Not to hate on the job done here. But I think most people today (or even back in 2013) could have gotten further with Blender from an artistic perspective.
Blender 2.4 was kind of terrible to use from an animation perspective, but the controls for simple objects / cycles rendering was never really that difficult IMO. The key was having the knowledge to stay on the "nice parts" of Blender.
Blender 2.6 and 3.0 these days have made great strides at being easier to use from an animation and rigging perspective.
------
The harsh shadow lines could be made easier if you used area-lamps (in Blender) for example. They probably exist in POV-ray, but a 3d GUI to move-and-place these objects around grossly simplifies the effort, compared to placing the objects around with a text-based format like POV-Ray.
> a 3d GUI to move-and-place these objects around grossly simplifies the effort, compared to placing the objects around with a text-based format like POV-Ray.
This can depend on what you're doing and what background you bring to the work. ie: Just recently, I've found myself more effective developing 3D print designs using a textual language than via a drag/drop UI. (But I bring a programming background, like to think in abstractions, and was making designs that were quite inorganic...)
I think POV-Ray was probably more impressive in 2000 than 2013, but even today I like the method of creating graphics via text rather than a GUI. That's probably because I'm a programmer rather than an artist, but it's why I still think fondly of POV-Ray even though it really has been surpassed in almost every way by Blender.
The text-based format is certainly superior for a programmer, who can write a script to output POV-Ray data.
Blender of course has a Python API, but the text for POV-Ray is almost certainly easier to think and use. After all, Python's "print" statement is just easier to use than going into Blender docs wondering what bpy.context.blend_data is exactly.
------
That's probably POV-Ray's biggest feature. It is truly a domain-specific language for describing 3d worlds and objects.
There were GUI's for POV-Ray too, Moray i used extensibly, and something for the Atari I forgotten what it was called, didn't use it much.
I started with my 8 MHz Atari, each test scene took at least 10 minutes for a stamp-size picture and then the sphere was in the wrong place and wasn't visible or the wrong texture. It took me soo long to get the hang of the coordinate system and the language. But fun it was!
I printed the documentation so I could read it while the computer was buzy rendering.
I also wrote a small program for my Atari Falcon that would turn of the graphics chip to give me more rendering speed. Since it was a shared bus it used up something like 25% CPU on full colour.
It's a different experience. POV-Ray is basically a programming language. If you want to generate geometry programmatically there's nothing quite like it. The "programming" features are primitive ("if" statements, loops, eventually they added macros that act like procedure calls...) but the way you describe geometry and textures is actually pretty nice.
POV-Ray is also very old. Turner Whitted's recursive ray tracing paper is from 1979. POV-Ray was released in 1991, but was based on an earlier raytracer called DKBTrace that was written sometime in the 80's. Blender is more modern and isn't built around the limitations of what was feasible to render in a reasonable amount of time on the computers of the 80's and 90's.
I was just telling my son about POV-Ray yesterday, since he's starting to learn to use Blender and I dabbled with POV-Ray as a teen. I'm quite impressed that it's still being actively developed 30 years later.
POV-Ray has a powerful set of solid geometry operations and is quite good for generating scene descriptions from another program and then invoking PV to render.
Obviously it's command line rather than a visual tool.
There was a nice GUI back then, developed Independent from the POVRAY renderer named "Moray".
More surprisingly, there is still a website online for the GUI:
Moray was my favorite 3D app back in the day, and then I fell out of 3D. Then recently I wanted to model a project I wanted to build out of glass, and tried to do so in Blender but couldn't figure out how to get refractive surfaces not to show polygons (I'm a Blender noob). Really made me miss working with Moray+POVray where the primitives were geometric shapes rather than polygons.
Moray is one of the first programs I paid for with my own money. I'm surprised that the site is still online.
The best use I ever made of it was modeling an addition to my paternal grandparents' barn before they built it. I entered measurements from tape measures and photographs to build up the basic shapes and the addition, before-and-after.
There was something addictive about seeing the final traced image build up line by line. I dissipated a lot of computer-hours in the 1990s with POV-Ray on a 486, then later a Pentium or Power Mac. I even showed my other grandfather how to run it on his FPU-less 68k Mac; now that was slow.
Oh my god, thanks for dislodging this long forgotten memory! I never would have remembered that I played with this program ages ago, you really made me reminisce :)
My computer graphics class used pov ray back in early 2000s. I remember I wrote a Perl script to generate my scenes to avoid typing things manually. I still may have the code somewhere.
I enjoy diving into POV-Ray these days. It's a fun way to get away from GUI-driven work I might do in other 3D software.
One cool thing about the experience of using POV is that you can scroll in one dimension and see the entire organization of elements that will become a sensory window into the imagination. If you can learn the syntax and options, you can effectively break down what would otherwise be many different sets of interfaces and gain an enhanced feeling of accessibility.
