correction by user jimmy: 1. the ps1 used the compact disc which held 700 megabytes of data. 2. n64 cartridges actually held 64 megabytes of data note by me: yes this is true and i forgot to mention the exact amount! but actually is more of a "could hold" rather than actually doing so. ps1 disks usually did get to the 700mb, but 64mb cards were super expensive to make, so only like 3 games or so actually did use 64mb cards. still thanks!
correction by user Bukachell: One important detail is vertex snapping for PS1 mainly if you're doing animations, since if I remember correctly the PS1 used decimals to calculate vertex positions or something but it wasn't able to do it as precisely as the N64 note by me: yeah pretty much, I forgot to mention that. vertex in the ps1 couldn't move sub pixel distances, so they either were in the center of a pixel, or they would snap to the center of the neighbor pixel. it was either that, or the vertex could only go to coordinates that were whole numbers. so like, a vertex was either in position 1, or position 2; and it couldn't be in position 1.5 per example. its one of those two, cant recall it from the top of my head RN
Just a few details: 1- The N64 antialiasing blurred the whole screen, meaning it didn't detect edges like MSAA (I think the closest way to approximate it would be a rough form of FXAA?) 2- It also used dithering like the PS1, but thanks to the aforementioned blurring it was less noticeable (but still definitely noticeable, so if you want to recreate the look you should definitely include it) 3- Many N64 games actually ran at a lower 256x224 but 320x240 is still a good compromise. 4- The texture filtering applied N64 is called "3-point filtering" and doesn't look anything like the default bilinear applied on PC (it has to be recreated with a shader) 5- The N64 could selectively leave some layers or textures unfiltered, which is useful so HUDs don't look blurry. 6- Not sure about this but wasn't the PS1 able to render more textured, lit polygons than N64? But still games couldn't take advantage of it because they had to subdivide large polygons to prevent textures from going crazy? (from lack of perspective correction). Anyway great video. Good stuff on the use of vertex texturing/shading I didn't know about.
the console manufacturers always state max polygons per second. if you want to get the polygons per frame you divide the number by fps, obviously. like by 30 for an adventure game or 60 for a racing game. this is what a scene would look like back than realistically.
Another thing about Crash Bandicoot's graphics are that the actual model for Crash is all vertex Shading, This allows him to be animated more fluidly with a higher polygon count by using the PS1's untextured polygon mode.
Just a couple of corrections: 1- The N64 also does use dithering when outputting images to the screen. It has a higher color depth (21-bits I believe) than PS1's (15-bits). 2- The N64's largest texture can be 90 x 90, but it only allows a very low depth as a tradeoff. I don't remember all the details (you can look up at the N64 Homebrew scene if you want to learn in depth), but this is how I memorized for my 3D artworks and projects: - 32x32 = Full color depth, unnecessary because it devours texture cache space - 48x48 (or 32x64/64x32) = High enough color depth, for detailing - 64x64 = Low color depth, most common depth (see next paragraph) - 90x90 = Very low color depth Color depth and resolution aren't necessarily locked with each other, but rather neither can be exceed if you use any of the listed samples I mentioned. ALSO the N64 also supports greyscale textures, which occupy half the space (if I'm not wrong) for the same amount of colors if you were to include hue on them. That's where you'll want to use vertex colors even more (Link's clothing in both OoT & MM is the most omnipresent example I can think of). PS: TOOLS FOR TEXTURES: Both GraphicsGale and Paint.net can lower the color depth of images. Pro Motion NG allows you to edit a color palette (up to 256 colors, you can also manually allow a max number of colors). But I only use them for doing the fine-tuning I just described on each program - I do the main job on Krita (I know, not the best option, but it's good enough and free).
One important detail is vertex snapping for PS1 mainly if you're doing animations, since if I remember correctly the PS1 used decimals to calculate vertex positions or something but it wasn't able to do it as precisely as the N64. Most N64 games used 32 x 32 textures instead of the maximum 64 x 64 resolution
yeah pretty much, I forgot to mention that. vertex in the ps1 couldn't move sub pixel distances, so they either were in the center of a pixel, or they would snap to the center of the neighbor pixel. it was either that, or the vertex could only go to coordinates that were whole numbers. so like, a vertex was either in position 1, or position 2; and it couldn't be in position 1.5 per example. its one of those two, cant recall it from the top of my head RN.
