Laserdisc backup on DVD: interlace or not?

[StephLG]

Hi,
I'm new to this forum and, after some research for posts about interlaced vs deinterlaced and IVTC, I start being confused :confused:

What I'd like to do is backup some NTSC Laserdiscs on DVD (maybe I'll keep the files on my HD since I watch movies with WinDVD and a DLP projector). I plan to capture the LaserDisc video/sound with my DV500 S-video input (Pinnacle's DV codec). Therefore, I'll end up with a 29.97fps interlaced avi file. I then plan to encode in MPEG2 using TMPGEnc v2.52 but I don't know what settings I must use.

Since I want the best possible quality, do I have to:
- de-interlace the avi file?
- use IVTC or Inverse 3:2 pull-down?

As I wrote above, I'll watch the movie with WinDVD and a 1024x768 DLP projector. This means WinDVD will do some resize and/or deinterlace operations. Will the result be better if I deinterlace during the encoding part or should I let WinDVD do the job?
If I de-interlace the file during the encoding, will my DVD be watchable on a regular TV?

As you can see, I've read too much articles in a too short time and I start mixing everything :scared: I've read about 23,97fps framerate which can save space to improve bitrate but I don't know if it can/should be done in my case.

Any suggestion/advice will be welcome :)

Thanx,
Steph (sorry for my poor English, I'm French ;) )

[rmatei]

Steph,
I'm not sure if you want to ultimately save to a DVD or keep watching from the PC.

If you don't want a DVD from it, and since your source is most likely telecined, you should do IVTC to end up with a 23.976 progressive video. No deinterlacing necessary unless your source is shot on video which would be surprising.

If you do want an NTSC DVD I'm not sure but most likely you don't need to do anything as far as fields are concerned because you're going from 29.97 interlaced (LD) to 29.97 interlaced (DVD). If you want a PAL DVD, then you would need to IVTC and speed up to 25 fps which is significantly more complicated.

[StephLG]

Hello rmatei,

thank you for your reply. Let me try to explain a little bit more what I'd like to do: my main goal is to create a DVD compliant file/structure to be played from the hard-drive, with WinDVD or PowerDVD, and displayed with a DLP XGA video-projector.

Then, if the file is not too big (but quality is my first priority), I may want to burn a real DVD to be able to play the movie in a standalone DVD player on a TV.

Now, let's go back to your answer; I'm not sure I understand everything well:
Originally posted by rmatei
If you don't want a DVD from it, and since your source is most likely telecined, you should do IVTC to end up with a 23.976 progressive video. No deinterlacing necessary unless your source is shot on video which would be surprising.

You say 'end up with a 23.976 progressive video' without deinterlacing!? I thought that if I captured the 29.97 interlaced LD movie with the DV500 I would end up with a 29.97 interlaced avi file. Does the IVTC filter produce a 23.976 progressive file?

You also stated:

If you do want an NTSC DVD...you don't need to do anything as far as fields are concerned because you're going from 29.97 interlaced (LD) to 29.97 interlaced (DVD).

Actualy, I'm not sure I understand IVTC. I thought one of its improvements for NTSC movies was to encode less frames per second, for a given bitrate, and therefore allow to allocate more bits per frame. This is interesting when you want to burn a DVD but if I keep the files on my drive, it's no use; I'd better do IVTC in the second case. Am I wrong?

Is 23.976fps mode compatible with the DVD standard? (I mean, are there a lot of NTSC DVDs encoded like that?).

I'm waiting for your advices. Then, I'll have questions about audio ;)

Steph.

[Atlantis]

I’m doing exactly what you want to do and I have DV500 plus too! I have several worthy laser discs that will probably never see the day on DVD thanks to the revisionism of their creators. So far I have made 2 DVD-Rs from my LDs (SW: ANH and SW: ESB) and I’m very pleased with the result. It looks very good and professional. With the audio commentary, subtitles and everything.

Yes you have to remove the 3:2 pulldown to have a 23.976 progressive frame file. In TMPGEnc you have an option to enable a flag that every commercial DVD has to apply 3:2 pulldown on the fly (BY THE PLAYER) and make an NTSC video for TV! However on PC you will get even better quality because it will show it in 23.976 progressive! Ask any question you like because I have done this and am doing this from start to end and we also both have the same video capture card!

