Jump to content

576p is not HD


Recommended Posts

Yup, either film or a digital video camera capable of doing 24/25p (I think The 4400 was done this way?).

Not sure about The 4400 (IMDB doesn't have any tech specs), but this is definitely the right way to do 1080@50i - it is then capable of being 'cleanly' deinterlaced for progressive display, and 'degrades' gracefully to 1080@50p (albeit at half the possible frame rate).

- Miles.

The process used by 1080i video cameras to record 1080 25P is called 'progressive segmented frame' (psf). The camera captures the whole image off the CCD each 25th of a second, although, usually, an electronic shutter limits the capture time to a 50th, to reduce blur. The resulting image is split into two fields and recorded as interlace, i.e. 1080i.

I think the 4400 was shot on video using 24p, (i.e. 29.97 psf, a sort of reverse video 3:2 pulldown).

Some one will correct me if I'm wrong here, but I think most progressive displays don't deinterlace 1080i to scale it. They simply treat each field as a progressive frame and scale the 540 lines to the display res.

Link to comment
Share on other sites



  • Replies 448
  • Created
  • Last Reply

Top Posters In This Topic

When next generation DVD movies are released on Blu-Ray/HD-DVD, they will most likely be in the format 1080p24 (or 1080sf24 I am not exactly sure what this format is), will current displays be able to accept this input or will players have to convert to say 720p or 1080i (can players interlace a progressive format?)

Link to comment
Share on other sites

When next generation DVD movies are released on Blu-Ray/HD-DVD, they will most likely be in the format 1080p24 (or 1080sf24 I am not exactly sure what this format is), will current displays be able to accept this input or will players have to convert to say 720p or 1080i (can players interlace a progressive format?)

I doubt they will be 24p. The 108024p or 24sf is really only a post production foemat. The comsumer realease will start with 108050i and as someone pointd out if the content was shot in a 'p' mode then the display would be a 'p' display, the only think being the 2 fields have the same image, but are presented in a time offset. Some displays will de-interlace etc.

Down the track there may be 24/25p discs but not for a while. When you see a 'PR' like "Presented in 24p", the cameras will in 'p' mode but the transmission has to converted back to interlace mode, with fields that are indentical.

Link to comment
Share on other sites

This is straight from my Terminator 2 Extreme DVD "Manual"

The previous transfer of T2 (the version that appeared on T2 The Ultimate Edition DVD) was created at the highest quality available at that time. However, since that time, many improvements have taken place in the realm of film-to-video transfers (also known as telecine).

Upon evaluation of the 1997 master elements, THX and Artisan determined that the picture quality and format of the high-definition (HD) images might not meet today's standards for HD distribution in the future. While video processing on this older master could have created a good quality image for DVD, time and effort would be spent on a master that had little future. THX is proud to take part in Artisan's decision to retransfer T2 using the latest telecine and post production technologies. This new film transfer has resulted in the superior images you will find on this DVD release.

Starting from the same interpositive (IP) print created from the original negative of the film used in 1997, THX supervised the transfer of the film into HD video. This new HD master could be called an "electronic film master" since it is an exact representation of the film recorded onto HD-D5 videotapes. The transfer and subsequent mastering were performed at International Video Conversions (IVC) in Burbank, California.

The technical format for the T2 electronic film master is 1920x1080 24PsF. Here is a brief explanation of what that means. 1920x1080 is the size of each video frame and is the current standard resolution of HD video. Each frame of the film is scanned into a frame of video that is 1,920 pixels wide and 1,080 lines high. Doing some math reveals that a picture size of 1920x1080 has an aspect ratio of 1.78:1 (1920 / 1080) which means that the width of the picture is 1.78 times the height of the picture. 1.78:1 is also known by the term 16x9 because 16/9 also equals 1.78.

Any movie that has an aspect ratio of 1.78:1 would fill a 1920x1080 frame size perfectly. However, T2 was released in an aspect ratio of 2.35:1 which means that the width of the picture is 2.35 times the height of the picture. Movies that have an aspect ratio larger than 1.78 would have to have black mattes (also known as letterboxing) added to the top and bottom portions of the screen in order to fit the enire width of the wider film frame into a 1.78:1 picture.

