1080/30p in 60i container - why?
 

I just wonder why Sony decided to implement a 1080/30p in 60i container? As far as I know it is much easier to remove interlacing in NLE than to interlace the footage that was shot progressive.

Also a question - is there a separate 1080/25p /50i version for EU or is the camera switchable?

The Euro version is 25p (50i) but its not 50i at all - each "field" is the same - the NLE and Blu-ray just thinks its 50i which it's not - your material will always feel like 25P - so keep those pans under control!

Thank you Paul for the reply. Now it starts to be even more confusing... So when I edit the footage (in Sony Vegas for example) and my NLE gets cheated so sees the 30p or 25p as an 60i or 50i - shall I force the NLE to see the truth..? I mean: shall I force it to treat this footage as a progressive? In other words: what timeline should I declare and use - the one for progressive or interlaced?

I know that AVCHD requires an interlaced format, but even so - what the heck was the purpose to cheat the AVCHD and Blue-rays etc. instead of just leaving the camera to record interlaced? Isn't it better in almost all cases to have native, interlaced material and de-interlace it only when you really need it? The biggest advantages, like the smooth panning etc., are lost when you shoot progressive! And de-interlacing is so easy, while interlacing the progressive material simply does not work. So what was the reason to put progressive recording into an interlaced container, instead leaving it just interlaced? Is it because the consumers market demands for the "film look", so progressive scanning became the 'golden rule'..? I don't get it, sorry.

A progressive frame in an interlaced signal is split by alternate lines of resolution into each field. Odds and evens. A progressive frame has arguably more resolution than an interlaced frame especially during motion as an interlaced frame is essentially two half resolution frames in sequence but displayed together due to image latency on the display device. So yes, motion is smoother in 60i because it's actually 60 frames per second although it's counted as a 30 frame signal.

"1080/30p in 60i container" is more properly known as psf - progressive, segmented frame. See Progressive segmented frame - Wikipedia, the free encyclopedia

In essence it contains exactly the same information as a progressive image, and has the same characteristics. The only difference between the two is the order in which the lines are presented, 1,3,5,7..... then 2,4,6,8 for psf, rather than 1,2,3,4,5,6 etc for progressive. Otherwise the lines are identical. All the lines in each frame are captured at the same moment in time. Hence translating from psf to p and back is a totally lossless process, simply a matter of line reordering.

In the case of interlace, the even lines are captured at a different moment in time to odd lines. De-interlacing, to form progressive frames from interlace, is NOT a lossless process, and not reversible.

The reason for psf is to carry a progressive signal on systems designed for interlace, such as broadcast transmission networks, So in the UK, the system is 1080i/25, but since most production is 1080p/25 most of the time it's actually 1080psf/25 that is being transmitted.. To a receiver, 1080i/25 and 1080psf/25 are identical signals technically - even if they may look different to the viewer in terms of motion..

Well, now it is clear, at least technically, thank you guys. But from the point of view of the videographer - the smoothness of the fast movements is better when the footage was shot interlaced or progressive? When for example there is a leaf or a flag trembling on a wind or a bird flapping wings - I can always see kind of a flickering on a footage that has been recorded using progressive-scan. So I rather don't like progressive, just because of the flickering effect, but maybe I should change my mind..?

Shutter speed is important. A leaf falling in 30p or 25p will look odd if the shutter is set too fast. Generally the shutter is set for twice the speed of the frame rate. 1/48th for 24p, 1/50th for 25p, 1/60th for 30p, 1/100th for 60i.

No simple yes or no to that. The reason (using 50Hz country rates) is that interlace ALWAYS implies 25 frames per second, 50 fields per second, and the appearance of movement "looks" like the field rate - 50fps.

With progressive, you can have 25 or 50 complete frames per second (25p or 50p) and 50p gives the same appearance to motion as interlace. Whereas 25p will give a more jerky effect - or "film-look".

