Rate Control Settings for new Staxrip...

[The.King]

Thanks for the input folks. I figured out why the .avs file wouldn't open or show the avisynth kind of icon - for some reason the .avs file was being treated as something else and vlc was the predominant file associated with opening it? I sorted the properties and as soon as the changes were implemented - (i.e. always open this file with 'notepad') all was good.

Thanks for the info on avspmod and vdub - but here's a quick Q....will either of these programs (in conjunction with avisynth) open up an alternative source file? (i.e. who the hell works with crappy x263 avi files nowadays anyway?) Most of my source files will either be original .vob files extracted from DVD, or will be larger .mkv or mp4 files. Can avisynth open these and deal with commands for these sources? When trying to open these earlier (albeit with Vdubmod) I kept getting an error due to the source file not being avi.

I'll keep plugging away for now and see how I get on - don't have a huge amount of time to dedicate to this at the minute, but don't want to put it on the back burner either.

Cheers.

[Selur]

for different sources there are different Source filters, see: http://avisynth.org/mediawiki/External_plugins#Source_Filters

i.e. who the hell works with crappy x263 avi files nowadays anyway?

there's no real reason not to use avi as a container, if the overhead doesn't bother you, you don't need vfr, your encoding and decoding routines support the container and the content you want to put in it. Some folks use x264 inside avi to be able to easier open it up with Virtual Dub,...

Cu Selur

Ps.: like mentioned before: VirtualDubMod has known bugs and is dead for quite some time, so using Virtual Dub instead is recommend. If you want to open other stuff than .avi with Virtual Dub etc. you need additional plugins for them, e.g. FFInputDriver. There exist quite a few additional Input PlugIns for Virtual Dub.

[poisondeathray]

A commonly used source filter in avisynth is FFMpegSource2 . It will open almost anything mostly independent of system installed codecs and splitters (for some file types you might need haali installed)
http://code.google.com/p/ffmpegsource/

For MPEG2 sources (e.g. DVD) , still the most consistent is DGMpegDec . You index your file with DGIndex and use MPEG2Source(). There are instructions in the package .

To learn about basic patterns - what is interlaced, what is telecined, what is 3:2 pulldown etc... start with this mini guide
http://neuron2.net/faq.html

[The.King]

there's no real reason not to use avi as a container

From all the guides and reports I've ever read, people seem to have their arguments as to why .avi is pretty poor in comparison to .mp4, .mkv as a container. I don't know enough about containers in that respect, but from viewing 350mb .avi files (of say a recent TV programme) and comparing it to a much smaller 225mb sized .mp4 file of the exact same programme - the difference in quality is completely night and day with the .mp4 file miles ahead in terms of quality. Previously, I always put that down to the x264 codec that people kept saying was the main difference between the two. But then again, I've also downloaded (just solely to see the difference) two completely different versions of the exact same movie - (1) a 2.4gb file of a movie BRRip 720p x264 Xvid which would have an .avi container and (2) a 1.7gb file BRRip 720p x264 AAC which would have an .mp4 container. Both movies were uploaded by the same person and both used an x264 codec but again, the difference in quality was significant. I'm sure there are a lot of reasons why that is the case (probably not just the container) but again, I don't know enough about that. XviD, DivX was always pretty poor quality to me anyway – hence why I want to learn this avisynth stuff and make quality encodes.

If you want to open other stuff than .avi with Virtual Dub etc. you need additional plugins for them, e.g. FFInputDriver. There exist quite a few additional Input PlugIns for Virtual Dub.

Thanks for that, I'll check out the additional plugins.

[The.King]

A commonly used source filter in avisynth is FFMpegSource2 . It will open almost anything mostly independent of system installed codecs and splitters (for some file types you might need haali installed)

This sounds like it might be the way to go so – thanks. I have haali installed as part of a codec pack anyway, so that should be around somewhere. After I download FFMpegSource2, can you tell me where this should go – folder-wise? Do I stick it in its own folder on my C drive (Program Files), or do I lump it into the avisynth folder I already have?

For MPEG2 sources (e.g. DVD) , still the most consistent is DGMpegDec.

After I download DGMpegDec, can you tell me where this should go – folder-wise? Do I stick it in its own folder on my C drive, or do I lump it into the avisynth folder I already have?

You index your file with DGIndex

as opposed to using Notepad, right? Is DGIndex a completely separate program/file compared to DGMpegDec? If so, is this also stored in a separate folder in my C drive?

and use MPEG2Source().

as opposed to using ‘avisource()’ in Notepad, right?