Actually technically the max possible resolution on the n64 is 64x128 if you go on n64squid you can see some info on all the supported texture formats and resolutions
n64 used floating point math which in basic terms is a mantissa (the significant digits) and an exponent (the position of the decimal point) this system could store fairly precise numbers and could also store really big/really small numbers but would lose precision the bigger/smaller they go the ps1 used a much simpler fixed point integer system where a 32bit integer is split in half with 16bits for the whole number and 16bits for the fraction this is much less precise than floating points but is also much cheaper on processing power so it was used a lot back in the day for video games
If you don't have an anti-pop for your mic you can use a highpass filter to cut those Ps from your recording. Or a compressor acting on the low frequencies can help:) nice video, very helpful information, thank you
this is a sincerely awesome video man, so insightful and easy to understand. I'm a total beginner and I had no trouble following along which I really appreciate! thank you so much for this TT__TT I never would've figured it out on my own ! (also your bunny character is so adorable ! and the chalkboard character that had the table of content looked like you plucked It right out of a real game.)
one of the reasons it's hard to get a solid figure on polygon count limits for the consoles is because fill rate and overdraw were generally more limiting than raw triangle count
yeh, the numbers I gave were like, mostly scene averages, like, how many triangles would generally be on any given scene per console, if im not mistaken the number used for the 64 was like, the average of all of the mario 64 areas including npcs, and a couple areas from the zelda games (plus maybe something else but I cant remember) but I couldn't get as good averages when it came to the play station, cause like, stuff varied a whole lot more in that console. cause like, yeah because of stuff with the rendering pipeline and whatnot, a scene with 100 triangles of the first type and another scene with the same ammount of triangles of the same sort could easily have very different prefformance. the actual practical architecture and modeling techniques used to count how many polygons are actually optimal for any given scene scapes me, cause I know virtually nothing about programming, and I sorta gotta go off of what I can see surface level as an artist.
Thank you for this detailed video, the amount of work to portray this information in this style must have been a lot, although it was well worth it as it was very entertaining.
Detailed video my ass. This is mostly regurgitated info from Wikipedia and some other retro game channels with little to no programming background for those systems. It doesn't even talk about recent Mario 64 modding achievements with Kaze Emanuar's work. His mod can handle 250k textured polygons per second on real hardware.
Colour depth differed between the two consoles, with PS1 having a lower depth than N64. Hence the dithering, which, thanks to NTSC artifacting in CRT displays, could result in a higher perceived colour depth and overall smoothness in presentation. PS1’s instability was caused by not really handling floating point numbers, like at all? You can actually see similar results in modern game engines by looking at objects rendered at great distances from the local origin of the scene (the middle of the game’s current map space, or coordinates 0,0,0). If you want an easy example of this, load up the original 2011 release of Skyrim and head to the southeast corner of the city of Whiterun, on the road between the farmsteads and Honningbrew meadery, then look to the west at the mountains on the edge of the world space. If you slowly pan the camera from side to side, you should see the snow caps on the mountain where it meets with the rocks there “warble” a little bit. This is because that location is incredibly far from the local origin of your current cell, so when the engine tries to convert the polygons’ XYZ coordinates to a vector (a value between 0 and 1), each frame gets vastly different and inaccurate approximations due to decimal rounding errors.
Can I just say I think the PlayStation one aged way better than the Nintendo 64 just because of how it draws textures it doesn’t it uses linear filtering and I think that single fact just makes everything look better than on the Nintendo 64 because it’s moved everything out making everything look bad
You would need to use arithmetic and do the wobbly calculations of vertexes and vertices in the 3D environment. It is not easy to pull off I think, but if someone manages to find the code for that exact calculation, it might be possible.
I think you forgot about affine texture mapping on the PS1 which is essential. Also I think you can just take 360 000 flat polygons per second as a reference and then count flat shaded poly's as 1, textured as 2 and gouraud shaded as 4 and add it all together. Other than that it's a good video, even though I knew most of it already I kinda forgot about some.
absolutely did forget that, and as you mention affine texture filtering, i just noticed that i also forgot to mention mip mapping in the n64 textures, as well as as that they specifically used 3 point texture filtering, rather than like, normal filtering. lmao, well anyways, could you paste your comment into the pinned comment of the video? thats where i expected corrections to go, well anyways, thanks for the correction!
