Is 30p really 1080p?

[David Ruhland]

On my HV30 I see that it says records in both 1080i and 1080p. I look on my menu setting and all I see is 1080i 30p and 24p. So if I shoot at 30 p is this considered 1080p?

I am still a bit confused on this 1080i 60 verses progressive mode, but tell me if I am getting it...Remember I came from the still photography world..

Here is my understanding of interlaced vs progressive:

Shooting in interlaced and editing to PROGRESSIVE in Vegas Pro is like shooting a RAW file in my still camera-it records all the data it sees.

Shooting in 30 p mode is like shooting High Resolution JPEGs in my still camera...it basically takes the info it thinks I want and throws the rest away

shooting 24p....Dont have a darn clue...

Am I understanding it correctly? and how do I shoot true 1080p in my HV30?

Merry Christmas!

[John Kilderry]

It's probably best to throw out what you think you know about this and start over.

It gets deeper than this, but this is a good starting point. 24p and 30p are frame rates per second progressive 60i is 30 frames per second with 60 interlaced fields which is what you would find on an average television set in the US.

24p is usually shot with a shutter speed of 1/48, 30p = 1/60, and 60i at aproximately 1/120.

The 1080 portion refers to the vertical resolution and can be either 1080i or 1080p depending if your final render is interlaced or progressive. The horizontal resolution would be a number like 1920, 1440, or 1280, so you'll see numbers like 1920X1080 or 1440X1080 or 1280X720.

So 24p and 30p is frames per second and 1080 refers to resolution.

Remember I said it gets deeper? Your HV30 really doesn't shoot true 24p or 30p it has to be deinterlaced during capture if you desire true progressive. Its resolution is 1440X1080i as most HDV video is.

[David Ruhland]

Glad I asked, but I am confused on the de interlacing...

" Your HV30 really doesn't shoot true 24p or 30p it has to be deinterlaced during capture if you desire true progressive"

Does that happen during CAPTURE or during RENDER? I am using Vegas Pro 8 and I only recall progressive setting on the properties menu.

[Chris Barcellos]

I have to disagree with the John's response above. It does not accurate portrait what is going on with each frame rate.

First, interlacing means that in 60i, the camera scans the chip 60 time per second. For the first scan, it will scan the screen every other line, and for the second scan, in scans the previously unscanned lines. Two of those scans make up was is essentially a single frame image. Thus 60i, combines the line scans to get a full single image, giving essentially 30 full frames a second.

HDV is set up to play back that way.

Enter 30p. It scans the chip fully in one pass, 30 each second. There is no interlacing of scans in this case, as each scan is a full image frame.

24p does the same thing, only at 24 frames per second.

In order to play on conventional TVs though, we have to get back to 1080i. In the case of 30p, the frames are easily turned in to 60 frames to conform with the HDV standard. But it is still true 30p, and most people misunderstand that.

With 24p, "pull down" is added by the camera chip to make the camera think it is recording 60i, again to conform to HDV standards. But, once again, that can be stripped away with your NLE to have you left with the true 24 frames per second.

As far as 1080 60i, I also disagree with John suggestion you use 1/120th of a second. IMHO 1/60 should be used for standard shooting. Getting in 1/120th may get you into a stacatto effect.

[Pete Bauer]

All HDV recorded to tape by this camera will be 60 fields/sec in compliance with NTSC standard.

The 30p will be what's often referred to as "progressive segmented frames" or psf, meaning one progressive frame taken from the imaging block will be split into two fields on tape. If treated as psf when put back together in your NLE, they will appear to be a progressive frame, not interlaced. That depends on your NLE's handling; I don't use Vegas but would expect it does that. For 24p, it will lay the progressive frame to tape as 2:3. Again, current NLEs ought to have presets to handle that in a way that's transparent to you.

EDIT: Chris B posted while I was typing. We're describing the same thing in different words, so between our two posts hopefully it'll be clear what the camera is doing with the signal. I also agree that 1/60th should be the normal shutter speed for 30p or 60i and 1/48th for 24p, while 120th would be for special purposes because the reduced amount of motion blur in each frame due to the higher shutter speed does cause that staccato effect.

