possible to frameserve from Virtualdubmod to HCenc?

[spiritgumm]

is there a workaround to get HCencoder to accept frameserve from Virtualdubmod ?
Trying to convert some interlaced Pal dvds to Ntsc, and tried Lordsmurfs guide (Virtualdubmod frameserve to Tmpgenc Plus). I only have Tmpgenc Express. Did a sample, and while it played fine, it looked poor. Thought HCenc might be an improvement over Tmpgenc Express.
I'd like to slow down the PAL (rather than keeping the same duration) to get the audio pitch back to normal (using Timewarp in Goldwave). So before trying Avisynth, wondering if I could get Virtuadubmod to work with HCenc.

[freebird73717]

See this link

[spiritgumm]

Thanks! I didn't think to search for just "Hc."
Do you happen to know where I put these two lines of script?

AVISource("filename.vdr")
ConvertToYV12()

[freebird73717]

open up notepad
type

AVISource("C:\path to your file\filename.vdr")
ConvertToYV12()

save it as
filename.avs

open that avs in HCenc

You have to have avisynth installed. HCenc only accepts input from avisynth.

[spiritgumm]

Thanks again! The good news is to works.
Bad news (slight tangent to my original post) is the HC results look similar to the Tmpgenc encode - a weird kind of bad compression - blocky but running along horizontal lines; noticeable when objects move up/down. FAVC did a perfect job of converting it (but I don't have a log of what it did), so obviously I'm not using the right filters in Virtualdubmod. The video source is interlaced PAL home recording and looks okay. Maybe I need a different deinterlace filter ?

[freebird73717]

I couldn't tell you. I haven't ever done an interlaced pal to ntsc conversion. Hopefully someone else can chime in.


You might just try the conversion with just straight avisynth
check this link
https://forum.videohelp.com/topic90816.html#244808

[spiritgumm]

I've tried some other deinterlacers but the problem persists, so it must be some other issue. I don't know if Avisynth would do anything different until I figure out what filters to add.
Here's a link to a demuxed video sample if anyone wants to diagnose it. I have some other PAL videos I'll try in the meantime.
http://www.megaupload.com/?d=LZ7D3SR0

[manono]

It's not interlaced; it's phase-shifted. Realign the fields and it's progressive again:

TFM(PP=0)#the field matcher part of TIVTC
LanczosResize(720,480)#or your favorite resizer
AssumeFPS(23.976)#or just keep it at 25fps and not have to slow the audio.

Deinterlacing it is the worst thing you can do to it. Forget frameserving out of VDub and learn how to do it right. It may actually be possible to do it in VDub by separating the fields, dropping the first field, and then weaving it back together into frames. But I don't use VDub for such things.

[spiritgumm]

I forgot to specify Lanczos3 in the "resize" filter; that improved the result so it's almost as good as the FAVC. Unfortunately my dvd player doesn't play PAL dvds, so I can only compare the original with the reencode sample on the PC.
I'll try to get Avisynth running and run that script. So Virtualdubmod isnt like a GUI for Avisynth? I only planned on doing a couple types of things (PAL conversion, and maybe logo removal) and figured it would be easier than scripting.

[manono]

So Virtualdubmod isnt like a GUI for Avisynth? I only planned on doing a couple types of things (PAL conversion, and maybe logo removal) and figured it would be easier than scripting.

It's not really a GUI. When using VDub filters you can see before and after. With an AviSynth script opened in VDub you can only see the after. Unless you open 2 instances of VDub, one unfiltered and the other filtered. The closest to a GUI that AviSynth presently has might be AvsEdit:

http://www.afterdawn.com/guides/archive/avisynth_tutorial.cfm

AviSynth has good logo removal filters also. They can be tricky to set up, but the pure VDub ones aren't all that easy either. And the easy ones aren't any good.

[spiritgumm]

manono, I ran Avsedit and AvsP to help with the process. The TFM filter wasn't in my Avisynth folder, so downloaded the TIVTCv105 filter set and copied the DLL into Avisynth plugin folder. I ran the script below but the whole clip wasn't getting processed. Any idea why?
DirectShowSource("C:\VTS_01_1.demuxed.m2v")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
TFM(PP=0)
LanczosResize(720,480)
AssumeFPS(23.976)

I read that I could get the audio converted by adding "true" to "AssumeFPS" filter: "AssumeFPS(23.976, true)".
I changed the source to the original un-demuxed clip, and played in MPC. While it wasn't in sync (sample was clipped with Dvdshrink & maybe the delay isn't incorporated in the avysinth output?), the audio was appropriately slowed down. Slowing the audio is one my goals since the pitch in PAL is noticeably off.
[I suppose I could just do the Pulldown conversion method, keeping the video duration the same, demux the audio and change the pitch without changing the audio duration in Goldwave.]

