Encoding DVD Movie - Does the DAR Really matter? and 710x480 vs 648x480 ?

[manofsteel31]

OK, so I have been encoding the video over and over in different ways and here is the best that I came up with.

FFMSindex and L-Shamsh Works could not handle the .mpls file but DirectShowSource would. However DirectShowSource Analyzed it wrong, so:

I muxed the mpls file into an MKV and used FFMSIndex to analyze it as Mostly Interlaced / Top Field First / TIVTC.

Then I used DirectShowSource to encode the .mpls file using the Analysis I got from the MKV.

It resulted in a 29.970fps video (like the original) and the chandelier scene looks better, however a few other scenes looks slightly worse (like in the room After the Chandelier Scene the Pictures on the wall look Blurry) ..... but its better than it was and at least its watchable now. Without another suggestion, its the best I can do at the moment. Here is my full encode, so let me know what you think.

[manono]

I think it should be bobbed using QTGMC. If you can't use QTGMC (why not?) then 'TDeint with EDI and BOB' is okay. Your 23.976fps parts look horrible (blended/blurry/flickering), the 29.97fps stuff looks okay, and the 36fps parts play jerky. And the bitrate is too low, I think.

720p can be 59.94fps for something like Blu-Ray.

[jagabo]

AviSynth script:

ffVideoSource("Original - 30fps (should be 29.970fps).mkv", fpsnum=30000, fpsden=1001, seekmode=0) 
AssumeTFF()
Yadif(mode=1, order=1) # or QTGMC()
BicubicResize(1280,720)

Encode at 59.94 fps.

[hello_hello]

I didn't realise the previous two posts were there when I posted this one, so here's another method.....

I had a bit more of a play with your sample. It's been badly converted from one format to another (or as jagabo said maybe from one frame rate to another) but the first part (outside shots of the clock) seem different to the section where the camera first heads through the door. I think there it might be interlaced as double frame rate de-interlacing does produce unique frames. Where the camera pauses on the back wall there's a "glitch" of some sort as it pans left toward the chandelier, where from then on it's back to the same frame blending, frame rate conversion as the outside part (the camera moving around the chandelier will never be perfectly smooth as the original frames have been butchered and it's no longer perfectly smooth), and then where the camera moves through the door into the second room I think it changes again.

I was wrong originally when I said it required IVTC. It gives that impression in places but after the conversions it was put through I think it's mostly progressive video that was then encoded as interlaced...... or something.

Anyway..... simple field matching seems to clean up all the combing and give you progressive video about as smooth as it's going to get, I think.

FFVideoSource("E:\Original.mkv", fpsnum=30000, fpsden=1001, threads=1)
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TIVTC.dll")
tfm(order=1)

If you use the above, you'll find that the very end part, after the camera moves into the second room and pans across the wall, it mightn't look perfectly smooth. That's because every one frame in five is repeated and that's the part I think convinced me it needed IVTC. If you apply IVTC though, the rest of it won't look as smooth as frames will have been removed.

Normally, when you de-interlace, the MeGUI option without Bob de-interlace to "half frame rate", or 29.970fps interlaced becomes 29.970fps progressive. The option with Bob give you full frame rate so 29.970fps interlaced becomes 59.940fps. In the case of your sample though, full frame rate de-interlacing mostly doesn't produce unique frames. Mostly, it results in every second frame being repeated because I think mostly, the sample isn't interlaced.

Attached is your sample re-encoded using the following script. I re-encoded it as 720p as 1080p seems like overkill given the lack of resolution.
See how that looks to you. Whether it'd work for the entire video, I don't know. You really picked a fun sample to start with.

LoadPlugin("C:\Program Files\MeGUI\tools\ffms\ffms2.dll")
FFVideoSource("E:\Original.mkv, fpsnum=30000, fpsden=1001, threads=1)
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TIVTC.dll")
tfm(order=-1)
Spline36Resize(1280,720)

So what is wrong with using the "DirectShow" Decoder? Which one do you recommend? In MeGUI I appear to have the following options (that I am aware of):
FFmpeg AC3 or Aften AC3. I have been using FFmpeg AC3 (though not sure of the difference).
Then for Decoders I have a choice of:
-NicAudio
-FFAudioSource
-DirectShow
-BassAudio
-LWLibavAudioSource
If I should not be using DirectShow, then which would you choose? Or is there a way to add another one which is better?

If you're setting up the audio decoder configuration, don't choose. Just leave the default selection and MeGUI will step through each one in the order it prefers until it finds one that can decode the audio in question.
DirectShow is usually the last resort but that's largely because DirectShow decoders are out of MeGUI's control. For example if you have ffdshow installed as a DirectShow decoder it'll work fine, but if you've enabled it's mixer or volume filters etc, something that'll alter the audio, MeGUI has no way of knowing. That's fine if you deliberately want to use those filters, but not if you don't want to. Plus there's no guarantee you'll have a DirectShow decoder installed for a particular format.

Another Question. I have always encoded by Bluray movies by selecting the ".mpls" file for the movie. OneClickEncoder and DirectShowSource can both handle this type of file, but it appears that when I try to open it in "FileIndexer" that neither FFMSIndex or L-Smash can. FFMSIndex errors out when trying to create the .ffindex file, and L-Smash completely crashes MeGUI when it tries to create its .lwi file. So if you are content that I always use the FileIndexer, then how do you recommend that I do it? Should I be using the .m2ts file instead or should I be using MKVMerge or MeGUI's Stream Extractor to mux the video into a MKV first? OR should I just use DirectShowSource for them since they are usually Progressive anyway and no Analyzing for Deinterlacing is typically needed?

