Demo to video

[Moon[Drunk]]

OK I want to make a video preferable in Darkplaces, but I guess FTE could work too, and I would like some advice from people who lately has some experience with this. A step by step guide from demo to video would be perfect...

I have searched the forum and found this sticky thread where Darkplaces is mentioned but looking at the last (unanswered) post by Mr. Burns where he more or less asks the same question as I am doing right now I think this is something that needs to be explained further.

[Seven]

Hello Moon[Drunk],

I am very excited about your thread and question...
I am looking forward to what you will show us. Cant wait.

I personally also use 3rd party programs when creating gameplay clips:
- Fraps to record .avi clips when playing darkplaces.
- virtualdub to edit & create the final video. Xvid codec to make its size very small.


But if you want to go with Darkplaces for recording, these are the cvars to adjust the recording properties:

cl_capturevideo is "0" ["0"] enables saving of video to a .avi file using uncompressed I420 colorspace and PCM audio, note that scr_screenshot_gammaboost affects the brightness of the output)
cl_capturevideo_demo_stop is "1" ["1"] automatically stops video recording when demo ends
cl_capturevideo_fps is "30" ["30"] how many frames per second to save (29.97 for NTSC, 30 for typical PC video, 15 can be useful)
cl_capturevideo_framestep is "1" ["1"] when set to n >= 1, render n frames to capture one (useful for motion blur like effects)
cl_capturevideo_height is "0" ["0"] scales all frames to this resolution before saving the video
cl_capturevideo_nameformat is "dpvideo" ["dpvideo"] prefix for saved videos (the date is encoded using strftime escapes)
cl_capturevideo_number is "1" ["1"] number to append to video filename, incremented each time a capture begins
cl_capturevideo_ogg is "1" ["1"] save captured video data as Ogg/Vorbis/Theora streams
cl_capturevideo_ogg_theora_bitrate is "-1" ["-1"] video bitrate (45 to 2000 kbps), or -1 to use quality only; higher is better; setting both to -1 achieves unlimited quality
cl_capturevideo_ogg_theora_keyframe_auto_threshold is "80" ["80"] threshold for key frame decision (0 to 100)
cl_capturevideo_ogg_theora_keyframe_bitrate_multiplier is "1.5" ["1.5"] how much more bit rate to use for keyframes, specified as a factor of at least 1
cl_capturevideo_ogg_theora_keyframe_maxinterval is "64" ["64"] maximum keyframe interval (1 to 1000)
cl_capturevideo_ogg_theora_keyframe_mininterval is "8" ["8"] minimum keyframe interval (1 to 1000)
cl_capturevideo_ogg_theora_noise_sensitivity is "1" ["1"] video noise sensitivity (0 to 6); lower is better
cl_capturevideo_ogg_theora_quality is "48" ["48"] video quality factor (0 to 63), or -1 to use bitrate only; higher is better; setting both to -1 achieves unlimited quality
cl_capturevideo_ogg_theora_vp3compat is "1" ["1"] make VP3 compatible theora streams
cl_capturevideo_ogg_vorbis_quality is "3" ["3"] audio quality (-1 to 10); higher is better
cl_capturevideo_printfps is "1" ["1"] prints the frames per second captured in capturevideo (is only written to the log file, not to the console, as that would be visible on the video)
cl_capturevideo_realtime is "0" ["0"] causes video saving to operate in realtime (mostly useful while playing, not while capturing demos), this can produce a much lower quality video due to poor sound/video sync and will abort saving if your machine stalls for over a minute
cl_capturevideo_width is "0" ["0"] scales all frames to this resolution before saving the video
cl_minfps_force is "0" ["0"] also apply quality reductions in timedemo/capturevideo
cl_video_brightness is "1" ["1"] brightness of video, 1 = fullbright, 0.75 - 3/4 etc.
cl_video_fadein is "0" ["0"] fading-from-black effect once video is started, in seconds
cl_video_fadeout is "0" ["0"] fading-to-black effect once video is ended, in seconds
cl_video_keepaspectratio is "0" ["0"] keeps aspect ratio of fullscreen videos, leaving black color on unfilled areas, a value of 2 let video to be stretched horizontally with top & bottom being sliced out
cl_video_scale is "1" ["1"] scale of video, 1 = fullscreen, 0.75 - 3/4 of screen etc.
cl_video_scale_vpos is "0" ["0"] vertical align of scaled video, -1 is top, 1 is bottom
cl_video_stipple is "0" ["0"] draw interlacing-like effect on videos, similar to scr_stipple but static and used only with video playing.
cl_video_subtitles is "0" ["0"] show subtitles for videos (if they are present)
cl_video_subtitles_lines is "4" ["4"] how many lines to occupy for subtitles
cl_video_subtitles_textsize is "16" ["16"] textsize for subtitles