Thought the "true" setting would save me a step when I eventually did the whole movie, but forgot HC doesn't process audio (I suppose another encoder would?). When I encoded this (with and without the "true" added), I got the entire video clip, unlike the demuxed clip above. Does that mean when I do the whole movie, I should use the complete muxed video?

Sample looked good in the PC, but not as good on TV when I burned it along with my VDub & FAVC conversions for comparison. What looks like a slight horizontal hiccup in the original video (about every 5 seconds) looks like a vertical hiccup now. You see horizontal lines (like interlaced) - Burr seems to jump at 5 seconds, and again at 10 seconds. So I might need another filter for that?

Wondering if there's an Avisynth/AVSedit equivalent to GSpot? I'm curious how you figured the video wasn't interlaced (Gspot, DGindex and visual inspection indicated it was interlaced). Of my other PAL movies, one is retail and shows as interlaced (an old American film); the other (which may have been retail) shows as progressive (a 1960's British film). I'd like to make certain before attempting converting them. Also wondering which situations require framerate to be 24 fps versus 23.976.

[manono]

Sorry, but I don't use DirectShowSource myself and don't recommend you use it. Get the DGMPGDec package and follow the instructions in the excellent included docs (make a D2V project file and use MPEG2Source):

LoadPlugin("C:\Path\To\DGDecode.dll")
LoadPlugin("C:\Path\To\TIVTC.dll")
MPEG2Source("C:\Path\To\Movie.d2v")
TFM(PP=0)
LanczosResize(720,480)

With PP=0 I disabled the post-processor. The encoding goes faster without it and it shouldn't be needed. However, if you cut out commercials and don't take care how you do it, or if there are cadence breaks, the result mught be some stray interlacing. If so, remove the PP=0 part of TFM()

I ran the script below but the whole clip wasn't getting processed. Any idea why?

DirectShowSource("C:\VTS_01_1.demuxed.m2v")

Is VTS_01_1.demuxed.m2v just the demuxed first VOB, or is it the entire movie? Scroll around and check. In any event, when making the D2V, load all the VOBs at once and you'll get a D2V for the entire movie and a separate file (or files) of the audio.

I process the audio separately from the video and bring them together when authoring, although it's certainly possible to handle the audio within the script. But someone else will have to help with that. You don't want your MPEG-2 encoder to handle the audio anyway, and many can't even if you wanted them to. HCEnc doesn't process audio (I don't think - never tried actually). In any event, Muxman requires the audio and video not be muxed.

I do the audio by first slowing it and making WAV file(s) using BeSweet and then reencoding for AC3. I believe the more 'modern' way is to use eac3to and its GUI to do the whole thing at one crack.

So I might need another filter for that?

A small sample of the output might help so I or someone else can understand what you're talking about. If it's interlacing, then try TFM with PP=0 left out and see if it improves. If there's a lot of interlacing then something's wrong. Maybe the entire video isn't like the sample you provided, or maybe DirectShowSource is screwing up somehow. Test the script in VDub(Mod) or AvsEdit before encoding.

I'm curious how you figured the video wasn't interlaced

The source frames were interlaced but when I separated the fields they came in pairs, each original and unique field being followed by a duplicate. That's the clue that tells you it's really from a progressive source. Separating the fields (or bobbing) is one of the things you do when trying to figure out what you have and how to treat it. It'll also tell you if there's field-blending.

Also wondering which situations require framerate to be 24 fps versus 23.976.

That sample was 25fps, so I'm not sure what you mean. You can slow it to whatever you like. It almost certainly was converted from a 24fps source, so, if you want to make your reencode 24fps, there's nothing wrong with that. But you won't be able to tell the difference between motion and audio when one's 24fps and the other's 23.976fps. 24fps for NTSC DVD requires a special pulldown. It's DGPulldown with the 'Custom' box ticked and 24->29.97fps filled in. It's different from standard 3:2 pulldown.

[spiritgumm]

Only thing I've been processing is the demuxed (M2V) 30 second clip. The "VTS" source was the original AV of that clip, used when I checked the audio filter. There are no commercials so I won't be editing (tho I'd like to get rid of the omnipresent watermark, and some brief onscreen text at the start of the movie). I usually mux with Muxman (but for clip testing, I use TDA since I'll get IFO at the same time).
This was a more general question, and it sounds like since I'm converting Pal to Ntsc, I should preferable use 23.796 (not 24 fps), correct?

