What are your thoughts on video captures?
category: general [glöplog]
This thread is too long to read, someone kkapture it and upload summary on youtube plz.
It not that bad, youtube supports HD now (even though it's only 480p), so if the capture is full of artifacts with crappy sound, blame the uploader.
Quote:
Youtube is really a terrible medium for demos with those tracks recorded at 8khz and the shitload of insulting artifacts on the videos themselves, i'm in favor of removing/banning all youtube links from pouet.
It not that bad, youtube supports HD now (even though it's only 480p), so if the capture is full of artifacts with crappy sound, blame the uploader.
480p is HD now? ... sigh
Since when was 480p HD? If that's right, they've gone from LD to SD. Which is a huge improvement, but still pretty crap.
Hitch: yeah, gamedev plus all the other creative scenes. I wonder if we could trick them into coding demos? Start up a 'gamedev' compo, where they have to make say an 'intro screen' for a game, non-interactive, showing the best effects + art they can do...
Hitch: yeah, gamedev plus all the other creative scenes. I wonder if we could trick them into coding demos? Start up a 'gamedev' compo, where they have to make say an 'intro screen' for a game, non-interactive, showing the best effects + art they can do...
Quote:
It not that bad, youtube supports HD now (even though it's only 480p)
What? It's been 720p for almost a year now.
What, you need more than 480p to watch those c64 demos ? :) Seriously, I still remember the times when 640x480 in truecolor was the epitome of happiness for most sceners. Or is it now just mindless pixel chasing ?
Times are a-changing. 480p looks pretty bad on a 1080p pannel. The bigger problem is the bitrate though.
psonice: you're thinking about outreach all wrong..
we dont need to come up with 101 ways to try and persuade new people to get interested in the scene. we just need the sceners who are around now to make good enough demos to make people _want_ to be a part of it.
we dont need to come up with 101 ways to try and persuade new people to get interested in the scene. we just need the sceners who are around now to make good enough demos to make people _want_ to be a part of it.
Smash: yeah, that's my point. 9 times out of 10 those people are going to come across that great demo on youtube, after watching a bunch of really good clips made in afterfx or maya or whatever, that's kind of hard to get past with anything 'traditional'. Something say half-interactive on the other hand, they would likely download and run realtime, then we're more likely to get them :)
By the way, what outsiders get to see from the scene is usually debris, elevated, chaos theory, maybe second reality (and kkrieger, linger in shadows, but let's talk about demos now). Now, the first thought of people watching debris or elevated is usually not "yeah, i could do that too, let's form a demogroup"...
tomaes> Do they ? Because i often have to watch the latest demos in lower then 1024x768 if i want to get any decent framerate, or enable AA (yes, my hw suck, i admit that). What i'm trying to say is that resolution is not that important, i always prefer fluid motion and good sync, and that goes the same for democaptures. Am i the only one ?
Democapture is obviously for people that want both fluid motion, good sync, high resolution, and not-so-high-end hardware.
Go back to 8-bit platforms... there playing a video _is_ a demo :)
(see Natural Bug Killer, Midline Process, or Phat 2 on Amstrad CPC. They are just about that.)
Broadband internet connections removed most of the interest for small-size demos. You can get the video in one minute more, so why bother. We have to find something else to keep the challenge going. It's not necessarily about doing something that can't be rendered to video, just about challenge.
On an 8-bit platform, doing something artistic was a challenge by itself.
On 16 bit, they started things like 'i have more dots than you', or 'look I can display more than 300 sprites per frame'
On 32bit, they had sizecoding, aka 'look how much I can fit in 64k'
Now it's 64-bit time and you have to find something else. Or just go back to just 'make something artistic' and forget about the challenge part. Realtime is no longer needed now we can just watch the video. So drop it. It already started with procedural graphics. Records can't be set because the hardware change too fast. Size-limits are pointless because no one cares anymore.
So, either what interest you in demoscene is constraints, and you should stick at a fixed platform and get the mos out of it. It can be Atari STe, or Pentium III 800MHz with some AGP gfx card. Or your own PC (but that is not good for sharing your prods).
If you're more in the artistic side, then the realtime constraint can be dropped, either by releasing videos or by waiting for more powerful hardware. Or by doing 'procedural videos'. Or whatever, I don't care, I'll stay in the first category.
Go back to 8-bit platforms... there playing a video _is_ a demo :)
(see Natural Bug Killer, Midline Process, or Phat 2 on Amstrad CPC. They are just about that.)
