Star Wars Conversions-Hit me with your best shot...

[LeedsStriker]

It's tripod in it. They're shit. Can i host it somewhere else?

[Karyudo]

@Gorgoloch
Next compile will use it. Just wish it didn't take a day to convert....

The beauty of Avisynth is that you generally don't have to render out your whole project to see the changes! Make a quick change to your script, reload in VDubMod, and see the results BEFORE you render.


@LeedsStriker

Yeah, i only have the full screen ones but the widescreen ones are out in sep.

No, only the SE trilogy is out in September! It's going to be available as both WS & FS NTSC Region 1 discs, by the way.

The stuff being worked on here is generally the original, pre-SE, HSF ("Han Shoots First") trilogy, which ain't comin' out in September -- GL has said it'll NEVER be made available! 'Course, he said that about an Episode I DVD, too...

What's itvc mean?

Inverse Telecine. That means to reverse the process of telecine, which is the film-scanning process that's necessary to display film material in a video (NTSC or PAL, anyway) format.


@flaninacupboard

Except if you're using PAL LD's, they're not IVTC'ed, the 24fps progressive is just sped up to 25fps

Strictly speaking, I'm still right: PAL uses 2:2 pulldown during telecine, so IVTC is still necessary. But you're right, for all intents and purposes: I was thinking of my usual 3:2 NTSC stuff when I wrote that chain down! PAL is trivially easy to make progressive, since every two fields come from a single film frame. None of this AA-BB-BC-CD-DD crap.


@psklenar (Pat)

+ If I understand correctly, the Canopus ADVC's have a built in Codec, thus eliminating the need for the HuffyUV in your list above?

True, but the Canopus encodes to DV, which isn't as good as HuffYUV. HuffYUV is Huffman encoded (like a Zip file), which means it's lossless compression (can be 24-bit colour). DV is lossless, and can have JPEG-type artifacts. The DV colour space is also 4:2:0 or 4:1:1 (depending on whether it's PAL or NTSC), which means you also lose some colour information. DV is good, of course, but HuffYUV is better!

[Mythos2002]

The advantage of Inverse Telecine is that you are eliminating frames which allows you to apply a greater bitrate to what is left. Film is roughly 24fps and video is needs to be at 29.97fps for standard NTSC TV's anyway. I don't know about PAL. When they convert film to video and go they add duplicate frames in order to achieve the 29.97fps rate.

Bitrate is applied per second and not per frame. So, if you are cutting out roughly 6fps, you are cutting out 21600 frames per hour of video that you don't need. Your bitrate can then be dedicated to the 24fps you are leaving because all or most of the frames you are taking out during the IVTC process are duplicates.

This is where it gets a little fuzzy for me. When you take your 24fps DVD and put it in your DVD Player, it will handle making it display at 29.97fps on your TV. On a PC, the video will remain at 24fps. I think this is also true for HDTV, but I'm not sure.

Hope this help.

Mythos

[flaninacupboard]

Huffy isn't any superior to DV if your final target is DVD. obviously if you're going for HD or Xvid as you end product huffy will look better.

[Karyudo]

Huffy isn't any superior to DV if your final target is DVD.

Yeah, it is. It's not much better, but it's better. You're right that it doesn't make a ton of difference when going to DVD -- because DVD uses a fairly-compressed 4:2:0 colour space -- but there is still the JPEG-like macroblock compression to worry about. Not to mention the 4:1:1 DV to 4:2:0 colour space conversion needed in NTSC. Or is it PAL? One of the two is 4:1:1 in DV. If you start with HuffYUV, it's like compressing straight from the source -- pure 24-bit colour, or 4:4:4!

Still, this is all splitting hairs, and reminiscent of Jesuitical pursuits like arguing over how many angels will fit on the head of a pin: after all, the source is LASERDISC! Not exactly 4:4:4 HD-CAM we're talkin' about...

