FMVs => just a thought

[whippersnapper]

So my favourite is still the ingame cutscenes... Won't be photo-realistic, but someone might be able to get the graphics into a state where it looks ok even with detail-zooming.

in my experience the extreme close-ups with Grims art look fabulous - EXCEPT for Cyborgs and some vtols. What doesn't look to good at all is the extreme zoom-OUTS or God Mode which i think is due to Warzone not utilizing dynamic lod or geometry instancing. If you cherry-pick the units and structs and stick to closups mostly you can get some really sweet imagery for cut-scenes imho. ::)

btw, really curious - what is the downside to outputing an .avi to be played back on the cross platform OSS VLC Media Player  ???

It's true that WZ would need to be substantially modified to accomodate external FMV-type displays, but the actual number of FMV's doesn't need to be great.

There were very few of the original videos that I don't think couldn't be done better just via scripted scenes.  Many of the intra-stage clips were simply the computer voice telling objectives, while showing a map (not the actual game map mind you, just a re-drawn version for the video) for direction.  All of those types of videos can be replaced by in-game scripted cutscene, and perhaps a voiceover.

For away missions, you can just give a (very) brief before it, and then hold off on the scripted cutscene until your original force lands at its destination

The only original videos I can remember that would be best served as being FMV's are the transition ones (intro, alpha to beta, beta to gamma, ending), and the ones that gave backstory (Reed's military involvement, for example).

If it's that big of a problem to incorporate video playback into WZ, then... why not show the videos outside of WZ?  Would it be possible to completely suspend (read:  pause) WZ's running to show a FMV, at which point the FMV "closes" and WZ resumes processing?  Or if in windowed mode, pop up a separate window for the sole purpose of showing the video, while not affecting the operation of the original WZ window?

Forgive me if that idea is bunk, my programming experience deals little with real-time situations .

those are excellent insights and analysis, imho.


______________________________________________________

[Chojun]

I am extremely interested in having the FMVs back in Warzone, even if that means recreating them or completely rolling our own new ones.

But is it just me or does it seem like the Linux folks are generally disinterested in having FMVs in the game?  I don't think they understand but the FMVs are what made Warzone (for me, anyway).  ???  Someone clarify this for me, please.

[themousemaster]

I am extremely interested in having the FMVs back in Warzone, even if that means recreating them or completely rolling our own new ones.

But is it just me or does it seem like the Linux folks are generally disinterested in having FMVs in the game?  I don't think they understand but the FMVs are what made Warzone (for me, anyway).   ???  Someone clarify this for me, please.

It's not that "the Linux folk don't want it", its that "the programming folk realize its a pain in the arse; and some of them happen to be on Linux"

[lav_coyote25]

yes the fmv / cutscenes will be a pain in a butt for any who do attempt... but... the rewards will be... not having to have this same conversation every year at the same time ( give or take a couple weeks... ::)  )  oh and also having the community all involved would also be something to look forward to.

[Hatsjoe]

from the looks of it yu already have some powerhouse volunteers - lav_coyote, chojun, yourself.

Don't think i can model and animate dr. reed by myself. A NASDA satellite firing a missile is one thing but a human talking head is something completely different. + you forgot the textures :p

[Kreuvf]

A NASDA satellite firing a missile is one thing but a human talking head is something completely different. + you forgot the textures :p

afaik it is wireframe only (you mean nexus, don't you?).

[cybersphinx]

It's not that "the Linux folk don't want it", its that "the programming folk realize its a pain in the arse; and some of them happen to be on Linux"

That's more like it, yeah.

I looked a bit at the sequence code. It seems to be not that difficult to add video playback, if two problems can be solved:

• Getting the video frames. Ideally we have a library that we feed the encoded video stream, and that gives back singe frames. Complete players are mostly useless for that, as they're a. generalized, while we only need to handle one format, and b. optimized, and that makes the code difficult to reuse.
• Getting the video frames onto the screen. That means uploading a frame into an OpenGL texture 25 times per second. I tested mplayer's OpenGL output, and on my system 640x480 was useable, while 1280x544 wasn't. I guess I'll post some instructions later to see what resolutions are playable on which system.