Broadband internet connections removed most of the interest for small-size demos. You can get the video in one minute more, so why bother. We have to find something else to keep the challenge going. It's not necessarily about doing something that can't be rendered to video, just about challenge.
On an 8-bit platform, doing something artistic was a challenge by itself.
On 16 bit, they started things like 'i have more dots than you', or 'look I can display more than 300 sprites per frame'
On 32bit, they had sizecoding, aka 'look how much I can fit in 64k'
Now it's 64-bit time and you have to find something else. Or just go back to just 'make something artistic' and forget about the challenge part. Realtime is no longer needed now we can just watch the video. So drop it. It already started with procedural graphics. Records can't be set because the hardware change too fast. Size-limits are pointless because no one cares anymore.
So, either what interest you in demoscene is constraints, and you should stick at a fixed platform and get the mos out of it. It can be Atari STe, or Pentium III 800MHz with some AGP gfx card. Or your own PC (but that is not good for sharing your prods).
If you're more in the artistic side, then the realtime constraint can be dropped, either by releasing videos or by waiting for more powerful hardware. Or by doing 'procedural videos'. Or whatever, I don't care, I'll stay in the first category.
Quote:
Democapture is obviously for people that want both fluid motion, good sync, high resolution
If I want:
- fluid motion (60 fps, not 24 or 25)
- high resolution (full 1680x1050 without compression artefacts)
- good sync (not a crappy video out of sync)
I will indeed prefer running the demo realtime on my GPU that can run most demos at a good framerate, although it's a few years old already.
Quote:
Size-limits are pointless because no one cares anymore
Maybe you don't but some people still do, TBC's latest productions prove it, among many others.
Some people still care about what you can do given a specific GPU and videos hide the real framerate and the technical challenge behind what's displayed on screen. The ratio ( stuff displayed / available power ) is still something that drives a lot of us.
That might surprise you but yes, a lot of people still care about that.
On a sidenote, the arguments like MODERN GPUS CAN DO EVERYTHING NOWADAYS AND DX AND OPENGL WILL JUST CODE IT FOR YOU!!1 are full of shit.
And what a pointless discussion...
I generally agree with maali on this one, except I actually have a supercomputer from space.
I seem to be watching fewer demo caps than ever, since I built a machine last year (quadcore/vista/gtx280), and even demos i used to watch as caps I watch in realtime (like relais and other silly nv only demos). About 98% of realtime PC demos I watch realtime, and only exist on my drives as such. I love seeing sharp pixels on my 1920x1200 monitor, rather than seeing blurry artifacts from a video. I also realized about three years ago that 4ks actually started to run for me rather than crashing, with my mouth on the floor watching elevated (omg another 4k runs, aside from the intro itself).
The only reason I have to watch a cap is on my laptop (with linux and the shitty open source driver), which has little more gfx power than a toaster after compiz.
I seem to be watching fewer demo caps than ever, since I built a machine last year (quadcore/vista/gtx280), and even demos i used to watch as caps I watch in realtime (like relais and other silly nv only demos). About 98% of realtime PC demos I watch realtime, and only exist on my drives as such. I love seeing sharp pixels on my 1920x1200 monitor, rather than seeing blurry artifacts from a video. I also realized about three years ago that 4ks actually started to run for me rather than crashing, with my mouth on the floor watching elevated (omg another 4k runs, aside from the intro itself).
The only reason I have to watch a cap is on my laptop (with linux and the shitty open source driver), which has little more gfx power than a toaster after compiz.
Quote:
Maybe you don't but some people still do, TBC's latest productions prove it, among many others.
Some people still care about what you can do given a specific GPU and videos hide the real framerate and the technical challenge behind what's displayed on screen. The ratio ( stuff displayed / available power ) is still something that drives a lot of us.
That might surprise you but yes, a lot of people still care about that.
Actually I don't care about PC demos and that modern stuff. Capture or not, I don't even watch them. I find the 8bit scene a lot more interesting for this kind of constraints. Why having a more powerful machine if you add artificial constraints to keep it interesting ? Of course having demos as big as the video capture won't help. And I hope some people still care about it and create some interesting linker/synth/whatever to get as much as possible shrinked into 4k. Else the scene would be dead (oh wait, it is already).