I usually rip using the HD Streams Extractor under the Tools menu. You select the appropriate mpls file (from the disc) and it'll display all the associated streams. You select the ones you want and the audio can be converted to another format on the way through if you like, but mostly the video is ripped to an MKV and the audio and subtitles and chapters etc are ripped to individual files. From there I open the MKV for re-encoding and mux it all together later on. I've never done it any other way. It requires AnyDVD HD running in the background doing the decrypting.
Even if you rip the whole disc first, I'd imagine you're still going to be better off opening the ripped files and extracting the required streams the same way, but if you have AnyDVD it makes more sense to rip and extract at the same time.

I think mpls files are kind of like DVD ifo files. They only point to the video and audio, but don't actually contain any, so while you might be able to open them via DirectShow, you can't index them as such.

[manofsteel31]

I think it should be bobbed using QTGMC. If you can't use QTGMC (why not?) then 'TDeint with EDI and BOB' is okay. Your 23.976fps parts look horrible (blended/blurry/flickering), the 29.97fps stuff looks okay, and the 36fps parts play jerky. And the bitrate is too low, I think.

720p can be 59.94fps for something like Blu-Ray.

OK, I encoded it the way you asked me to. I encoded it as "Interlaced / Top Field First / TDeint with EDI and BOB". It said the output was going to be 59.94fps, but for some reason it ended up as 60fps instead. I would also like to note that this is a 7-8Minute Video and it took '5 HOURS" to encode it this way. That is absolutely insane considering I am running a i7-4770K clocked at 3.9Ghz with 32GB of RAM. I can usually do a 2 hour movie in 2 hours of encoding time, or 4 hours when using 2-Pass. But 5 Hours for a 7Minute Video??? Why does "TDeint with EDI and BOB" add so much extra time?

Oh, and QTGMC is not a standard option in MeGUI and I am not sure how to add it or use it.

Anyway, the video is attached to this post. I kept it at 1920x1080 because I prefer it that way.

You really picked a fun sample to start with.

Yeah, Sorry about that, LOL. I wasnt trying to find an overly complicated video, it just turned out that way. Of course, I guess if I can figure out one as complicated as this then the rest should be a piece of cake, lol.

LoadPlugin("C:\Program Files\MeGUI\tools\ffms\ffms2.dll")
FFVideoSource("E:\Original.mkv, fpsnum=30000, fpsden=1001, threads=1)
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TIVTC.dll")
tfm(order=-1)
Spline36Resize(1280,720)

I think you meant to say "tfm(order=1)" right, since it should be Top Field First?

Oh, and are these the settings you used to encode the video you posted because it still looked really choppy in the chandelier scene.

Just leave the default selection and MeGUI will step through each one in the order it prefers until it finds one that can decode the audio in question.

I dont understand this. When I go into the FFmpeg AC3 and click "Load Defaults" the decoder changes to the first on on the list which is "NicAudio". Are you saying that leaving it on NicAudio will allow it to try all possible decoder options but leaving it on "DirectShow" will not? I dont understand that.

I usually rip using the HD Streams Extractor under the Tools menu. You select the appropriate mpls file (from the disc) and it'll display all the associated streams. You select the ones you want and the audio can be converted to another format on the way through if you like, but mostly the video is ripped to an MKV and the audio and subtitles and chapters etc are ripped to individual files. From there I open the MKV for re-encoding and mux it all together later on. I've never done it any other way. It requires AnyDVD HD running in the background doing the decrypting.
Even if you rip the whole disc first, I'd imagine you're still going to be better off opening the ripped files and extracting the required streams the same way, but if you have AnyDVD it makes more sense to rip and extract at the same time.

I think mpls files are kind of like DVD ifo files. They only point to the video and audio, but don't actually contain any, so while you might be able to open them via DirectShow, you can't index them as such.

Yeah, .mpls files seem to be some type of Pointer-Files (for lack of a better term). I dont really understand them myself and some Bluray disc have more than one of them that point to the exact same .m2ts file and look identical in HD Streams Extractor but the .mpls files themselves are actually different sizes. I dont get that. I have come across some .m2ts files that play with NO AUDIO if you try to play them directly, but play fine when played from the .mpls file, even though MediaINFO shows the audio is clearly part of the .m2ts file. Again .... baffled.

I always use AnyDVD HD to rip the entire bluray to my Harddrive. That makes it easier for me.

OK, so you like to deal with muxed video as apposed to the .mpls files. Thats usually fine except of course for my current video which keeps showing a different FPS when muxed.

I have the "dtsdecoderdll.dll" file in my eac3to folder in MeGUI thinking I needed it to extract and convert DTS audio, but then I came across a post the other day saying that it might not even be working unless I activated it so I followed these instructions. I got a confirmation message after doing it, but no idea if it will help at all with encoding DTS audio directly. Here is what I found:

Simple method to install ArcSoft DTS Decoder (1.1.0.7) in eac3to:

(From Steel's instructions):

1. Copy the following files to your eac3to directory.

ASAudioHD.ax
checkactivate.dll
DtsDec.dll
dtsdecoderdll.dll
MagCore.dll
MagPCMac.dll
MagUIEngine.dll
MagUIInter.dll

2. Run regsvr32.exe ASAudioHD.ax from a command line in your eac3to folder.

Have you done this yourself?

[hello_hello]

LoadPlugin("C:\Program Files\MeGUI\tools\ffms\ffms2.dll")
FFVideoSource("E:\Original.mkv, fpsnum=30000, fpsden=1001, threads=1)
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TIVTC.dll")
tfm(order=-1)
Spline36Resize(1280,720)

I think you meant to say "tfm(order=1)" right, since it should be Top Field First?

No, I left it at -1 because that specifies a varying filed order. If top field first works better then use it, but I suspect it won't.

Oh, and are these the settings you used to encode the video you posted because it still looked really choppy in the chandelier scene.

It is really choppy in the chandelier scene. It's been butchered. De-interlacing that section to double frame rate (Yadif with Bob) doesn't look any better to me because even though there's twice as many frames there's still lots of repeated frames so it's still choppy. De-interlacing might look a tad smoother than just using tfm() but due to the de-interlacing it's probably a tad more blury too. Step through the frames in MeGUI's preview during that section (without any de-interlacing enabled) and you'll see how uneven the motion is. The best you can probably hope for is not to make it any worse.

