It should be "add" if you do not have an existing entry (check with gspot , under seq ext)
color in the sequence_display_extension should describe entries for primaries, transfer, matrix - it does not affect the display_horizontal_size . I would leave it uncheckmarked
The sequence_display_extension is part of the official mpeg2 spec , not necessarily the DVD spec. The example given for display_horizontal_size for "PAL" Rec601 is 702 width. But in practice, it does not work most of the timeBecause I am making an ITU-compliant DVD, while it will play correctly on all DVD players through analog, I am trying to find a way to make it at least possible to play on digtal quipment that is capable of extracting and using the cropping information.
for 625-line "PAL" the example given is display_horizontal_size 702
Sample aspect ratio information is provided by means of aspect_ratio_information and (optional)
display_horizontal_size and display_vertical_size in the sequence_display_extension(). Examples of
appropriate values for signals sampled in accordance with Recommendation ITU-R BT. 601 are given in
+ Reply to Thread
Results 31 to 60 of 96
So it should look more or less like this. Or should the "Correct Seq..." not be checked and only the box with SDE should be used?
length correction should not be checkmarked unless you used old version of CCE to encode . You must have replied before I added that edit above
Because the PAR (Pixeal Aspect Ratio) of the active picture is 4600:3159 according to ITU-R BT.601, the player should display the active picture of 702x576 as 16:9 for an undistorted view. My old DVD player with SCART output did this correctly, while my blu-ray player with HDMI stretches the 720x576 (including the borders) to 16:9 which means a distortion of the movie of approx. 2.5%.
Last edited by Sharc; 24th Aug 2020 at 17:19.
Like I said, the lazyness of the manufacturers, or they had simply new staff that had no idea what REC.601 was when they were designing new digital equipment...
If anamorphic ws is not acceptable then create non-anamorphic. My understanding of that is that for PAL one takes the 576 vertical pixels cropped 72 top and bottom to leave 432 pixels.
So 432 / 315 * 512 = 702
and 432 / 9 * 16 = 702
I'll leave it to you to calculate how many horizontal pixels to start with to end up with a display of 702
Had I wanted to export widescreen as letterbox Vegas can easily do that and the active image wold indeed be contained in 702x432 pixels at 128:117 Sampel Aspect Ratio.
It was inferred in a reply which you later modified your opinion in a subsequent post following a reply from Sharc that corrected you.
I just fail to see why you need to go down this ITU road. There must have been good reason why studios changed. Possibly since analogue playback devices were no longer being manufactured.
I need it, because ITU is standard and otherwise it would be non-compliant and not backwards compatible. Just because peope ignored standards does not we have to comply with the ignorants.
The reason was perhaps that the manufacturers of the TVs and PC software gave a middle finger to ITU and were too lazy to implement additional rules for the playback of SD material
I am not about to get in to an extended 'discussion' with you.
But it seems my, and most on here's, opinion of what constitutes a compliant dvd is different to yours.
I do not require discussion, only solution to the problem presented, though any insight is appreciated.
A compliant DVD is compliant to ITU REC.470 and 601, which means 128:117 pixel aspect ratio for 4:3 video and 512:351 for 16:9 (or apparently 4600:3159 if top and bottom halflines are taken into precise calculations).
My question regarded how to force 128:117 / 512:351 SAR in AVS2DVD and possible cropping of digital blanking in MPEG2, not whether it's worth doing considering the incorrect design of digital playback hardware and software.
I need strict compliane with ITU standards, even if it means incorrect playback in incorrectly designed devices, even if they are common majority.
How does one create a DVD with this compliance? Is it merely cropping the file (the black bars) to 702x576 and
resizing to 720x576?
Or does the DVD itself also have to have some small pillar boxing?
I think most don't really care because the difference in aspect on the screen is so small,
most people can't even see it.
And in the CRT days, with the TV's indeterminate overscan (probably 5 - 10% on average), renders the whole
absolute precision thing almost moot
Actually it's either shooting video that is already compliant or converting to the compliant specification where the active image is 702x576 padded with blanking to 720x576 frame, however DV video uses the full 720 line of samples, while still having the ITU compliant pixel aspect ratio of 128:117, or 512:351 by using the full length of 53,333 Ás line instead of typical 52 Ás line, which might havre resulted in the confusion that the active image is in the 720 rather than 702 pixels.
The properly encoded DVD should have the digital blanking and in most cases DVDs manufactured before 2006 have them.
The standard actually takes overscan into accout, which is why the 18 pixels are the equivalent of nominal analog blanking.