Can I also ask – when using Vdub, can you tell me what the two small screens refer to? Are they both identical, is one used for the original source and the other to show changes made, is one for certain fields and one for another?

Finally, can I ask this one – when assessing whether the source video is progressive, interlaced, telecined, field-blended, etc., and the guide suggest the following…
MPEG2Source("file.d2v")
AssumeTFF() # or AssumeBFF().
SeparateFields()

……Does that mean that if the source file is called TEST for example, my commands would look like this on three separate lines…
MPEG2Source("TEST.d2v")
AssumeTFF() # or AssumeBFF().
SeparateFields()

Or like this on three separate lines…

MPEG2Source("TEST.d2v")
AssumeTFF("TEST.d2v") # or AssumeBFF("TEST.d2v")
SeparateFields("TEST.d2v")?

I'm just trying to learn if something always needs to go inside the brackets, or if that's just a way to finish each command - with ()

To learn about basic patterns - what is interlaced, what is telecined, what is 3:2 pulldown etc... start with this mini guide

Thanks for this, I’ll keep reading.

(Sorry for all the quotes folks, I'm just trying to understand everything and keep it clear in my head..........which is slowly turning to mush)

As ever, serious thanks to poisondeathray & Selur for you time, patience and help with this - you'll make an encoder out of me yet....

[Selur]

Do I stick it in its own folder on my C drive, or do I lump it into the avisynth folder I already have?

Normally it's recommend to put each filter in it's own subdirectory inside the avisynth/plugins folder and load it directly via 'LoadPlugin(...)' to avoid conflicts.
Conflicts mainyl occur since .dll and .avsi files inside the plugins folder get automloaded directly, so if you have two versions of a Plugin e.g. RemoveGain.dll and RemoveGainSSE2.dll which share the same interface in the plugins folder you will get a crash when calling the filter e.g. RemoveGrain().

Is DGIndex a completely separate program/file compared to DGMpegDec?

DGMpegDec is the name of the package, which contains DGIndex and DGDecode. DGIndex is used to index content, DGDecode is to decode the content by using the index created with DGIndex. MPEG2Source is the name of the Filterinterface that is used to use DGDecode. see: http://neuron2.net/dgmpgdec/DGDecodeManual.html

Does that mean that if the source file is called TEST for example, my commands would look like this on three separate lines…
MPEG2Source("TEST.d2v")
AssumeTFF() # or AssumeBFF().
SeparateFields()

This is the right one. There's a hidden variable.
MPEG2Source("TEST.d2v") <> last = MPEG2Source("TEST.d2v")
AssumeTFF() <> last = AssumeTFF(last)
last = SeparateFields(last)

I'm just trying to learn if something always needs to go inside the brackets, or if that's just a way to finish each command - with ()

if you don't enter a parameter inside the brackets and the input wants a 'clip'-variable Avisynth automatically feeds it the last-variable.

-> you should read the Avisynth getting started guides

[The.King]

[QUOTE=Selur;2205231]

you should read the Avisynth getting started guides

If you actually closed your eyes, convinced yourself for 5 seconds that you've never used anything other than Staxrip for a while and therefore all of this avisynth DGIndex stuff was complete and utter "alien" to you, what you just wrote sounds like a foreign language. It baffles me that what you wrote makes perfect sense to you and no doubt to a lot fo people. For someone like me - it's like your keyboard just had a meltdown.

Right, I've downloaded all the DGMpegDec DGIndex FFInputDriver Vdub http://code.google.com/p/ffmpegsource/ & avspmod etc. and installed all of them in my C-drive.

I opened avspmod and typed my first line as I'm using a vob file which was: MPEG2Source("C:\Desktop\folder\myclip.d2v"). I closed this and saved it as test1.avs (as avspmod wouldn't give me the option to save it as any other file extension.)

Yet, when I try to open (load) this with either VirtualDub or with DGIndex, I get an error on both counts because I'm not leading the command with 'avisource' - considering I'm not using an avi source. Opening with DGIndex I get: This is not a DGIndex project file! Opening with VirtualDub I get: Avisynth open failure: Script error: there is no function named "MPEG2Source" (C:\Desktop\folder\test1.avs, line 1)

I am slowly losing the will to live. Is there any point studying/reading about this for hours/days/weeks when I can't even get the file to open? In order to stick with this, I need a little incentive or help to just get me going and show me it can be worth it to keep trying, so what I hoping is this....