Just leave the default selection and MeGUI will step through each one in the order it prefers until it finds one that can decode the audio in question.

I dont understand this. When I go into the FFmpeg AC3 and click "Load Defaults" the decoder changes to the first on on the list which is "NicAudio". Are you saying that leaving it on NicAudio will allow it to try all possible decoder options but leaving it on "DirectShow" will not? I dont understand that.

If you leave it on NicAudio MeGUI will start with that and if it fails, it'll try the next decoder, then the next etc, until it successfully opens the audio. The log file will tell you which decoders it tried and which was successful. I'm not sure if the order MeGUI tries them in is exactly the same as they're listed, but if you pick a different decoder (ie DirectShow) it'll try that one first. That's why the option is labelled "preferred decoder". There's no guarantee it'll be used, but if the audio can be opened with the selected decoder, it will be.

Yeah, .mpls files seem to be some type of Pointer-Files (for lack of a better term). I dont really understand them myself and some Bluray disc have more than one of them that point to the exact same .m2ts file and look identical in HD Streams Extractor but the .mpls files themselves are actually different sizes. I dont get that. I have come across some .m2ts files that play with NO AUDIO if you try to play them directly, but play fine when played from the .mpls file, even though MediaINFO shows the audio is clearly part of the .m2ts file. Again .... baffled.

I really don't understand them myself to be honest.

Simple method to install ArcSoft DTS Decoder (1.1.0.7) in eac3to:

(From Steel's instructions):

1. Copy the following files to your eac3to directory.

ASAudioHD.ax
checkactivate.dll
DtsDec.dll
dtsdecoderdll.dll
MagCore.dll
MagPCMac.dll
MagUIEngine.dll
MagUIInter.dll

2. Run regsvr32.exe ASAudioHD.ax from a command line in your eac3to folder.

Have you done this yourself?

No, I've never tried it. I didn't think it was free. Eac3to will still decode DTS without it, as will MeGUI, and foobar2000 will do likewise. I don't know what the differences are between the decoders they us and the ArcSoft decoder. I probably did at one stage, but I've forgotten.

[manono]

I would also like to note that this is a 7-8Minute Video and it took '5 HOURS" to encode it this way.

Use Yadif then. Much faster. If you're going to complain about the length of time the encoding takes, then you won't like QTGMC either.

[manofsteel31]

I would also like to note that this is a 7-8Minute Video and it took '5 HOURS" to encode it this way.

Use Yadif then. Much faster. If you're going to complain about the length of time the encoding takes, then you won't like QTGMC either.

Sorry if it sounds like I am complaining, but I legitimately thought something was wrong because I was not prepared for it to take that long considering the short length of the video. I have never used "TDeint with EDI and BOB" before and only did so under your advice. However, did you take a look at the resulting video I uploaded (at the bottom of my last post)? I would certainly like your input to know if that is the result you were expecting.

[jagabo]

did you take a look at the resulting video I uploaded (at the bottom of my last post)?

The chandelier scene was smooth, aside from the one jerk from the duplicate interlaced frame in the source.

Attached is a quick encoding of the script I posted earlier. I used QTGMC(preset="fast") instead of Yadif(). I tried to attach this sample in that earlier post but uploads weren't working then.

[manono]

However, did you take a look at the resulting video I uploaded (at the bottom of my last post)?

Yes. That's why I said the 23.976fps parts looked terrible and the 36fps parts (beginning at about 42 seconds in) played jerky (missing every 5th frame). I suggested the EEDI way because it's the best of the bobbers you have easily available. I usually use QTGMC for such things

You're working with 1080p and slow filters. That combination will bring the fastest computer to its knees. You could try and get the multi-threaded version of AviSynth going, I suppose, but it might not be easy.

[manofsteel31]

However, did you take a look at the resulting video I uploaded (at the bottom of my last post)?

Yes. That's why I said the 23.976fps parts looked terrible and the 36fps parts (beginning at about 42 seconds in) played jerky (missing every 5th frame). I suggested the EEDI way because it's the best of the bobbers you have easily available. I usually use QTGMC for such things

You're working with 1080p and slow filters. That combination will bring the fastest computer to its knees. You could try and get the multi-threaded version of AviSynth going, I suppose, but it might not be easy.

Your comment about the 23.976 parts looking terrible was in regards to the video I posted in POST #91. I encoding the video again after that, the way you asked me to, and posted it in POST #95 (at the bottom). Did you take a look at the one in post 95 yet?

The file is named "00043-Interlaced-TopFieldFirst-TDeintWithEDIandBOB.mkv".

[manofsteel31]

did you take a look at the resulting video I uploaded (at the bottom of my last post)?

The chandelier scene was smooth, aside from the one jerk from the duplicate interlaced frame in the source.

Attached is a quick encoding of the script I posted earlier. I used QTGMC(preset="fast") instead of Yadif(). I tried to attach this sample in that earlier post but uploads weren't working then.

Thats not bad. The motion seems smoother, at least in the one chandelier scene.

When I search for QTGMC I end up here:

http://avisynth.nl/index.php/QTGMC

I am not sure which package to download, or how to add it into MeGUI.

[manono]

Your comment about the 23.976 parts looking terrible was in regards to the video I posted in POST #91.

You're right. Sorry. No, I haven't looked at the EEDI one.

[manofsteel31]

Your comment about the 23.976 parts looking terrible was in regards to the video I posted in POST #91.

You're right. Sorry. No, I haven't looked at the EEDI one.

I was going to ask if you ever plan to take a look at it for me, but it no longer matters since my TV is not even able to play it (which is where I primarily watch all my videos). When I try and play it I get a "Not Supported Video Codec" Error message.

