 

You need to get down to the metal - Risc ASM - and as rewarding as it is, it's a tedious and painful process. 

Well that's why I'm starting off simple with QB64 so I can wrap my head around programming in general... It makes for a great prototyping tool, it's compatible with Windows, Apple, and Android thus becoming a gateway for some indie stuff, and whatever I make with it I can port down to the Jaguar; I'll get to the Motorola 68000 eventually. Although I mentioned QB64 earlier, I also have a copy of "GFA-BASIC" for Windows and an "Atari ST 520" to sort of help me alone (also to control some old synthesizers I own); would be great for experimenting with some things in the 3D department, but before I get to that, I will have already done so in QB64. It's basically a process of gaining understanding, I am more of an artist after all than I am a programmer although I did do some HTML once.

 

Here's a link of some books I ran into on programming 3D on the Motorola... I know the information is old, but still some good today for what I need.

 

Real Time 3D GFX for Atari ST

 

Amiga 3D Graphic Programming in (GFA) BASIC

 

4 minutes ago, Clint Thompson said:

The ability to do something in assembly doesn't translate to something great or anything at all in most cases since it apparently takes a decade of learning just to implement and another decade to produce anything worthwhile from it. Most people don't have the spare time or willingness to dive that deeply just to try and get theoretical results from a 25-year old console but more power to you or anyone else that does I guess. I've only been around 25-years, what's another 25. (I mean, I'll be dead before then but you know, whatever lol)

 

 

@philipj - Ideas mean nothing if not acted upon and you can't prove anything if you don't act. You're the only one stopping you from creating something cool at this point. Even if it sucks at least you tried. You've either got to shit and get off pot or die trying ? I hear there's this cool software package that lets you create you're own games on the Jaguar that doesn't require 10+ years of experience, but I've probably mentioned it x times before:

It'll be about 19 or 20 for me with most of those years taking a backseat to programming in favor of real life, but planning for the of future making games just in case. I'm at the point now I just want to know how to program and make games now Jag or no Jag, but no need in letting all of that knowledge of the system go to waste. I was a hair away from letting go till I sold my collection and regretted doing so afterwards. I'm in it for the long haul this time; I still have the "Gorf Lessons", but we both know how he felt about the 68000 (Turn it off)... Since he isn't around any more I got to go a different route in regards to learning Assembly Language. I probably will use those tools you link me to as a starting point for Jaguar programming once I have a good footing on programming the Motorola then use the Reboot tools as a kick-start. Some ground work is already in motion believe it or not; just have to make time for stuff and stay committed. Also I've been eyeballing the "Atari 5200" lately, but that's for another topic.

[WAVE 1 GAMES]

46 minutes ago, Clint Thompson said:

The ability to do something in assembly doesn't translate to something great or anything at all in most cases since it apparently takes a decade of learning just to implement...........

 

[Lost Dragon]

Factor 5 developed proprietary speech
compression and sound tools for the N64 and because of these tools, Nintendo granted them access to highly proprietary Microcode information and development environments...

 

That's the version of events that's been portrayed in my experience. 

 

Factor 5 'earnt' the access rather than just convinced Nintendo.

 

 

Rare were also allowed to use their own custom code on N64:

 

Chris Sutherland of Rare talking  of the headaches the N64 memory handicaps gave in an interview discussing the original Banjo-Kazooie;
"From a software perspective, we pushed the memory of the system very hard. As you move the camera around the map in Banjo Kazooie, the machine is constantly throwing out of memory things you can't see and pulling in the scenery that appears into view. This gave us major memory fragmentation issues. We used a proprietary system that "reshuffled" memory continuously as you played to eliminate the fragmentation. I'd doubt many N64 games of the time did anything like that—overall it meant we could dedicate a higher number of polygons to the characters and backgrounds than many other games at the time managed."

Edited September 1, 2019 by Lost Dragon

[Lost Dragon]

Whilst not a coder,  i do enjoy reading industry figures talking of their experiences. 

 

Glenn Corpes of then Bullfrog had some sound advice:

 

 

"This is _the_ big problem, everyone wants to start out by beating Quake 
(or at least writing a Quake clone), what they are forgetting is that Quake took two years to write plus the experience of having written  Doom, which was built on Wolfenstien and so on.

 

 

I'd say it's  almost  impossible for one person to start from nothing and produce a quake standard engine, even if they are a genius with access to every piece of info on how to do perspective texture mapping, BSPs etc... I'd go further and say that the one guy in a million who can do the above still has a one in a million chance of being able to go on to turn it into a game.

 

 

 

My advice  is to write something 2D ,rewrite the game logic a couple of 
times, have some fun, who knows you may even come up with an original 
twist, think about the 3D stuff later.

 

Edited September 1, 2019 by Lost Dragon

[Lost Dragon]

Factor 5 talking about Custom Sound Drivers they and Rare developed for the N64:

 

IGN64: Before this interview, you let me hear some music demos on the N64 that surpassed even the quality of Rare's music. What's your secret?
Factor 5:

Completely new, homemade drivers from top to bottom. In that respect, we owe Nintendo a big thank you for their cooperation. It's not like Nintendo isn't aware of some of the sound questions. They know that sound can count for a lot in a game. And I'm sure Nintendo isn't all that pleased with the sound that comes from many of the third-party games, either. We probably did the same thing that Rare did and wrote new sound drivers from the ground up. We're using our own sample format, our own drivers -- from the first to the last line of code we're basically doing everything different from the way SGI had originally planned it. And then you can do a lot with the sound.