Karyudo

[Karyudo]

The advantage of Inverse Telecine is that you are eliminating frames which allows you to apply a greater bitrate to what is left. Film is roughly 24fps and video needs to be at 29.97fps for standard NTSC TV's anyway. I don't know about PAL. When they convert film to video they add duplicate frames in order to achieve the 29.97fps rate.

Close, but not quite:

Film is exactly 24 fps. Film is telecined to PAL using 2:2 pulldown at 25 fps -- in other words, 24 fps film just sped up 4% when converting to PAL. Film is telecined to NTSC using 3:2 pulldown at 23.976 fps -- or, 0.1% slower. No duplicate frames are added, exactly, but some film frames last for 2 NTSC video fields, and others last for 3 fields. This gives 4 film frames at 23.976 fps spread over 5 video frames at 29.97 fps. (One side note is that, inevitably, one of every four film frames is spread across two video frames.) This 3:2 pulldown can be reversed via IVTC to leave 23.976 fps progressive video frames. You could make your computer-based AVI 24 fps, but that's sort of pointless, considering the audio is already adjusted to match the 23.976 fps speed of the telecined video.

This is where it gets a little fuzzy for me. When you take your 24fps DVD and put it in your DVD Player, it will handle making it display at 29.97fps on your TV. On a PC, the video will remain at 24fps. I think this is also true for HDTV, but I'm not sure.

The part you're missing is the TFF (top field first) and RFF (repeat first field) flags that are set in the MPEG-2 bitstream, to allow the DVD player to effectively telecine the 23.976 fps frames stored on the disc to 29.97 fps NTSC video on the fly. You're right: HD works the same way. Well, at least 1080i ATSC HD does.

Karyudo

(P.S. I'm quite the little post whore today, aren't I?)

[Mythos2002]

Thanks for correcting me on that. I like to have all the accurate information possible for the future.

As a side note, it looks like DVD-Lab Pro is available in the trialware stage. If it can do everything it claims effectively, it looks like it is the authoring package to go with. It should be out in May or June for purchase. I've been happy with the Standard 1.4.2 Beta so far.

Mythos

[dphirschler]

Man, I want that software bad (DVD Lab Pro). But I am in the hole already with this project. I am going to have to go with the freeware tools. So far it has been IFOedit for authoring, but I may switch over to DVDauthor with the GUI. It seems pretty powerful too. Problem is, I cannot find a guide for it. I may look again today for a good guide.


Darryl

[Gorgoloch]

If I'm using IVTC do I want to ue 3:2 pulldown as I'm going to TV? Or does that just bodge the whole thing up? I think I just used that and I'm getting some flicker in motion. Really annoying. It's like I had few pints before watching.

Also, I'm getting these faint vertical lines on the TV. Anyone have any clue where those came from? Pretty sure I didn't check off 'Add Annoying faint Vertical Lines' in any of my processing.

G-Loch

[Mythos2002]

Just use the 3:2 pulldown when playback on the Video Tab and not the one on the Advanced Tab. Also, make sure your framerate says 23.976 fps (internally 29.97 fps). Also on the Advanced Tab under IVTC, choose Flicker Prioritized and Deinterlace None. Use the Source Range filter to encode a small clip for testing. I don't know about the faint lines, but that should take care of the flickering.

Hope that helps.

Mythos

[Gorgoloch]

Thanks, running the test now. Figured I had some setting a bit off.

The lines I'm not sure about. Not sure if they come in during the burn. Since the file is too big to write using DVD Workshop, I write an ifo, then use DVD Shrink to get it on DVD. A nice little trick that someone here mentioned. So maybe there's something there. Or maybe SOmething happened to my AVI when I put the Subtitles in. Have to keep testing.

Going to be a long day. (fun, but long)

Thanks again.

[Gary Lightfoot]

Hi all,

Has anyone got a large size "Star Wars" logo like that you see at the beginning of each movie (the one that travels back)? I've a couple, but I need to manualy paint out the insides of all the letters so I can make it just black and white, and it's quite tedious and time consuming.

I'd like a plain black and white image if possible, or one with the least amount of color shading so it can be done easily.

