its not easy. The whole reason Xuggler has still no player available is that people have tried ever since, but still have sync issues.
And overall I dont have the time, skill or knowledge to build an own decoder/video player / whatever =/
Come to think of it when I tried MJPEG a little while back it was too slow if you wanted a reliable 15fps or better. (Bearing in mind the wide disparity in CPU power available on desktops)
I was using a 1.6GHz mobile CPU at the time (the one I developed Revenge of the Titans on) - not surprisingly it was around half the speed at that resolution (doing sound as well, and rendering) - so it could just about cope with 60Hz. Trouble is I wanted rather higher resolution and it got rather a lot slower.
This is very true Ok, I admit it, I was doing hi-def stuff at Sony at the time Ignore me. Nevertheless the space issue is a bit of a showstopper if you’ve got any reasonable amount of video.
Uh, no? Isn’t 23.976 or something the standard frame rate of real movies? I was gonna check it up, but after encountering strike pages on 3 different sites I gave up. This strike IS starting to bothering me now. -_-
23.97fps is only used by standard film format - there are a whole bunch of other standard frame rates, compounded with dropframe and non-dropframe versions of the fractional rates, progressive vs. interlaced, etc. You wouldn’t believe what a quagmire it is. For any normal sane computer video you will be using “30p”, which is 30fps, non-interlaced (“p” for “progressive scan”).
And 50p and 60p formats are not at all uncommon in the hi-definition world, btw!
The Hobbit movie is getting shot at 48fps. Which I think is really about time IMO, 24fps cases visible flicker effects for me on faster pans. Unfortunately is still going to be 3d.
What i don’t understand is why interlaced hasn’t died its deserved death. The 1970s called and wants is crappy interlaced CRTs back.
Don’t worry, nobody in the industry understands why it still exists either. Actually we do, it’s because of the vast pile of legacy equipment out there.
JavaFx 2 supports “FLV containing VP6 video and MP3 audio”
does anyone know a tool that an do this easily ? I remember that creating a flv video is not so hard, but VP6 specifically… I rarely could ever select that and when I could VP6 wasnt one of them.
Just would like to know that, if I actually could render every video in this flv codec, how would javafx 2 perform…
[quote]it’s because of the vast pile of legacy equipment out there
[/quote]
btw same reason people use windows in general, and C++ for game development.
No, it’s not the same reason, it’s because the A/V situation on Linux is still a massive clusterf*ck that shows no signs of ever getting resolved. For audio, assuming the driver manages to work, now add on the desktop environment’s choice of mixer, whether you need superuser perms to even use it, and whether the distribution managed to completely cock it up or not. For video, nvidia’s looking pretty good, everything else you can forget about, and so on.
I do all my work dev work in Linux, but it’s all network server stuff. I do my hobby dev stuff in Linux too, but I don’t bother holding out any hope it’ll run on any machine but mine without a lot of debugging work. Using Java is a good headstart, but once you depend on the external environment, you’re still hacking through the brush.
Even if many Linux’ are screwed there are still many other options, people just use windows because it has been the standard for a long time and all the good software is written for it.
Only mac changes the game a little bit, with awesome software like final cut pro, only on mac
C++ same story - only reason its used so much is that legacy code and libraries are written in C++
I hate C++ with a passion. But it’s still the only viable option for balancing performance vs. work involved in creation for high performance situations.
For codecs the gap (should be) narrow. Sequential memory accesses, not pointer intensive, have the option to push stuff off to the GPU, etc. etc. Not having access to non-scalar SIMD and non-temporal hints (and/or dropping to ASM for special cased) for stuff left on the CPU side is where any real difference will come in.