If you ever get the time, can you write out what my commands should be to open the 'myclip' file which is a .vob file and tell me as you go, what program opens it, if avspmod should be used to write the commands, are they saved as .avs or .d2v files, etc.?

All I'm trying to achieve is to be able to load the file successfully and see if I can tell whether the source is interlaced, telecined etc. If I can achieve that, I should have enough to go on to start studying this thing. I just might need a few valium first

[Selur]

there is no function named "MPEG2Source" (C:\Desktop\folder\test1.avs, line 1)

this happens because you didn't understand what I wrote

I try it in a more simple step-by-step version:

• Index your vob files
  1. start DGIndex (double click on the DGIndex.exe file)
  2. select you content (File-> open) e.g. VTS_01_1.VOB
  3. this should open a File List where you can add additional files; normally all VTS_01_X.VOB files belong together)
  4. if all the sources that belong together are listed in the 'File List' exit the 'File List'-window by pressing OK
     (the DGIndex window should resize now and with the slider at the bottom you should be able to move through your source)
  5. enable Audio->Output Method->Demux all tracks (this way DGIndex will create the project file and extract the audio streams at once)
  6. press File->Save Project and specify a name of the project e.g. MyTest, than hit 'Save'
  7. now DGIndex should start creating the .d2v project file and the audio files, once it's finished the Information-window that should open should indicate Staus->Remain->FINISH
  8. close DGIndex
• create a .avs File that loads the video
  1. open notepad (Start->Run..->notepad)
  2. write the following text:
     

     LoadPlugin("Path to the DGDecode.dll")
     Mpeg2Source("Path to the .d2v Project")

     so it should e.g. look like this:
     

     LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
     Mpeg2Source("G:\TestPoject\MyTest.d2v")

  3. save the file with a .avs extension e.g. MyTest.avs (File->Save, select Save as type (*.*), enter the name and hit Save)
  4. close notepad
• Now open up Virtual Dub and select the .avs file as source (File->Open video file...)
  this should now allow you to see the content of your material in Virtual Dub, if not Virtual Dub should complain about something
• open your .avs file with notepad again
  add: 'bob()' to the content, so it should look like:
  

  LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
  Mpeg2Source("G:\TestPoject\MyTest.d2v")
  bob()

• save the file and close notepad
• open the file again in Virtual Dub
  Now the number of frames should have doubled and you can inspect where movement occurs if you scroll through the file to identify what sort of interlacing pattern your file shows.

Hope this helps.


Cu Selur

[The.King]

[QUOTE=Selur;2205249]

Hope this helps. Cu Selur

It MOST certainly does help! Thank you soooooo much.

From your last post, now I understand how the sequence of events works in order to load the file.

You were on a roll there - I was hoping you were going to write more and more and more and more.......

I think I will have to play around with the other commands though: 'MPEG2Source("TEST.d2v") <> last = MPEG2Source("TEST.d2v")
AssumeTFF() <> last = AssumeTFF(last) last = SeparateFields(last)' because the 'bob ()' thing didn't really seem to do anything. I read somewhere (sorry, been reading a lot lately so don't remember where) something about a, b, c, d, or a, a, b, b, c, c, d, d, or something - to understand what source I was dealing with. After sticking in your bob() command and as I scroll through (frame by frame) every 2nd or 3rd frame is a little blurry when there's a lot of movement, but I don't know what that means. I still don't know why I have two screens in Virtual Dub either.

Anyway, I'll have to start reading up now to see what I can learn about this - but please feel free if you have a few minutes here and there to post a few codes/scripts that could help.

Again, serious thanks for having the patience to help - I appreciate that for someone like you, this is very simple and boring, but it's all new to me.

[Selur]

the main question is if when skipping through the frame the frame content changes.
for normal interlaced content the content should change each frame and if you give each different frame another character from the alphabet you would have something like: a, b, c, d, e, f, g, .....
since you are following the DG-way and you might want to change your script to:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")
AssumeTFF()
SeparateFields()

save the .avs file and open it in Virtual Dub, if you now scroll fast through the content and it's jumping back-and-forth during movement you got the wrong field-order and need to change to:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")
AssumeBFF()
SeparateFields()

if you now have the right field order you write down characters until you find a pattern.
(if an image is just a bit blurier than the next you count it as the same, what counts is real movement, so choose a section of your film where movement happens)

About which pattern means what there are probably other folks can help out.
For normal (non-anime or norm-converted) material you only need to decide:
a. is the content interlaced? This is simple, load the content with e.g.