TIA.

Gary.

[Gorgoloch]

Thanks Mythos. Just finished the test. Seemed to do the trick.

G-Loch

[dphirschler]

SW logo. You can probably get a nice vector out of the font. Pop that into Illustrator, convert to outlines, and there you go.


Darryl

[Gary Lightfoot]

Sounds like a good idea - where can I get the font?

Cheers.

Gary.

[vhelp]

Man, I want that software bad (DVD Lab Pro). But I am in the hole already with this project. I am going to have to go with the freeware tools. So far it has been IFOedit for authoring, but I may switch over to DVDauthor with the GUI. It seems pretty powerful too. Problem is, I cannot find a guide for it. I may look again today for a good guide.


Darryl

Hay dphirschler,

Liquid217 has a guide. I think its at his site. However, I dont' think or not too
sure that it's ben revised according to his latest GUI guild though but
you can ask him over on the DVD Author thread that Baldrick started.
I can't remember the link at the moment though.

-vhelp

[vhelp]

@ all.

Anybody else up for posting their samples yet ??

I'm still testing/debugging the best method (ADVC-100 vs. Analog cap card)
and so far, I'm still positive about my ATW card, because of the added color
space. It's a bit more richer, IMO.

Maybe I should post a small sample (though its not SW, I'm still waiting for
mine to arrive via e-bay) - - can't wait.
--> which one.. topGun; SGirls; Mohigans; bInstinks ??

-vhelp

[vhelp]

.
.
One more thing. On the topic of "hatches" or "checkers" or "dot crawl" or
whatever you all refer to it as..

I think that this is truely a Capture card related issue (including dv devices)
and not something related to wires or sw filters or what-have-you.

I purchased very expensive cables just for this and even desinged to
eliminate "dot crawl" but I still have the above issues. The cable is a short
one also.. only 1 meters long.
.
.
I've fed it into my ATW capture card and my ADVC-100 box, and I still get the
same effect. This is purely a capture device problem.
TV's and capture card previews do NOT show this problem. Therefore, I
have concluded that it's a capture device problem.
.
.
It is also my personal opinion that no-one truely knows what is the cause of
this issue (other than the one I just stated above) add to that, that no-one
knows how to eliminate them, once and for all

So, the search goes on, to rid these issues. However, I beleive that PAL
does not suffer this. I have not seen any mention of these issues w/ PAL
sources.. or did I miss anyone ?

Until then, I will continue my quest for the answer we all are seeking. . .
-vhelp

[Perro Grande]

VHelp,

The thing that bugs me about the crosshatching is that I actually FOUND an explanation of it, and can't find it again....

I've been able to repeat that darn crosshatching on both my capture devices AND my big TV. I've dropped back to using composite wherever I can't go component video... Frustrating!

I *believe* it has something to do with the Luminance (Y) signal bleeding on to the Chrominance (C) signal. However, I don't know this for certain. Surely, though, major capture devices like Canopus and ATI would not systemically have issues with SVideo without there being a HUGE swell of commentary on this.

I did just have a thought, though -- while writing this... Do you have both SVideo and Composite connected at the same time on EITHER end of the circuit? If the output signals are not properly isolated from one another, this could be the path by which Y is meeting C before it should... I might try a few experiments with this idea....


I'm going to see if I can post some clips somewhere. I'm presently working on Return of the Jedi. I also have some max bitrate clips done with my Panasonic set-top box.

[Perro Grande]

Well, here is more for the mystery...

This is captured from LD using SVideo (the same source and LD player that produced my checkerboard example earlier). The difference is that this time I ran it through the DV Passthrough option on my Sony Camcorder instead of the Canopus box...

No checkerboard...



ADDENDUM: I just captured via the Canopus box w/SVideo and had no checkboard pattern. ARGH! Now I'm REALLY confused! The only difference between then and now is the computer/firewire card for tonight's test is different (my main capture computer is busy). But if it is somehow related to the computer or the firewire card, why does composite video work perfectly?

