what is the simplest
- record the output of a dx9render
- end the recording after 2 minutes.
- loop and play the output on top of a live video stream
- and repeat
im using dx9texture->asvideo->compress->writer to record and save the 2 minute recordings as a video file…
next i will access the video files sequentially pushing and popping files off the file spread of 10 filename slices or slice that represent the video output (textures… filenames etc)… and loop them over the video. are there any other possible options i am not aware of?
i dont see another way, saving the output as video and reload it with auto generated filenames should work.
imho saving singleframe sequences as dds could be more performant.
woei i dont understand what your describing. how would saving individual frames a spread of textures be faster?
using dx9texture->asvideo->compress->writer i had too many dropframes and overallperformance became unbearably low. maybe it’s just my comp that’s too slow.
i was more lucky recording frames. since vvvv can run much faster than 25fps i kind of use the time inbetween drawing the videoframes to save a pic (correct me, if i’m talkin bullshit) - see attached
framerecorder.v4p (12.8 kB)
but how would you load all the pictures automatically without stopping performace? thats what hype wants to do… only thing i can imagine is to load one picure per frame… but 2 minutes of video will hardly fit into memory…
hum, why would you have to load all the pictures at once?
playing dds sequences i just use dir-getslice-filetex, without preloading the whole sequence, since dds loading is fast enough for not notable loadingtime. even a few layers of sequences should not be a problem.
or did get something wrong? (save dx9 output, and looping it over the delayed output and record again)
hm heres a repatched version using videoin… how do you save a series dds sequences? i can save one sequence… and i could save m ore using the counters as means to rename but the playback isnt as smooth as videos… im sure i didnt make the best example though perhaps some one can add on?
FrameRecorder.v4p (19.2 kB)
from your current patch i guessed, that each sequence is delayed one frame to the preceeding sequence. if i’m correct with this, it would not be necessary to record eg 8 separate sequences, just accessing 8 pics at once and shifting this through the sequence.
however i get the feeling, i didn’t really get your problem though :)
overlay.v4p (7.4 kB)
i am using video files as opposed to sequenced dds files. how do i play a the last recorded file and fade the video (from alpha 1 to alpha 0 in 30 seconds) over the videoin output. there may be multiple recorded files fading simutaneously over the videoin output depending on the time intervals at which the videos are recorded… i am setting the max amount of recorded videos to 10. after the 10 video the 1 is wrote over. here is my patch so far.
NetLab.v4p (21.8 kB)