VideoHelp Forum
+ Reply to Thread
Results 1 to 16 of 16
Thread
  1. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    I've done some quite successful PAL to NTSC conversion in the past but have ran into trouble on my latest one. I've taken these steps:

    - Extracted PAL DVD VOBs with DVDDecrypter
    - Demuxed the M2V and AC3 streams in TMPGEnc Plus
    - Opened the M2V in VirtualDubMod, deinterlaced, resized to 720x480, changed frame rate to 23.976 (for 3:2 pulldown encoding later) and saved as lossless AVI
    - Converted AC3 to WAV
    - Opened new AVI file in VirtualDub, noted exact length, stretched the WAV audio to the precise length in Goldwave

    I then play the AVI in VirtualDub with the audio stream from the new WAV file to check the sync and it starts off fine, only to slowly drift off towards the end by up to half a second.

    I have done these exact steps before on other PAL discs and it has worked out fine. Does anyone know what could be going wrong here?
    Quote Quote  
  2. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    -30-
    Last edited by sanlyn; 19th Mar 2014 at 06:34.
    Quote Quote  
  3. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    The source is a UK TV series that was shot and transmitted from PAL video. The DVD is interlaced. I more or less followed this guide here:

    http://www.digitalfaq.com/guides/video/convert-pal-ntsc.htm

    I have done it before with PAL DVDs and had perfect results.
    Quote Quote  
  4. Originally Posted by GavSalkeld View Post
    Does anyone know what could be going wrong here?
    No, not even knowing the series, much less having the DVD. But I wouldn't do the audio conversion as you do (actually, I wouldn't do any of it the way you do). Stretching audio like that is a recipe for out of synch audio unless the audio is exactly the same length as the video. Sometimes audio ends before the video. I think if you checked you might discover the audio being shorter than the video.

    There are 2 possible ways to solve the problem, one quite easy. Assuming there's no opening delay (did you check?), encode for 720x480 and 25fps and you won't have to do anything at all with the audio. If the series was created at 25fps you'll also be keeping the video speed the same (not slowing it by over 4%) and (maybe) keeping the audio pitch correct. After doing the encoding for DVD run the MPV through DGPulldown set for 25->29.97. Rather than 3:2 pulldown, it'll apply a 3:2:3:2:2 pulldown.

    The other way is not to stretch the audio as you usually do but by the ratio of 25 to 23.976. eac3to can do that PAL2NTSC conversion (although I use BeSweet for the job).

    And your method is real screwy anyway. Did you ever say how the lossless (?) AVI you created in VDubMod is reencoded to MPV? And, as sanlyn suggested, maybe deinterlacing isn't the best thing to do. Would you care to make available 10 seconds of the PAL M2V, a piece with steady movement?
    Last edited by manono; 19th Feb 2014 at 03:30.
    Quote Quote  
  5. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    Originally Posted by manono View Post
    Stretching audio like that is a recipe for out of synch audio unless the audio is exactly the same length as the video.
    Well there was no delay on the VOBs and as per that link you can stretch the audio to precisely the right length in Goldwave.

    After doing the encoding for DVD run the MPV through DGPulldown set for 25->29.97. Rather than 3:2 pulldown, it'll apply a 3:2:3:2:2 pulldown.
    Ah yes, forgot about that tool. Why is that method more preferable?

    Did you ever say how the lossless (?) AVI you created in VDubMod is reencoded to MPV?
    That's easy, I just use TMPGEnc, on the 23.976fps (internally 29.97) setting with 3:2 pulldown, non-interlace for a 480p M2V file. Why would deinterlacing it be bad?