Sheesh -- the more I dig into this problem, the less I understand it...

[dphirschler]

I believe this is a decent description of dot crawl.

http://www.cybertheater.com/Tech_Reports/Comb_Filter_Tut/guide_comb_filters.html


Darryl

[FulciLives]

Well, here is more for the mystery...

This is captured from LD using SVideo (the same source and LD player that produced my checkerboard example earlier). The difference is that this time I ran it through the DV Passthrough option on my Sony Camcorder instead of the Canopus box...

No checkerboard...

*** IMAGE DELETED TO SAVE SPACE ***

ADDENDUM: I just captured via the Canopus box w/SVideo and had no checkboard pattern. ARGH! Now I'm REALLY confused! The only difference between then and now is the computer/firewire card for tonight's test is different (my main capture computer is busy). But if it is somehow related to the computer or the firewire card, why does composite video work perfectly?

Sheesh -- the more I dig into this problem, the less I understand it...

Damn that is STRANGE !!!

If I'm not mistaken doesn't Canopus post a list of recommended 1394 PCI boards for compatability with their products such as the ADVC-100

If so then perhaps this is an issue as to why they do that.

- John "FulciLives" Coleman

[dphirschler]

Back on the subject of fonts. Does anyone know offhand the font used on the "The Definitive Collection" box? It's sort of a narrow slightly squared sans font with an intersecting center horizontal line.


Darryl

[viande]

anyone tried this filter?
http://home.earthlink.net/~tacosalad/video/dotcrawl.html

[Gary Lightfoot]

Will you have my babies??

I don't know wther to kill you or kiss you - it actualy seems to cure my problem!! I'd tried almost everything else, and resigned myself to the fact this artefact was going to have to stay...

I'll have to test it with various scenes that have it, and see if it causes any other side effects, but it does look like the answer to my only 'issue' with my conversions.

Many Thanks!!

Gary

[flaninacupboard]

Kill him! he will be the first martyr in our cause!!!!!

(obviously i am joking. no sand in my shoes or towel on my head)


[obviously i'm -not- joking, but this isn't a political statement]


{ok, it is really, but just ignore it} :P

[dphirschler]

This complicates things for me if it works. According to the docs, I will have to run the virtualdub filter before the inverse telecine operation. Problem is, almost everything I do is being done in AVIsynth. The only solution I can think of is to run the filter on my video, then save it (gasp! uncompressed?), then run the AVIsynth script.


Darryl

[Gary Lightfoot]

I've been using TMPGenc for all of my processing, except I have used virtualdub to convert from my captured avi to huffyuv avi for space saving purposes (lots more gigs left!).

I can now use vdub to remove the dots and make the original avi smaller all in one go. This filter is looking pretty good so far.

Has anyone used a sharpen filter to see if that improves the image clarity at all? I had been using a little noise reduction to remove grain and picture noise (20 1 20 in tmpgenc) as I couldn't use sharpen because it would enhance the dots and some edge hatching. Now I can afford to add about 20 20 in the sharpen filter.

Blimey - I'm thinking of re-doing all the ESB avi files to make the image better than it already is... That'll mean another load of ovenight renderings to mpg2!!

Gary.

[Gorgoloch]

Quick question. I still haven't ventured to try AviSnth (probaly should) but when I used the Subtitle filter in Vdub I couldn't do a direct stream copy, so I used the same codec that the the file was captured in (Panosonic DV CODEC), I tried Huffy, but it was all screwed up (may try the older version). If I use the same codec, I would asume that I'm getting the same compression and quality, but knowing a bit about computers, I know that that isn't always the case. I had a faint vertical line issue that I think was due to this conversion, so I'm wondering if going to Huffy would solve that. But mainly I'm wondering if I use the same codec if I'm losing anything, or if it's 'bad' in any way.

I'm curious as this dot-crawl filter may come in handy.


G-Loch

[LeedsStriker]

I done A New Hope. 1.3 gigs. 640x480 mp3 sound. Looks ok for what i want. Can I put it on kazaa to show u all?