Originally I didn't have the DGdecode. dll loaded, so I got an error when scripting with D2V. I followed the AvsP demo of "inserting source" video directly, and it utilized Directshow. Since that script worked, I went with it.
DGdecode.dll is now loaded so script works with D2V file, and ran clip with TFM(). Unexpected good news, the video plays a little smoother than previous tests. Also the spots at the horizontal hiccups/bumps don't look interlaced which made image seemingly jump up. Still would be nice to eliminate bumps altogether. Would "bobbing" those spots would be helpful to determine what filter to use?

I use Belight for wave<>ac3. However, I see (after converting a longer clip from my other progressive PAL movie) that slowed-down audio (in Goldwave) to the new video length, and syncing near the start of the clip, won't guarantee it's synced thruout. That clip kept getting more unsynced over time. Hopefully there are some methods/tips to process audio so it's more easily synced? For example (and I'm just brainstorming), if it's possible to output audio-only in VDUBmod using AssumeFPS(23.976, true), would that audio be overall synced (after the initial correction)?

Current script:

LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\DGDecode.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
MPEG2Source("C:\New Folder (7)\VTS_01_1.d2v", cpu=0)
TFM()
LanczosResize(720,480)
AssumeFPS(23.976)

[manono]

(but for clip testing, I use TDA since I'll get IFO at the same time).

I don't understand. Muxman also gives you the IFO and BUP files. That's one of the things an authoring program does.

I should preferable use 23.796 (not 24 fps), correct?

I would (and do).

Still would be nice to eliminate bumps altogether. Would "bobbing" those spots would be helpful to determine what filter to use?

Since I don't know what you mean by 'bumps', I can't offer a solution. I've seen plenty of movies on DVD where at scene changes the video 'jumps' momentarily before settling back down, but I don't know if that's the kind of thing you're talking about. A sample of the source, the 'bumpy' output, and the script used (if different from the one you included) would be helpful.

Hopefully there are some methods/tips to process audio so it's more easily synced?

Done properly it should always be in synch. I make the WAV files using BeSweet and its ability to slow the audio from 25->23.976fps (there's a preset). I don't use Goldwave so I have no idea where it's going wrong. You said you tried to match the audio with the video length, but a better method would be to slow it by a certain percentage, perhaps also taking the delay (if there is one) into account. Most WAV editors have that ability.

There are no commercials so I won't be editing (tho I'd like to get rid of the omnipresent watermark, and some brief onscreen text at the start of the movie)

We can work on the watermark later. First we have to get you a synched and converted movie. If you'll provide a sample that includes a very dark background where the watermark is, perhaps a night scene, or a fadein/out, but only the watermark itself really has to have black behind it, I can work up a script for it. Or you could open the script in VDub(Mod), use one of the VDub logo removers (if you're already comfortable with them) to get rid of it, create a lossless AVI (like Lagarith), and then feed that into the encoder using an AviSynth script and AVISource. As for the text, if it's at the very beginning, you can use the [ button of DGIndex when making the D2V file to cut out that part.

[spiritgumm]

My mistake, dunno why I associate Muxman with outputting plain vob without ifo/bups; maybe because most of my other wares do so. I mux the big, finished files in Muxman. I mux short test clips in TDA because I can preview them, and combine with other clips to compare after burning.

I should mention the watermark is semitransparent. It's in the top left thru entire movie. Would you need another clip (as you described, with dark background or scene change), or is the current uploaded clip enough to work from? Or I can experiment with Vdubmod's included logo remover (or another one written for it). I only just started using Vdubmod and don't know much about it. I stopped using it when you and Freebird recommended Avisynth.
The text at the beginning of the movie is over the opening credits, so it too needs to be "removed" rather than edited out.

Inre audio, the Pal>Ntsc guide (Vdubmod, Goldwave, Tmpgenc) I used only indicated entering the new duration in Goldwave. Maybe it's suppose to get synced when Tmpgenc encodes the avi, and since I was using HC & Muxman instead, it didn't sync correctly.
Followed your suggestion to change framerate in BeSweet (first tried BeLight but duration oddly stayed same). Duration increased, but volume decreased. Did a couple runs, tweaking gain and boost settings but can't keep the volume same. Any recommendations to do that? Does a PC only use one ac3 decoder, and I need to configure that instead?