And some more thoughts:

Any proprietary video thingy is useless, since a. we won't get the necessary code to play it, and b. it's limiting the number of people that can contribute (since you need to aquire those tools either illegally or by paying money, and most won't be crossplatform). For the same reason, I'd like only free tools like Blender used, though I guess I can't convince everyone of that.

For the container format, something that supports several audio streams (for different languages) like Matroska might be nice.

Getting an external video player to work right on all platforms might be more trouble than it's worth.

[DevUrandom]

Cyber got it very accurately... Exactly my opinion...

[Watermelon]

I think the brief window with a static picture and some nice voice acting(like kage's homeworld idea) will be better and easier than full video(at least size-wise),I think we have some good 2D/3D artist(e.g:the new logo by Terminator,the new models for headquarter and stuff by rush2049 and Lixtan),but we dont have the 'quanity' to do a huge project like remaking the fmv's imo...

just an idea:
1.voice file offset 0 display pic 1
2.voice file offset somenumber display pic 2
3.voice file offset somenumber2 display pic 3
...

[Giel]

WARNING LOTS OF WORDS HERE, READ ON YOUR OWN RISK

• Getting the video frames. Ideally we have a library that we feed the encoded video stream, and that gives back singe frames. Complete players are mostly useless for that, as they're a. generalized, while we only need to handle one format, and b. optimized, and that makes the code difficult to reuse.

I agree there. It would most likely be best to use an open source/free software (i.e. GPL compatible) library which we can feed a stream of encoded video data and then outputs a sequence of frames.

• Getting the video frames onto the screen. That means uploading a frame into an OpenGL texture 25 times per second. I tested mplayer's OpenGL output, and on my system 640x480 was useable, while 1280x544 wasn't. I guess I'll post some instructions later to see what resolutions are playable on which system.

Well as for how to get the frames on the screen. I personally would have to look for a tutorial just to get started with that. Although uploading a frame to video memory and only then displaying it doesn't seem all that efficient to me. By-passing video memory all the way and throwing it on screen immediately seems like the faster way. (Although I'm now using my experience with OpenAL and soundhardware memory so this might be way off.)

Any proprietary video thingy is useless, since a. we won't get the necessary code to play it, and b. it's limiting the number of people that can contribute (since you need to aquire those tools either illegally or by paying money, and most won't be crossplatform). For the same reason, I'd like only free tools like Blender used, though I guess I can't convince everyone of that.

You only forgot to mention that proprietary software is bad (my opinion) and those who enforce others to use it are just pure evil (still my opinion).

[offtopic-example]According to that definition both of my current two schools (yup doing two educations atm) are evil for explicitly forcing me and requiring me to use M$ software (e.g. one of them requires me to use M$ Visual Basic!!!! Argh!!! And that while the whole education has no aspect for which the usage of M$VB provides any benefits).[/offtopic-example]

For the container format, something that supports several audio streams (for different languages) like Matroska might be nice.

Sounds good to me, I like Matroska quite a bit personally.

Getting an external video player to work right on all platforms might be more trouble than it's worth.

Honestly, I wouldn't know how to begin using it on just one platform, let alone three (e.g. M$Win, GNU/Linux, Mac).

So aside from getting the stuff to actually be displayed on the screen I think I could implement this all.

Now only if I had the time. Which you won't have a lot of when doing two educations in parallel. Maybe when I'm done with my intern-ship (beginning of April) I'll have enough time to spare. That's a big maybe cause my final exams (of one education) will start pretty soon after finishing my intern-ship. So the biggest time-window which I can be certain of actually having it will be from the start of June to the end of August (summer break).

[Hatsjoe]

OK so you say it would be easier to let warzone do the frame processing than to implement some sort of REAL video player. Also this would have no legal obligations about use of playback software since you're going to code some sort of a frame to video compiler. For modelling and animation work this doesn't make any difference so i guess that's our best option yet. 

Edit: Hmm two sides here apparently. Some say feeding loose frames and displaying them one at a time and others say feeding processed movie files

Wouldn't the latter provide legal difficulties because of codec use, player use, ... ? Cause those are the problems that prvious dicussion have been going about me thinks

[Hatsjoe]

afaik it is wireframe only (you mean nexus, don't you?).