[Atlantis]

For 3:2 pulldown removal, I have tried almost everything. None but one is good believe me. Even TMPGEnc inverse telecine is not good. All these softwares try to predict the phase and sometimes do a bad job of that and you will always end up with some interlaced frames, even worse some try deinterlace filters to deinterlace those frames. So at the end it will not be that good or smooth and you don’t have the exact frames. The best solution is to specifically tell what is the phase and not leave the decision to a software that do it on the fly. I use this Avisynth script. Forget about all the other methods!

OpenDMLSource("xxxx")

DoubleWeave()

#Pulldown(0,2)
#Pulldown(1,3)
Pulldown(2,4)
#Pulldown(0,3)
#Pulldown(1,4)

Only two of these five Pulldown should work and you should use only one of them and the rest 4 should be on commentary #

[StephLG]

Hi Atlantis, and thank you for joining this thread.

You know what? The LDs I absolutly want to backup on DVDs (or at least on my HD with a DVD structure for my Home-Theater) are: StarWars Special Ed. trilogy! I own the last edition (NTSC, Widescreen, Dolby Digital AC3). So, I couldn't find better specialist than you.
Here is what I had planed to do:
- capture with DV500 under Premiere 6, with 'Pinnacle Systems DV500 - NTSC' setting (full res 720x480, 29,97fps, 44,1KHz 16-bit stereo sound). This will produce 3 avi files (3 disc sides) that I can join without any other processing but undesirable frames removal at the beginning and end of each part of the movie (what should I use to join? Premiere?)
- for the sound, I will re-record the Left+Right chanels directly from the SPDIF output of my LD-Player (what frequency must I choose? 44,1KHz or 48KHz?), and then the 4 other chanels from the pre-amp output of my HT amp (again, 44,1KHz or 48KHz?) with the following method: Right+Center, Right+Sub, Right+SurLeft, Right+SurRight so that I'll be able to align the 4 tracks with the analog Left+Right recorded with the video. I don't know yet what sound software I'll use for this (any idea?).
Once I have my 6 wave files, I'll encode them in DTS fullrate (a friend of mine owns Minnetonka Surcode DVD DTS software) to not too much quality. By the way, what file type should I choose: DTS Wave (.wav), DTS Padded (.dts) or DTS Compact (.cpt)?

So, at this point, I'll have a (huge) avi file without sound and a 5.1 chanels sound file.
At this point, I really need your help. Now,I'm sure I have to IVTC to go back to 23.976fps. But, to be sure to understand well: my avi file is, and will stay, interlaced? is that right? It's 29.97fps interlaced and will become 23.976fps interlaced, right?

To be honest, I don't understand the AVISynth trick you're talking about. I'm not familiar with this tool. I know it can be used to frame-served from Premier to TMPGEnc. Is that what you advise me to do? I'll need help on that point.

One more question: I want to encode the video in MPG2 DVD compliant but without sound, since I process the sound separatly in DTS. Will the video still be synchronized with the sound? What (simple)DVD authoring software should I use?

I hope I didn't ask too much :rolleyes: It's not very easy to explain all this stuff, especially in a foreign language ;)

Thanx,
Steph.

[Atlantis]

Ok! We should first go step by step. You have to answer few questions before I can reply correctly.

01) How much free disk space do you have? Are they in NTFS?
02) What windows do you have? What version of driver do you use for DV500?
03) Do you want your final MPEG file to be anamorphic or non-anamorphic?

Few quick answers. You have to capture the audio in 48, that is DVD spec but unfortunately I don’t know anything about DTS or 5.1 audio. My DVDs are stereo only! I cannot get 5.1 sound from my LD player.

No the resulted 23.976fps file is progressive, not interlaced! That’s the point. You do the 3:2 pulldown removal to get the frames the way they looked in the theater, progressive. There is no 23.976 interlaced movie!

You don’t need to understand the AVISynth script, just use it!

By the way I live in France but I’m not French. I still prefer to talk in English because my French is not that good and I’m more comfortable in English!

By the way, why the special editions! Greedo shooting first? Censored blasters! No way! ;-)

[StephLG]

Originally posted by Atlantis
01) How much free disk space do you have? Are they in NTFS?
80Gb UDMA33 for capture, 18GB U2W SCSI for editing. All NTFS.
I've already successfully captured more than 2hours of TV program (about 30Gb) without a frame dropped.

02) What windows do you have? What version of driver do you use for DV500?

Windows 2000 Pro Sp2, DV500 driver v3.0

03) Do you want your final MPEG file to be anamorphic or non-anamorphic?