In the original uploaded clip, two examples of the horizontal bump are at frames 125-126-127, and 513-514-515 (frame numbers according to Vdubmod viewer):
1) Bottom half of frame shifts slightly left, 2) then top half shifts left, 3) then the whole frame shifts right and back to "normal."

[manono]

I should mention the watermark is semitransparent. It's in the top left thru entire movie. Would you need another clip (as you described, with dark background or scene change)

I checked the sample before posting and there wasn't any background dark enough to use for creating a mask needed for the logo remover I have in mind. Other logo removers can use what's there. It just depends on the logo remover you use. I'll work with the sample I already have. In the meantime, have a look at AlanHK's XLogo guide:

https://forum.videohelp.com/topic333648.html

It can work with translucent logos, but I think that LogoTools is better with them, and for that one you need as black a background behind the logo as possible. Even one frame is enough.

The text at the beginning of the movie is over the opening credits, so it too needs to be "removed" rather than edited out.

Oh, that's tough, especially if it's opaque and especially if there's important stuff beneath and around it, and especially if the text changes from time to time. About all you can do is blur it and anything mixed in with it.

Did a couple runs, tweaking gain and boost settings but can't keep the volume same. Any recommendations to do that?

I find using Hybrid Gain works OK. But even so, sometimes I'll open the resulting WAV file in Audacity and adjust the volume. The boost makes everything more-or-less equally loud. I can find no way to adjust the boost.

In the original uploaded clip, two examples of the horizontal bump are at frames 125-126-127, and 513-514-515 (frame numbers according to Vdubmod viewer):
1) Bottom half of frame shifts slightly left, 2) then top half shifts left, 3) then the whole frame shifts right and back to "normal."

I'll have a look later. I'm at a different computer at the moment.

[manono]

I checked the frames you mentioned. Good find. Leaving off the post-processing allows interlacing to slip through. Making it:

TFM()

fixes them. There are still slight jumps there, but you can see that the film shifts for a frame. Something happened in the capture to digital or something. I don't understand that stuff very well.

Here's an XviD of the sample using XLogo. I think I could probably do a better job of it with LogoTools, but AlanHK is really the expert with this logo remover.

xlogo.avi

[spiritgumm]

I wanted to see how well the framerate-adjusted audio would sync, so I converted one of my other PAL dvds (it's already progressive, possibly retail, so was an easier project), and the audio synced up perfectly! Kind of shocking - expected at least some minor sync issue. I used Ac3-Ac3 to change framerate in BeSweet, but I'll try Ac3-Wave and boost volume in a wave editor.

I uploaded a 10 second clip from the beginning of the movie with watermark and onscreen text. Watermark is on a blackish background. The clip ends when the onscreen text fades out. The text is on same black background, so blurring will be fine. The sample you did looked pretty darn good.
http://www.megaupload.com/?d=HTYI1ZUI
I'll check out the logo removing instructions in the next couple days when I get a chunk of time.
Thanks a million for your help!!

[manono]

Thanks for the new sample, but I couldn't come up with anything better. Here's the script I used on the earlier AVI. The crop/resize/addborders have been changed for the NTSC DVD:

LoadPlugin("C:\Path\To\DGDecode.dll")
LoadPlugin("C:\Path\To\TIVTC.dll")
LoadPlugin("C:\Path\To\XLogo.dll")
MPEG2Source("C:\Path\To\Test.d2v")
TFM()
XLogo("LogoFinal_x_32_y_24_2.bmp", X=32, Y=24, alpha=40)
Crop(20,0,-8,-6)
LanczosResize(700,480)
Tweak(Bright=3,Cont=0.90,Coring=False)
AddBorders(10,0,10,0)
AssumeFPS(23.976)
#ConvertToYUY2()#use for CCE, not for HCEnc

Because you had blown out whites and some crushed blacks, I also used Tweak on it. Feel free to adjust anything you like to suit yourself. Here's the BMP I used:

logofinal_x_32_y_24_2.zip

Unzip it and stick it in the same folder as the movie files, adjust the paths at the top of the script, and everything should work right out of the box. You can get the XLogo filter and a nice doc here:

http://avisynth.org/warpenterprises/

If you were talking about removing the word 'Subtitles' from the video, I decided an overlay would be simpler than a logo remover. I added this to the bottom of the script:

A=Trim(0,54)
B=Trim(55,145)
C=Trim(146,0)
D=B.Crop(610,0,0,-444)
E=Overlay(B,D,X=484,Y=50)
A+E+C

Here's a pulled down M2V of that second sample:

xlogo.m2v

[spiritgumm]

Wow, thanks for all the work you've done on this! It'll give me a big head-start for when I review the logo-removal instructions. I was actually still stuck on the audio volume problem (think I finally solved it by using AftenGUI).
Dunno if this is a temporary or permanent result of today's site update, or my server, but your BMP and M2V links aren't working for me.