    I am still working with it so will let you know how it goes.
    Quote Quote  
  6. Originally Posted by GavSalkeld View Post
    Well there was no delay on the VOBs...
    Okay, good.
    ...as per that link you can stretch the audio to precisely the right length in Goldwave.
    The right length being what? Did you read what I wrote? Your method won't work if the original video and audio lengths are different. I'm not saying they are different, only that it's a likely explanation for your problem.
    Why is that method more preferable?
    Again, did you read what I wrote? If the series was shot at 25fps then by doing it that way you keep it the original length, the speed the same, the audio pitch the same and, best of all, you can use the original audio without stretching it and with no synch problems.
    Why would deinterlacing it be bad?
    For one thing, you lose half the fluidity of motion. For another, unless and until a sample is made available we don't know for sure what you have and it's possible deinterlacing is exactly the wrong thing to do. Maybe everything you said about it is true. But you are here asking for help. You could even keep it interlaced at the same time converting it to NTSC and keeping it the original length and using the original unstretched audio. I'm not necessarily recommending doing it that way, though. When converting interlaced PAL to NTSC, there is no perfect solution.
    Quote Quote  
  7. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    The right length being what? Did you read what I wrote? Your method won't work if the original video and audio lengths are different. I'm not saying they are different, only that it's a likely explanation for your problem.
    Well if the A/V lengths were different, the audio wouldn't start off in sync and then drift as the show went on.

    There are 2 possible ways to solve the problem, one quite easy... Encode for 720x480 and 25fps and you won't have to do anything at all with the audio. After doing the encoding for DVD run the MPV through DGPulldown set for 25->29.97
    OK, so take the original 720x576i M2V stream and just resize it to 720x480i, saving it as a new AVI to re-encode to M2V?
    Quote Quote  
  8. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    -30-
    Last edited by sanlyn; 19th Mar 2014 at 06:34.
    Quote Quote  
  9. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    I will trying the DGPulldown method - at least it means I don't have to mess around with the audio. Thanks.
    Quote Quote  
  10. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    -30-
    Last edited by sanlyn; 19th Mar 2014 at 06:35.
    Quote Quote  
  11. Originally Posted by GavSalkeld View Post
    Well if the A/V lengths were different, the audio wouldn't start off in sync and then drift as the show went on.
    Nonsense, that's exactly what it would do. You said it was half a second off by the end, right? Say the original video is exactly 30 minutes long and the audio 29 minutes 59.5 seconds (and they begin in synch). First you stretch the video and then stretch the audio to match the new video length. What do you think will happen?

    You could easily prove my theory (or disprove it) by finding out the exact lengths of the original audio and video.

    OK, so take the original 720x576i M2V stream and just resize it to 720x480i...
    No. If you do a straight resize, and if it's really interlaced, that will completely ruin it. I thought you said you had been deinterlacing it. But no, that's not the way to convert it to NTSC and keep it interlaced. It's more like this:

    Yadif(Mode=1)
    LanczosResize(720,480)
    ChangeFPS(59.94)
    AssumeTFF()
    SeparateFields()
    SelectEvery(4,0,3)
    Weave()
    Quote Quote  
  12. Member
    Join Date: Aug 2002
    Location: United States
    Search PM
    Originally Posted by manono View Post
    No. If you do a straight resize, and if it's really interlaced, that will completely ruin it.
    OK, that's what I figured.
    Quote Quote  
  13. Member FulciLives's Avatar
    Join Date: May 2003
    Location: Pittsburgh, PA in the USA
    Search Comp PM
    I just wanted to mention that I've gotten lazy in my old age and these days I tend to use AVStoDVD which will almost always do a perfectly good PAL to NTSC conversion.

    99% of the time it does the DGPulldown method but I think that is because I mostly have progressive PAL video in the source materials that I deal with but once I had a PAL DVD that was interlaced and it did something that looks a lot like the script manono posted (although slightly different):

    Code:
    Import("C:\Program Files (x86)\AVStoDVD\Lib\A2DSource.avsi")
    LoadPlugin("C:\Program Files (x86)\AVStoDVD\DGIndex\DGDecode.dll")
    LoadPlugin("C:\Program Files (x86)\AVStoDVD\Lib\LeakKernelDeint.dll")
    
    Video = MPEG2Source("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_05_0.d2v")
    # Audio is frameserved by AviSynth just for Preview and Edit purposes.
    Audio = A2DAudioSource("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_05_0 T80 3_2ch 384Kbps DELAY 0ms.ac3", CacheFolder="C:\Users\BRUCEW~1\AppData\Local\Temp")
    
    Video = Video.ConvertToYV12(interlaced=true)
    Video = Video.LeakKernelBob(1,7,false,false)
    Video = Video.Spline16Resize(720,480)
    Video = Video.ChangeFPS(59.94)
    Video = Video.SeparateFields().SelectEvery(4,1,2).Weave()
    
    AudioDub(Video, Audio)
    With a PAL video source that is progressive I get this:

    Code:
    Import("C:\Program Files (x86)\AVStoDVD\Lib\A2DSource.avsi")
    LoadPlugin("C:\Program Files (x86)\AVStoDVD\DGIndex\DGDecode.dll")
    
    Video = MPEG2Source("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_01_0.d2v")
    # Audio is frameserved by AviSynth just for Preview and Edit purposes.
    Audio = A2DAudioSource("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_01_0 T80 2_0ch 192Kbps DELAY 0ms.ac3", CacheFolder="C:\Users\BRUCEW~1\AppData\Local\Temp")
    
    Video = Video.ConvertToYV12()
    Video = Video.Spline16Resize(720,480)
    #Using DGPulldown/HCenc to upsize FPS
    
    AudioDub(Video, Audio)
    AVStoDVD is just a front end to several programs. It uses HCenc for MPEG-2 video conversion.

    - John "FulciLives" Coleman
    "The eyes are the first thing that you have to destroy ... because they have seen too many bad things" - Lucio Fulci
    EXPLORE THE FILMS OF LUCIO FULCI - THE MAESTRO OF GORE
    Quote Quote  
  14. Member FulciLives's Avatar
    Join Date: May 2003
    Location: Pittsburgh, PA in the USA
    Search Comp PM
    Out of curiosity I took the same source as above (the interlaced PAL DVD source) and told AVStoDVD to deinterlace it and this is the script it made:

    Code:
    Import("C:\Program Files (x86)\AVStoDVD\Lib\A2DSource.avsi")
    LoadPlugin("C:\Program Files (x86)\AVStoDVD\DGIndex\DGDecode.dll")
    LoadPlugin("C:\Program Files (x86)\AVStoDVD\Lib\LeakKernelDeint.dll")
    
    Video = MPEG2Source("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_05_0.d2v")
    # Audio is frameserved by AviSynth just for Preview and Edit purposes.
    Audio = A2DAudioSource("C:\Users\BRUCEW~1\AppData\Local\Temp\VTS_05_0 T80 3_2ch 384Kbps DELAY 0ms.ac3", CacheFolder="C:\Users\BRUCEW~1\AppData\Local\Temp")
    
    Video = Video.ConvertToYV12(interlaced=true)
    Video = Video.LeakKernelDeint(1,10,false,false)
    Video = Video.Spline16Resize(720,480)
    #Using DGPulldown/HCenc to upsize FPS
    
    AudioDub(Video, Audio)
    I'm curious manono if you think that makes sense and which would be better?

    Also notice the line: Video = Video.LeakKernelDeint(1,10,false,false)

    AVStoDVD had a default of 10 but said it could be set anywhere from 0 - 255

    Any ideas on what number to pick there?

    - John "FulciLives" Coleman
    "The eyes are the first thing that you have to destroy ... because they have seen too many bad things" - Lucio Fulci
    EXPLORE THE FILMS OF LUCIO FULCI - THE MAESTRO OF GORE
    Quote Quote  
  15. Originally Posted by FulciLives View Post

    I'm curious manono if you think that makes sense...
    Yes, if you deinterlace it that makes sense. Except the deinterlacer/bobber being used is old now.

    Which is better? Deinterlacing first before conversion as opposed to keeping it interlaced and adding dupe fields?

    Originally Posted by manono View Post
    When converting interlaced PAL to NTSC, there is no perfect solution.
    Personal preference, I guess. I'm not opposed to deinterlacing to 25fps and running DGPulldown afterwards. Others might not like it.
    Quote Quote  
  16. Member FulciLives's Avatar
    Join Date: May 2003
    Location: Pittsburgh, PA in the USA
    Search Comp PM
    Yeah I originally did it the first way but it made an interlaced NTSC DVD but it's for a friend who still has an old TUBE TV so I figured he wouldn't care but now I'm thinking I'll do it again by deinterlacing and using DGPulldown. I'm thinking that will look better on a flat screen HDTV (and better option for my own NTSC copy of said movie).

    - John "FulciLives" Coleman
    "The eyes are the first thing that you have to destroy ... because they have seen too many bad things" - Lucio Fulci
    EXPLORE THE FILMS OF LUCIO FULCI - THE MAESTRO OF GORE
    Quote Quote  



Similar Threads