I wish it could be anamorphic but the LDs are 4:3 letterboxed (with black bars). Can I convert it to anamorphic? I've already tried this with a short video; cut some pixel at top and bottom and changed the pixel aspect ratio to 1.2 but the result was bad (due to resize, I guess).


No the resulted 23.976fps file is progressive, not interlaced! That’s the point. You do the 3:2 pulldown removal to get the frames the way they looked in the theater, progressive. There is no 23.976 interlaced movie!

You mean that the 3:2 pulldown removal also deinterlace the video?


You don’t need to understand the AVISynth script, just use it!

I'll still need your help on that point. I didn't get what you said with 'Only two of these five Pulldown should work and you should use only one of them and the rest 4 should be on commentary #'


By the way I live in France but I’m not French. I still prefer to talk in English because my French is not that good and I’m more comfortable in English!

No problem!


By the way, why the special editions! Greedo shooting first? Censored blasters! No way! ;-)
I know they censored like that and I agry it's stupid; but these are the only version I own, and they are AC3 5.1.

[Atlantis]

OK, it’s good that you have a lot of free space. Because you can make intermediate avi files. That’s what I do. Yes you can achieve good looking pseudo-anamorphic files. That’s what I do because I have a widescreen TV.

We will go step by step.
1) First capture a small video.
2) Install VirtualDub and huffyuv avi codec.
3) Isntall Avisynth 1.05.
4) Write this in a text file.

OpenDMLSource("xxxx.avi")

DoubleWeave()

#Pulldown(0,2)
#Pulldown(1,3)
Pulldown(2,4)
#Pulldown(0,3)
#Pulldown(1,4)

xxxx.avi is the name of your captured avi file. Rename your text file to a file with extension .avs and put it in the same folder as avi and open the avs in VirtualDub. Can you see the video? If yes, I will explain the next step.

[StephLG]

Hi Atlantis,

I've just captured a 2min sequence from 'The Return Of The Jedi' (speedbike sequence) with Premiere (NTSC 29.97fps, 0.9 pixel aspect ratio).
If I open the file itself in VirtualDub, it seems it can't play the video: audio is Ok but no video. If I use your script: video is perfect but the sound is all messed up, like if the file was too big. Maybe it's normal.
For your information, the video can be played in media player but with no sound (because of Pinnacle DV Codec).

So, I'm waiting for further instructions ;)

[Atlantis]

Don’t worry about the audio it is going to work! OK now make 5 copies of the same avs file and in each one put only one pulldown without #. Like this

File1

OpenDMLSource("xxxx.avi")

DoubleWeave()

Pulldown(0,2)
#Pulldown(1,3)
#Pulldown(2,4)
#Pulldown(0,3)
#Pulldown(1,4)

File 2

OpenDMLSource("xxxx.avi")

DoubleWeave()

#Pulldown(0,2)
Pulldown(1,3)
#Pulldown(2,4)
#Pulldown(0,3)
#Pulldown(1,4)

And so on

Then open each avs file in virtualdub and check which one gives good result.
This means that you should see progressive frames and no interlaced frames.

2 of those avs files should work well. Let me know when you did it.

[StephLG]

Hi Atlantis,

Ok, I ran your scripts and only the first and third one gave a perfect video (I can see interlaced frames in the others).

Just a few questions:
- I used Huffyuv codec; is it completely lossless? I used the default options:
YUV2 Predict Median(Best)
RGB Predict Gradient(Best)
Always suggest RGB format for output

Are you OK with that?

- my original 2min footage, captured with Pinnacle DV codec is 445 440Kb large; the IVTC resulting files, with Huffyuv codec, are 1 077 596Kb and 1 077 322Kb large. Why have they a (slightly) different size? I can see in VirtualDub that the first one has one more frame (2886 vs 2885).

- between 1st and 3rd setting, should there be one better than the other?

- will I have to do this test for every side of the LD captured or will one of these two settings be OK for all my captures (3 per episode)?

Steph.

[Atlantis]

Ok, you just went ahead quickly. I was waiting so that you finish the last steps. Now we are going into the next steps.

Of the 2 avs files that look good, you can use any of them. Just check the start and end of your capture to be sure that it looks progressive because sometimes they change the phase in the middle.

Now open one of the good avs files in VirtualDub. Edit the movie so that there is no black frames at the beginning or at the end of the capture, only the movie. (Not this time, when you completely captured one side!)