Because of letterboxing, in a frame size of 1920x1080, a 2.35:1 picture would have used the entire resolution width of 1,920 pixels, but would only use approximately 815 lines of the 1,080 available lines. The rest of the 265 lines would be used and wasted by the black mattes. A decision was made early to optimize the new master by transferring the movie FULL FRAME by scanning each film frame so that it filled the entire 1920x1080 area, thus increasing the resolution of the image. While this master is not formatted properly, it utilises every single line for the actual picture content and maximises the ability to perform needed restoration.

24PsF (or Progressive Segmented Frames) describes two things: the frame rate and how the picture was created. The number "24" means that the electronic film master plays back at 24 frames-per-second, just like film. In fact, the electronic film master matches the film, frame for frame--one frame of film equals one frame of video.

"Progressive Segmented Frames" means that each frame of video is created progressively , just like film. In order to understand what progressive means, here are the differences between the two ways of creating video frames: progressive and interlaced. The majority of standard definition consumer televisions create frames of video using interlacing, while many HD consumer televisions are capapble of creating video progressively.

With interlaced scanning the video frame is scanned as two seperate fields, thus:

* Top field (odd scan lines) scan line numbers 1, 3, 5, 7, 9, 11, etc. followed by

* Bottom field (even scan lines) scan line numbers 2, 4, 6, 8, 10, 12, etc.

* The video frame does not always capture the exact same content that a film frame contains

* Motion does not appear as smooth; diagonal lines are jagged

With progressive scanning the video frame is scanned in one complete pass thus:

* The complete video frame as: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, etc.

* The video frame caputres the exact same content that a film frame contains at any instance in time

* Motion is smoother; diagonal lines are smooth

* Video frames are easier to work with since they are complete frames and no parts of frames

you can find out more details at http://www.thx.com/T2

Link to comment
Share on other sites



Thanks Davo, so why wouldn't the film studios produce their new Blu-Ray/HD-DVD HD film transfers in 1920x1080 24PsF?

24PsF is only used in telecine and production. The reason it's called PsF rather than just PF (progressive frame) is because it is 'segmented', i.e. a single progressive frame is divided into two fields for recording. The purpose of PsF is to allow the use of facilities designed for interlace to be used for low frame rate (i.e. 24 or 25 fps) progressive. It's also a way for Sony to flog their ageing 1080i HD system to the film industry, rather than develop a proper progressive HD System. Ultimately films will be shot on super high resolution true progressive systems.

When a DVD is made the HD master is converted to either 480i or 576i. For NTSC systems that means a conversion to 30fps, i.e. 3:2 pulldown. For us it means speeding things up by 1 fps to 25fps. A PAL DVD is effectively 576 25PsF.

Link to comment
Share on other sites

I should also add that the T2 manual is a bit confusing. You don't get jaggies from film transferred as interlace, only from interlace video originated material.

And the reason why the format commonality will not include 24P is the ongoing, and highly unproductive, film vs video format war that has raged for at least 20 years.

Link to comment
Share on other sites

Yup, either film or a digital video camera capable of doing 24/25p (I think The 4400 was done this way?).

Not sure about The 4400 (IMDB doesn't have any tech specs), but this is definitely the right way to do 1080@50i - it is then capable of being 'cleanly' deinterlaced for progressive display, and 'degrades' gracefully to 1080@50p (albeit at half the possible frame rate).

- Miles.

The process used by 1080i video cameras to record 1080 25P is called 'progressive segmented frame' (psf). The camera captures the whole image off the CCD each 25th of a second, although, usually, an electronic shutter limits the capture time to a 50th, to reduce blur. The resulting image is split into two fields and recorded as interlace, i.e. 1080i.

I think the 4400 was shot on video using 24p, (i.e. 29.97 psf, a sort of reverse video 3:2 pulldown).

Some one will correct me if I'm wrong here, but I think most progressive displays don't deinterlace 1080i to scale it. They simply treat each field as a progressive frame and scale the 540 lines to the display res.

Sony are planning HDCAMs with 1080i/150hz and 1080i/180hz formats to aid image quality particularly for fast motion like sports coverage instead of trying to do progressive formats. Their consumer display technology is focussed on 1080p at the moment though since the internal digital deinterlacing is a seemless process.

Link to comment
Share on other sites



Yup, either film or a digital video camera capable of doing 24/25p (I think The 4400 was done this way?).

Not sure about The 4400 (IMDB doesn't have any tech specs), but this is definitely the right way to do 1080@50i - it is then capable of being 'cleanly' deinterlaced for progressive display, and 'degrades' gracefully to 1080@50p (albeit at half the possible frame rate).