[manono]

They work now. Try again.

[Admin]

Yeah, you're right. And I tested the links after posting. I wonder why the first AVI link still works, though. Anyway, I'll try again. It won't let me upload the pulled down M2V, only the MPV. Very odd:

I added m2v to the allowed attachment list. And all old forum attachments links will also work fine now.

[manono]

Thanks, Mr. Admin. More editing to do.

[spiritgumm]

They work now. Try again.

yep, got them. I heard the links were offline from a plasma leak in the port nacelle.

Backtracking before I get into the logo removal, I'm hoping you can instruct (or direct) me on bobbing to determine interlace or telecine formats. I've seen a couple guides (one on Doom9 using Virtdubmod) but they didn't mention bobbing, and by their reckoning, my video appears to be interlaced rather than field-shifted. They recommended the Telecide filter (there's no TFM() filter for Virtdubmod).

There's 3 other movies I'd like to work on, and I'd like to figure out their true video properties. Hopefully these attachments will work.
Appears to be progressive Pal:
omnibus.m2v

Appears to be interlaced Pal:
obsession.m2v

Appears to be interlaced Ntsc. It's a dark, poor multi-generation copy, all edges have lines:
Bus_stop.m2v

[manono]

I'm hoping you can instruct (or direct) me on bobbing to determine interlace or telecine formats.

You don't even have to bob it. Just adding SeparateFields() to the script is usually enough. But bobbing using Yadif(Mode=1,Field=1) sometimes makes it easier to see. If you see blends/ghosts/double images, it's field blended and you then have to figure out the real framerate and unblend it. If each unique field is followed by a duplicate (and the source frames look interlaced), it's field-shifted.

They recommended the Telecide filter

The Telecide filter will also fix field-shifted video. It does the same as TFM from TIVTC. All filtering should be done in AviSynth (Telecide is part of Decomb). Don't use any filters in VDub if you can possibly help it.

Omnibus is progressive.
Obsession is progressive.
There's not enough movement in the Bus Stop sample to be entirely sure, but I'm fairly certain it's field-blended from an NTSC source and has to be unblended:

Yadif(Mode=1,Order=1)
SRestore(Frate=23.976)

With a better sample I'd know for sure.

[spiritgumm]

I converted the Omnibus and Obsession PAL videos successfully, but haven't done the first PAL video with logo yet. I needed some small triumphs before the slightly tougher ones.

Couple stupid questions: how does interlaced video look when Separatefields() is used?- do you see "lines" (horizontal gaps)?
If you don't see lines, but objects in motion in the video appear to go back and forth while advancing fields, what does that mean? No matter what the source, will Separatefields always produce images that are half in height?

I have uploaded another sample of "Bus stop." Just a reminder, this video is Ntsc (not Pal):
VTS_01_1.demuxed.m2v
You can see some motion near the end as the seated man stands up. My goal is mainly to brighten the dark video. I also cleaned up bottom of frame, even though it's not visible on TV. Thought it would be good practice for me.
Will the audio have to be framerate-changed if it's being deblended? If deblending isn't correct, how would I know?
Here's the script with your suggestions:

LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\DGDecode.dll")
LoadCplugin("C:\Program Files\AviSynth 2.5\plugins\yadif.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\mt_masktools-26.dll")
Import("C:\Program Files\AviSynth 2.5\plugins\srestore.avs")
MPEG2Source("C:\Documents and Settings\Dave\Desktop\bus samp\VTS_01_1.d2v", cpu=0)
tweak(0,1,50,3)
crop(0,0,0,-12)
addborders(0,0,0,12)
Yadif(Mode=1,Order=1)
SRestore(Frate=23.976)

[manono]

Couple stupid questions: how does interlaced video look when Separatefields() is used?- do you see "lines" (horizontal gaps)?

If you're talking about something shot using interlaced 30fps video cameras, then during movement, after separating the fields, every field is unique. Although I've seen it a few times, the fields themselves aren't supposed to have any interlacing. You only get interlacing when 2 fields from 2 different points in time are woven back together.

