proper viewing for converting 1280x720 to SD 720x576

J

jeremiah

Member
Hello,
I have posted this question on a few other sites and have yet to get a response, so I am hoping that someone here can help me.

I am trying to convert HD 1280x720 files into SD 720x576 with the 16:9 option chosen making them 1024x576.

This is for offline editing using the dvcpro pal codec.

These are CG animation shots that are submitted to me by the artists at the HD resolution.

It almost seems to work, however the SD image produces ever so slight black vertical pillars on either side and this causes the image to squish in a few pixels compared to the original HD image.

Is this how it is supposed to look? Or am I doing something wrong here in the conversion?

I wanted to do this to avoid using prores, as my system has no kona card and I have a limited amount of space.
 
Is the final output going to be DVD? Before dragging your CG shots into the sequence, change the sequence settings to the appropriate preset. Next, drag your shots into the sequence and when it warns you about the sequence settings not matching the content and asks if you want to change the sequence settings to match, select no. This may fix your problem, but if you still see some slight pillars, just double-click the shot in your sequence and in the Viewer, click the Motion tab. Adjust the Scale value (using decimals if need be) until the pillars go away. This will preserve the aspect ratio. You will lose a few pixels off the top and bottom (equal to or less than the number of black pillar pixels), but there's no way around that. You either have to squish or zoom.

If the image is squished or stretched when you drag into the sequence, you can also use the settings in the Motion tab to stretch or squeeze it to the right aspect ratio.

The real problem you're going to run into is that FCP's downscaling is terrible. I would actually advise that you edit in HD, then Export as Quicktime Movie (making sure to check the "make Quicktime movie self-contained" option) and downscale with Handbrake (100% quality) or something before compressing with Compressor. Alternatively, you can downscale and compress using Handbrake all in one go.

If you can't edit at the HD resolution, then follow the instructions above but do the downscale before the edit. When you finish the downscale, send the output back to FCP.

It's been a while since I've dealt with all this stuff so hopefully I've got my facts straight. And hopefully the new FCP fixed the awful downscaling.
 
Sounds like it could be a square pixels versus rectangular pixels issue...What did the graphic artist render the animation in?

Also (I was under the impression that), Standard Definition in PAL land is 720x576, whether 4:3 or 16:9. It is simply a matter of pixels being packed into the 4:3 'box' and being unsquished on playback.

Just thought I would share some ideas, since you have little feedback on the issue.


Jason
 
Thanks for the replies.
Sorry, I would have got back on here earlier, but I assumed that I would receive and email notification when something got posted.

To answer some questions, this is going to be for broadcast, however I am not doing the actual online edit. My only reason I was trying to avoid HD was because of the size of prores files, as this is a 26 episode series 22 mins long each and I receive updates of these shots everyday, so there is going to be A LOT of footage and I'm still not entirely sure how much space I will be allocated.

Oddly enough, the problem seems to have corrected itself. They only difference was that I created a droplet in compressor with my SD settings. The weird thing is, while they show up when you open it in QT, upon import into FCP they disappear. So there is something very quirky going on there and makes me a bit nervous.

But yes, PAL is always 720x576, I just wanted to give the pixel dimensions when unsquished, but looking at it now, that might seem confusing.
 
You must log in or register to reply here.
Back
Top