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")

and see if you see interlacing lines in scenes where movement is present
b. if the content seems to be interlaced is it normal interlacing or telecine
(c. if it's neither telecine/3:2-pulldown or interlaced a whole sack of problems opens )

Cu Selur

[The.King]

[QUOTE=Selur;2205270]

since you are following the DG-way and you might want to change your script to:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")
AssumeTFF()
SeparateFields()

save the .avs file and open it in Virtual Dub, if you now scroll fast through the content and it's jumping back-and-forth during movement you got the wrong field-order and need to change to:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")
AssumeBFF()
SeparateFields()

Yep, I tried all that, all it did was make the two screens go very skinny? All I can tell as I scroll through the footage is that - if I move frame-by-frame, then frame a is clear, frame b has a shadow of the movement (almost as though it's where frame c will end up) and then frame c ends up thre. Frame d nothing seems to happen and then the pattern seems to be repeated.

a. is the content interlaced? This is simple, load the content with e.g.

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")

and see if you see interlacing lines in scenes where movement is present
b. if the content seems to be interlaced is it normal interlacing or telecine
(c. if it's neither telecine/3:2-pulldown or interlaced a whole sack of problems opens )

I don't see any interlacing lines anywhere, but if I play only the left screen - the quality is really bad - if I play both screens at the same time, the quality is better. I'm sure this is normal, but as I don't know why I have two screens to begin with - it makes no sense.

I'll keep reading up on it.

Thanks.

[poisondeathray]

I don't see any interlacing lines anywhere, but if I play only the left screen - the quality is really bad - if I play both screens at the same time, the quality is better. I'm sure this is normal, but as I don't know why I have two screens to begin with - it makes no sense.

left screen is "before" or input , right screen is "after" or output . If you apply a vdub filter, you should see the result on the right side


This is an unexpected result . They should be the same if you have no filters applied. Options=> preferences => display => uncheckmark use directx for display

Yep, I tried all that, all it did was make the two screens go very skinny?

You're viewing fields not, not frames , because you used separatefields() . A field is like 1/2 a frame

I don't see any interlacing lines anywhere

You shouldn't see any "interlacing lines" or combing when the fields are separated, that' s normal.

Only when 2 separate fields are weaved together into a frame, and they are different moments in time will you see "interlacing lines" . That is one defintion of what "interlace" is : consecutive fields represent different moments in time

[Selur]

You shouldn't see any "interlacing lines" or combing when the fields are separated, that' s normal.

but if you don't see any interlaced lines when only using:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")

in your script, than your content is progressive and not interlaced,...

[poisondeathray]

You shouldn't see any "interlacing lines" or combing when the fields are separated, that' s normal.

but if you don't see any interlaced lines when only using:

LoadPlugin("G:\Hybrid\avisynthPlugins\DGMpegDec\DGDecode.dll")
Mpeg2Source("G:\TestPoject\MyTest.d2v")

in your script, than your content is progressive and not interlaced,...

Yes , that's true

But if you do see "interlaced lines" with only that, it doesn't necessarily indicate it's interlaced content (it might be field shifted progressive content, have 3:2 pulldown, etc... for example)

[Selur]

yes, but it helps to show if the content is progressive or not, if it's not progressive the 'fun' starts and the whole 'pattern search' begins,..

[hello_hello]

The.King,
Maybe you should try using MeGUI. One of the reasons I like it, is it doesn't attempt to hide any of it's inner-workings. It uses the same tools as Selur described previously to do the actual work, ie DGIndex to index the vob files and extract the audio, then it creates an AVISynth script which it uses for encoding. It might be a good way to learn how the various "tools" work, and to learn how AVISynth scripts are created.
As MeGUI simply creates an AVS script and saves it to your hard drive, you're free to modify it in any way you choose before running an encode, so it's kind of like "the best of both worlds" in that respect. The GUI helps automate the process but you still have "manual" control. MeGUI will also analyse the file before encoding to add the appropriate de-interlacing etc to the script. Admittedly it doesn't get it right 100% of the time (I doubt many GUI's do) but it's pretty reliable (I disable MeGUI's "force film" setting in it's options).

As I live in PAL-land I don't have to worry about IVTC and similar NTSC problems. PAL is either interlaced or progressive so it's not hard. If I'm encoding an NTSC DVD though and I'm not 100% sure MeGUI is getting it correct, I run it through AutoGK to check. AutoGK is about the only GUI I've used which seems to get it right all the time as it analyses the video quite thoroughly. If AutoGK disagrees with MeGUI then I copy the IVTC/de-interlacing from the script AutoGK creates and use it in MeGUI's script. That's my lazy way of getting it right as I don't encode many NTSC DVDs.