- Miles.

The process used by 1080i video cameras to record 1080 25P is called 'progressive segmented frame' (psf). The camera captures the whole image off the CCD each 25th of a second, although, usually, an electronic shutter limits the capture time to a 50th, to reduce blur. The resulting image is split into two fields and recorded as interlace, i.e. 1080i.

I think the 4400 was shot on video using 24p, (i.e. 29.97 psf, a sort of reverse video 3:2 pulldown).

Some one will correct me if I'm wrong here, but I think most progressive displays don't deinterlace 1080i to scale it. They simply treat each field as a progressive frame and scale the 540 lines to the display res.

Sony are planning HDCAMs with 1080i/150hz and 1080i/180hz formats to aid image quality particularly for fast motion like sports coverage instead of trying to do progressive formats. Their consumer display technology is focussed on 1080p at the moment though since the internal digital deinterlacing is a seemless process.

Don't you mean 1080/50p and 1080/60p? The rate you metion will be bloody bandwidth hungry.

Link to comment
Share on other sites

Don't you mean 1080/50p and 1080/60p? The rate you metion will be bloody bandwidth hungry.

No, I most definitely don't. We are talking an all new recording format. I don't know a time frame on this as Sony has not divuluged one yet.

The formats you describe are what is called 'Super slow motion'. That is record at mutples of the frame rate and then replay at 1080i/25 frames, and you have proper slow mo, not step frame slow mo.

Sony have in the works a 1080/50p camera which will be twice the bandwidth with really high quality images for post production. The problem is it will be so good the old 'film die hards' won't like it because there will be no 'flicker'.

Link to comment
Share on other sites

Don't you mean 1080/50p and 1080/60p? The rate you metion will be bloody bandwidth hungry.

No, I most definitely don't. We are talking an all new recording format. I don't know a time frame on this as Sony has not divuluged one yet.

The formats you describe are what is called 'Super slow motion'. That is record at mutples of the frame rate and then replay at 1080i/25 frames, and you have proper slow mo, not step frame slow mo.

Sony have in the works a 1080/50p camera which will be twice the bandwidth with really high quality images for post production. The problem is it will be so good the old 'film die hards' won't like it because there will be no 'flicker'.

The presentation I saw on 1080i/150/180 did not speak of super 'slowmo', just better image quality. No mention was made of 1080p cameras being developed for professional use.

Link to comment
Share on other sites



Their consumer display technology is focussed on 1080p at the moment though since the internal digital deinterlacing is a seemless process.

It might be seamless for film originated, or 24/5 PsF, but it's not for 1080 50i.

Sort of begs the question why Sony would focus on 1080P for domestic displays when are they are so keen on 1080i for production? Why interlace to de-interlace?

Link to comment
Share on other sites

Their consumer display technology is focussed on 1080p at the moment though since the internal digital deinterlacing is a seemless process.

It might be seamless for film originated, or 24/5 PsF, but it's not for 1080 50i.

Sort of begs the question why Sony would focus on 1080P for domestic displays when are they are so keen on 1080i for production? Why interlace to de-interlace?

Good question, perhaps they feel it's easier to work with. Owen is the expert on the 1080i/p deinterlace question, but my understanding is that quite flawless for 1080i 50/60, the feedback on Sony's 1080p Qualia rear pro seems to indicate this. Check it out on AVSForum..

Link to comment
Share on other sites

Their consumer display technology is focussed on 1080p at the moment though since the internal digital deinterlacing is a seemless process.

It might be seamless for film originated, or 24/5 PsF, but it's not for 1080 50i.

Sort of begs the question why Sony would focus on 1080P for domestic displays when are they are so keen on 1080i for production? Why interlace to de-interlace?

Good question, perhaps they feel it's easier to work with. Owen is the expert on the 1080i/p deinterlace question, but my understanding is that quite flawless for 1080i 50/60, the feedback on Sony's 1080p Qualia rear pro seems to indicate this. Check it out on AVSForum..

The reason I say it's not flawless is because once something is shot interlaced it cannot really be de-interlaced. Deinterlacing is a bit like interpolating to a higher resolution.

The easiest way to explain this is to imagine (theoretical) material shot in 1080 50P. To convert it to 1080 50i you have to drop half the vertical resolution (every alternate line) from each frame. When the image is static that's OK, becasue the two fields will still combine to produce a full resolution image. But when the image moves you've only got half the resolution, because the fields don't line up anymore.