As far as I'm aware, there were also games on the PS1 that used Mip Mapping, but I don't know if the PS1 natively supported it, or if some developers implemented it into their games themselves.
@@Definitely_a_Fox wait rlly? Do you have any examples? I have never seen a ps1 game with any kind of texture filtering, and im really curious to see that
oh did i misspell the subtitles? or was it a problem in the way i was speaking? i was kinda rushed on the voiceover and captioning haha and thanks a lot :) worked hard on it
ok so, basically, i make the model in high poly using primitives, and then I re-mesh it in very very low poly as one solid piece. ua-cam.com/video/CuQzPDs99yM/v-deo.html here is a good tutorial on that. the poly count is more or less the point in between the amount of ps1 models and n64 models then my texturing uses very standart uv mapping, mostly in the n64 style, with most of my textures being 16*34 in resolution. hope that helps
actually it has both if im not mistaken. the n64 uses an old equivalent of the thing that most modern games have on their options labeled as "quickAA" (or just as the "anti aliasing" node in blender) which is like, not *true* anti aliasing in the modern sense of the word, its more of a filter than anything else, but it is a lot more complicated than a simple blur. but the n64 does also have a flat horizontal blur, if I'm not mistaken it is a blur of half a pixel? as in like, the image is rendered in 320p horizontally (as an example) then it is scaled linearly to 640p, and then those pixels are squished so that they occupy the same space as the original 320p. so the result is like, every even pixel horizontally is the original color, and every odd pixel is the average color of both the pixels to its sides. or at least thats how I think it works? I'm not 100% sure. the n64 probably did things slightly differently internally, but that's what would be the closest recreation that can easily be done with modern art tools for non programmers. but correct me if im wrong.
hey great tutorial! please consider a 100hrtz hipass filter on your voice (an EQ that cuts off low end) to tame some of those plosives POPs. a pop filter can help some
sip, pero son bastante molestos de usar, al final me termina saliendo mejor grabar mi voz, aunque tome algo de tiempo y sea algo dificil de entender y gracias, estoy feliz que te guste :) subire mas videos similares eventualmente.
The ps1 polycount is wild... Becouse it could output actualy a lot more than you would guess. But they were such shitty poligons that every actual 3d game on ps1 used some form of tesselation so the textures dint became a mess. Funny is that there were some isometric ps1 games that becouse of the lack of perspective had amazing detailed scenes, wild arms 2 comes to mind
3:53 Useless fact: Sega Saturn cannot render tris, it renders quads instead since Sega considered better at the time plus Saturn was mainly made for 2D games and the whole 3D high-res 60fps thing was thrown in after they found out that Sony is making the PlayStation.
is there a name for this technique? I would love a more indept tutorial on how this approach works and how to apply it to different models and model shapes!
Please take this video with a massive grain of salt, as it mostly regurgitates info from Wikipedia and other retro game UA-cam channels that they themselves mostly regurgitate it from Wikipedia. The whole N64 texture size thing is mostly a myth. Proof is several N64 games that have motion blur framebuffer effects. The way they do it is that they load chunks of the two framebuffers into the texture cache and blend them there before storing the result into the framebuffer. It seems like the N64 can actually handle any size up to 1024 pixels horizontally or vertically if it fits in the texture cache. I believe you can actually get up to 1024x1024 textures on the N64 by loading them 4K chunk by 4K chunk. However you also want to subdivide it into multiple long horizontal rectangles for each texture load for rendering efficiency. If you want accurate N64 style graphics, I'd suggest you Saauren and Kaze Emanuar's videos as resources for documentation.
yeah honestly, I would have adored to touch on the dreamcast, but like, by itself my research for this video was pretty... not amazing as it is. and the dreamcast just sorta worked completely differently, it was a very very unique system, the way it managed lighting and materials and texturing, something completely different from anything before or after it, not quite a modern (ish) material based rendering like that of the gamecube or the ps2, but also way more sophisticated than the stuff on the ps1 or the n64, far beyond the stuff I understand how to use. not to mention how inconsistent (tho rather high) the polygon counts were on it.
efectivamente camarada :3 hablo español, y mi hacento se notaba bastante en este video, ya no es tan pronunciado como entonces, pero ciertamente aun lo tengo.
