Deinterlace method for source project

megabit wrote on 12/31/2013, 6:42 AM

I hardly ever deal with interlaced video, hence lack of any proven workflow established. I'm doing an 8-camera MC edit, and only cameras used were HD (EX1) - the rest is SD. To lessen the burden during MC edit, I'd like to pre-render those 2 cameras' material into SD, so that the common SD properties can be used for the MC project.

I'm rendering the UFF 1920x1080 into the Avid DV, 50 Mbps, PAL "intermediary", which should result in LFF 720x576.

Question:

Which, if any, deinterlace method should I set my source HD project to - blend, or deinterlace?

Thanks, and a Happy New Year to everyone

Piotr

Comments

JohnnyRoy wrote on 12/31/2013, 7:01 AM

> "Which, if any, deinterlace method should I set my source HD project to - blend, or deinterlace?"

You always want to set it to Blend Fields. This will take the existing fields and blend the edges together to create a single frame made up of data from both fields.

You would only use Interpolate Fields if Blend doesn't work. For example, if the motion is so fast that the difference between the two fields still shows interlace artifacts. Interpolate fields fixes this by throwing away one field. That means you throw away 1/2 of your vertical resolution! 1080p effectively becomes 540p and then Vegas will double each field to get back to 1080p. This is less desirable and should only be used as a last resort.

You never want to set your Deinterlace Method to None if you have any interlaced video in your project. The None setting is really to disable the deinterlacer because you want to use your own 3rd party deinterlacer like BCC Deinterlace. Then it's up to you to deinterlace any video that requires it. Otherwise there is no reason to ever set the deinterlace method to None even in a project that is all progressive because it will only be used when needed.

~jr

megabit wrote on 12/31/2013, 7:15 AM

Thanks Johnny.

One thing though: I realize that by Interpolating I'm effectively getting only half the resolution of HD, but does it make any difference in this particular scenario of downrezzing (I'm going from 1080p to 576p (PAL), anyway)...

Piotr

JohnnyRoy wrote on 12/31/2013, 7:42 AM

You probably won't be able to tell the difference when downrezzing. If it looks good to your eyes, that's all that really counts.

~jr

john_dennis wrote on 12/31/2013, 9:12 AM

@jr

"[I]1080p effectively becomes 540p and then Vegas will double each field to get back to 1080p.[/I]"

1080i?

JohnnyRoy wrote on 12/31/2013, 12:19 PM

Yea, should have been 1080i becomes 540p... good catch John.

~jr

PeterDuke wrote on 12/31/2013, 6:28 PM

If you are converting HD interlaced to SD interlaced, you should set the project properties deinterlace method to either "interpolate" or "blend fields". It doesn't matter which because the results will be identical. (I have confirmed it using the FC /B command).

If you set the deinterlace method to "none", Vegas will convert a full frame at a time rather than a field at a time, something you don’t normally want unless the fields have no motion.

You should not convert SD UFF to LFF or vice versa, if you can avoid it, because it will degrade quality slightly. However, if you start with HD it probably won't make much difference, but I would keep the same field order in any case, just to be on the safe side.

EDIT

If you are mixing down-rezzed HD with LFF SD, then it would be better to make the former HD LFF as well.

VidMus wrote on 12/31/2013, 7:22 PM

@ JohnnyRoy

If two thirds of my videos are progressive and one third is interlaced I will render the interlaced videos to an intermediate uncompressed format making them progressive and then use them in an all progressive project. I have done this with a theater video where 3 cams were progressive and 2 were interlaced. This made for great final results.

The intermediate files were huge but for a one time project that did not matter.