Best of luck,
Seven

[Henry]

Free yourself from engine specific settings, commands and cvars.
Use the engine your are most comfortable with for playback the demo. And use engine independent recording software. That is the most logical way.
You will need to edit your recorded video file(s) anyway!

Baker linked to several fraps alternatives if you do not want to use it.
After that use virtualDub or ffmpeg to edit, cut, merge, transcode, compress or whatever your video file to bring it to a nice small size to be able to upload it to youtube or somewhere else.

All those software tools are freeware and work fine.

[dfsp_spirit]

Fraps never worked fine on my machine, although I have seen it work on the PC of a friend. All the videos it takes on my machine have like 5 fps, and it makes my machine very slow.

Although I'm an open source fan, I am now using Mirillis Action (which is commercial and closed, like Fraps). I tried many free screen capture tools, and they were all absolutely terrible. They had bad performance, would not capture in usable formats, would not work with directx games, would not work with openGL games, came bundled with trashware in the installer, or all of the above. I was also unable to get them to output formats suitable for Youtube export.

[Spike]

I actually have a plugin for fte that directly uses ffmpeg. recording straight to .mp4 with x264+aac is quite simple with it. it should also be possible to use in-engine rtp streaming too, both playback and capture. screw fraps
but yeah... video encoders and patents and distribution and gpl and paranoia and gah. thankfully I don't live in the us or germany, but it limits where I can upload it to, and whom I can really distribute it to. plus there's apparently still a few issues. and I'm lazy, hence why I didn't mention it earlier.

[Baker]

In the console type cl_capture and then press TAB. Should show you a list of commands.

I don't know if I would recommend using built-in DarkPlaces video capture though. Unlike ezQuake or JoeQuake that use a "real codec" that compresses video, DarkPlaces does uncompressed video or very poorly compressed video as far as I have ever been able to tell.

Like in under 10 seconds and you'll be hitting a few GB, no kidding.

I'd probably use video capture software instead with DarkPlaces. I haven't been paying attention to recent video capture software but apparently Fraps is still one of the better free alternatives Are there any Alternatives to Fraps [Solved] - Performance - Apps General Discussion


(Apparently I said the same thing as Spike, haha. Just slower at clicking submit.)

[Spike]

in DP, there's a load of cl_capturevideo_* cvars. apparently configure it how you want and then use:
cl_capturevideo 1; playdemo foo.dem
I *think* it'll use avi by default, but with no real compression. I don't know any of the details.

FTE uses capture* cvars to configure it.
capturedriver avi; capturecodec ""; to configure it to write an avi with no video compression (you can try h264 if you have x264vfw installed, or other fourccs if you have vfw-compatible encoders installed)
capturedemo foo.dem bar.avi; to then capture the demo
(note that capturedemo will adjust playback timings, removing the need to record in realtime so you can use more cpu intensive encoders, or just finish faster).

in both cases, you can then use some external tool like ffmpeg to transcode it to use a real video compression codec.
alternatively, both engines support raw .tga output, which can also be fed into an encoder, but using many input files instead of a single one is generally more messy.