N64 can render 8000 tris with texture filter, dynamic lights and reflex effect. Ps1 about 3500 without the effects. The N64 had some cars with 50 mb too. Look conker bad fur day, is really beautiful. The PS1 is better in storage, and only this. Oh yeah, and 64 can handle more resolution too. You can compare with Rayman 2 and hot wheels for example. The flat colors in PS1 is very cool, but it's just happing because the console don't running the game with much effects or resolution. Both consoles is very different, both has a beautiful styles but you forgotten some details and miss others
i was talking about the average game for the most part, not really the hardware pushing outliers like conker's or pokemon stadium 2 I do plan to go over specifics on console per console basis in future videos tho
Not everyone is a native speaker, it's quite a common mistake for a foreigner. I myself still pronounce lot of words in odd ways even though I've studied English for well over 10 years now.
PS1/PSX/N64 didn't have ambient occlusion, they used pre-baked lighting using lightmap grid, its very efficient way to fake lighting, but it does increase file sizes so it was very low res
that seems to be an alternative to just hand painted vertex colors, but still the same idea at the core. the system I talked about in the video isnt real ambient oclussion either, so much as just a cheaper baked in form of it. if im not mistaken they were both used at the same time sometimes? i think? if im not mistaken that is how they did it in spyro the dragon right? I could be mistaken tho. but yeah, big F on my part at forgetting to mention that.
corrections thread:
pls comment here any corrections you wanna make
or things you think i should have added into the video!
correction by user jimmy:
1. the ps1 used the compact disc which held 700 megabytes of data.
2. n64 cartridges actually held 64 megabytes of data
note by me: yes this is true and i forgot to mention the exact amount!
but actually is more of a "could hold" rather than actually doing so. ps1 disks usually did get to the 700mb, but 64mb cards were super expensive to make, so only like 3 games or so actually did use 64mb cards. still thanks!
correction by user Bukachell:
One important detail is vertex snapping for PS1 mainly if you're doing animations, since if I remember correctly the PS1 used decimals to calculate vertex positions or something but it wasn't able to do it as precisely as the N64
note by me:
yeah pretty much, I forgot to mention that. vertex in the ps1 couldn't move sub pixel distances, so they either were in the center of a pixel, or they would snap to the center of the neighbor pixel. it was either that, or the vertex could only go to coordinates that were whole numbers. so like, a vertex was either in position 1, or position 2; and it couldn't be in position 1.5 per example.
its one of those two, cant recall it from the top of my head RN
Just a few details:
1- The N64 antialiasing blurred the whole screen, meaning it didn't detect edges like MSAA (I think the closest way to approximate it would be a rough form of FXAA?)
2- It also used dithering like the PS1, but thanks to the aforementioned blurring it was less noticeable (but still definitely noticeable, so if you want to recreate the look you should definitely include it)
3- Many N64 games actually ran at a lower 256x224 but 320x240 is still a good compromise.
4- The texture filtering applied N64 is called "3-point filtering" and doesn't look anything like the default bilinear applied on PC (it has to be recreated with a shader)
5- The N64 could selectively leave some layers or textures unfiltered, which is useful so HUDs don't look blurry.
6- Not sure about this but wasn't the PS1 able to render more textured, lit polygons than N64? But still games couldn't take advantage of it because they had to subdivide large polygons to prevent textures from going crazy? (from lack of perspective correction).
Anyway great video. Good stuff on the use of vertex texturing/shading I didn't know about.
the console manufacturers always state max polygons per second. if you want to get the polygons per frame you divide the number by fps, obviously. like by 30 for an adventure game or 60 for a racing game. this is what a scene would look like back than realistically.
Some N64 games utilized extra boundaries for some textures for added detail. With some being 64x128 or 128x64. Though most would be 64x64
I really really love how you chose to present this video, it's incredibly charming and fun to watch. Brilliant work
thanks!
RIP that one guy wanting to make the 3DO game
Huh?