Since it is encoded from the same source video, in the same container (MKV), using the same x264 codec as all my other trial encodes of this particular video, why is this one not supported? The only difference is the Deinterlacer used (i think you mentioned earlier that it might be a problem, but I am just trying to understand why). The "codec" should refer to "x264/h264" right? So is the TV just spitting out a generic error message (blaming the codec) because it does not know how to read a video deinterlaced in that method?

Just curious.

[hello_hello]

It shouldn't have anything to do with the de-interlacing as such, unless the frame rate ends up too high. The TV's media player likely supports h264's High Profile, Level 4.1 as a maximum. It might try to play anything exceeding that, or it might spit it out, but Level 4.1 support seems fairly standard these days. http://en.wikipedia.org/wiki/H.264/MPEG-4_AVC#Levels
If you de-interlaced to 60fps, maybe the TV will play it at 720p but not 1080p. I think mine plays higher frame rates at 1080p but I'd have to check.
Edit: I remembered the topic being discussed a while back and I experimented at the time. Looks like my TV won't play frame rates above those officially supported by Level 4.1 but the Bluray players will.
https://forum.videohelp.com/threads/367248-mov-files-perfect-playback-via-USB-but-not-H...=1#post2348419

I didn't download your sample from post #95 as it's large and my internet connection is quite slow, but you could check the profile/level of the encoded video with MediaInfo. Use MediaInfo's View/Text mode and post the details here if in doubt.

Honestly, if you can re-encode that video at 1080p and at 720p and pick a difference between them running fullscreen on your TV I'd be surprised. If there is any difference it'd have to be very marginal.

[manofsteel31]

It shouldn't have anything to do with the de-interlacing as such, unless the frame rate ends up too high. The TV's media player likely supports h264's High Profile, Level 4.1 as a maximum. It might try to play anything exceeding that, or it might spit it out, but Level 4.1 support seems fairly standard these days. http://en.wikipedia.org/wiki/H.264/MPEG-4_AVC#Levels
If you de-interlaced to 60fps, maybe the TV will play it at 720p but not 1080p. I think mine plays higher frame rates at 1080p but I'd have to check.
Edit: I remembered the topic being discussed a while back and I experimented at the time. Looks like my TV won't play frame rates above those officially supported by Level 4.1 but the Bluray players will.
https://forum.videohelp.com/threads/367248-mov-files-perfect-playback-via-USB-but-not-H...=1#post2348419

I didn't download your sample from post #95 as it's large and my internet connection is quite slow, but you could check the profile/level of the encoded video with MediaInfo. Use MediaInfo's View/Text mode and post the details here if in doubt.

Yeah, I already know about encoding to certain levels. For example, I usually encode using 5 Reference Frames which is fine for most of my videos, but when i get to larger sizes like 1920x1040 or 1920x1080 I encode with 4 reference frames instead in order to keep the Level down.

I know my TV cant handle Level 5.0+, but I thought it could handle up to Level 4.2. Apparently I was wrong as I just looked up the specs (while writing this response) and found out it maxes out at Level 4.1. That probably explains it considering the level of the "EDI and BOB" encode resulted in Level 4.2.

I also never really knew (or paid attention to the fact) that Frame Rate directly affected the level. I think that because almost all my videos are 23.976 that I really only paid attention to frame size and # of Reference frames used to determine the Level. Now I know that Frame Rate can be a problem too. I will keep that in mind, although I dont plan on encoding much at 60fps. This was a Unique Situation...... at least I hope it was.

Anyway, just to test the theory, I RE-encoded the (already encoded) "EDI and BOB encoded video" two times. Once to 1920x1080 (which really didnt change anything) and again to 1280x720. The 720p version played fine on my TV, but the 1080p version still wouldnt because it was still at Level 4.2 obviously. Assuming you think it looks good, I will encode it properly (as apposed to the re-encode) to 1280x720p, but didnt want to wait another 5 hours (for the encode to run) just to test this theory.

Since you said you have slow internet, I made a sample of the 720p re-encoded video for you to look at. Its about the first 1m13s and only 25MB so hopefully thats small enough for you. Please take a look and let me know what you think.


***Just for fun, I looked up the specs on a fairly new Samsung 4K TV (model#UN55JU650DFXZA) just to see up to what Level it could handle. It did NOT note the level specifically, but said it could read h264 videos up to "4096X2160 @ 24fps and 3840X2160 @ 30fps". After looking at the h264 wikipedia page, I can see that is at least Level 5.0 but I am not sure if it meets the requirement for Level 5.1 or not. Regardless, it certainly can NOT handle Level 5.2 videos. Thats strange as I would think newer technology would be able to handle the max 5.2 Level. It makes me wonder what devices (TV or otherwise) actually can. However, as a plus the TV in question did say it handles the new h265 encodes up to 4096x2160 @ 60fps, so thats good in terms of looking towards the future. I have never encoded anything using x265 personally because I didnt think i could play it on anything but my computer, which is NOT where I like to watch videos.

[hello_hello]

As well as frame rate and ref frame restrictions, there's also bitrate restrictions for the various levels. If you set the appropriate Level in MeGUI's x264 encoder configuration (check "show advanced settings" in order to set it manually), the encoder itself will make sure the number of reference frames doesn't exceed the allowed maximum for the video resolution. If you're using a x264 speed preset that would normally use a higher number of reference frames than allowed, setting the Level manually should limit it so you don't need to worry about it yourself.
If you select DXVA as the target playback device, that's the same as setting High Profile, Level 4.1. The command line should look like this:

program --level 4.1 --vbv-bufsize 78125 --vbv-maxrate 62500 --output "output" "input"

The vbv settings control the bitrate (technically the video buffer). Most hardware players don't support the full vbv settings for Level 4.1 even though they officially support level 4.1
I don't think I've ever encoded a video where the bitrate climbed high enough for it to be an issue though (not counting when I deliberately encoded that way for testing) but for stutter free playback using my TV's media player I needed to use the following when I tested:

program --level 4.1 --vbv-bufsize 50000 --vbv-maxrate 50000 --output "output" "input"

If you set the Level to 4.1 and encode using a higher resolution or frame rate than Level 4.1 supports, I don't think the encoder will refuse, but you'll find a warning in MeGUI's log file. The TV may then be willing to play the 1080/60p encodes, or not...... mine won't but the Bluray player will.