Consider the following quick sample mpeg2 (dvd-compliant) which is sourced from a DV original - cropped 18 pixels but then 18 pixels added back.
While this, if I read you correctly, confirms to your 'idea' of compliance, those 702 pixels will not, nor can not, end up displayed as 768 pixels.
If the active image is contained in 702 pixels intended for 4:3 aspect ratio, then it's correct and compliant.
As you can see it is expanded properly to square pixels, they can and should be displayed as such, however the blanking is not cropped, as it does not have it encoded in Sequence Extension.
Just because you have software and hardware written by morons that have no idea about ITU does not mean I have identical software.
Oh and it's clearly analog video converted to digital, so there was no need to crop pixels and pad them with black, it was already acptured like that, because like I said, REC.470 spefifies active image as 52 Ás and according to REC.601 analog signal is sampled at 13,5MHz and the 53,333 Ás of the line is sampled so it ends up with blanking sampled in 18 pixels either way.
Well I did it that way, and you are correct - it is obvious - that the source was VHS but transferred to PC via an ADVC, since the blanking did not centre the image (it rarely does)
Now it did puzzle me where you got your width of 787 pixels from but you(or your player) have simply manipulated the video by stretching it to achieve 768 pixels from the active 702. But if your player can do that then surely your problem is solved except that no stand-alone dvd player has such an over-ride.
And neither will sequence extension help sine 702*576 is NOT a compliant size for dvd.
And as a moron I apologise for the sample which was created in...........
I must see if I can find some really early dvds from my collection - I have been buying since 2000 - to see what was going on back then.
The 702 just describes is the active image area. It's 702 in 720 . The display_horizontal_size describes this area
You can't "force" anything in MPEG2 because of the IEC 131818 MPEG2 specs. Only certain values are "allowed" or legal. It's up to your hardware or software to interpret it correctly .
There is no aspect ratio information in MPEG2 bitstream, expect 4:3, 16:9 , 1:1, 2.21:1 . You can't enter different SAR values or crop values like you can with AVC.
The only other possible way according to the MPEG2 specs is using the sequence_display_extension entries (display_horizontal_size, display_vertical_size) and to offset the area, the picture_display_extension ( frame_centre_horizontal_offset, frame_centre_vertical_offset) , which of course is rarely implemented correctly in anything
Maybe Sharc can rip one if the old DVD's that play correctly on SCART , and examine the stream to see how it's done there. Did it actually use those additional entries? or does it just "assume" a value? What about modern DVD's , how does it play them ?
PAL DVD 720x576, 16:9, anamorph
Active picture: 704x576
Borders of 8 pixels left and right
GSpot reports for seq ext: 540x576
540/576*64/45 = 1.33333..... = 4:3
540/576*4/3=1.25 = 720/576 (= Storage Aspect Ratio)
I don't know whether GSpot calculates something or really reads the seq ext from the stream.
My old SCART DVD playback infrastructure which played the movie undistorted (circle test) is gone, so I can't retest. The DVD player was a 'no name'. It was just frustrating to find the newer blu-ray infrastructure playing the movie with 2.5% distortion (16:9 including the borders, means movie 2.5% squashed). Oh well, Modern Times, I thought.
To the OP:
A workaround would be to resize the source (assuming 1:1 SAR) to 704x576, or resize to 702x576 for ITU SAR and add a 2 pixles border left or right for DVD format compliance. Author the DVD as 704x576, 16:9. All players should play this as 16:9 and the deviation to a 'true ITU' compilation would be minimal.
Maybe this is even the reason why 704 got included in the DVD standard at all; just guessing.
Last edited by Sharc; 26th Aug 2020 at 02:35. Reason: 'Notes' added
As mentioned earlier, playback of ITU-compliant video on DVD players via analog output results in correct picture, beucause analog signal is created correctly and correctly received by the TV.
Theoretically physically cropping 720 to 704 with 1 pixel of blanking at each side left would result in the smallest distortion, if interpreted as 4:3, or 16:9 from full 704 line.
As for the screenshot, it's simply implementation of proper aspect ratio that takes the ITU specification into account.
Vegas, as I wrote somewhere at the beginning. So I only need to crop the resulting 720x576 frame to 704x576 with 702 active pixels and the remaining 1 on each side is already padded black.
1. Remove (just to clear what might already be there)
2. Add 540 (540 is what some of my old DVD originals suggest, some report 720, most leave it empty. None however reported 704 or 702)
After this GSpot should indicate 540x576 for seq ext.
I didn't try. No idea whehter this will solve the problem. DVD specs and strict ITU compliance have been in conflict since the beginning.