Now add a null transform filter and crop all the black bars on the top, bottom, left and right. You should get a picture around 704 x 276.

Now add a Temporal Smoother (3) filter.

Now add a resize filter Bicubic 720 x 364.

Now choose the Huffyuv codec at best. It’s a perfect lossless codec, ideal for editing and intermediate works.

Now save your avi file.

After that you have saved your files for example
Part_1.avi
Part_2.avi
Part_3.avi

Create a text avs file and add these lines in it

a = OpenDMLSource("Part_1.avi")
b = OpenDMLSource("Part_2.avi")
c = OpenDMLSource("Part_3.avi")

AlignedSplice(a, b, c)

Now you can open this avs file in TMPGEnc and encode your entire movie in one part which gives the best result. In TMPGEnc use the 3:2 pulldown flag with the picture set to 16:9. Size 720 x 480. 2 Pass VBR and it is very important that in the “advanced tab” you choose source aspect ratio 1:1 VGA and Video arrange method: Center.

Voila, you have now a perfect anamorphic picture and the result of months of testing by me in just 2 days!

[StephLG]

Thank you Atlantis!

I won't do this before 2 weeks 'cause I already have a 25Gb movie to encode before I can get rid of it and have a empty 80Gb hard drive.

I'll keep you informed if I encounter problems, and to tell you how it looks like.

Thanks again.
Steph.

[tyee]

Great thread Atlantis! I'm doing the same thing now and have a couple of questions too -

[QUOTE]Originally posted by Atlantis
Of the 2 avs files that look good, you can use any of them. Just check the start and end of your capture to be sure that it looks progressive because sometimes they change the phase in the middle.


Q1. What do we do if the phase is changed in the middle. I guess we have to search and find where, then use VirtualDub to save these two sections, then redo the changed phase section again. Is this correct?



Now open one of the good avs files in VirtualDub. Edit the movie so that there is no black frames at the beginning or at the end of the capture, only the movie. (Not this time, when you completely captured one side!)

Now add a null transform filter and crop all the black bars on the top, bottom, left and right. You should get a picture around 704 x 276.

Now add a Temporal Smoother (3) filter.

Q2. Does this filter reduce detail? Is it really necessary for laserdiscs?



Now add a resize filter Bicubic 720 x 364.

Now choose the Huffyuv codec at best. It’s a perfect lossless codec, ideal for editing and intermediate works.

Q3. I just tried this with a small section of a laserdisc, then tried playing it back on my P3-800, and the playback in Mediaplayer was very random, meaning one frame would show, then after a couple of seconds, another frame would show.Not smooth at all. Is this normal for a huffy encode. Is a huffy encode a totally uncompressed avi, and my computer doen't have the throughput speed to view it?



Now save your avi file.

After that you have saved your files for example
Part_1.avi
Part_2.avi
Part_3.avi

Create a text avs file and add these lines in it

a = OpenDMLSource("Part_1.avi")
b = OpenDMLSource("Part_2.avi")
c = OpenDMLSource("Part_3.avi")

AlignedSplice(a, b, c)


Q4. Can't VirtualDub open many avi files at once so you could edit them all as one, then frameserve to your encoding program (CCE or TMPEG)?

Thanks, Atlantis, keep up the good tutorial work! Do you like TMPEG better than CCE?

Tyee

[Atlantis]

Hi tyee!

Q1) If the phase is changed in the middle then you have to create 2 avi files for one side. You have to cut the bad part for each file. As I said you have to create only good usable avi files because later you will append them and feed them into TMPGEnc with no re-cutting!

Q2) It’s easy. Try once with the filter and once without it and decide what you like. But believe me, it is better to use it for an analog capture in DV!

Q3) Yes that’s because your computer cannot handle it. But it’s not important. You want to feed it to TMPGEnc.

Q4) Yes but never do that. I will say why. That alone took me few months and trying different softwares for finding a solution. For the moment the only solution in the world that I have found is avisynth. There is absolutely no other solution and I will be glad to learn if anyone can find a different solution. Different softwares have different problems. With VirtualDub you will get a slight out of synch audio after each part. Because of the way VirtualDub appends and I have talked to its creator and Avery Lee confirmed this (audio is a little shorter than video). Also Premiere is incapable of handling 23.976 avi files.

I don’t like CCE because each time I encounter a different error with it but if you like it and think that it works, go ahead!