[The.King]

Maybe you should try using MeGUI.

Hmmmm - that might be interesting alright, thanks. I'd really love to have a crack at this avisynth thing - especially after all the help I've received from Selur and poisondeathray, but there really is 'an awful lot' to it and aside from the time required, there's no decent video tutorials, etc. to speed up the process of learning. It's like - come home from work and spend a few hours reading 'another language' I don't even understand. I'm sure it's easy enough and I would do it in my sleep once I get the hang of it - but like most things, it would be a million times easier if someone just sat down with you for 10 minutes and ran through the process! I'm going to have a look at Hybrid too, just to see if I can make sense out of that, or if its uber technical too.

It's a bit bonkers to think that there's loads of mega technical guys out there who know how to write their own programs/GUI's and spend hours/days/weeks - probably months doing it, but none of them can use a screen capture program and upload a short tutorial to teach people how to use them. As the saying goes - 'a picture can paint a thousand words!'

As I live in PAL-land I don't have to worry about IVTC and similar NTSC problems. PAL is either interlaced or progressive so it's not hard.

I live in PAL land too and the DVD's I'm trying to encode are PAL also, so that's good to know thanks.

If I'm encoding an NTSC DVD though and I'm not 100% sure MeGUI is getting it correct, I run it through AutoGK to check.

Never used that before, so I'll have to check it out - hopefully I won't need it for the PAL DVDs. Thanks for your input.

[Selur]

If I'm encoding an NTSC DVD though and I'm not 100% sure MeGUI is getting it correct, I run it through AutoGK to check.

Last time I checked they use the same interlacing analysis routine, so the results shouldn't differ,...

It's a bit bonkers to think that there's loads of mega technical guys out there who know how to write their own programs/GUI's and spend hours/days/weeks - probably months doing it, but none of them can use a screen capture program and upload a short tutorial to teach people how to use them. As the saying goes - 'a picture can paint a thousand words!'

May be it's bonkers, but on the other hand why do not other people care about it and create videos and write guides?

Cu Selur

[The.King]

May be it's bonkers, but on the other hand why do not other people care about it and create videos and write guides?

That's also a good question, but I'd take a 'stab in the dark' by saying: "probably because they don't know enough about it?" I know if it were me (and I had spent ages actually creating a new program that I thought was really good - [I can't even figure out avisynth, so fat chance of that ever happening ] I would want to justify the time I spent creating it, by making a short tutorial video (screen capture) so that others could learn how to use it the way I intended it to be used. i.e. if there was a 'proper' tutorial video out there (youtube for example) of avisynth, DGIndex, etc - chances are - a lot of GUI's would become obsolete as everyone would know how to use it 'properly' and would see the results. Same could be said for Hybrid (which I'm going to have a look at now) and other programs.

I wouldn't claim to speak for anyone only myself, but if a good video tutorial of avisynth was out there (using all associated programs and showing some frequently used commands, etc.) I know I'd be using it. I'd love to use it - it does sound really good, but the fast paced world we live in doesn't really allow for hours/days reading up on something which is quite complex to a newbie, and then spending hours/days/week trying out what was learned to see if it works. If the truth be told - I guess that's why there's a lot of very average encodes out there? And I'm not even a brand new newbie - I've been using Staxrip for years (after trying AviDemux, AVS Video Converter, HandBrake, HDConvertToX, etc.) I think my brain would be fried if I was trying to use avisynth from scratch

[hello_hello]

Last time I checked they use the same interlacing analysis routine, so the results shouldn't differ,...

I can promise you they do sometimes. Just today I encoded an old PAL DVD. MeGUI decided it was partially film and required IVTC.... or something which didn't sound too likely. AutoGK confirmed is just needed de-interlacing.

After a long battle with one particular NTSC DVD I discovered MeGUI's "force film" option somehow changes the way the DGIndex file is written if "force film" is used. I can promise you once the DGIndex file has been altered, all the analysis in the world isn't going to get it right. I banged my head against the desk for a few hours while encoding that one, trying to work out why MeGUI wasn't getting it right and why copying AutoGK's de-interlacing to MeGUI's script wasn't fixing the problem..... until I eventually started playing with the DGIndex files and replaced the one MeGUI created with the one AutoGK created. MeGUI's "force film" option is now permanently disabled.