I understand that you still have limits and challenge on modern platforms (like, selecting the good function to do something in OpenGL's hundred of ways). The problem is not there. The problem is there is no common reference platform. So saying "this effect runs at full framerate" may be valid on your computer, and not on mine. And the other way arounf for some other effect. Demos are about pushing the hardware/software/whatever you use as a platform beyond its known limits. If you can't even define this platform properly, how can you push it ? Either you do things that run only for you, or multiplatformism can become a challenge too (a bit like Planet Hively did). Or you can go to some big demoparty and work on their reference demoboxes. But then, no one can be sure to watch your demo the way you intended it. So, why bother running it as executable ? Maybe you can get better quality this way. Maybe it'll get worse than the video, depending on the demo, your hardware and some random fuzz-factor.
And yes. Pointless discussion. This is pouet BBS :)
I don't think it's pointless at all.. but I don't expect it to go anywhere :)
The scene is in decline, if nothing changes it'll slowly slide into oblivion. I doubt it would disappear, but it'd be more like the 8 bit scene - there, doing cool stuff, but a fraction of what it used to be and pretty unknown in general.
I'm sure at some point we'll hit another big turning point and the scene will rise again. I mean, we've survived the death of the amiga, the end of dos, and 3d accelerators (all of which were supposed to kill the scene), surely we can survive youtube? :D
The scene is in decline, if nothing changes it'll slowly slide into oblivion. I doubt it would disappear, but it'd be more like the 8 bit scene - there, doing cool stuff, but a fraction of what it used to be and pretty unknown in general.
I'm sure at some point we'll hit another big turning point and the scene will rise again. I mean, we've survived the death of the amiga, the end of dos, and 3d accelerators (all of which were supposed to kill the scene), surely we can survive youtube? :D
Now go coding on the vectrex. The vector display means no video can reach the quality of the real thing. Or use a Laserium, which will get you color and a big screen.
Pushing hardware to it's limits is one of the defining factors of the demoscene, but I think for me the draw has always been the 'seeing something that has never been done before' aspect of it. Personally, I would f'in love to see a demo specifically designed to whip the crap out of, say, a quad sli GTX 295 rig, even if it meant I'd be relegated to watching a vidcap of it.
Quote:
I guess this is one of the big reasons for macs getting more popular - you buy it, turn it on, and get on with it.
Yeah cos you can't just go and buy a PC that does that. Uhuh.
well you could.. but you'd likely have to update it (same for mac here of course), update dx, update drivers too. A month later you'll might well have to do the same to watch the next big demo. There's plenty of advantages against that (easy/cheap upgrades, tweaking stuff to make it run faster, ...) but if you can't be arsed with all the hassle that goes with it you're left with just annoying updates + fixing stuff. Horses for courses etc. :)
that bullcrap. the only shit that requires semi-monthly updates are those d3dx9_ .dlls that luckily more and more groups just provide with the release itself so that you dont have to manually update that. as for having up-to-date drivers is not really relevant for demos, rather for gaming (cos of the builtin performance cheats/tweaks) the chance a certain demo uses some kind of trick that was buggy in some old driver is as big a chance a certain mac demo requires exotic python libraries.
So why is the advice for people with a problem running a demo usually "try the latest drivers"? And why does that often fix the problem? (At least it's helped for me numerous times before, which is why I mentioned it in the first place..)
Good to hear people are bundling the d3dx dlls though. Wasn't the reason for not doing so that microsoft tells you not to in the license or something? Not that it's a good enough reason to actually not do so.. and those dlls were always easily found on various download sites instead of grabbing the whole dx package.
Also, I don't remember seeing a mac demo requiring python libraries, it's sdl normally ;) (and yeah, seeing "failed to find sdl library" when I run a demo gives me a huge case of can't be arsed too)
Good to hear people are bundling the d3dx dlls though. Wasn't the reason for not doing so that microsoft tells you not to in the license or something? Not that it's a good enough reason to actually not do so.. and those dlls were always easily found on various download sites instead of grabbing the whole dx package.
Also, I don't remember seeing a mac demo requiring python libraries, it's sdl normally ;) (and yeah, seeing "failed to find sdl library" when I run a demo gives me a huge case of can't be arsed too)
I haven't updated my graphics drivers for more then a year (ati 4870) and I have only updated DirectX once because I wanted the latest SDK. Still I can run every ATI compatible demo without problems.
psonice: "latest drivers" is more relevant for e.g. glsl-based productions where the compiler is in the driver. (thread crosslink!!) for d3d-based demos its much less often a problem.
and no, ive hardly ever had to update for a demo except for d3dx in the past few years either.
and no, ive hardly ever had to update for a demo except for d3dx in the past few years either.
smash, do you mean the d3d shader compiler gets updated through windows update or something?