I'm trying to use Canopus Procoder to convert this PAL DVD:
2.35:1 Original Aspect Ratio
Average Bitrate: 6.53 mb/s
PAL 720x576 25.00 f/s
...to an NTSC DVD. It was NOT my intent to change the aspect ratio. For the source and destination, I selected 16x9. When I watch my DVD, the top and bottom black borders are significantly larger than they were when viewing the original DVD and the actual movie display seems smaller because of this. What am I doing wrong? The quality of the conversion is fine, its just the aspect ratio doesn't look the same as the source.
+ Reply to Thread
Results 1 to 16 of 16
The size before encoding to NTSC (in case 2.35:1 inside 16:9) should be 720x360 with (60+60) borders, did you provide this? In case your original video was anamorphic 16:9 you could also resize it without cropping borders to full 720x480 with disabled 'keep aspect ratio' (if there's such a setting in Canopus).
How did you resize and did you crop original borders?
Thanks for the response. I did NOT resize or crop anything. I simply selected the source video. Procoder detected that it was 16x9 PAL and I configured the destination to be 16x9 NTSC and did the conversion. Can you elaborate on what I need to do to maintain the proper size?
Is there some option like 'resize to full screen'? If you didn't crop then everything (incl. borders) can be resized from 720x576 to 720x480 to keep original proportions. In some encoders there's an option 'keep AR' that should be disabled at resizing (try to find it), otherwise your detected 16:9 image with internal borders can get extra letterboxing inside full 720x480 matrix.
Can you take a quick look at my attached image file. I took some screen shot of some of the various options I see. The "Aspect Ratio Code" option for both the source and target can be changed to either 4x3 or 16x9. The source has a "Video Frame Size" of 720 x 576 and a display size of 540 x 576 if that helps.
I'm striving to get as close to what the original size is as possible without cutting off any of the picture.[/img]
Lots of stuff wrong there, but the main one that jumps out is that you reencoded for 4:3, and that accounts for the extra black.
You also added in a bunch of duplicate frames to go from 25->29.97fps, which will make the resulting DVD play jerky during certain kinds of movement. Much better would to have been to frameserve into Procoder using an AviSynth script and encode for 25fps. Then you can apply a 25->29.97fps pulldown afterwards using DGPulldown. And I don't know why you switched the field order, but maybe it doesn't make any difference with this progressive source. But first things first, I guess.
In that example (looks like from help file) the input PAL video is a 4:3 image letterboxed inside 16:9 frame and the settings used are for conversion it to full-screen 4:3 NTSC video. The numbers of display size (as I can guess) correspond to cropped size of source PAL image and come from crop settings. That part is resized to 720x480 and encoded.
In your case 2.35:1 (if you don't crop and make new borders that is also possible) you must have both frame size and display size 720x576, otherwise look for cropping somewhere enabled and disable it. Then the display size of full 720x576 will be resized to 720x480 at encoding. AR setting should be 16:9 for both source and destination.
What manono says about framerate conversion is absolutely true. Hi, manono!
Did I read it wrong, Alexander? In his very first post he says the source DVD is 16:9, and in the screen he showed his source setting as also 16:9. In his first post he says he set both source and destination for 16:9, which he clearly didn't. All the PAL DVD versions of Lost Highway I've tracked down (English, Italian, French, German, Australian) really are 16:9.
I supposed the screen capture was not for that very video (2.35:1) and 16:9 had been used for its destination (4:3 in the picture). Also numbers 540x576 in source settings point at 4:3 pillarboxed image (unlikely a default setting, looks like an example of conversion from manual), strange. Let's wait for more detail from tubularusa.
I'm sorry about the confusion, but when I made the screen shots I didn't change anything in the target like I did the first time. I was merely showing you what the options are. When I selected the original PAL video and created a target DVD, I did in fact change the target to 16x9, figuring it should match what the source had indicated when I added the video to the source tab.
Knowing this, what would I have to change to get the aspect ratio correct and the same as what the original movie looked like? Then I'll worry about the frame rates. I've never had a problem doing it this way, but maybe I'll try you way if it is not too difficult. Maybe you can walk me through that once we figure out he borders.