AutoGK also seems to spend time re-analysing the video if it decides the source is some sort of hybrid, whereas MeGUI doesn't (Or at least it doesn't spend the same amount of time on it). I've not paid a lot of attention to it.... looking for optimum hybrid thresholds or something to that effect..... and I'm almost 100% sure I've seen AutoGK change it's mind after the second analysis on a few occasions. Before the second analysis AutoGK sees the source as a hybrid, after the second one it decides it's really pure film..... that sort of thing..... but I'm definitely 100% sure MeGUI has given me a 29.970 fps DVD encode in the past whereas for the same DVD AutoGK has given me a 23.976 fps encode, and after checking each quite thoroughly it was once again AutoGK which got it right.

And for reasons I don't understand MeGUI always seems to fail analysing video if it's being decoded via DirectShow. AutoGK doesn't seem to have the same problem..... so while I don't encode via DirectShow a great deal, once again I rely on AutoGK to check the video when MeGUI can't.

One day I'm going to learn how to create AVS templates for MeGUI which mimic the clever way AutoGK resizes. Automatically cropping video which is close to 4:3 to exactly 4:3, and "smart resizing".... similar to the way AutoGK automatically calculates the correct height so there's no need to do anything but chose the width. One day......

[The.King]

I'm a bit lost/dubious of this AutoGK thing. While I'm sure it is a good program, it looks a little pre-historic to me and only outputs to .avi? It's also telling me that my OS is Vista and not Win7 64-Bit? Anyway, this is what it gave me...

AutoGK 2.55
OS: Windows Vista (6.1.7600).2
Job started.
Indexing input for preview.
Processing file: C:\Users\Desktop\xxxxxxx.xxxxxxxx - Season 1\01x01\VTS_01_VOBID_003_1.VOB
Source resolution: 720x480
Found NTSC source.
Source aspect ratio: 4:3
Output will contain 92121 frames
Job finished. Total time: 24 seconds 

First off - the DVD is definitely a PAL DVD (it even says so on the DVD box) so not too sure how it's reading an NTSC source? Whilst the AR of the source .vob may be seen as 4:3, it definitely plays (in DVD player or on PC) as a widescreen 16:9?

However, if you're only using AutoGK to get the avisynth script, this is how is reads:

LoadPlugin("C:\PROGRA~2\AutoGK\DGMPGDec\DGDecode.dll")
LoadPlugin("C:\PROGRA~2\AutoGK\filters\autocrop.dll")
movie = mpeg2source("C:\Users\Desktop\xxxxxx.xxxxxx - Season 1\01x01\agk_tmp\VTS_01_VOBID_003_1.d2v")
audio = DirectShowSource("C:\Users\Desktop\xxxxxx.xxxxxx - Season 1\01x01\agk_tmp\VTS_01_VOBID_003_1 T80 2_0ch 192Kbps DELAY 0ms.ac3")
audio = DelayAudio(audio,0)
movie = AudioDub(movie,audio)
cropclip = autocrop(movie,mode=0,wmultof=4,hmultof=4,samples=10,aspect=0,threshold=34,samplestartframe=0,leftadd=0,rightadd=0,topadd=0,bottomadd=0)
fixed_aspect = 0.888888888888889
c_width = width(cropclip)
c_height = round(height(cropclip) / fixed_aspect)
input_par = float(c_width)/float(c_height)
input_par = (input_par > 1.4) || (input_par < 1.25) ? input_par : (4.0/3.0)
out_width = 704
out_height = round(float(out_width) / input_par)
hmod = out_height - (floor(out_height / 16 ) * 16)
out_height = (hmod > 4) ? (out_height + (16 - hmod)) : (out_height - hmod)
new_aspect = (float(out_width) / float(out_height)) / fixed_aspect
autocrop(movie,mode=0,wmultof=4,hmultof=4,samples=10,aspect=new_aspect,threshold=34,samplestartframe=0,leftadd=0,rightadd=0,topadd=0,bottomadd=0)
LanczosResize(out_width,out_height)

Anyway, most of the above may as well be Swahili to me, but I don't see any command line saying anything about deinterlacing, IVTCing, Decombing, etc. etc.

I've extensively tried MeGUI & Hybrid and the resulting encodes are no better (or worse) than what Staxrip can achieve. Hence why I keep coming back (get drawn back to) avisynth and hoping that a decent tutorial becomes available from someone who really knows how to operate it properly. Anyone - please?

Thanks.