If you don't set a level manually, the encoder will do it for you according to the resolution and frame rate etc.

The TV you looked at appears to support level 5.1 to me. Level 5.0 doesn't officially include the 4k resolution of 4096×2160 (UHD isn't really 4k, it's 3840×2160). It'll be a while before I start using h265 myself. From what I understand the x265 encoder is still very slow and not all that much better than x264 yet so it's not really ready for prime time, and very few hardware players support h265. One day I'll no doubt be using it, but for the moment I'm sticking with x264.

The sample looks pretty good to me. I think maybe it's a little smoother than jagabo's QTGMC encode. It's never going to be perfect as the video is a mess, but that looks pretty good. I wouldn't encode it at 1080p myself as aside from the frame rate issue there's simply nowhere near 1080p worth of picture detail. I doubt you'd lose any by resizing down to 720p and I'd rather have 720p at 60fps than 1080p at 30fps, at least for that video.

[manofsteel31]

Yeah I noticed looking closer at the Wiki page that there are indeed bitrate restrictions as well, but considering the max bitrate I have been using is 3,105Kbps, I dont think that is something I need to worry about right now. Its really the frame size and reference frames that most effect me (and apparently frame rate given the recent 60fps video).

OK, I see that in the x264 configuration you can change the "AVC Level" so I could specify it as "Level 4.1" if I wanted. Currently I have it set to "Unrestricted/auto guess". The real question is, if I were encoding a 1920x1080p video for example and my settings were set to 5 Ref Frames, would MeGUI ONLY change the Ref Frames down to 4, OR would it also change any of my other settings? I have completely customized all the Advanced settings in the x264 Configuration based on setting another encoder uses (who's work I really like, so I modeled my one style after his), so I dont want any of those other settings to change. I am just worried it might change something else as well. If so, it is not that big of a deal for me to simple change the Reference Frames from 5 to 4 every time I encode a a video with a frame size of 1920x1040 or 1920x1080. I just want to be sure before I change anything.

Just curious .... what is the max level your Blu-Ray player can handle? 5.0, 5.1 or 5.2? I doubt it can handle 5.2, but then again I dont currently know of any device that can beyond a computer (although I am sure they exist).

Glad to hear you think the video looks good. I was watching a few different reference points in my various encodes of the video (the chandelier scene motion, and the banner over the archway in the first roughly 10 second, and even the pictures on the walls in the room after the chandelier scene), and all points looked pretty good I think. So I guess since you agree, I will consider it good enough.

While dealing with this video, I have been encoding some other content as well, and I find non-progressive video to be very challenging for me to encode.

When the video is analyzed as Progressive, then all I do is crop and resize if needed, select the bitrate and DONE! Very easy.

However when it does NOT get analyzed as progressive, I have a hard time determining how to handle it. I always try to use whatever settings MeGUI chooses for me based on it's Analysis, but I am not always happy with the end result, so I feel that it is being Analyzed Wrong.

In fact, when I have a .VOB video file from a DVD and I analyze it, then I mux the same orginal .vob into a .MKV then then analyze the MKV, I get Completely Different Results from each analysis. How is that possible since the video is the same, its just in different containers? At that point I encode the video twice based on each result and try and chose which one is best. However sometimes neither one looks good, at which point I have NO IDEA what to do. I try "guessing" and randomly picking different combinations, essentially encoding the same video many times in different ways to get a result I am happy with. Obviously there is a much better way since people like you (@hello_hello) as well as @manono and @jagobo seem to be able to figure out exactly what it needs just by looking at it. I am not sure how you all do that.

Some of my videos result in a blurry/shaky way during motion scenes that gives me a headache to watch. Others end up with "ghosting" effects, and then others have the horizontal lines over fast motion.

Here is an example: I was trying to encode Season 1 of the Simpsons.

For the first episode, I opened the VOB with File Indexer and DGIndex reads the video as "Progressive" yet when I analyze it I get the recommended settings of "Mostly Film / Top Field First / TIVTC". I added in the "Minimal Noise" filter option and then just cropped, resized and encoded. It seemed to look OK to me, so I moved on to episode #2.

For episode #2 I again loaded and analyzed the episode in the same way, but this time the recommended settings were "Partially Interlaced / Varying Field Order / Yandif". I thought it was a little strange that the settings were so different from the first episode but I encoded it that way anyway. However I did not like the output as it had "Color Ghosting". I might have just made that term up, LOL, so let me explain. There was no actual ghosting, but in a few scenes where the baby maggie was moving her head (where there was normally ghosting in the original VOB) I could see the Yellow color from her head move to the next frames position before the outline of her head. So it was like the color was ghosting. Regardless of the reason, I was not happy with it.

So, I muxed the original VOB into a MKV and loaded that with FFMSIndex, and after analyzing that I was recommended with "Mostly Interlaced / Top Field First / TIVTC". However that resulted in actual Ghosting of the video. So that didnt work either. So at that point I did not know what to do. I couldnt believe that the analysis for the VOB and MKV was do different and that neither option worked.

So I started google searching for the simpsons season 1 encoding questions and came across another thread on this site talking about it. manono was actually one of the people who answered this guys question and he told him to use:

Yadif(Mode=1)
Srestore()

I tried to look up how to use "Srestore()" and how to implement it into MeGUI, but all I could find was the "avsi" file and I think i need a "avs' or "dll" file, not really sure. So instead, I tried "Partially Interlaced / Top Field First / Yandif" which I thought was the same except of course for the srestore. However the result ended up with the "color ghosting" again.