I think it's wireframe with some visual effects applied. but you'd still need to model and animate the wireframe. this should probably be done by someone more experienced than me.

BTW: Can i use a student licensed version of 3Ds Max to recreate those files? Cause that's all i got

[cybersphinx]

I think the brief window with a static picture and some nice voice acting(like kage's homeworld idea) will be better and easier than full video(at least size-wise),I think we have some good 2D/3D artist(e.g:the new logo by Terminator,the new models for headquarter and stuff by rush2049 and Lixtan),but we dont have the 'quanity' to do a huge project like remaking the fmv's imo...

Oh, there might be enough people if we concentrate on remaking those cutscenes that need to be videos (EDIT to continue what I should have finished in the beginning:), and not spending time on things that can be done at least almost as well in the engine. Of course, you need a good team and a good plan.

WARNING LOTS OF WORDS HERE, READ ON YOUR OWN RISK

Gee, thanks for the warning, I almost thought it was just dirt on my screen. Hey, that wasn't nearly as much as Kage writes sometimes, now I'm dissapointed!

Well as for how to get the frames on the screen. I personally would have to look for a tutorial just to get started with that. Although uploading a frame to video memory and only then displaying it doesn't seem all that efficient to me. By-passing video memory all the way and throwing it on screen immediately seems like the faster way. (Although I'm now using my experience with OpenAL and soundhardware memory so this might be way off.)

You are way off. Everything that's shown on screen comes from video memory (unless you intend to program the VGA/DVI-connector yourself, if that's even possible). Perhaps you meant texture memory, then I'll kind of agree with you. But I don't know any other way to get images on screen in OpenGL (and we don't want to change to something else, as that'll be about as messy as calling an external player).

You only forgot to mention that proprietary software is bad (my opinion) and those who enforce others to use it are just pure evil (still my opinion).

That was quite deliberate, I only wanted to list objective reasons. As the problem is manpower (as usual), I guess a model done in 3D Studio is better than no model. But perhaps we can do at least the actual animation etc. in Blender.

Also this has legal obligations.

What kind of legal obligations?

EDIT again: If someone wants to start making some videos: Don't bother with actual file formats etc., you can make demo renderings or stuff however you like. Just give us a Blender file to render all frames at our desired resolution and a matching sound track, and I'm almost positive we can encode it into a useable format and implement playback in the engine.

[Hatsjoe]

What kind of legal obligations?

I meant NO legal obligations. As any bitmap or jpeg or png or whatever can be freely distibuted
Have to be more carefull next time or this will lead to problems :s sry

EDIT again: If someone wants to start making some videos: Don't bother with actual file formats etc., you can make demo renderings or stuff however you like. Just give us a Blender file to render all frames at our desired resolution and a matching sound track, and I'm almost positive we can encode it into a useable format and implement playback in the engine.

I ve done the "project comms" in max. Just the version that says "briefing commencing" it's just a 2.4 sec file but There are some other longer comms that use the exact same animation only longer so all those are actually done (if you like it that is) Chojun is going to host the file asap. So stay tuned. And i do believe that the best way to produce videos is by all using the same modelling tools etc. but if i have to do it in blender you can count me out cause that would take me 10's of hours to get used to the prog. However I believe that we can use max for a certain FMV and Blender for another. As long as both artists are working closely together since the end product of those people will just be a bunch of unprocessed images.

[cybersphinx]

EDIT again: If someone wants to start making some videos: Don't bother with actual file formats etc., you can make demo renderings or stuff however you like. Just give us a Blender file to render all frames at our desired resolution and a matching sound track, and I'm almost positive we can encode it into a useable format and implement playback in the engine.

Ah, damn, perhaps that promised a bit much... I won't go running to implement video playback for a two second clip. Investigating if and how it's possible will take some time anyway. But if you get a serious team together that's ready to produce some videos, you don't have to care about what format to encode the videos to, what resolution to use, what bitrate and other technical stuff. At least if you use a program that's freely available. I don't care what you use for modelling, but if the animation can all be done in Blender, that'd make getting useable results very easy. (And I guess all rendering should be done with the same program, to have a consistent look and feel to the movies.)