The problem with AR you describe could come from setting 4:3 instead of 16:9 (and then authoring as 16:9) but if you are sure it was 16:9 for encoding target...The slider is in correct position; the strange thing is display size for the source video (unless it simply indicates size of program's preview window). Is there any filter for manual resizing?
The help file (in the extract you provide) reads that scaling (resizing) in Canopus keeps AR. If it is kept in pixels then it will be OK in PAL=>PAL or in NTSC=>NTSC case but this would change displayed AR in case PAL=>NTSC without correction for different target. In other encoders like TMPGEnc Xpress you can disable 'keep AR' to get proper fullscreen resizing.
Originally Posted by tubularusa
If you think a different way would be better, is there a procedure on doing what you suggested?
1. For importing MPEG2 to AviSynth you'll need to create . d2v project file in DGIndex (part of DGMPGDec package, the second part is mpeg decoder for AviSynth, DGDecode.dll). To do it, drag all VOB files from re-encoded VTS to program window, confirm to import them (in project they will be joined in alphabetic order). Preview video to know from info window if it is progressive (possible if film was sped-up for PAL framerate, most favourable case for resizing) or interlaced, to know how to resize it in AviSynth. With default settings (video/field operation/honor pulldown flags and audio/output method/demux all tracks) use 'file/save project'. D2V file will be created and audio demuxed.
2.Install AviSynth 2.5.7 (if not yet) and create a script (a plain text file with .avs extention) with the following text (substitute your files data):
LoadPlugin("path/DGDecode.dll") MPEG2Source("path/your_file.d2v")# this loads your video LanczosResize(720,480)#if your video is progressive #(otherwise it should be previously deintelaced) Letterbox(60,60)#covers top/bottom with fresh black borders
In encoder use custom settings for target 25fps and 720x480, 16:9, progressive frame, elementary stream; if there is a setting for GOP (group of pictures) size, set it to 12. Do a short encoding and test in a software player capable of forcing 16:9 AR to be sure.
In the end use DGPulldown software with 25 to 29.97 setting, it works fast and only sets display framerate flags.
Correction: quotes added to paths in script
When I first opened up my original m2v file (decrypted dvd) in dgindex.exe, here are the properties when I selected preview:
Frame Size 720x576
Aspect Ratio 16:9
Frame Rate 25fps
Frame Type Interlaced
Frame Struct - Frame
Field Order - Top
When I saved the project as a d2v file, the video information box on the right said changed and said Frame Type progressive. Why did this change? Which really is it?
At this point I'm not sure what to do. Your instructions say you need to do something different for interlaced, but I'm not sure if I need to or specicially how I would do that.
Here is the script I have so far:
MPEG2Source(C:\Lost Highway\VIDEO_TSVTS_01_1 - 0xE0 - Video - MPEG-2 - 720x576
(PAL) - 16~9 - Letterboxed.d2v)# this loads your video
LanczosResize(720,480)#if your video is progressive
#(otherwise it should be previously deintelaced)
Letterbox(60,60)#covers top/bottom with fresh black borders
First of all, I apologize for not putting in quotes paths for loaded plugin and opened file (this would call error message). If video is interlaced, deinterlace it before resizing e.g. using Yadif, a new fast working plugin (http://avisynth.org.ru/yadif/yadif.html):
LoadPlugin("C:\Program files\dgmpgdec149\DGDecode.dll") LoadCplugin("path\Yadif.dll") MPEG2Source("path\your_file.d2v") Yadif(order=1)#if TFF LanczosResize(720,480) Letterbox(60,60)
I don't know why it shows progressive only after saving project, maybe the video contains mixed interlaced/progressive material or just intro is interlaced (then readings should change during preview; you can preview different parts if set ranges). You can also explore the script for progressive in VDub using right arrow key for frame by frame navigation. In case of progressive video there shouldn't be any interlaced frames (in resized material such frames give noticeable artifacts), otherwise use the second script.