So then I tried to encode episode #2 with the same settings that seemed to work for episode #1 "Mostly Film / Top Field First / TIVTC". The result seemed to fix the ghosting and color-ghosting problems and was better than all the other encodes. So am I to assume that I should use the same settings on all the other episodes from the same season 1 disc, regardless of how MeGUI analyzes them?

I swear this is confusing me. I wish all video was Progressive so things could be much easier. I wonder if all iTunes content is progressive or not because if so then they would be a "easier to deal with" encoding source even though I am sure the initial bitrate will not be as good as the DVD source.

Anyway, so how am I supposed to determine how to deinterlace/detelesine/ etc a video if I can't trust MeGUI to analyze it correctly?

[manono]

When the video is analyzed as Progressive, then all I do is crop and resize if needed, select the bitrate and DONE! Very easy.

An automatic analyzer may be wrong if the source is the least bit problematic. It may get it right only on the pure progressive ones.

For the first episode, I opened the VOB with File Indexer and DGIndex reads the video as "Progressive"

Meaning what? You ran the Preview and it showed as 'Progressive' all the way? Or you had a look at the bottom of the D2V you made and it was 100% film?

... then I mux the same orginal .vob into a .MKV then then analyze the MKV, I get Completely Different Results from each analysis. How is that possible since the video is the same, its just in different containers?

Because you put it in a different container. You should learn to analyze these things yourself, particularly if you like to encode peculiar sources from companies other than the major movie studios.

Doing the encoding from the MKV is just asking for trouble when the source is a DVD. The sooner you learn that the better off you'll be.

[hello_hello]

Yeah I noticed looking closer at the Wiki page that there are indeed bitrate restrictions as well, but considering the max bitrate I have been using is 3,105Kbps, I dont think that is something I need to worry about right now. Its really the frame size and reference frames that most effect me (and apparently frame rate given the recent 60fps video).

I assume you've been specifying a 3105kbps average bitrate? Even so though, with that sort of average bitrate the peak bitrate is not going to be be a problem.

OK, I see that in the x264 configuration you can change the "AVC Level" so I could specify it as "Level 4.1" if I wanted. Currently I have it set to "Unrestricted/auto guess". The real question is, if I were encoding a 1920x1080p video for example and my settings were set to 5 Ref Frames, would MeGUI ONLY change the Ref Frames down to 4, OR would it also change any of my other settings?

MeGUI doesn't limit the number of reference frames according to the level, the encoder does. I don't think the encoder limits anything else, which is why when you select Level 4.1 in the encoder configuration MeGUI adds --vbv-bufsize 78125 and --vbv-maxrate 62500 to the command line, because the encoder doesn't even enforce those limits without being told to.
The only thing MeGUI does automatically I can think of, is to set the keyint according to the frame rate, although you can even over-ride that if you like. Thinking about it, MeGUI would probably set keyint to 600 for 60p video (it's generally a maximum of 10 seconds, so 60 x 10 = 600). That may effect the ability of some players to play 60p video. I'm not sure I've tested it. You could try setting "GOP Calculation" to "fixed' and the maximum GOP size (which is the -keyint setting) to 300 in the encoder configuration see if that makes a difference, although I doubt it will. I think I've always left it on "fps based" and let MeGUI worry about it.

I have completely customized all the Advanced settings in the x264 Configuration based on setting another encoder uses (who's work I really like, so I modeled my one style after his), so I dont want any of those other settings to change. I am just worried it might change something else as well. If so, it is not that big of a deal for me to simple change the Reference Frames from 5 to 4 every time I encode a a video with a frame size of 1920x1040 or 1920x1080. I just want to be sure before I change anything.

To be honest I don't think there's any magic settings and I've seen lots of ridiculous recommended settings on the internet. The general rule of thumb is to use an appropriate x264 tuning, pick a quality (or bitrate if you're living in the dark ages ) and the slowest speed preset you can stand. Any tweaking from there tends to require tweaking for each individual video.

Edit: I looked at the settings you used for the sample you uploaded and if that's what you normally use then you're mostly not far off the defaults. I don't think there's any changes likely to hurt the quality in a noticeable way and none that'd improve it either. Your --keyint settings are odd though. I wonder what the logic there is? (keyint=576 and keyint_min=55)

Just curious .... what is the max level your Blu-Ray player can handle? 5.0, 5.1 or 5.2? I doubt it can handle 5.2, but then again I dont currently know of any device that can beyond a computer (although I am sure they exist).

The Samsung player (about 4 years old) seems to handle 4.2 as long as it's encoded as 4.1 (ie if I encode 1080/50p it'll play it as long as I set the level to 4.1 when encoding). The Sony Bluray player in this house (about the same age) will play 4.1 and 4.2. I haven't tested anything higher.

While dealing with this video, I have been encoding some other content as well, and I find non-progressive video to be very challenging for me to encode.

When the video is analyzed as Progressive, then all I do is crop and resize if needed, select the bitrate and DONE! Very easy.

Isn't it?
Living in PAL-Land I've generally got it easier. Video tends to be progressive or interlaced, and there's nowhere near the butchery to deal with that's inflicted on NTSC at times. There is a form of IVTC needed very ocassionally, but mostly, it's de-interlace or don't de-interlace. Much easier.

However when it does NOT get analyzed as progressive, I have a hard time determining how to handle it. I always try to use whatever settings MeGUI chooses for me based on it's Analysis, but I am not always happy with the end result, so I feel that it is being Analyzed Wrong.

You should trust your eyes and if you're not sure MeGUI is getting it right, upload a sample here. I've done it a few times when I'm stumped. Eventually you learn to work it out for yourself (most of the time).

In fact, when I have a .VOB video file from a DVD and I analyze it, then I mux the same orginal .vob into a .MKV then then analyze the MKV, I get Completely Different Results from each analysis. How is that possible since the video is the same, its just in different containers?