I think he means 3DS but idk
3do was a system blueprint they was sold to brands to create a multimedia device that could play photo cds games movies and music from the company 3d0
@@spacedud7911nuh uh, 3DO is a real console
@@spacedud79113DO is a different gaming console, it’s time to do a google search
Another thing about Crash Bandicoot's graphics are that the actual model for Crash is all vertex Shading, This allows him to be animated more fluidly with a higher polygon count by using the PS1's untextured polygon mode.
iirc he has a texture for the fur on his back but the rest is entirely vertex shaded
You know what the FAC thing is *vine boom*
*no fac* killed me
i like how you showed what the environment looks like at each step. Made it easy to see what did what.
Just a couple of corrections:
1- The N64 also does use dithering when outputting images to the screen. It has a higher color depth (21-bits I believe) than PS1's (15-bits).
2- The N64's largest texture can be 90 x 90, but it only allows a very low depth as a tradeoff. I don't remember all the details (you can look up at the N64 Homebrew scene if you want to learn in depth), but this is how I memorized for my 3D artworks and projects:
- 32x32 = Full color depth, unnecessary because it devours texture cache space
- 48x48 (or 32x64/64x32) = High enough color depth, for detailing
- 64x64 = Low color depth, most common depth (see next paragraph)
- 90x90 = Very low color depth
Color depth and resolution aren't necessarily locked with each other, but rather neither can be exceed if you use any of the listed samples I mentioned.
ALSO the N64 also supports greyscale textures, which occupy half the space (if I'm not wrong) for the same amount of colors if you were to include hue on them. That's where you'll want to use vertex colors even more (Link's clothing in both OoT & MM is the most omnipresent example I can think of).
PS: TOOLS FOR TEXTURES: Both GraphicsGale and Paint.net can lower the color depth of images. Pro Motion NG allows you to edit a color palette (up to 256 colors, you can also manually allow a max number of colors). But I only use them for doing the fine-tuning I just described on each program - I do the main job on Krita (I know, not the best option, but it's good enough and free).
One important detail is vertex snapping for PS1 mainly if you're doing animations, since if I remember correctly the PS1 used decimals to calculate vertex positions or something but it wasn't able to do it as precisely as the N64.
Most N64 games used 32 x 32 textures instead of the maximum 64 x 64 resolution
yeah pretty much, I forgot to mention that. vertex in the ps1 couldn't move sub pixel distances, so they either were in the center of a pixel, or they would snap to the center of the neighbor pixel. it was either that, or the vertex could only go to coordinates that were whole numbers. so like, a vertex was either in position 1, or position 2; and it couldn't be in position 1.5 per example.
its one of those two, cant recall it from the top of my head RN.
Actually technically the max possible resolution on the n64 is 64x128 if you go on n64squid you can see some info on all the supported texture formats and resolutions
n64 used floating point math which in basic terms is a mantissa (the significant digits) and an exponent (the position of the decimal point) this system could store fairly precise numbers and could also store really big/really small numbers but would lose precision the bigger/smaller they go
the ps1 used a much simpler fixed point integer system where a 32bit integer is split in half with 16bits for the whole number and 16bits for the fraction this is much less precise than floating points but is also much cheaper on processing power so it was used a lot back in the day for video games
Love the video! Your way of storytelling is great :)
If you don't have an anti-pop for your mic you can use a highpass filter to cut those Ps from your recording. Or a compressor acting on the low frequencies can help:) nice video, very helpful information, thank you
oki, next video i'll do that with the audio, didnt know thats what the high pass was for
I've recently become fascinated with early 3D technology. This video was so fun! ^^
this is a sincerely awesome video man, so insightful and easy to understand. I'm a total beginner and I had no trouble following along which I really appreciate! thank you so much for this TT__TT I never would've figured it out on my own ! (also your bunny character is so adorable ! and the chalkboard character that had the table of content looked like you plucked It right out of a real game.)
this is such a helpful and in-depth look at the specifics of low poly design. It helped me massively! Thank you!