If you combine this with an appreciation for abstraction-focused scene modeling, you can reach this point of really amazing emotion where you realize you can create or model _anything_ with the provided tools, as long as you can embrace the need for abstraction as its own sort of user-mindset technology.
The closest analog that comes to mind outside of 3D would be MacPaint, MS Paint, Grafx2...if you've ever seen someone seriously plan their home garden or design a home theater in one of those apps, and then build the damn thing in real life, you may know what I'm talking about.
Outside of that...maybe something like watching a concept designer use a bic crystal pen to sketch out an entire world on newsprint.
With POV-Ray I find myself honing in on something like an intuitive feel for how much abstraction to employ to simply and effectively depict what I want to depict, given any time constraints. It's really cool to reach that point no matter the tool. I'm sure a lot of people feel similar feelings about their favorite programming tools as well.
I actually made it several years ago, probably around 2007, I only re-rendered it and uploaded it to YouTube recently. I always felt there was a rather painful skill ceiling in POV-Ray. Sure, people have done some really amazing things with it, but going from something fairly simple like this to something really impressive required a level of magic and time I could never get out of POV-Ray.
In the early-mid 90s, on the Canadian children's TV channel YTV, between shows they'd air early 3D animations under the name "Short Circutz". This is definitely giving me similar vibes to some of those - this [1] in particular. Thanks for setting off my nostalgia neurons :)
Unrelated but might trigger a few memories on here:
I managed to get a copy of povray on one of those freeware cd-roms that computer stores often sold for a few dollars (since the internet wasn't a thing yet and getting it from compuserve was tricky). I think I sunk years of my life rendering beginners geometry, all the usuals: snowmen (spheres!), tables of still life geometry, basic animal characters and lots of snow scenes or rippling water sunsets (povray was particularly good for these.)
That's probably where I got it from too, as I used it pre-Internet.
It's funny to think I had to leave my 286 on all night to get a single 320x200 render like those in OP's link. Now you can render scenes like that in realtime.
Lol, seems I have somehow actually been remembering the forerunner DKBTrace or something. Sometime around 1989 I had it on my Amiga from a Fred Fish PD software disk. Even on my buffed out A1000 (68020 14Mhz 4MB) it took a long time to make a picture. So since I was student working for the university computing services department... I shanghaied a whole user room of Sun 3/50s (it was closed for the night) and had like 20 workstations each doing a single slice of an image. Much faster. Then my account was disabled and I had to go talk to Security about misuse of computer resources. heh. Good times.
Hi got into ray tracing back in the early 1990s as a teenager. The community back then was very small and we all had ridiculously horrible computers. But it was still a lot of fun and I remember doing cool things such as adding a fog command to the ray tracer, which taught me a lot about programming.
For some reason, it’s just really captivating to build a scene using declarative code, rather than doing so visually in an editor like Blender.
On a side note, you can basically do the same thing in 2D with the SVG format, and you can even alter the properties programmatically when used within a webpage.
There's a lot for Blender, and 3d courses generally take you through the same iterative approach of simpler geometry that is gradually made more complex to teach higher concepts.
The amount of comments comparing POV-Ray to Blender is really puzzling to me. The comparison is apples to oranges.
In fact you could (and probably still can) use POV-Ray to render your Blender scene. Blender is not a renderer - it just ships with some integrated ones and integrations for various external ones.
As a joke I once wanted to post a ray traced image per day, except the images were actually real photographs of things composed and lit to look like it was a ray traced scene. I still wonder if anyone would have noticed, and how.
I work on robotic space missions and we have an old janky tool that does modelling of lunar insolation for upcoming lunar rover missions (for solar panel insolation, prediction of ice, etc). It was with an odd mix of horror and delight that I looked under the hood and found POV-Ray. I played with POV-ray as a teenager (30s now) and was amazed to see it still in use.
POV-Ray was so awesome when a friend from work and I were playing around with it from the CD-ROM source code included with the book. I think it was the texture generation capabilities that was the most fascinating parts for me. Rendering then as now was super slow, as is always the case because we want the best we can get on hardware of the day so we only raise expectations and pay in render times.
This is pretty off-topic but I've been trying, very unsuccessfully, to port POV-Ray to WASM as a very low-priority side project for a while now. Of course there are better, more web-native tools to build ray tracers with but I really like the POV-Ray SDL and would love to play around with it in the browser.
It might be misguided but the ffmpeg builds in WASM made me think it might be possible but I reached the end of Google/SO trying to get the Boost libs to link while running emconfigure with the existing codebase.
I'm totally winging it so if someone has any ideas that could set me on the right path it would be greatly appreciated.
40 comments
[ 3.4 ms ] story [ 92.5 ms ] threadBlender 2.4 was kind of terrible to use from an animation perspective, but the controls for simple objects / cycles rendering was never really that difficult IMO. The key was having the knowledge to stay on the "nice parts" of Blender.
Blender 2.6 and 3.0 these days have made great strides at being easier to use from an animation and rigging perspective.