The analysis can only analyse the decoded video, so how it's decoded can make a difference.
Speaking of which, go into MeGUI's options and look for the "force film" option and disable it. The idea there is if DGIndex decides 95% or more (by default) is "film" it'll ignore any pulldown flags and output progressive video. Sometimes that's great, but it changes the analysis too. I disable that option. If you do change it, you need to re-index with DGIndex and let it create a new index file. And always index vob files with DGIndex rather than remux as MKV first. Always.

At that point I encode the video twice based on each result and try and chose which one is best. However sometimes neither one looks good, at which point I have NO IDEA what to do. I try "guessing" and randomly picking different combinations, essentially encoding the same video many times in different ways to get a result I am happy with. Obviously there is a much better way since people like you (@hello_hello) as well as @manono and @jagobo seem to be able to figure out exactly what it needs just by looking at it. I am not sure how you all do that.

Experience. The other guys are better at it than me because I live in PAL-Land.

Some of my videos result in a blurry/shaky way during motion scenes that gives me a headache to watch. Others end up with "ghosting" effects, and then others have the horizontal lines over fast motion.

It sounds like maybe hybrid video not being handled the best way. Just post a sample of the original.

Here is an example: I was trying to encode Season 1 of the Simpsons.

Animation is hard to analyse because it tends to have sections which are really (for example) 12fps with every frame repeated, then sections where every frame is different etc and the analysis doesn't always know what to make of it.

So I started google searching for the simpsons season 1 encoding questions and came across another thread on this site talking about it. manono was actually one of the people who answered this guys question and he told him to use:

Yadif(Mode=1)
Srestore()

It sounds like the video was converted from one frame rate to another using frame blending, which is fairly horrible. The analysis can't help with that. SRestore is designed to undo frame blending and it can do a very good job, although I don't think I've used it for animation myself. I think manono would probably be the resident SRestore expert.

I tried to look up how to use "Srestore()" and how to implement it into MeGUI, but all I could find was the "avsi" file and I think i need a "avs' or "dll" file, not really sure. So instead, I tried "Partially Interlaced / Top Field First / Yandif" which I thought was the same except of course for the srestore. However the result ended up with the "color ghosting" again.

You need to put the avsi file in the Avisynth plugins folder (the installed Avisynth, not MeGUI's Avisynth plugins folder) and then it'll be loaded automatically. You also need to put the required plugins in the same folder (dlls). They're listed on the SRestore Avisynth page but I've attached them to this post to get you started. RemoveGrain.dll is specified, but the zip file has RgTools.dll as it's a newer replacement. Use it instead.
With everything in the attached file in the plugins folder, add SRestore() to a script and cross your fingers. If you don't get an error, you've done it all correctly.

For the record, you'd get QTGMC working the same way. In fact......
Because I'm an incredibly nice guy and finding the right plugins can be confusing as there's often several versions, I've put all the required plugins for QTGMC (hopefully) in the zip file attached to this post too.

I swear this is confusing me. I wish all video was Progressive so things could be much easier. I wonder if all iTunes content is progressive or not because if so then they would be a "easier to deal with" encoding source even though I am sure the initial bitrate will not be as good as the DVD source.

All the itunes video I've seen is progressive.
For TV shows at least, the bitrates are around 4200kb/s for 720p and 1000kb/s more for 1080p. That doesn't seem to vary much. If itunes hosts standard definition video I know nothing about it.

Anyway, so how am I supposed to determine how to deinterlace/detelesine/ etc a video if I can't trust MeGUI to analyze it correctly?

Post a sample.....

[manofsteel31]

Meaning what? You ran the Preview and it showed as 'Progressive' all the way? Or you had a look at the bottom of the D2V you made and it was 100% film?

I was saying that When I select "AviSynth Script Creator" then "File Indexer", there is "File Information" listed within that "File Indexer" Windows which lists the "Codec, Scan Type and Container" although that information seems Sketchy at best since it sometimes changes when its actually Analyzed. I was NOT claiming the video was Progressive because its not, I was just trying to list all the information I noticed.

I assume you've been specifying a 3105kbps average bitrate?

Yes, I believe it is "Average" Bitrate. In the x264 config screen I have 2-pass selected which asks for a "Bitrate" and even though it is not labeled as "average" I believe it is.


Also, thanks for the tip about setting the AVC Level. I will keep it in mind, but at the moment I will probably just continue to adjust the Reference Frames manually depending the the frame size I am encoding to.

Edit: I looked at the settings you used for the sample you uploaded and if that's what you normally use then you're mostly not far off the defaults. I don't think there's any changes likely to hurt the quality in a noticeable way and none that'd improve it either. Your --keyint settings are odd though. I wonder what the logic there is? (keyint=576 and keyint_min=55)

Yes, I use the same settings for everything I encode. I cant answer the question about the "keyint" settings. Like I said, I "borrowed" the settings I currently use from another encoder whose work I have been very happy with. I simply increased the bitrate over what he/she uses. I am not sure if I should really be using those same settings for everything from SD Animation to HD Live-Action, but I have so far (because I dont know any better, lol) and everything has been working out pretty good so far.

go into MeGUI's options and look for the "force film" option and disable it. The idea there is if DGIndex decides 95% or more (by default) is "film" it'll ignore any pulldown flags and output progressive video.

OK, I can try that. Are you saying it should Analyze the video wore accurately when the option is disabled?

And always index vob files with DGIndex rather than remux as MKV first. Always.

OK, manono just said the same thing. I did it because you told me that is how YOU like to encode Blu-ray movies ...... you mux the .m2ts file into a MKV first then analyze and encode it. So I figured I would try the same thing with the VOB file. I wont do it to VOB files anymore since you both said not to, but I have to ask WHY its ok for a Bluray and not a DVD???