So it follows that in general, 1080 is better than 720, progressive is better than interlace, and 50 fps is better than 25 fps. Hence, the best system obviously becomes 1080p/50? Unfortunately, even now, that's still too difficult to implement in most equipment (without cutting corners, which some consumer cameras are starting to do), so you have to choose which two of the three factors are most important, so it becomes a practical choice between 720p/50 and 1080i/25 if "smooth motion is important. If it's not, and some directors actually prefer the "film-look" of 25p, then the best system is unarguably 1080p/25 - which can be carried as psf25 over an interlaced system.

I do agree with everything what you say. And shutter speed should be just right, but generally speaking - the fps is more important, as the higher the frame rate, the better the smoothness. 25p gives 25 complete frames per second. 50i is NOT 50 frames per second, but FIELDS per second, so since the complete frame is built of two fields - it is also 25 frames per second. In todays world, most of the display devices require stream of deinterlaced images, so my main question is:
- is it better to shoot progressive
or
- shoot interlaced and then deinterlace it in NLE. In other words: what will bring better results as the smoothness of fast moving objects is the priority?

In answer to that question alone, 50p will look the same as interlace. They will both look better than 25p in terms of smoothness of motion. Yes, 1080i/25 is 25 frames per second, 50 fields per second, but because the fields correspond to different moments in time, the eye is presented with 50 different images (motion wise) per second. If both fields of a frame correspond to the same moment in time, (as with the VG10) that's 1080psf/25.

I'm confused by what you say about "shoot interlaced and then deinterlace it in NLE". That will get you a derived 25p - with less smooth motion - and why not then just shoot in 25p in the first place? In terms of the display, best to just feed it interlace, and let it do it's own de-interlacing to drive the screen. Or feed it 50p - but that then means you must work in 720p/50.

When interlaced footage is played (on a computer screen for example) the odd/even lines looks ugly. Therefore it has to be de-interlaced first in some non linear editing comp. program. That is what I meant. And I was not talking about VG10 but in regards of general rule.

Well, thank you David for your kind explanations. And for confirming that 50 fields per second brings better results, because the impression of a watcher is of better smoothness. However, the 50i/60i in a technical specification of the Nex-VG10 might be very confusing to buyers and I was very much confused with it.

My understanding is that Sony Vegas is the only NLE that sees VG10 footage as 50i, hence the fact that you're finding the need to deinterlace. Other NLE's eg FCP ignore the interlaced wrapper and see it as 25p. Indeed FCP will convert the footage to Proroes 25p footage when logging and transferring automatically. As far as I'm aware this process does not involve deinterlacing, just discarding the redundant wrapper. This being the case, maybe you could use a program like mpeg streamclip to get rid of the wrapper prior to bringing it into your NLE.

As for why SONY did this in the first place, my understanding is that it was so you could burn BLU-RAY discs in the source format, as they only work with 50i footage...

1) the deinterlacing in an HDTV is usually far better than the one you would use on your computer -- unless you are willing to wait a very long time.

2) what you see on your NLE has no relevance to what your video will look like.

3) when you pause in an NLE you never see both fields, so you wont see interlacing.

4) when you play video in an NLE you should not see interlacing either. You might see them IF you create a progressive sequence and drop interlace video into it. That is your error.

5) if you need progressive, deinterlace only on export.

6) when you play interlace on your computer it is always deinterlaced so it should look fine.

Bottom-line, I'm not sure why you think you can see combing on your computer!

Now about the VG10. Every NLE I've tried sees the video as INTERLACED. As long as you edit in an interlaced Sequence, you will be fine AS LONG AS YOU EXPORT TO INTERLACED MEDIA. If you want progressive output, you must be sure you or you streaming service does NOT deinterlace the video.

Therefore, it is better to edit as progressive -- which is what the video really is -- as long as you can ret-tag the clips in a Bin or in a Sequence as PROGRESSIVE. You can export as either progressive or interlaced (PsF) video.

If your NLE can't do this, it may auto-deinterlace (BAD) if you use any FX that scale video.

PS: iMovie will automatically deinterlace unless you use my iMakeFullHD software that batch re-tags clips as progressive.

Steve, I hate to disagree but my version of FCP (studio 2) converts to prores 422 on log and transfer and sees the footage as 25p by default. Would FCP be auto deinterlacing on log and transfer?

And both products come from Sony...