[Punisher5.0]

On 8/30/2019 at 10:37 PM, A Black Falcon said:

Most N64 games use the default microcode, but a couple of developers convinced Nintendo to allow them access to the custom microcode, to make their own changes to it.  The two external developers who are best known for custom microcode are Boss Games (as seen in World Driver Championship and Stunt Racer 64) and Factor 5, as seen in Indiana Jones and the Infernal Machine and Star Wars: Battle of Naboo.  I'm not sure if Rogue Squadron had custom microcode, probably phillipj was thinking of Battle for Naboo would be my guess.  Battle for Naboo is a pretty amazing looking game and one of the best-looking games of the generation.

I dont know about the microcode but Stunt Race 64 was probably the best looking N64 game. It looks close to Dreamcast level. Actually now that I think about it, it looks better than Speed Devils (SD is a lot better game though).

[VladR]

You are being ridiculous now.

 

The effort to create some base running 3D engine on jag is measured in days or weeks. Not months or years.

 

My first RoadRash prototype with textured buildings and road in C took about 4 days. Took me another 2 days to implement curves.

 

When I switched to GPU, the initial experimenting took couple days then within a weekend I had a 3D terrain with runtime Level of details.

 

The decision to not release any of those builds is a different matter.

 

11 hours ago, Clint Thompson said:

The ability to do something in assembly doesn't translate to something great or anything at all in most cases since it apparently takes a decade of learning just to implement and another decade to produce anything worthwhile from it.

 

@philipj

[Lost Dragon]

Well to be fair ValdR,  this being the Jaguar section of the forum, if anyone wants ridiculous,  it is in the right section

[VladR]

Now that I have a basic working 3D engine also on 4 MHz Lynx, with detailed breakdown of each pipeline stage ( in terms of performance :cycle count), I computed some rough average cycle count for a test Quake 3d scene.

 

And the 4MHz Lynx can handle the flatshaded Quake scene in a higher (slideshow) frame rate than Jaguar can handle that complex terrain from N64.

 

Now, I don't want to be accused of not offering any solution, so here it is:

 

Reduce the Framebuffer resolution on Jaguar from 320x200 by a factor of 16x to 80*50 (and just make OP scale it to full screen) and you could have a very playable frame rate even with texturing.

 

Good luck trying to hit anything in such giant pixel mess

[Greg2600]

I never understood the fixation on graphics and/or sound alone, as if that makes a game successful or not.  My favorite N64 racing game was  Rush 2: Extreme Racing USA, and the graphics were terrible!

[JagChris]

Beetle battle!

[philipj]

8 minutes ago, Greg2600 said:

I never understood the fixation on graphics and/or sound alone, as if that makes a game successful or not.  My favorite N64 racing game was  Rush 2: Extreme Racing USA, and the graphics were terrible!

 

I agree... One of my favorite games for the "Playstation 1" was "VS. by THQ". This game has some of the worst graphics I've seen on the system, however it's got one of the most unique fighting engine I've ever played. Compare it with "Fight For Life" for the Jaguar and I think a better fighting engine could've been implemented for FFL. I remember playing VS and wishing they'd put more work into the game, the fighting engine is very engaging once you get over the learning curb.

 

 

[WAVE 1 GAMES]

6 hours ago, VladR said:

Now that I have a basic working 3D engine also on 4 MHz Lynx, with detailed breakdown of each pipeline stage ( in terms of performance :cycle count), I computed some rough average cycle count for a test Quake 3d scene.

 

And the 4MHz Lynx can handle the flatshaded Quake scene in a higher (slideshow) frame rate than Jaguar can handle that complex terrain from N64.

 

Now, I don't want to be accused of not offering any solution, so here it is:

 

Reduce the Framebuffer resolution on Jaguar from 320x200 by a factor of 16x to 80*50 (and just make OP scale it to full screen) and you could have a very playable frame rate even with texturing.

 

Good luck trying to hit anything in such giant pixel mess

 

Doen't matter I'm about to put all of that to shame. And my game isn't even true 3D. It's fake 3D.

 

 

#TheAliensAreComing

[Clint Thompson]

8 hours ago, VladR said:

You are being ridiculous now.

 

The effort to create some base running 3D engine on jag is measured in days or weeks. Not months or years.

 

My first RoadRash prototype with textured buildings and road in C took about 4 days. Took me another 2 days to implement curves.

 

When I switched to GPU, the initial experimenting took couple days then within a weekend I had a 3D terrain with runtime Level of details.

 

The decision to not release any of those builds is a different matter.

 

We must be talking about different things altogether then, like different people in a completely different reality. You were responding to Philip suggesting him to jump on the "RISC ASM" bandwagon to implement his theories as though it's a routine walk in the park yet remark how tedious and painful it is. So which is it, a 4-day stroll down rainbow road to 3D glory or a tortuous and difficult offset in life after a decade of experimenting and finally getting the grips of the hardware only to still not have actually released anything...