The world needed this video
it's always interesting how people explain this
so now people can make ARGs and Creepypastas!
one of the reasons it's hard to get a solid figure on polygon count limits for the consoles is because fill rate and overdraw were generally more limiting than raw triangle count
yeh, the numbers I gave were like, mostly scene averages, like, how many triangles would generally be on any given scene per console, if im not mistaken the number used for the 64 was like, the average of all of the mario 64 areas including npcs, and a couple areas from the zelda games (plus maybe something else but I cant remember)
but I couldn't get as good averages when it came to the play station, cause like, stuff varied a whole lot more in that console. cause like, yeah because of stuff with the rendering pipeline and whatnot, a scene with 100 triangles of the first type and another scene with the same ammount of triangles of the same sort could easily have very different prefformance.
the actual practical architecture and modeling techniques used to count how many polygons are actually optimal for any given scene scapes me, cause I know virtually nothing about programming, and I sorta gotta go off of what I can see surface level as an artist.
Thank you for this detailed video, the amount of work to portray this information in this style must have been a lot, although it was well worth it as it was very entertaining.
thank u uwu
Detailed video my ass. This is mostly regurgitated info from Wikipedia and some other retro game channels with little to no programming background for those systems.
It doesn't even talk about recent Mario 64 modding achievements with Kaze Emanuar's work. His mod can handle 250k textured polygons per second on real hardware.
@@fungo6631 ok im waiting for your video now
@@owenelliott5742 Kaze Emanuar and Sauraen already made videos about the topic.
No cap, the female rabbit character in PS1 or N64 graphics looks absolutely adorable though.
thanks!, I put a lot of time into the model :)
"full fac"
sorry that had me rolling xD
I under stand how much work go's into this thank you very much for the work I support your work
DUUDE! thank you so much for all this information in one video
really pure gold here
Colour depth differed between the two consoles, with PS1 having a lower depth than N64. Hence the dithering, which, thanks to NTSC artifacting in CRT displays, could result in a higher perceived colour depth and overall smoothness in presentation.
PS1’s instability was caused by not really handling floating point numbers, like at all? You can actually see similar results in modern game engines by looking at objects rendered at great distances from the local origin of the scene (the middle of the game’s current map space, or coordinates 0,0,0).
If you want an easy example of this, load up the original 2011 release of Skyrim and head to the southeast corner of the city of Whiterun, on the road between the farmsteads and Honningbrew meadery, then look to the west at the mountains on the edge of the world space. If you slowly pan the camera from side to side, you should see the snow caps on the mountain where it meets with the rocks there “warble” a little bit.
This is because that location is incredibly far from the local origin of your current cell, so when the engine tries to convert the polygons’ XYZ coordinates to a vector (a value between 0 and 1), each frame gets vastly different and inaccurate approximations due to decimal rounding errors.
Now id love to see sum games made this way
Wow thank you for making this video!
Can I just say I think the PlayStation one aged way better than the Nintendo 64 just because of how it draws textures it doesn’t it uses linear filtering and I think that single fact just makes everything look better than on the Nintendo 64 because it’s moved everything out making everything look bad
eh I think its pretty nostalgic tbh, I kinda love the blurry look of the n64. but thats just my personal taste.
Naaah, the PS1 looks like 3D DOS games in software mode.
You would need to use arithmetic and do the wobbly calculations of vertexes and vertices in the 3D environment. It is not easy to pull off I think, but if someone manages to find the code for that exact calculation, it might be possible.
CDs have a 700mb capacity and DVDs 4.7GB with a few exceptions being bigger in some cases.
I think you forgot about affine texture mapping on the PS1 which is essential. Also I think you can just take 360 000 flat polygons per second as a reference and then count flat shaded poly's as 1, textured as 2 and gouraud shaded as 4 and add it all together.
Other than that it's a good video, even though I knew most of it already I kinda forgot about some.
absolutely did forget that, and as you mention affine texture filtering, i just noticed that i also forgot to mention mip mapping in the n64 textures, as well as as that they specifically used 3 point texture filtering, rather than like, normal filtering.
lmao, well anyways, could you paste your comment into the pinned comment of the video? thats where i expected corrections to go, well anyways, thanks for the correction!
As far as I'm aware, there were also games on the PS1 that used Mip Mapping, but I don't know if the PS1 natively supported it, or if some developers implemented it into their games themselves.
@@Definitely_a_Fox wait rlly?
Do you have any examples? I have never seen a ps1 game with any kind of texture filtering, and im really curious to see that
Ah, clarification.