You need to put the avsi file in the Avisynth plugins folder (the installed Avisynth, not MeGUI's Avisynth plugins folder)

Okay ............. ??? Are you saying that the AviSynth that comes with MeGUI is not good enough and that I need to install Avisynth separately as well? While asking you this question I noticed that I do indeed have it installed separately (although i dont remember doing it). Why do I need two instances of the same program (one installed and then another that portable and part of MeGUI)? I noticed that the MeGUI version contains a lot of plugins, but the installed version does not. Should I move them all to the installed version? How does MeGUI know which instance to look at?

For the record, you'd get QTGMC working the same way. In fact......
Because I'm an incredibly nice guy and finding the right plugins can be confusing as there's often several versions, I've put all the required plugins for QTGMC (hopefully) in the zip file attached to this post too.

lol ... Thank You for taking the time to collect all the needed files and share them with me to make it easier.

If itunes hosts standard definition video I know nothing about it.

Yes, they host a lot of 480p SD content as well.

Post a sample.....

Okay. I will post a few shortly to get some opinions on them. Thanks.

[hello_hello]

Yes, I use the same settings for everything I encode. I cant answer the question about the "keyint" settings. Like I said, I "borrowed" the settings I currently use from another encoder whose work I have been very happy with. I simply increased the bitrate over what he/she uses. I am not sure if I should really be using those same settings for everything from SD Animation to HD Live-Action, but I have so far (because I dont know any better, lol) and everything has been working out pretty good so far.

Using the same setting for (Simpsons type) animation and live action doesn't necessarily make sense to me. x264 has an animation tuning that bumps up the number of reference frames (and maybe B frames) and changes the deblock and psy defaults to something more suited to animation as well as a film tuning and a grain tuning for really grainy stuff. I use the film tuning for most stuff (it retains a bit more fine detail compared to the default settings) along with the slow speed preset (mostly).

go into MeGUI's options and look for the "force film" option and disable it. The idea there is if DGIndex decides 95% or more (by default) is "film" it'll ignore any pulldown flags and output progressive video.

OK, I can try that. Are you saying it should Analyze the video wore accurately when the option is disabled?

If the force film option is enabled and it's used because 95% or more of the video is seen by DGIndex as being film (23.976 progressive) it'll ignore any pulldown or interlaced flags and always output progressive video. The analysis can't do anything because it only sees progressive video. IVTC filtering won't work properly.

Mostly the force film option works fine because the whole movie is film and maybe the studio crap at the start isn't film..... that sort of thing.... but if you want to be 100% accurate you should disable it. I do because I've had experiences in the past where 99% of a movie is "film" with pulldown flags that can be ignored, but a few small sections have encoded pulldown and they couldn't be IVTCd properly with "force film" enabled.

And always index vob files with DGIndex rather than remux as MKV first. Always.

OK, manono just said the same thing. I did it because you told me that is how YOU like to encode Blu-ray movies ...... you mux the .m2ts file into a MKV first then analyze and encode it. So I figured I would try the same thing with the VOB file. I wont do it to VOB files anymore since you both said not to, but I have to ask WHY its ok for a Bluray and not a DVD???

Good point. I'd not really looked at it like that. Much Bluray video isn't mpeg2 anyway (it's h264 or VC1), but when it is mpeg2 every Bluray I've encoded has been progressive so using MKV/FFMS2 hasn't been a problem. DGIndex is far more reliable for mpeg2 though when it's interlaced or NTSC.

You need to put the avsi file in the Avisynth plugins folder (the installed Avisynth, not MeGUI's Avisynth plugins folder)

Okay ............. ??? Are you saying that the AviSynth that comes with MeGUI is not good enough and that I need to install Avisynth separately as well? While asking you this question I noticed that I do indeed have it installed separately (although i dont remember doing it). Why do I need two instances of the same program (one installed and then another that portable and part of MeGUI)? I noticed that the MeGUI version contains a lot of plugins, but the installed version does not. Should I move them all to the installed version? How does MeGUI know which instance to look at?

If Avisynth is installed there's a registry entry telling it where to find the "installed" Avisynth plugins folder and any plugin inside (except C plugins) will auto-load when Avisynth runs. Even when MeGUI is using the portable Avisynth. It still auto-loads the "installed" Avisynth plugins folder if there is one.
MeGUI can use either the "installed" or it's own portable Avisynth. Most programs can only use the "installed" Avisynth. Having it installed allows other programs to open scripts (ie MPC-HC).
The plugins in MeGUI's avisynth plugins folder are loaded in each script by MeGUI when it needs them. You can copy them to the "installed" Avisynth plugins folder (I wouldn't move them) and it won't matter. They'll auto-load, but because MeGUI loads plugins from it's own plugins folder it'd pay to make sure if there's a plugin in the MeGUI plugins folder, or a plugin MeGUI uses, don't put a different version in the "installed" plugins folder. Hopefully that makes sense.

MeGUI uses the portable Avisynth+ (it's very similar to Avisynth 2.6). For maximum plugin compatibility install AVIsynth 2.6 (the latest version) or Avisynth+.
I have Avisynth 2.6 installed myself, but if you have the old Avisynth 2.5.8 installed remove it and install the new version because some of the newer plugins aren't compatible with Avisynth 2.5.8.

Some plugins are "C" plugins that can't be auto-loaded by the standard Avisynth. To get around that, I load them with an avsi script in the "installed" Avisynth plugins folder. As an example, copy the version of yadif MeGUI uses to the installed plugins folder (it's a C plugin), then to auto-load it put something like this in an avsi file and call it something like yadif.avsi if you like. Obviously change the pathe to match the location of the plugins folder on your PC.

Load_Stdcall_Plugin("C:\Program Files\AviSynth\plugins\yadif.dll")

Then you should be able to use Yadif without loading it in a script as MeGUI does.

Aside from Yadif, I only have two other C plugins. Most plugins aren't C plugins and will auto-load without help.