[tyee]

Hi Atlantis
Thanks for the reply. I've been playing in the last few hours with a laserdisc and experimenting. Here's my avsynth script for testing -

LoadPlugin("H:\Rip\MPEG2Dec\MPEG2DEC.dll")
OpenDMLSource("N:\jp1.avi")
DoubleWeave()
Pulldown(1,3)
BicubicResize(704,458,0,0.6,8,68,704,344)
TemporalSmoother(2,1)
AddBorders(8,11,8,11)
ResampleAudio(44100) # CCE 2.5 'crashfix' for Athlons

It seems to work great except there is quite a bit of noise still in the picture. I'll bump up Temporal Smoother to 2,2 and see what happens. As you can see here, I'm using avisynth to do all the resizing but I used VirtualDub to get the crop figures along with the "FitCD" program (which is excellent) for calculating resize values for anamorphic encoding. If the order is wrong here please let me know. Let me know if using VirtualDub is better than AVSynth for noise reduction?

Here's a simple equation to use for calculating the anamorphic resize values. I think your values might have a small error but on screen it won't be noticible I think.

VirtualDub Cropped Aspect ratio/1.1 = New Resized Anamorphic Aspect Ratio/.825

This equation is for NTSC only. I think you might have PAL where you are so for PAL change the 1.1 to .9152542 and change the .825 to .6864407. I just tried it both ways and the final answer of the New Resized Anamorphic Aspect Ratio is the same!
In your movie you said you get a cropped value of 704 x 276, then the equation becomes 2.55/1.1 = New/.825. New Resized Anamorphic Aspect Ratio should then be 1.913. You resized to 720 x 364 = 1.978. Not much different and probably not noticeable but I thought I should mention it. I've spent many hours getting a handle on these aspect ratios and thanks to "FitCD", mb1, and this great web page -

http://www.mir.com/DMG/aspect.html

I now think I understand it all.

The 1.1(NTSC)and .9152542(PAL) and .825(NTSC) and .6864407(PAL) are of course the pixel aspect ratios of nonanamorphic (4/3)and anamorphic (16/9) DVDs. This is a very slick equation and agrees with the "FitCD" program. Oops, I forgot to say I capture my Laserdisc via a Firewire Sony Digital Media Converter which takes S-video/audio from Laserdisc and converts it to DV format that's stored as my source .avi file to work with. That's all I've got for capture but it's OK. Off I go to more experimenting!

Tyee

[Atlantis]

About avisynth filters I don’t know, you have to do some tests yourself. The 720 x 364 size comes from commercial DVDs. I have checked several 2.35:1 aspect ratio anamorphic DVDs and they all are more or less this size. So I use this size.

[StephLG]

Hi you two,

I couldn't wait to complete the job with my2min footage so I did it!

It worked perfectly well but I'm not sure I choose the right settings in TMPGEnc (there's so much of them :( ).
I choose:
VBR 2-pass, max 7000, average 5000 and min 4000
DC component precision 9bits
Motion search precision High Quality (slow)

My original 445Mb DV file results in a 80Mb Mpeg2 file. It took 20 minutes to encode on my Dual-PIII 700 Xeon (2Mb cache/Cpu), a rather old setup but which used to be quite high-end about 1 year and half ago :rolleyes: .
I thought it was a rather high bitrate but I find the result in PowerDVD a bit soft, with visible pixels. Maybe the resize did that.
Here are some captures (sorry for the size but I didn't want to lose some more quality)

Before TMPGEnc:
http://mapage.noos.fr/stephane.leguevel/before.jpg

After TMPGEnc:
http://mapage.noos.fr/stephane.leguevel/after.jpg

Don't bother for the aspect ratio; the capture is non-anamorphic.

What do you think?

[rmatei]

I think you guys would be better off at the DVD encoding forum. So, thread moved. :)

[easy2Bcheesy]

This all seems very very complicated. I have produced my own home-made Star Wars Trilogy and Indiana Jones Trilogy DVDs and used a far, far simpler technique.

The Star Wars Trilogy came from the PAL LD Special Edition. The Indiana Jones movies came from Sky+ (UK digital satellite TV system, kind of like TiVo).

In both cases I simply captured at 25fps, interlaced.I used the Matrox RT2500 for capture and TMPEG for encoding into MPEG2. I tried CCE 2.5 but was simply bogged down with lots of strange errors. Only the CCE Premiere plug-in worked, but 1 hour of footage took 36 hours to encode using this.

