MeGUI aspect ratio & cropping problems ??

[3thman]

Hi, I'm sorry to ask many times about cropping, but I wasted many days to understand their principles & solve their problems.

I asked before and now about processing of my vids that I record by satellite card as transport stream files.

Their settings are like that:
A chuck : 3.4 MB http://www.mediafire.com/?5mwzewmm1w1

Video :::::::

Format : MPEG Video
Format version : Version 2
Bit rate mode : Variable
Bit rate : 2 235 Kbps
Nominal bit rate : 2 700 Kbps
Width : 720 pixels
Height : 576 pixels
Display aspect ratio : 4:3
Frame rate : 25.000 fps
Standard : PAL
Resolution : 8 bits
Colorimetry : 4:2:0
Scan type : Interlaced
Scan order : Top Field First

Audio :::::::

Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Format_Settings_Mode : Joint stereo
Bit rate mode : Constant
Bit rate : 128 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz

- I process my vids as following: ( all with the aid of Avisynth scripts)

1 - I may encode them with Xvid with 5 target quantitizer directly.

2 - Crop some lines from the bottom to remove the ads bar or remove the black bar in some vids.

3 - Deinterlace by Donald Graft's Smart Deinterlacer , remove logos then export.


I asked before in this forum about cropping & encoding in VirtualDub. As I said, I encode with Xvid. I learned that square pixel encoding is the best way and as above my vids are 4:3 720x576 PAL vids, so I resize them to 720x540 or 640x480 to maintain the DAR as 4:3 or it will appear as 5:4 or some thing like that.


Recently, I has found that x264 codec in MeGUI is very good , more stable than Xvid , Gives me higher quality , lower size than that of Xvid when using constant refractor mode, but now I face aspect ratio problems again.

The chuck and the settings are above, please tell me the best way to crop the vids, resize, maintan aspect ratio and your suggestions to enhance my vids to get a smooth video with high quality.

For information I deinterlace with LeakKernelDeint.

I tried tried clever anamorphic encoding and mod16, but I dont know what to do.

I'm sorry for prolixity.

[poisondeathray]

As a general rule, deinterlace before you crop, otherwise you can get big problems cropping interlaced content. (you can crop in multiples of 4 before deinterlacing safely)

If you crop then resize and use square pixels (e.g. 640x480), you don' t need anamorphic encoding, it doesn't apply (anamorphic encoding refers to non-square pixels)

The problem with your sample, is that the "news ticker" (the scrolling credits overlay) are interlaced (50 fields per second) , but the main content underneath is 25 frames per second progressive

[3thman]

Thank u so much 4 reply, there is some thing that I don't understand , is there any difference in the commands' arrangement in the script?

Generally, this is my script:

LoadPlugin("C:\Program Files\megui\tools\dgindex\DGDecode.dll")
DGDecode_mpeg2source("F:\Life learned me.d2v", info=3)
LoadPlugin("C:\Program Files\megui\tools\avisynth_plugin\ColorMatrix.dll" )
ColorMatrix(hints=true, interlaced=true, threads=0)
LoadPlugin("C:\Program Files\megui\tools\avisynth_plugin\LeakKernelDeint. dll")
LeakKernelDeint(order=1,sharp=true)
crop( 0, 4, 0, -132)
Spline64Resize(720,412) # I don't know what to do here, to calculate : 540 - 132 + 4 = 412 or 540 - 132 - 4 = 404 or other?

Did I do the right? and what about Mod16 ? is it better ? What are ur recommendations here ?

Another question, If I want to remove logo and I was lazy to build mask , can I remove the logo by region remove filter in VDub (with this script), encode with lagarith, then re-encode with x264 codec in MeGUI ? If it it was right way which color mod should I use?

[poisondeathray]

Definitely mod16 is better for encoding (more efficient) . Use a value that you think makes it look close to correct (ie. circles look like circles not ovals)

I don't know about vdub region remove filter, I've never used it before

[3thman]

OK, I can understand now, but I don't ask here about the filter, but about encoding with lagarith then encoding with x264, is it a good way ? and generally what color mod should I use in Lagarith ?

Thank u again.

[poisondeathray]

You should use the same colorspace (YV12) as the original video, but for lagarith, you should always leave it in "suggest RGB mode", which will automatically use YV12 if your source is YV12

But if you filter in vdub, it will change to RGB (you use full processing mode instead of fast recompress mode), and you have to convert it back to YV12 when encoding in megui

[3thman]

Another question, to use mod16 should I change input DAR from 1.3675 to 1.333 ?

1.3675 gives 720x528 & 1.333 gives 720x544 which is closer to 720x540 non mod16 , which of them is more efficient?

[poisondeathray]

A frame dimension that is evenly divisible by 16 will always be more efficient (mod16)

I don't know what those settings do for the avisynth creator in megui. I usually manually make scripts