The game in question is Spyro, and the devs absolutely came up with their own solution.
I don't know if any other games did it.
That's one way to make me feel old 👍
U can say was instead of whas, sir. Nice video i really enjoyed it and entertained!
oh did i misspell the subtitles?
or was it a problem in the way i was speaking?
i was kinda rushed on the voiceover and captioning haha
and thanks a lot :) worked hard on it
Thanks for your vid! This will help a lot w making any retro games
this video was very interesting
Hey man I love the video. By any chance how did you make you model the way it is? I love the style of it.
Like your process
ok so, basically, i make the model in high poly using primitives, and then I re-mesh it in very very low poly as one solid piece. ua-cam.com/video/CuQzPDs99yM/v-deo.html here is a good tutorial on that.
the poly count is more or less the point in between the amount of ps1 models and n64 models
then my texturing uses very standart uv mapping, mostly in the n64 style, with most of my textures being 16*34 in resolution.
hope that helps
9:37 good one dude :D
vey underrated !! thank you
thanks :)
it was really helpful, thanks a lot!
Me saying yes to everything then getting hit with the, "me neither!"
Haha yeah, man. Totally, we young people gotta stick together hah yes hahaha
one thing to note about N64AA, its not true anti aliusing, its more like a soft blur.
actually it has both if im not mistaken.
the n64 uses an old equivalent of the thing that most modern games have on their options labeled as "quickAA" (or just as the "anti aliasing" node in blender) which is like, not *true* anti aliasing in the modern sense of the word, its more of a filter than anything else, but it is a lot more complicated than a simple blur.
but the n64 does also have a flat horizontal blur, if I'm not mistaken it is a blur of half a pixel? as in like, the image is rendered in 320p horizontally (as an example) then it is scaled linearly to 640p, and then those pixels are squished so that they occupy the same space as the original 320p. so the result is like, every even pixel horizontally is the original color, and every odd pixel is the average color of both the pixels to its sides.
or at least thats how I think it works? I'm not 100% sure. the n64 probably did things slightly differently internally, but that's what would be the closest recreation that can easily be done with modern art tools for non programmers.
but correct me if im wrong.
Great video, thanks for sharing the knowledge!
hey great tutorial! please consider a 100hrtz hipass filter on your voice (an EQ that cuts off low end) to tame some of those plosives POPs.
a pop filter can help some
oh haha yeah, ill look into that, it feels like ive tried everything to get better audio, but it feels like not much works, so yeah i'll check on that
Yooo, subscribed.
Great and entertaining tutorial, thank you
Sigue subiendo videos amigo, me encantan, ¿no hay loquendos en inglés?
sip, pero son bastante molestos de usar, al final me termina saliendo mejor grabar mi voz, aunque tome algo de tiempo y sea algo dificil de entender
y gracias, estoy feliz que te guste :) subire mas videos similares eventualmente.
Here's before watching the full video to see if it was rendered obsolete by Kaze Emanuar's recent Mario 64 modding work.
The ps1 polycount is wild...
Becouse it could output actualy a lot more than you would guess.
But they were such shitty poligons that every actual 3d game on ps1 used some form of tesselation so the textures dint became a mess.
Funny is that there were some isometric ps1 games that becouse of the lack of perspective had amazing detailed scenes, wild arms 2 comes to mind
The N64 has better polygon processing than people generally assume. In fact, polygon count with later microcodes was the least of your concern.
3:53 Useless fact: Sega Saturn cannot render tris, it renders quads instead since Sega considered better at the time plus Saturn was mainly made for 2D games and the whole 3D high-res 60fps thing was thrown in after they found out that Sony is making the PlayStation.
would this work tutorial with something leaning more twoards the graphics of the original DS?
is there a name for this technique? I would love a more indept tutorial on how this approach works and how to apply it to different models and model shapes!
Thanks for this. Like for real, you helped me so much and im sure other feel the same.
Could you do something like this for like wii sports style art?
I luv the 90s!
3:50 ese es el modelo de sonic r, el de sonic jam esta mas pulido y limpio
You forgot to explain how to make the PS1 styled games look jittery
Heya is there any way I could commision you to make one of these for me? Name your price man it looks great!