[The.King]

Ok, so I've just re-read all posts here on this matter and I think fromreading back over some of them that we got off on a too-advanced footing rightfrom the get-go. I may have been finding the 'quick-fix' to encoding for 6+yrsnow with Staxrip (which is a great little program), but I am prepared toeducate myself (with some much needed help from you guys here) on how to dothis properly. More importantly, (and I think this is a good plug) if Ieventually figure out 'how to' with avisynth, Vdub, x264, etc. I WILL post aproper tutorial of commands and how to 'get started' giving the information Iso badly need now.

I have read over some responses from Selur and poisondeathray and I've figured out - by applying anot-too-clear method that...........during a fast paced scene of my .vob, thatif I was to give a letter to each kind of sequence (a=clear still image,b=slightly interlaced still image and c=heavily interlaced image), I end upwith this.....

ABCCBA ABCCBA ABCCBA....

Not a million per cent sure what that actually means I have as a source file - but I just wanted to say to Selur and poisondeathray (being the main two guys to weigh-in throughout) that I want to stick with this (with your help) and figure out how to perfect my encodes.

[Selur]

I've extensively tried MeGUI & Hybrid and the resulting encodes are no better (or worse) than what Staxrip can achieve.

can't say much about that,.. Hybrid with avisynth extension offers QTMC which should for interlaced PAL content clearly create better results than StaxRip with Yadif,...

About your sequence, can't say much about it since you chosen differentiation doesn't help. :P
If you source is a PAL DVD from an NTSC original series you might have a lot of fieldblending and frameblending in your source, which would require srestore/restore24 or similar (see: http://avisynth.org/mediawiki/External_filters#Fieldblending_and_Frameblending_removal ).

btw. it might help to google for:
Avisynth "NAME your series"
since some series are known for totally messed up versions (Captain Future, some of the Simpsons DVDs, Babylon 5,...)

Cu Selur

Ps.: if you cut of a < 30 second clip of your series which includes movement and upload it we and others can look at the footage and probably say more,..

PPs.: not sure how good Google translate from german to you language is, but http://home.arcor.de/scharfis_brain/ExotischesInterlacing/ (german) might be interesting to understand some of the BS that people do when converting from NTSC to PAL.

[The.King]

Ps.: if you cut of a < 30 second clip of your series which includes movement and upload it we and others can look at the footage and probably say more,..

yep, that sounds like a good idea at this point. Any quick idea on how I cut a 30 second clip from the original .vob file - and - can I upload it here or do I need to join some funny site, upload it there and place a hyperlink in here?

PPs.: not sure how good Google translate from german to you language is, but http://home.arcor.de/scharfis_brain/ExotischesInterlacing/ (german) might be interesting to understand some of the BS that people do when converting from NTSC to PAL.

That looks like a really good link, (and Homer looks exactly like most of my B's & C's) but I don't think Google translate will be up to the job. Will give it a try anyway, thanks.

[Selur]

Any quick idea on how I cut a 30 second clip from the original .vob file

Cuttermaran, MPEG2Cut, DGIndex,... there are a bunch of tools that can cut mpeg2.

can I upload it here

Should be no problem to attach the file here,.. (and don't forget to google)

Cu Selur

[hello_hello]

I'm a bit lost/dubious of this AutoGK thing. While I'm sure it is a good program, it looks a little pre-historic to me and only outputs to .avi? It's also telling me that my OS is Vista and not Win7 64-Bit? Anyway, this is what it gave me...

It all seems a bit odd. How was the DVD ripped to your hard drive?
I'm not sure I've ever seen vob files labelled "VTS_01_VOBID_003_1.VOB" so that seems a bit strange. Anyway.....

I've seen some aspect ratio oddness with DVDs in the past. How do the other programs you've tried identify the DVD in question (PAL or NTSC)? Instead of opening the vob file directly with AutoGK, try opening the appropriate IFO file instead. That might fix any misidentification of the type of video. If it still gets the aspect ratio wrong, then once you've opened the vob file, use CTRL+F9 to open AutoGK's hidden options. You'll find a box to tick to over-ride the aspect ratio and you can select 16:9.

I've never seen AutoGK misidentify PAL video as NTSC, ever. It's claiming the resolution is 720x480 which is definitely NTSC. Is that the resolution according to the player you're using? AutoGK won't analyse 23.976fps video, as far as I recall. It shouldn't need to as it can only be progressive. It'll only analyse the video if it's 29.970fps NTSC or 25fps PAL.

