I ripped my collection of Looney Tunes Golden Collection DVDís using MakeMKV for use with my Plex media server. For some odd reason, when these videos are played via Plex or any other PC media player, jagged lines are visible on the edges of moving objects. At first I though perhaps this was related to an interlacing issue, however after closer examination this appears to be something different. The really odd part is that if I burn these clips back to DVD and play them on my STB DVD player, they look perfect with no hint of these jagged lines. Is my DVD player doing some kind of post processing which my PC isnít? Iíve included a screenshot below.
Any help is appreciated.
[Attachment 44896 - Click to enlarge]
+ Reply to Thread
Results 1 to 25 of 25
It looks like it's being deinterlaced instead of IVTCed (inverse telecine) . Those jaggies are likely deinterlacing artifacts
Plex or other media players without exhibiting these jaggies. Furthermore, if Iím forced to use VLC, then that essentially renders my Plex server of no value. Is it safe to assume then that when played through my STB DVD player, it is using a much better deinterlacer?
Just not sure what the ultimate solution is...
It usually IS safe to assume that (or assume that the deint in your TV is doing the job, and is probably even better than may be what your STB DVD player can do).
You could rip your dvds via AnyDVD, DvdFab, etc to Vobs and run those through DGIndex, and then run the D2V through a deint in AVISynth and out to an encoder as progressive (e.g. x264). That would get you as clean AND as cross-compatible a file as you could expect (though with 1 generation down).
Thanks for all the replies. That said, I tried the suggestion above of playing these videos through VLC on my 4th gen Apple TV and low and behold they look perfect with no jaggies and perfectly smooth when panning. I can only assume that VLC is properly deinterlacing these videos. So at this point, the question is why Plex is not deinterlacining these same video files as I believe Plex is capable?
Are you sure your other DVDs are really interlaced? Most movies are encoded progressive with pulldown flags. Players will just play the progressive frames, ignoring the pulldown flags (that instruct the player how to convert those progressive frames to interlaced video).
Following the soft pulldown flags would result in interlaced frames being sent to the renderer. Ignoring the pulldown flags results in progressive frames being sent to the renderer.
If they were really encoded as 23.976fps with soft pulldown, it's easy enough to strip out those flags, return the video to progressive 23.976fps, and then remux with the audio as an MPG. No reencoding is done and then Plex won't have any trouble playing them. I do this from time to time when uploading what was originally DVD video to YouTube. I used to do it when using Plex to play DVD video from my computer through a Roku box and to the television.
1. Demux audio and video - PGCDemux or DGIndex.
2. Strip out the pulldown flags - run the M2V through DGPulldown set for 23.976->23.976
3. Remux as MPG - Imago-MPEG Muxer.
Last edited by manono; 12th Mar 2018 at 18:18.
Last edited by jagabo; 12th Mar 2018 at 18:29.
You're right and I should have been more precise in how I worded it. Maybe, instead of saying "2. Strip out the pulldown flags" it could have been, "2. Remove the 3:2 pulldown". The RFF flags all become 'false' and the TFF flags all become 'true'. It's the progressive (and 23.976fps) output that's the main thing, so Plex can't screw it up.
MediaInfo still shows the clip as interlaced and when I play it in MPCHC the renderer shows it is displaying at 48 fps, not 24 fps. So MPCHC is bob deinterlacing. I didn't notice any deinterlacing artifacts but the clip I tested with (something that was sitting around) wasn't very sharp so it would have been hard to see them.
I don't know how to tell how it's rendered, but I just did a test myself and everything seems in order:
General Format : MPEG Video Format version : Version 2 File size : 13.2 MiB Duration : 17s 893ms Overall bit rate mode : Variable Overall bit rate : 6 206 Kbps Video Format : MPEG Video Format version : Version 2 Format profile : Main@Main Format settings, BVOP : Yes Format settings, Matrix : Custom Format settings, GOP : M=3, N=12 Duration : 17s 893ms Bit rate mode : Variable Bit rate : 6 206 Kbps Maximum bit rate : 9 500 Kbps Width : 720 pixels Height : 480 pixels Display aspect ratio : 4:3 Frame rate : 23.976 fps Color space : YUV Chroma subsampling : 4:2:0 Bit depth : 8 bits Scan type : Progressive Scan order : Top Field First Compression mode : Lossy Bits/(Pixel*Frame) : 0.749 Time code of first frame : 00:04:11;04 GOP, Open/Closed : Open Stream size : 13.2 MiB (100%)
That was the M2V after being run through DGPulldown. After muxing to MPG it's identical except for the additional audio information. I also use MPC-HC.
Last edited by manono; 13th Mar 2018 at 12:54.
Maybe this is the render information? 23.976fps, right?
I don't see that. Checking them I find a couple giving the framerate and it's not 47.xxx but 23.xxx.
Yeah, sure, but the video is progressive. If you follow the process I described and use a soft telecined DVD as a source, you get a fully progressive video, where you said it shows as interlaced. My MediaInfo posting earlier showed scan type as progressive. So does about any other app I use to have a look. And, if it's progressive, nothing should be either deinterlacing it or bobbing it. So, I don't understand.
I installed GSpot to test and got the same as you for the video for which I had removed the 3:2 pulldown. Then I encoded a short video in HCEnc as progressive 23.976 with no 3:2 pulldown applied. GSpot still showed it as interlaced and 47.952 fields per second. My conclusion is that GSpot is dead wrong. Here's a portion of the HC-Enc log:
frames: 0 - 1174
aspect ratio: 4:3
chroma format: 4:2:0
max. bitrate Kb/s: 9500
pass: 1 (Constant Quant)
constant Q: 4.500
closed gops: no
VBV check: yes
VBV size (bits): 1835008
scene change det: yes
interlaced: no, TFF
goplen,B-pic: AUTO 12
scan method: zigzag
chapter frames: 1
time code: 0 0 0 0
SMP active: yes
matrix: FOX3, adaptive
luminance gain: yes, level 2
adaptive quantization: yes, strength 2
deadzone quantization: auto
Maybe Ahwman can test Plex by removing pulldown as described. He might not want to as there's work involved. I don't use it any longer and don't want to set up a test with it. But I think that's the only way to know for sure if his videos look good played through Plex.