------
The harsh shadow lines could be made easier if you used area-lamps (in Blender) for example. They probably exist in POV-ray, but a 3d GUI to move-and-place these objects around grossly simplifies the effort, compared to placing the objects around with a text-based format like POV-Ray.
They do: http://www.povray.org/documentation/view/3.6.2/38/
> a 3d GUI to move-and-place these objects around grossly simplifies the effort, compared to placing the objects around with a text-based format like POV-Ray.
This can depend on what you're doing and what background you bring to the work. ie: Just recently, I've found myself more effective developing 3D print designs using a textual language than via a drag/drop UI. (But I bring a programming background, like to think in abstractions, and was making designs that were quite inorganic...)
Blender of course has a Python API, but the text for POV-Ray is almost certainly easier to think and use. After all, Python's "print" statement is just easier to use than going into Blender docs wondering what bpy.context.blend_data is exactly.
------
That's probably POV-Ray's biggest feature. It is truly a domain-specific language for describing 3d worlds and objects.
I started with my 8 MHz Atari, each test scene took at least 10 minutes for a stamp-size picture and then the sphere was in the wrong place and wasn't visible or the wrong texture. It took me soo long to get the hang of the coordinate system and the language. But fun it was!
I printed the documentation so I could read it while the computer was buzy rendering.
I also wrote a small program for my Atari Falcon that would turn of the graphics chip to give me more rendering speed. Since it was a shared bus it used up something like 25% CPU on full colour.
Comparing POV-Ray to Cycles or Octane would make more sense.
POV-Ray is also very old. Turner Whitted's recursive ray tracing paper is from 1979. POV-Ray was released in 1991, but was based on an earlier raytracer called DKBTrace that was written sometime in the 80's. Blender is more modern and isn't built around the limitations of what was feasible to render in a reasonable amount of time on the computers of the 80's and 90's.
Looks like it's being modernized to build with c++11 compilers
Obviously it's command line rather than a visual tool.
http://www.stmuc.com/moray/
The best use I ever made of it was modeling an addition to my paternal grandparents' barn before they built it. I entered measurements from tape measures and photographs to build up the basic shapes and the addition, before-and-after.
There was something addictive about seeing the final traced image build up line by line. I dissipated a lot of computer-hours in the 1990s with POV-Ray on a 486, then later a Pentium or Power Mac. I even showed my other grandfather how to run it on his FPU-less 68k Mac; now that was slow.
One cool thing about the experience of using POV is that you can scroll in one dimension and see the entire organization of elements that will become a sensory window into the imagination. If you can learn the syntax and options, you can effectively break down what would otherwise be many different sets of interfaces and gain an enhanced feeling of accessibility.
If you combine this with an appreciation for abstraction-focused scene modeling, you can reach this point of really amazing emotion where you realize you can create or model _anything_ with the provided tools, as long as you can embrace the need for abstraction as its own sort of user-mindset technology.
The closest analog that comes to mind outside of 3D would be MacPaint, MS Paint, Grafx2...if you've ever seen someone seriously plan their home garden or design a home theater in one of those apps, and then build the damn thing in real life, you may know what I'm talking about.
Outside of that...maybe something like watching a concept designer use a bic crystal pen to sketch out an entire world on newsprint.
With POV-Ray I find myself honing in on something like an intuitive feel for how much abstraction to employ to simply and effectively depict what I want to depict, given any time constraints. It's really cool to reach that point no matter the tool. I'm sure a lot of people feel similar feelings about their favorite programming tools as well.
i like using code instead of a gui for some 3d stuff
I actually made it several years ago, probably around 2007, I only re-rendered it and uploaded it to YouTube recently. I always felt there was a rather painful skill ceiling in POV-Ray. Sure, people have done some really amazing things with it, but going from something fairly simple like this to something really impressive required a level of magic and time I could never get out of POV-Ray.
[1] https://www.youtube.com/watch?v=ScvIbcNV38k
Alongside using Enlightenment as their window manager.
It was always cool to check their desktops.
It's funny to think I had to leave my 286 on all night to get a single 320x200 render like those in OP's link. Now you can render scenes like that in realtime.
For some reason, it’s just really captivating to build a scene using declarative code, rather than doing so visually in an editor like Blender.
The free stuff is here: https://www.blender.org/support/tutorials/
After growing past that, the paid stuff is inexpensive and well made: it: https://studio.blender.org/welcome/ ($10 a month.)
In fact you could (and probably still can) use POV-Ray to render your Blender scene. Blender is not a renderer - it just ships with some integrated ones and integrations for various external ones.
I had no idea, back then, just how much impact the process of thinking in CSG would have on my nascent mind.
A truly great piece of software for a wain cutting their teeth with math and programmatic construction.
It might be misguided but the ffmpeg builds in WASM made me think it might be possible but I reached the end of Google/SO trying to get the Boost libs to link while running emconfigure with the existing codebase.
I'm totally winging it so if someone has any ideas that could set me on the right path it would be greatly appreciated.