To fit the movies onto single DVD-Rs, I used two pass VBR, set to an average of 4800, with a maximum of 6500 and a minimum of 2000. Sonic Foundry Softencode was used to bring the audio into the AC3 format, with a 320 bitrate here.

Now I'm not saying that all this telecine stuff won't produce a better image, but I'm watching the movies on a TV as opposed to a projector or monitor, so I very much doubt you'll see too much difference in quality.

[StephLG]

Hi,

easy2Bcheesy, this method may seem complicated but it's actually quite simple when you have the scripts.

Since your version of StarWars was PAL, you didn't need to do IVTC; for NTSC, on the other hand, it is usefull: you end-up with progressive frames and encode less frames (23,976/s instead of 29,97/s).

[Atlantis]

That’s the best you can get with a non-anamorphic transfer. If you don’t like it, don’t resize it and keep it non-anamorphic but personally I want anamorphic even pseudo-anamorphic because on my widescreen TV it looks a lot better than a non-anamorphic movie with zoom on and all the visible scan lines!

[Atlantis]

Well this telecine stuff does make a lot of difference on NTSC films because you use the same amount of bitrate instead of for 29.97 interlaced frames for 23.976 progressive frames. Not only 23.976 is less frames but also it’s progressive and compress a lot better.

By the way, about your PAL movies. It’s nomal that it’s easier because PAL doesn’t need any inverse telecine. I don’t know why you used interlaced because PAL films are normally progressive. They don’t do any transformation on a movie for PAL transfer, only it plays faster. You could use progressive frames.

[talman]

StephLG:

Can you please explain how you can get the AC3 off your laserdiscs? I have lots of LD's that I'd love to be able to preserve the 5.1 sound on. Thanks!

[StephLG]

talman:

Actually, you can't really preserve the AC3 5.1 sound of your LDs since the 4 channels that are encoded in the analog RF signal have to be decoded in hardware (by your preamp or a separate decoder). There is no PC software for AC-RF decoding :( This means you have to record most of the channels (those coded in the RF signal) from the analog out of your preamp, then sync/adapt the 5.1 channels and re-encode them in a DVD compliant format. I chose DTS because it can have a higher bitrate than AC3 and I don't know what quality an AC3 compression would produce on an already decoded once analog stream. Maybe it's worth trying.

So, what i'm going to do (it'll be very fastidious but I think it's worth trying) is:
- play the movie once to record the video and the analog R+L channels
- play the movie again to record the digital R+L channels, from the SPDIF out of my LD player. In this case, I'll preserve the original quality.
- play the movie 4 other times to record, out of preamp, R+Center, R+SurroundRight, R+SurroundLeft, R+Sub. I record the Right analog channel each time so that I can synchronize the 4 channels with the stereo R+L recorded with the video.
- after having synchronized the 6 channels with the video, I'll encode separately the video (MPG2 in TMPGEnc) and the sound (Surcode DVD DTS)

It'll be long and I'll need some drive space but I want to try.

[easy2Bcheesy]

Thanks Atlantis, I might well do a test with a few minutes of Star Wars and see if it's worth the effort of re-doing it all. I'm very happy with my transfers as is to be perfectly honest, but I like the sound of resizing the image into an anamorphic format - as you say, I could see a tangible improvement in image quality doing that.

PAL LDs don't support AC3, so that's something I can't explore really :(

[Haywire]

Has anyone else used a different capture card than the DV500 with s-video in? Like the Matrox G450-eTV or similar? I think the matrox only does 702x480 (Not 720!?) I like this thread but a Pinnacle DV500 costs around $1000 cdn where a Matrox is half the price..

[Atlantis]

Actually DV500 is not that good. I was unaware of NTSC DV problems. Later I found out that DV NTSC has some serious problems with red vertical lines. It gives a step looking. With DV500 you are capturing analog and then converted to DV. If you have a cheap analog capture card and a fast enough computer, you can easily capture and achieve even better results than DV500. You should use huffyuv for capture!

By the way there is a picture attached to show the NTSC DV artifact. Steph, I’m also wondering if you see the same thing with red lightsabers when you capture. Is it possible for you to post an image of a scene with a red lightsaber? Thanks!

By the way I have attached the file but I don't see it, why?

[StephLG]

Atlantis, I haven't captured a scene with lightsaber yet. I go in vaccations on Sunday but, if I have time, I'll try to post a picture with a red one.