[John Kilderry]

Originally Posted by Chris Barcellos

I have to disagree with the John's response above. It does not accurate portrait what is going on with each frame rate.

First, interlacing means that in 60i, the camera scans the chip 60 time per second. For the first scan, it will scan the screen every other line, and for the second scan, in scans the previously unscanned lines. Two of those scans make up was is essentially a single frame image. Thus 60i, combines the line scans to get a full single image, giving essentially 30 full frames a second.

HDV is set up to play back that way.

Enter 30p. It scans the chip fully in one pass, 30 each second. There is no interlacing of scans in this case, as each scan is a full image frame.

24p does the same thing, only at 24 frames per second.

In order to play on conventional TVs though, we have to get back to 1080i. In the case of 30p, the frames are easily turned in to 60 frames to conform with the HDV standard. But it is still true 30p, and most people misunderstand that.

With 24p, "pull down" is added by the camera chip to make the camera think it is recording 60i, again to conform to HDV standards. But, once again, that can be stripped away with your NLE to have you left with the true 24 frames per second.

As far as 1080 60i, I also disagree with John suggestion you use 1/120th of a second. IMHO 1/60 should be used for standard shooting. Getting in 1/120th may get you into a stacatto effect.

I agree, Chris, and was trying to keep things simple so David could chew off a bit at a time. I tried to explain that the subject gets deeper.

[Predrag Vasic]

Bottom line of this is: there are very few reasons for shooting interlaced. In most cases, shooting progressive will yield better results and will be easier to work with. The easiest of them all will be shooting 30p. Your Vegas will easily de-interlace that HDV stream and re-compose original 30 progressive frames. As the guys explained before, HDV will always encode video into an interlaced stream (in order to stay compatible with the old, analogue, Standard-Def NTSC standard), even if the video capturing was done progressively. Luckily, with 30p, each frame is split into two fields and these two are encoded one after the other in that 60i stream. Vegas simply recomposes those two fields into a proper single frame, same as it was captured.

The only major advantage of 24p is that 'cinematic' look. The disadvantage of it is, in order to stretch 24 progressive frames across 60 interlaced fields, the encoding Canon 'DiGiC' chip chops up those progressive frames into two fields each, and then spreads the resulting 48 fields across 60 interlaced fields using 2:3 formula (essentially duplicating every fifth field). To remove those duplicated fields from 24p shoehorned into 60i, you'll need some other software; Sony Vegas CAN remove them, but the 60i footage needs to have 'pulldown flags' embedded in it, so that Vegas knows which field is duplicate. I believe only HV40 inserts those 'pulldown flags'; HV 30 and older don't.

So, shoot 30p, have Vegas de-interlace footage and edit merrily. You get the highest image quality, clarity, sharpness and low-light sensitivity.

[Taky Cheung]

If you shoot 30p, you don't need to deinterlace AT ALL. It is already progressive. Even if it is the segmented progressive frames, the two interlaced frames carry the same content. That is still progressive.

[Predrag Vasic]

Originally Posted by Taky Cheung

If you shoot 30p, you don't need to deinterlace AT ALL. It is already progressive. Even if it is the segmented progressive frames, the two interlaced frames carry the same content. That is still progressive.

I don't think it is entirely correct. My understanding is that HV-40 (and other HDV camcorders) don't encode entire frames. They encode only FIELDS, and each field only contains one HALF of an entire FRAME. This stuff often gets confusing (fields fs. frames). The point being, odd fields will only containing odd horisontal lines, while even fields will contain even horisontal lines. You will then have to deinterlace this content in order to re-compose original progressive FRAMES from the two FIELDS containing two complementing halves of those original, progressive frames.

The legacy of old, standard-def, analogue broadcast systems (NTSC and PAL) is still rearing its ugly head in HD. While there is absolutely NO reason for interlaced content in HD (nor is there any difference between NTSC and PAL -- HD is HD, everywhere in the world), these ancient practices of chopping up a perfectly decent progressive frame and shoving it into two consecutive halves continues to give headache to newcomers to HD...

