Why is everything interlaced?

[Alex DeLarge]

Every commerial video I have is interlaced, I can understand it when the framerate needs to be changed, but why interlace something that was filmed at 25 fps and encoded at 25 fps? The number of problems I've had with interlacing when there wasn't any need to do it in the first place and progressive frames would have been fine!

[guns1inger]

Pretty much everything shot at 25 fps was shot on video, and was interlaced at that point. It is only the past few years that anything would have been shot progressive at 25 fps. Anything shot on film would be shot at 24 fps, then either speeded up or telecined to 25 fps and transfered to video as interlaced 50 fields per second for TV broadcast.

Many commercial DVDs that are 25 fps interlaced are in fact progressive sourced, and each field pair is identical.

[Alex DeLarge]

Interesting, do you know how to convert 24 fps progressive into 25 fps interlaced? I tried this with TmpgEnc, but it only interlaces every other second and leaves a duplicate inbetween.

[Alex_ander]

Apply DGPulldown 24=>25 after encoding at 24, this will work similar to usual telecining (without duplicating full frames or blending).

[guns1inger]

Usually the framerate is sped up to 25 fps first, so there is no need to duplicate frames. This is why there is often a 4% running time difference between an NTSC and a PAL release of the same film. You could do something like this loading the video via Avisynth, and having avisynth speed it up for you before frameserving it to tmpgenc. You also have to alter the speed of the audio to keep sync, and if you really want to do it properly, pitch shift the audio back down to compensate for the speed increase. Something like Sound Forge will do this in one go. Audacity would require two passes.

[Alex DeLarge]

Thats how its done? They just speed it up? HAHA, thats weak!

[guns1inger]

Actually, it's a great way of doing it because it prevents the need for duplicate or blended frames. Makes for a much smoother outcome, if done properly. Compare a PAL video transfer to an NTSC video transfer and you will see how much better it was.

You can always get the NTSC version and watch it on a progressive player and TV at 23.976 fps, although the trade off is a lower resolution video with often noticeably reduced quality over the PAL version (more so for 4:3 material than 16:9)

[Alex DeLarge]

I guess, I'm surprised how things still look natural though being sped up by 4%, you speed audio up by 4% and it sounds like Mickey Mouse, hence the pitch shift.

While your here, I have some video that was filmed at 24 fps, then interlaced to 30 fps for American broadcast. Then blended and interlaced down to 25 fps. Is there any way of restoring them to 30 fps?

[guns1inger]

That is one for the likes of manono. It will require avisynth skills outside of my range, if indeed it can be done at all. Your ultimate goal should be to get back to the original 24p if possible.

[manono]

While your here, I have some video that was filmed at 24 fps, then interlaced to 30 fps for American broadcast. Then blended and interlaced down to 25 fps. Is there any way of restoring them to 30 fps?

I think you're pretty much screwed. How are you so sure that it was telecined to 30fps before being converted to 25fps. It's much more common to fieldblend 24->25fps, which is bad enough, but which can also be undone. Got a sample?

If it's really 30->25fps, your best bet may be the AviSynth function CDeblend, followed by a decimator, but I don't think there's much chance of removing all, or anywhere near all, of the blending.

[Alex DeLarge]

Well, you can see two out of every five-ish frames blurred together like there would be with 24 to 30 then blended.

Actually, I think your right, I can't ecen imagine how it could work, but I thought I'd ask.