If you don't see lines, but objects in motion in the video appear to go back and forth while advancing fields, what does that mean?

The field order has been reversed. If you have it as TFF, then it's supposed to be BFF. And vice-versa.

Just a reminder, this video is Ntsc (not Pal):

I know. It's been field blended from an NTSC source, as I said above, and not from a PAL source. My script from the previous post is what you use.

Will the audio have to be framerate-changed if it's being deblended?

No, the length stays the same.

If deblending isn't correct, how would I know?

If you don't see blends/ghosts/double images after separating the fields or bobbing the video, it's not field blended.

will Separatefields always produce images that are half in height?

Yes.

[spiritgumm]

When I ticked thru the "Bus stop" clip, I couldn't really tell if it was blended or if each field was unique (ie interlaced). It seemed like the latter. I know you've said it's uncommon for PAL dvd to be interlaced, but since this was an Ntsc, how common is interlacing here? Would a homemade dvd recording be interlaced? I have no idea how common (or what typically generates) interlacing here.

Inregards to converting progressive Pal (for those 2 videos I did), I only changed framerate & resize. Did I omit a filter?

Separate question, and let me know if I should edit out and post elsewhere:
Have a Pal progressive xvid (25 fps) in 704x288 (2.40:1). No borders around it. This was converted from a Pal dvd, but the movie was originally from USA and now 4% faster. I'd like to restore to correct slower run time, and burn for Ntsc dvd. The audio was mp3, so I used Belight to convert to wav, then framerate to 23976, then to ac3.
The film DAR is suppose to be 2.35:1, but I've read sometimes people round it to 2.40 when converting to avi (they apparently used Virtualdub). I went thru old threads and cobbled together this script:

Addborders(8,96,8,96) #fill out to 720x480
LanczosResize(720,480)
BlindPP(quant=8, cpu2="xxxxxx", moderate_h=45, moderate_v=55) #someone used this to reduce macroblocking
AssumeFPS(23.976)

I did a couple samples in HC (set to 16:9 - well, made sense at the time), and it looked okay. I noticed the sides got slightly cut off until adding the "8" to borders. Any script suggestions/corrections?

[manono]

When I ticked thru the "Bus stop" clip, I couldn't really tell if it was blended or if each field was unique (ie interlaced). It seemed like the latter.

I already told you what it is and the script to use on it.

I know you've said it's uncommon for PAL dvd to be interlaced, but since this was an Ntsc, how common is interlacing here?

Movies aren't interlaced. If you see any, either it's been telecined (NTSC, and you IVTC it) or it's been blended (and you unblend it).

Would a homemade dvd recording be interlaced?

Anything shot on video will be interlaced.

Inregards to converting progressive Pal (for those 2 videos I did), I only changed framerate & resize. Did I omit a filter?

That sounds OK. If you change the framerate, then of course you also have to slow the audio.

The film DAR is suppose to be 2.35:1, but I've read sometimes people round it to 2.40 when converting to avi

Most of them are 2.39:1, so 2.40:1 is close. Your source is 2.44:1, which is a bit strange, but maybe he cropped into the video so the aspect ratio is correct. When you're dealing with some downloaded stuff, there's no real way to know what it's supposed to be. About all you can do is try and find something round (a moon is good) and start measuring. Even on a DVD there's no real way to be sure it was done correctly.

Your script is all wrong. That's not how it's done. Use FITCD to find the correct Resize/AddBorders.

[spiritgumm]

Here's the resizing script from Fitcd. I did a sample and it looked okay.
BicubicResize(688,256,0,0.6,0,0,704,288 )
AddBorders(16,112,16,112)

For my own knowledge, I researched but couldn't figure what the last 4 entries in BicubicResize do (specifically "0,0,704, 288"). That is, what does "src" mean and what's being done to these parameters src_left, top, height, width? They can't be cropping as in Bilinearesize, can they?

When encoding, when does it really matter to set HCenc to 16:9 or 4:3? I tried both for the sample clip I did above, and they looked the same on TV, regardless of dvd player setup.

When I ticked thru the "Bus stop" clip, I couldn't really tell if it was blended or if each field was unique (ie interlaced). It seemed like the latter.
I already told you what it is and the script to use on it.

Yep, and I'll most gratefully use all the scripts you've been giving me!! I was stating I still couldn't detect interlacing/blending for myself, and the questions that followed were so I could get a mental checklist to deduce it when I can't see the difference. I don't want to keep nagging you (and the board) with every iffy video I need to fix.