(An avatar that is)
sure dude, you can writte to me on discord and we'll talk it over there, i've already sent you a frend request
Thanks bro
Please take this video with a massive grain of salt, as it mostly regurgitates info from Wikipedia and other retro game UA-cam channels that they themselves mostly regurgitate it from Wikipedia.
The whole N64 texture size thing is mostly a myth. Proof is several N64 games that have motion blur framebuffer effects.
The way they do it is that they load chunks of the two framebuffers into the texture cache and blend them there before storing the result into the framebuffer.
It seems like the N64 can actually handle any size up to 1024 pixels horizontally or vertically if it fits in the texture cache.
I believe you can actually get up to 1024x1024 textures on the N64 by loading them 4K chunk by 4K chunk. However you also want to subdivide it into multiple long horizontal rectangles for each texture load for rendering efficiency.
If you want accurate N64 style graphics, I'd suggest you Saauren and Kaze Emanuar's videos as resources for documentation.
NICE! thanks!
they dont mention dreamcast, because at that point we get to high quality models
yeah honestly, I would have adored to touch on the dreamcast, but like, by itself my research for this video was pretty... not amazing as it is. and the dreamcast just sorta worked completely differently, it was a very very unique system, the way it managed lighting and materials and texturing, something completely different from anything before or after it, not quite a modern (ish) material based rendering like that of the gamecube or the ps2, but also way more sophisticated than the stuff on the ps1 or the n64, far beyond the stuff I understand how to use. not to mention how inconsistent (tho rather high) the polygon counts were on it.
dude holy fuck thank you!
iny you linked to your dashboard
no shut up get out of here aaagggghhhh
0:13 same
It just lacked a subtitle in Portuguese but apart from that it's wonderful, nice vídeo.
sorry for that, even if i had tried, my Portuguese is quite bad, and using a translator would have given very bad results, but maybe eventually
@@HiimIny no problem
5:25 ... I'm interested..?
Like!
tscherektahs :D
Um abraço do Brasil 🇧🇷
y un abraso de colombia 🇨🇴, amor parceiro!
the way he says fac..(pls dont get mad)
How can I get in contact w you??
you can write to me to this discord "hi im iny#1463"
orgullo colombiano
I like 3K
I didnt live the 90s soo idk how it looks like ;-;
Tú acento me dice que hablas español
efectivamente camarada :3
hablo español, y mi hacento se notaba bastante en este video, ya no es tan pronunciado como entonces, pero ciertamente aun lo tengo.
"🅱️dyogam*
why do you say characters like that
Because they have an accent?? That’s obvious blud 😭🙏🏻
N64 can render 8000 tris with texture filter, dynamic lights and reflex effect.
Ps1 about 3500 without the effects. The N64 had some cars with 50 mb too. Look conker bad fur day, is really beautiful. The PS1 is better in storage, and only this. Oh yeah, and 64 can handle more resolution too. You can compare with Rayman 2 and hot wheels for example. The flat colors in PS1 is very cool, but it's just happing because the console don't running the game with much effects or resolution. Both consoles is very different, both has a beautiful styles but you forgotten some details and miss others
i was talking about the average game for the most part, not really the hardware pushing outliers like conker's or pokemon stadium 2
I do plan to go over specifics on console per console basis in future videos tho
🖤🥰
Is pronounced karacter
Why do you say "chair" icters, isnt it pronounced "k" aracters
I've heard it pronounced both ways I think, tho I could be wrong
but yeah i just pronounced as its written, I'll try to pronounce it right next time
Not everyone is a native speaker, it's quite a common mistake for a foreigner. I myself still pronounce lot of words in odd ways even though I've studied English for well over 10 years now.
PS1/PSX/N64 didn't have ambient occlusion, they used pre-baked lighting using lightmap grid, its very efficient way to fake lighting, but it does increase file sizes so it was very low res
that seems to be an alternative to just hand painted vertex colors, but still the same idea at the core.
the system I talked about in the video isnt real ambient oclussion either, so much as just a cheaper baked in form of it.
if im not mistaken they were both used at the same time sometimes? i think? if im not mistaken that is how they did it in spyro the dragon right? I could be mistaken tho. but yeah, big F on my part at forgetting to mention that.