If you want to know exactly what AutoGK decided to do, look at the log file. It'll tell you if AutoGK decided the video was interlaced etc, but there's definitely something odd happening with the vob file in question. It does seem like it's the result of some particularly dodgy NTSC to PAL conversion. As Selur suggested, upload a small sample of the original video if you can.

I've extensively tried MeGUI & Hybrid and the resulting encodes are no better (or worse) than what Staxrip can achieve. Hence why I keep coming back (get drawn back to) avisynth and hoping that a decent tutorial becomes available from someone who really knows how to operate it properly. Anyone - please?

What are you trying to achieve in terms of "better"? MeGUI simply uses AVISynth for frameserving, which includes using some noise filtering plugins, but it really only encodes the video "as-is" (aside from the usual cropping and resizing). So in that respect, as they all use the same x264 encoder, then with the same x264 settings the result shouldn't be terribly different to StaxRip or Hybrid (although I've not used Hybrid myself so I'm assuming there). In fact if the output from each GUI was significantly different then there's probably something wrong.

I personally avoid using filtering when encoding unless the source is particularly bad, but maybe if you can describe what you're trying to achieve? There's an abundance of plugins for AVISynth which you can manually add to the script MeGUI creates, but you need to know the type of plugin you're wanting to add and the result you're wanting to achieve.

[The.King]

It all seems a bit odd. How was the DVD ripped to your hard drive?

Was ripped with DVDDecrypter

How do the other programs you've tried identify the DVD in question (PAL or NTSC)?

They all see it as NTSC (even though it's an Irish TV program and, the DVD box states 'PAL - DVD9 x 2' and I doubt that anyone State side has even heard of it.)

Instead of opening the vob file directly with AutoGK, try opening the appropriate IFO file instead.

I get the following error if I try to open the IFO file.

Your input can't be opened:
VLC is unable to open the MRL 'dvd:///C:/Users/Desktop/xxxxxx.xxxxxx%20Season%201/01x02/'. 
Check the log for details.

The actual media info I get from vlc is:

Stream0
Type:Video
Codec:MPEG-1/2 Video (mpgv)
Resolution:720x480
Framerate: 59.940059
Decodedformat: Planar 4:2:0 YUV
 
Stream1
Type:Audio
Codec:A52 Audio (aka AC3) (a52 )
Channels:Stereo
Samplerate: 48000 Hz
Bitrate: 192 kb/s 

If you want to know exactly what AutoGK decided to do, look at the log file.

Log file says:

Analyzing source.
Source has percentage of interlacing in motion areas: 58.20
Source has percentage of telecined patterns: 9.27
Source has percentage of progressive patterns: 5.91
Source has percentage of interlaced patterns: 84.82
Source is considered to be hybrid (mostly NTSC).
Looking for optimal hybrid thresholds.
Found threshold of: 0.01
Output will contain 92121 frames

As Selur suggested, upload a small sample of the original video if you can.

See attached.

What are you trying to achieve in terms of "better"?

See attached.

Hope someone can shed some light on this for me and perhaps write a short script for me which I could amend (for correct file locations) and apply to see if I get a better encode?

Thanks in advance.

[The.King]

I probably should point out that (although it's most likely my settings) if I play the 24.46mb sample in Media Player Classic, aside from the AR it looks ok. If I play it with vlc, you can clearly see the interlaced lines.

The encode is pretty much identical no matter what player you use.

[poisondeathray]

This is a poor example to start with because it's field blended . It would be easier to learn on more "typical" textbook example . These types of patterns will fool any autoanalysis tool.

If you follow the earlier instructions by separating the fields, the field blends will become readily appearant

Mpeg2Source("VTS_01_VOBID_003_1 [clip for upload].d2v")
AssumeTFF()
QTGMC(preset="faster", sharpness=0.8)
SRestore(25)

The original frame rate was 25 (which make sense since it was original Irish TV program), it's field blended PAL to NTSC => 29.97

[hello_hello]

It's definitely NTSC regardless of what it might say on the box, so AutoGK wasn't wrong there. As poisondeathray says it's fieldblended to convert it from 25fps to 29.970fps. Using his method to restore it back to 25fps is no doubt the way to go.

I ran your sample though both AutoGK and MeGUI and both see it as purely interlaced NTSC, which I'm sure is correct. Neither will automatically remove the fieldblending so they'll just de-interlace it and output progressive 29.970fps.

I don't have QTGMC installed so I can't add poisondeathray's script to MeGUI myself, but looking at poisondeathray's sample it obviously did a pretty good job of restoring it to the original 25fps.