VideoHelp Forum
+ Reply to Thread
Results 1 to 11 of 11
Thread
  1. edit: See my second post for current direction. This first post will quickly become redundant as I try different approaches/learn more.

    I'm pretty damn confused, been kind of working my way through some "simple" stream extracting and remuxing.
    (for some background: Have a DVD with good quality video stream, have mkv with subs i need and also superior DTS audio stream)

    So I was all happy for a while, I used DVDdecrypter to get the video stream out as a .m2v file. Converted the SRT from the mkv to .sup, changed the palette colours to be compatible, remuxed 2 audio streams from the mkv with my new subs + good video. However, the source is still in need of some post processing, sharpening etc.(which the current torrented rip uses on their mkv release to good effect. Their file size was 1.3gb. I want to archive my own definitive version with a space limit of DVD-9.

    So the easiest way I found to apply some filters was simply loading my m2v into virtual dub. Then using a few filters (I don't really understand AviSynth or how to integrate it) and these came out pretty well on the previews, nice added sharpness.

    At which point i hit a roadblock. How on earth do I get this post processed file back into .m2v stream format? I really don't know the next step,. Can re-encode straight to m2v? Or can I do it through x264/avi to m2v? I couldn't find much/anything on converting these formats to m2v. So am a bit lost.

    Thanks in advance on any advice/pointers!
    Last edited by transcendent; 29th Jun 2012 at 13:34. Reason: helping
    Quote Quote  
  2. There are a few different methods you could use. You could use TMPGenc or TMPGenc Mastering Works to open the m2v file directly and apply filters, cut & edit, etc. and then export as a dvd compliant mpg file. Then you could use the program's built in demuxing tool to demux the mpg file back into m2v.
    Quote Quote  
  3. Originally Posted by transcendent View Post
    How on earth do I get this post processed file back into .m2v stream format?
    Since you seem unwilling or unable to learn any AviSynth, save it out as a lossless AVI (Lagarith or whatever), and then reencode that to M2V. Doing it that way could conceivably save you some time in the long run if the filters you use are very slow. They'll only have to be run once while MPEG-2 encoding is usually 2 or more passes.

    https://www.videohelp.com/tools?toolsearch=lossless+video+codec&Submit=Search&convert=&...y=Name&hits=50
    Quote Quote  
  4. Thanks for the advice guys, I had a look at TMPGenc but didn't really like the built in filters they had. I much prefer the ones written on http://neuron2.net/ by David Graft. I then ended up having a better look at AVIsynth, I'm still playing around: I got the d2v indexed file from his DGMPGDec program.
    So I wrote and AVS script saying MPEG2Source (or something like that) and got it displaying in the latest version of virtual dub, I then got his Msharpen filter for AVIsynth and dropped the DLL into the plugin folder, edited my script with an extraline with some test parameters, reloading the script...and...nothing really happens. I get the feeling I'm doing something wrong, i know virtualdubmod had an internal script editor but it complained about YV12 stuff which seems to be native to normal virtualdub in fast recompress mode.

    So I'm still working it out, but it would be nice to get some confirmation on whether I'm meant to be reloading the script everytime i make an edit...and some hint on how I get a realtime preview when I do reload the script. I put the filter parameters on something extreme to show but it seems to be showing just the source instead of any real time changes...

    Oh yeah, one more thing. When I look at the video normally, say, frame by frame in MPC-HC, it's PAL in progressive frame by frame format. However, when loaded into virtual dub via avs script, when i'm going through the frames, I seem to see the interlacing every few frames.(same happened while previewing on TMPGenc). Could someone explain what is going on here. It seems like I'm not meant to "force film" or "ignore pull down flags" as they will screw up fps sync. Do I need to do something regarding Decomb filters /IVTC deinterlacing on the source pre post processing filtering?

    Many thanks in advance, I know these questions must get very repetitive.. but I am trying!
    Quote Quote  
  5. Originally Posted by transcendent View Post
    When I look at the video normally, say, frame by frame in MPC-HC, it's PAL in progressive frame by frame format. However, when loaded into virtual dub via avs script, when i'm going through the frames, I seem to see the interlacing every few frames.
    You probably have a deinterlacer on in MPC-HC. VDub shows it without any filtering (unless you have some already in the scrip).
    It seems like I'm not meant to "force film" or "ignore pull down flags" as they will screw up fps sync.
    Never Force Film a PAL source. Almost never Ignore Pulldown Flags. Always make the D2V using Honor Pulldown Flags.
    So I'm still working it out, but it would be nice to get some confirmation on whether I'm meant to be reloading the script everytime i make an edit
    Yes, anytime you change the script you have to reopen it in VDub to be able to see the changes you've made. Post your full script if it seems to be doing nothing. To make a script do something, add in "Tweak(Sat=0)" and watch your color movie turn black and white.
    Quote Quote  
  6. Ok making good progress now. Script appears to be working. Regarding the interlacing, the PAL was telecined 2:2:2 etc:3 pull down (what ever that means lol) to convert from 24 to 25fps. So I used:

    AssumeTFF()
    Telecide()

    And now there are no interlaced extra frames in my raw anymore

    Onwards!
    Quote Quote  
  7. Originally Posted by transcendent View Post
    And now there are no interlaced extra frames in my raw anymore
    Well, there wouldn't be, even if Telecide isn't called for, because it has a post processor included by default. It will deinterlace any interlaced frames remaining. You might try:

    Telecide(Post=0)

    and see if just the field matching alone does the same thing. Also, if there's something peculiar about this source, you might make a small sample available for us to have a look.

    If it's really NTSC to PAL without field-blending, after applying Telecide, going to a place with movement and advancing a frame at a time, there should be a duplicate frame every second (25 frames). You could remove it with:

    Decimate(25)

    after the Telecide line.
    Quote Quote  
  8. yeah for some reason earlier vdub had chopped off the last few seconds of my video, don't really know why.
    I played with the telecine commands, post=0 seemed to work fine.
    This is what I was using before through trial and error:

    LoadPlugin("C:\Users\Samantha Carter\Desktop\dgmpgdec158_SSE\dgmpgdec158\DGDecode.dll")
    LoadPlugin("C:\Program Files (x86)\AviSynth 2.5\plugins\MSharpen.dll")
    MPEG2Source("C:\Users\Samantha Carter\Desktop\avisyntheoepro\eoe.d2v")
    AssumeTFF()
    Telecide(guide=0,post=0)
    MSharpen(threshold=13,strength=90,mask=false,highq =true,debug=false)
    MSmooth(threshold=2,strength=2,highq=true,chroma=t rue,mask=false,show=false,debug=false


    I can simplify all of this to:
    MPEG2Source("C:\Users\Samantha Carter\Desktop\avisyntheoepro\eoe.d2v")
    AssumeTFF()
    Telecide(post=1)
    MSharpen(threshold=13,strength=90,mask=false,highq =true,debug=false)
    MSmooth(threshold=2,strength=2,highq=true,chroma=t rue,mask=false,show=false,debug=false)

    (I quite like post=1, and yeah, i can get rid of the explicit argument declaration but nvm for now :P)



    So I took screenshots for your viewing pleasure^^ I don't really understand the interlacing scheme: These are 6 frames first set without Telecide, second with Telecide)


    Click image for larger version

Name:	eoe-non-IVTC.jpg
Views:	124
Size:	184.3 KB
ID:	12932

    Click image for larger version

Name:	eoe-with-IVTC.jpg
Views:	151
Size:	184.8 KB
ID:	12933
    Quote Quote  
  9. I know it took you some time to create the pictures, but pics are pretty much useless. If you have questions about the source you should upload a short 10 second sample. Cut them using DGIndex and the [ and ] buttons to isolate a small piece. Then 'Save Project and Demux Video'. Upload the resulting M2V here or at Sendspace or MediaFire.
    Quote Quote  
  10. rofl. *is a noob*
    I just figured it was a bit odd having 3 interlaced frames followed by 3 progressive, it seemed somewhat irregular.

    Anyhow, for kicks:
    Here's a clip of the video in transport form:
    https://dl.dropbox.com/u/14943889/remuxing/eoe_clip.demuxed.m2v

    and if for some reason you need it:

    https://dl.dropbox.com/u/14943889/remuxing/eoe_clip.d2v
    the indexed thing from that program

    I'm currently using Henc with the avs script and compressing it to m2v. Hopefully it'll be ok...
    Oh yeah, 1 weird peculiarity I came across with reauthoring the DVD, the only program which didn't give me silence on the DTS track, was ReJig. IFO edit/muxman wouldn't let me hear the DTS track for some reason. I think ReJig probably just wrote the ifo properly and I didn't have the skill to make it work in ifoedit though. *shrugs*
    Quote Quote  
  11. That was interesting, thanks. There's all kinds of interlacing patterns in there, but the IVTCs straighten it out easily. You might add in 'Guide=2' to your Telecide line as it's used to help with the field matching of PAL progressive video.

    Muxman accepts DTS audio so, you're right, something strange is going on. Did you check the Muxman log? Did it actually accept the audio as input?
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!