What you say about the analog DV500 input being converted to DV with artefacts is interesting. When I first tried to capture my LDs with the DV500, I thought that I could use HuffYUV codecs under Premiere but I realized that DV500 forces the use of their DV codec, since it's done at the source in hardware by the card itself.
I also have a Miro PCTV Rave card in another PC; maybe I should try to capture with it, in S-video and using HuffYUV codecs. The files will be bigger (445Mb for 2min in DV vs 1Gb for the same 2min in HuffYUV).

About inserting your picture, it has to be hosted on the internet, not on your hard drive, maybe that's why. If you have 'IMG code' set to OFF in your preferences, you'll see hypertext links to your pictures (like mine on the previous page); if 'IMG Code' is set to ON, the pictures will be merged in the post.

[Atlantis]

Yes I know that for a http link or an image link the image should be on the net but I was not talking about them. I was talking about the attachment option. There is an attachment option that seems to allow you to attach certain files under a certain size. I did that but it doesn’t work. Actually it says the file is already attached, I don’t know where you can see it!

[tyee]

Hey Atlantis
I just completed my first LD transfer to DVD. StarWars3. Check out this AVIsynth script -

LoadPlugin("H:\Rip\MPEG2Dec\MPEG2DEC.dll")
AVISource("Side12.avi") ++ AVISource("Side34.avi") ++ AVISource("Side56.avi")

video=Trim(0,50347) ++ Trim(51064,93988) ++ Trim(97392,143491) ++ Trim(144255,193167) ++ Trim(193386,0)
#return video

seg1=video.DoubleWeave().Pulldown(0,2).Trim(0,68559) #vary pulldown,pulldown change'frame'=>'to'
#return seg1

seg2=video.Trim(85700,0).DoubleWeave().Pulldown(1,4).Trim(0,6057) #search for original 'frame'+1 =>'from'
#return seg2

seg3=video.Trim(93273,0).DoubleWeave().Pulldown(0,2).Trim(0,36879)
#return seg3

seg4=video.Trim(139373,0).DoubleWeave().Pulldown(1,4).Trim(0,30329)
#return seg4

seg5=video.Trim(177286,0).DoubleWeave().Pulldown(1,4).Trim(0,8798)
#return seg5

seg6=video.Trim(188286,0).DoubleWeave().Pulldown(0,2).Trim(0,0)
#return seg6

total=seg1 ++ seg2 ++ seg3 ++ seg4 ++ seg5 ++ seg6

#2.35 movie
return total.BicubicResize(704,368,0,0.6,8,103,704,274).AddBorders(8,56,8,56).ResampleAudio(44100)

#1.85 movie
#return total.BicubicResize(704,458,0,0.6,8,68,704,344).AddBorders(8,11,8,11).ResampleAudio(44100)


Perfect audio sync without using Huffy temp files!!

I use CCE. This script allows me to do the following steps-

1. Load the *.avi''s I captured via DV board, then delete the frames of side changes, black areas,etc.
2. Start to check from the beginning of the movie for the correct pulldown (put a # character just after the 'pulldown(x,x)' text in all the 'segx= lines)', then find where the pulldown changes and enter this in the 'trim(0,to)'
3. Then remove the # character in that line to view the changes if desired, then do the same for the rest of the movie.

The secret is when you find out where the pulldown changes and enter it in the 'trim(x,x)' area, this exact same frame must be found in the original unpulldowned video. An example will illustrate how to do this -

Let's use these lines of the script -

seg2=video.Trim(85700,0).DoubleWeave().Pulldown(1,4)#.Trim(0,6057)
return seg2
#seg3=video.Trim(93273,0).DoubleWeave().Pulldown(0,2).Trim(0,36879)

Ok, the 85700 is the unpulldowned video frame where the last pulldown ended. This was discovered in the 'seg1='line above. So initially you put a # character just after the 'pulldown(1,4) like I've shown, so the 'trim' part is not processed yet. Then load this script into Vdub and view it to see if the pulldown you chose was right.(Obviously you hash out all the lines below the 'return seg2' line so they are not processed yet).

Change the pulldown until it is right, then find out the frame where it next changes. Note this last good frame without interlacing and enter it (6057). Now remove the # character after the 'pulldown(1,4)' to enable trim to take effect.