[Taky Cheung]

I think you are confused with HD vs SD to Interlace vs Progressive. :) In both HD and SD video, they could be either progressive or interlaced. The segmented progressive frame is a good way to make progressive video compatible to all the TVs in the world.

HV30/HV40 shoot progressive wrapped in an interlaced stream. But there is no need to deinterlace that footage. =)

[Chris Hurd]

Originally Posted by Taky Cheung

If you shoot 30p, you don't need to deinterlace AT ALL.

Originally Posted by Predrag Vasic

I don't think it is entirely correct.

Actually yes, Taky is indeed quite correct.

When recording HDV in progressive scan mode, the frame is still separated into
two different fields, just as it is during interlace. The big difference, however, is
that those two fields were acquired at exactly the same time. Since there is no
difference in the timing of those two fields, there is no need to de-interlace.

When recording in interlace, the two fields are separated in the timing of their
acquisition, by 1/60th of a second. That is the circumstance and reason for
de-interlacing in post.

In short: no matter whether the recording happens in normal interlace or in
progressive scan mode, each frame will be composed of two separate fields.
Since those fields were acquired simultaneously in progressive scan mode,
there is no need to de-interlace them in post. Hope this helps,

[Predrag Vasic]

I don't think that is correct. Interlaced stream can contain source material that may have originated as progressive frames. However, because the stream contains a sequence of fields (each of which only contain a half of an image), it will be interlaced. In order to re-compose original progressive frames, the stream must be de-interlaced. However, since the interlaced fields represent two halves of a single progressive frame (rather than two independent fields, separately acquired at different times), deinterlacer does NOT need to do any processing to the fields before recombining them into a single progressive frame.

Deinterlacing is a process that re-composes frames from independent fields. It doesn't matter whether those fields were acquired in sequence (the way they are on standard-def, analogue cameras), or if they came from a single frame that was broken up into two fields. Whenever a stream is interlaced (i.e. containing sequence of fields, rather than full frames), such sequence needs to be de-interlaced in order to be put onto a progressive timeline.

If the source material was captured as truly interlaced (i.e. each field captured after another), the deinterlacer must employ some clever image processing (blending, selective blending, or more advanced processing) in order to neutralise motion artifacts inherent in interlaced video. If the source was captured as progressive, interlacer can just do the simple 'weave', which combines two successive fields into a single frame.

The only time you wouldn't want to de-interlace the footage that was shot progressive and encoded in an interlaced stream (30p into 60i) would be if your target device knows how to automatically detect and de-interlace such footage. Many modern HDTV sets can now detect 30p (even 24P) source inside a 60i stream and deinterlace, some TVs even doing the inverse telecine (removing 2:3 pulldown) on the 24p in 60i material.

For editing on a progressive timeline (30p or 24p), deinterlacing must be done, even if you did shoot 30p (wrapped into 60i).

[Chris Hurd]

Originally Posted by Predrag Vasic

I don't think that is correct.

For the last time... yes, it is indeed correct.

Originally Posted by Predrag Vasic

The only time you wouldn't want to de-interlace the footage that was shot progressive and encoded in an interlaced stream (30p into 60i) would be if your target device knows how to automatically detect and de-interlace such footage.

But that's the whole point -- the NLE knows how to automatically detect progressive
scan video and put two fields together (which is not the same thing as de-interlacing,
by the way). As long as there is a capture setting for HDV 1080p30, then nothing
further needs to be done.

Originally Posted by Predrag Vasic

For editing on a progressive timeline, deinterlacing must be done, even if you did shoot 30p (wrapped into 60i).

Actually, no -- that is quite wrong. For progressive scan HDV video, assuming that the
proper capture settings for HDV 1080p30 are used, the NLE already knows how to put
two fields together to make up the frame. If the video is captured into the NLE properly,
then there is no need to de-interlace, as has already been pointed out several times in
this discussion.

Once again: de-interlacing is the process by which the *difference in time between two
fields of one frame* is abated. In progressive scan video recorded to HDV cassette,
there is no difference in time between the two fields of a given frame... thus, no need
to de-interlace.

We're starting to go in circles, so we're now at the conclusion of this thread. Thanks all,