Instead of messing with the MKV, just tell VSFilter to not pre-buffer the subs (i.e. uncheck the 'Pre-buffer subpictures' option in the config). That tends to fix the problem of subs not showing up. Also make sure that it isn't a case where the default subtitle stream is actually a dummy so that the
real subs don't show automatically and annoy the user - although this would be more relevant to AMVs (currently, this is how I treat subtitles in my own videos, and Bakadeshi used this tactic on his most recent release as well; I'm not aware of other instances of it, though), fansubs wouldn't really apply here.
Otherwise, extract the subtitle stream (and leave it in ASS format), then use:
Code: Select all
[source loader; DirectShowSource or FFmpegSource with MKV file]
TextSub("subtitles.ass")
Which will hardcode the subs to the video. Use of TextSub requires a copy of VSFilter.dll to be in AviSynth's plugins directory or loaded with LoadPlugin.
PS.. how the heck do they get their captured video's to look soo nice ?? I'm yet to figure that out also T_T .. maybe some New kind of Editor that no ones ever heard of before .. a script that keeps colors from gradiating ....
gradfun2db is a debanding filter. Insanely-high quality encodes tend to be the product of the encoder using similarly high-quality sources, and for recent shows, more than likely taking an aggressive IVTC measure (very sensitive TFM/TDecimate settings or something completely insane like using
YATTA to do it manually) but otherwise simply doing light filtering elsewhere so as not to overfilter the source.
The actual RAWs probably originated either from A) R2 DVD rips which were obtained the way most groups get their RAWs, B) the few true high-def series that have aired (in which case, the source is usually downsampled to standard-def resolutions from the high-def RAW), or C) R2 DVDs themselves which are in the encoder's physical possession.
Using higher bitrates also doesn't hurt (and is actually the instance where using gradfun2db will be the most use - heavy compression will negate the benefit, but on high-bitrate encodes that would retain fine details like that instead of creating them, it helps).
For instance, this is my script when working with rather clean DVD footage (which I use as a template script for DGIndex, hence the __vid__ part):
Code: Select all
MPEG2Source("__vid__", ipp=true, cpu=4)
deint = TDeint(mode=2, mtnmode=3, blim=100)
TFM(order=-1,mode=5,PP=7,field=-1,slow=2, clip2=deint)
TDecimate(mode=1)
AssumeFPS(24)
deen("w3d",2,8,8,8,8)
gradfun2db(2.0)
ConvertToRGB24()
The AssumeFPS and ConvertToRGB24 lines are only in there because of Premiere. Afterwards, I use a pretty much maxed-out encoding profile in MeGUI, and use crf - aka Constant
Quality - 18 (although I can see the benefit of using that as a rough first pass and then using 3-pass to refine it, but I'm not that masochistic).