Now the trick is to find the actual frame in the original interlaced movie equivalent to this pulldowned 6057 frame. It's easy, just use this equation, 85700 + 1.25(6057) = 93273. I open another instance of Vdub to view the original interlaced movie, just to confirm that this original frame number (93273)is correct. It may be a frame or two out but you can get the exact frame with Vdub (Use Cntl-G, to instantly go to that frame, left/right arrows to scroll).

Now enter this value in the 'seg3=' line as you see above, and keep repeating this for the whole movie.

Now once all the frame numbers are filled in, you can view the entire movie in Vdub and 'File-Save wave' to get the synch audio. Then run pulldown for the video if required, then convert .wav audio to mp2 or ac3. Mux in desired application. Done!

This seems like a lot of work but it isn't now that you have the script to work with.

I tried it with MpegI-Layer 2 audio last night and got perfect synch. I'll try it with ac3 audio when I can.

Tyee

[Atlantis]

That’s great for people who don’t have enough space but for me space is not an issue. I want speed. Because I want to encode in 2-pass, I want my video to be all ready before the encoding phase. If you use that script for a 2-pass, it will take longer time.

[Atlantis]

Do you have DV500plus also?

[tyee]

Atlantis
Unfortunatly, I only have Studio DV capture board with a Sony DVMC to convert S-video to DV format. I only did a one pass VBR on this movie and the average bitrate was about 2500bps. I noticed some funny vertical squiggley lines in faces now and then ( I have a 100" dlp projector so I see all flaws) but I'm not sure what is causing that. I did do it anamorphic but I'm not sure if that has anything to do with it. I used Angel's matrix also which is meant for a lower bitrate.

Last night I let it run again to do a Multipass VBR, 1 pass selected, (actually 2 passes including vaf file generation). Tonight I'll view it and see if it's any better. Trying Andreas matrix now.

CCE 2.5 supposedly doesn't properly mux encoding matrix into MPEG2 header for decoding application to use, so this could be part of the squiggley line problem also??

Do you use Multipass 2 pass selected or Robshot's method??

Tyee

[Atlantis]

No, I'm using the new TMPGEnc plus. Never liked CCE!

[Atlantis]

Tyee, well it’s still NTSC DV. I’m wondering do you also get the stepping problem I posted in the image above, around the red lightsabers and blasters?

[tyee]

Atlantis
I think I did notice that problem on my first version. I didn't get a chance to view my second version last night, so I'll look again tonight.

Tyee

[StephLG]

Hi guys,

for your information, here is a capture from my MPG2 encoded sequence; crop and bicubicresize have been made with Avisynth v1.07; no temporal filter used.

http://mapage.noos.fr/stephane.leguevel/esb2.jpg

Looks like I have the same stepping problem as Atlantis.

[Atlantis]

Yours is less because of the color I think. Why the lightsaber in your picture looks more purple than red? Have you touched the tint?

[easy2Bcheesy]

Laserdisk has a lower vertical resolution than DVD. Might this explain the stepping? Perhaps the capture process is introducing duplicate lines to give you 480 vertical lines? Just a thought...

[StephLG]

Originally posted by Atlantis
Yours is less because of the color I think. Why the lightsaber in your picture looks more purple than red? Have you touched the tint?

I haven't touched anything but 'Brightness' when capturing, because I though the picture on my PC monitor was too dark.
When I'm ready to capture the whole movie, I'll send the DV500's blue box video out signal to my TV to control and tune all this (Tint, Contrast, Brightness, etc...).
Maybe the color shift comes from PowerDVD or VirtualDub (I don't remember which one I used to capture this frame).

[teko77]

I am trying to transfer some of my NTSC-laserdiscs into DVD-R, but I am hard time trying out the option of TMPGEnc.

I am capturing directly into MPEG2.

I presumably now have an interlaced MPEG2-source, could someone please give information regarding the options I should use in order to transferring the NTSC-capture into DVDR.

I have a few open questions, e.g. is it Inverse 3:2 pulldown I should select, or under "advanced tab" is it the Inverse Telecine I should use?


Thank you for any help, I sure wish there would be a TMPGEnc-guide
for the operation somewhere...



Teko

[Chauncey]

Thank you guys for posting all this stuff. I have captured a laserdisc and performed the pulldown in Avisynth as indicated in this thread.

I have a question about CCE. In the video options there is a place to put a check if the frames are progressive. Should I have this box checked when encoding a 23.976 avi? What exactly does checking this box do?

I've already encoded without the box checked, and the video on the DVD looks great. I am considering re-encoding if there will be a noticeable difference.

Again thanks for you help.