Interlacing and vegas question

Mike M. wrote on 8/7/2010, 5:06 PM

I'm a bit confused as to what happens during import, edit and rendering interlaced files.

For example, let's say I have an interlaced file, import it into Vegas, edit it and then render it to be used as a mpeg compliant DVD. Does Vegas do all the interlacing conversion behind the scenes...........or............do I have to blend the fields and then re-interlace?

Also, what would be the case if you had a combination of interlaced and progressive files and then rendered to interlaced?

Thanks,

Mike

Back to post

Comments

jabloomf1230 wrote on 8/11/2010, 3:18 PM

1080 60i AVCHD used to be a "novelty act", but on newer hardware, it's more or less editable. Now, it's replacement in the cutting edge gimmick zone is 1080 60p AVCHD. But the format does have some legitimate uses. First, you can crop it to 720 60p, which is helpful for both creating close ups and improving the appearance of motion stabilization. The usual problem with motion stabilization is that you end up having to either mirror the edges or magnify the center of the shot. It's also good for slow motion, as you can slow everything down to 30p or 24p.

It won't be too long before every camcorder at the mid to high range will be able to do 1080 60p. But then the high end will probably be 1080 60p 3D.

farss wrote on 8/11/2010, 4:10 PM

One thing to be aware of will all this progressive resolution is it can create serious problems displaying it on interlaced systems. I can shoot 1080p with around 1,000 lines resolution and 720p50 with 700 line v res and it can be quite a nightmare to go from that to standard definition as there's way too much vertical resolution.
I shot one short piece in a space with dark brown brick walls that had fine traces of white in the bricks. It needed quite a bit of work to go from the 720p50 to SD 50i without it being a shimmering mess.

Bob.

PeterDuke wrote on 8/11/2010, 5:50 PM

When you are reducing the number of samples, whether it be audio or image (where we call the samples "pixels") you should heed the Nyquist-Shannon theorum (also known by other names). Its consequence is that you should low pass filter the signal (blur an image) before downsampling or you will get artefacts generated. After downsampling an image you may then apply a modest amount of sharpening.

I'm sure that Vegas, Photoshop and other editors would already take this into account, but their algorithms may not be optimum for some images.

Edit

Note that interpolation is a form of lowpass filtering.

farss wrote on 8/11/2010, 10:14 PM

"I'm sure that Vegas, Photoshop and other editors would already take this into account, but their algorithms may not be optimum for some images."

Photoshop doesn't have to deal with this problem. Vegas and every other NLE makes some attempt at wrangling this problem however it is extremely difficult to effectively wrangle the problem in software and is the cause of many, many posts and angst in every forum for just about every NLE. Even the quite expensive scaling hardware used by broadcasters is not 100% effective judging by the amount of aliasing and line twitter I see on television from time to time.

Bob.

PeterDuke wrote on 8/11/2010, 10:53 PM

Hang on! Video is only a series of still images, called frames. Scaling a frame (if we discount interlacing for the moment) has the same issues as scaling a still image in Photoshop. Broadcasters are constrained to work in real time but we video editors aren't. We are always trying to balance between too much smoothing/blurring (possibly with heavy-handed sharpening) on one hand or two many aliasing artefacts on the other.

farss wrote on 8/11/2010, 10:59 PM

No, interlaced video is not the same as a series of still images which is what progressive scan video is. I'd also point out that by definition still images do not portray motion.

There is around an order of magnitude greater complexity to systems that handle moving images compared to still images.

Bob.

Edit:

If you want to try to get a handle on the issues this article here: http://www.mentallandscape.com/Papers_siggraph90.pdf is probably a good starting point.

PeterDuke wrote on 8/12/2010, 1:36 AM

Thanks for the reference. However I wasn't intending to go into the difficulties with interlaced video, as I said.

My point was that when reducing the resolution of an image, whether still or part of a video you should first low pass filter it or you will get such things as jagged sloping lines. Video exacerbates the problem of down sampling for several reasons.

farss wrote on 8/12/2010, 1:57 AM

Agreed the problem is easier with progressive scan display however my post to which you replied initially did say "One thing to be aware of with all this progressive resolution is it can create serious problems displaying it on interlaced systems."

Even if the source is progressive when displayed on an interlaced system you can run into problems.

Bob.