The other problem with this theoretical conversion is that the refresh rate of 1080i is 50Hz but each field is only refreshed every 25th of a second. If the 1080P original has detail that resides on only one line that line would flicker annoyingly at 25Hz, so the conversion would need to soften the image to avoid this.

So, as you can see, a theoretical conversion of 1080p to 1080i will result in up to a 50% loss of potential resolution.

The only "flawless" 1080i to 1080p conversion is 1080 25PsF. If 25fps is what is to be displayed it would make a lot more sense to record and compress it as 25 frames progressive.

Link to comment
Share on other sites

I should also add that the T2 manual is a bit confusing.  You don't get jaggies from film transferred as interlace, only from interlace video originated material.

I think what they're referring to is the difference between the old tranfer on the ultimate dvd to the new transfer on the extreme dvd

that would make perfect sense

Link to comment
Share on other sites

The reason I say it's confusing is that what they say is correct, but not in the context of a telecine transfer, e.g:

"Progressive Segmented Frames" means that each frame of video is created progressively , just like film."

Wrong, PsF means it's recorded as 1080i, i.e. interlace.

"With interlaced scanning the video frame is scanned as two separate fields, thus:

* Top field (odd scan lines) scan line numbers 1, 3, 5, 7, 9, 11, etc. followed by

* Bottom field (even scan lines) scan line numbers 2, 4, 6, 8, 10, 12, etc.

* The video frame does not always capture the exact same content that a film frame contains

* Motion does not appear as smooth; diagonal lines are jagged"

This would be true in the case of a film camera and a video camera recording the same scene, but this is not the case here. A telecine will capture both fields from the same frame. Jaggies are an artifact of interlaced origination, not interlaced telecine. You shouldn't get jaggies on either an interlace or 24 PsF transfer.

I know I'm being a bit pedantic, but THX should explain things correctly.

Link to comment
Share on other sites



The reason I say it's confusing is that what they say is correct, but not in the context of a telecine transfer, e.g:

"Progressive Segmented Frames" means that each frame of video is created progressively , just like film."

Wrong, PsF means it's recorded as 1080i, i.e. interlace.

"With interlaced scanning the video frame is scanned as two separate fields, thus:

* Top field (odd scan lines) scan line numbers 1, 3, 5, 7, 9, 11, etc. followed by

* Bottom field (even scan lines) scan line numbers 2, 4, 6, 8, 10, 12, etc.

* The video frame does not always capture the exact same content that a film frame contains

* Motion does not appear as smooth; diagonal lines are jagged"

This would be true in the case of a film camera and a video camera recording the same scene, but this is not the case here.  A telecine will capture both fields from the same frame.  Jaggies are an artifact of interlaced origination, not interlaced telecine.  You shouldn't get jaggies on either an interlace or 24 PsF transfer.

I know I'm being a bit pedantic, but THX should explain things correctly.

It then goes on to say:

The theatrical version of the film and the individual special edition scenes were all tranferred and put through the DRS process using the above methods to 1920x1080 24PsF video. Because the HD version was similar to film in regards to being 24 frames per second progressive, editing the special edition version together was much easier than dealing with interlaced vdeo masters.

The 24PsF format enables scenes to be added due to the 24-frame-per-second video rate. Prior HD formats were 30-frame based which meant the film images were recorded with a 3:2 sequence in order to fir 24-frame-per-second film on a 30-frame-per-second video format. The 24PsF format allows seamless editing and post production to take place without the 3:2 sequence, implenting this only after scenes are added. The produces a perfect 3:2 sequence integrity which is critical for the MPEG-2 compression forDVD

blah blah...

From my understanding, they're not saying that "Progressive Segmented Frames" means that the film is shot as progressive, rather the video is progressive, just like the film is progressive - that's what they're saying

Link to comment
Share on other sites

Good question, perhaps they feel it's easier to work with. Owen is the expert on the 1080i/p deinterlace question, but my understanding is that quite flawless for 1080i 50/60, the feedback on Sony's 1080p Qualia rear pro seems to indicate this. Check it out on AVSForum..

The easiest way to explain this is to imagine (theoretical) material shot in 1080 50P.  To convert it to 1080 50i you have to drop half the vertical resolution (every alternate line) from each frame.  When the image is static that's OK, becasue the two fields will still combine to produce a full resolution image.  But when the image moves you've only got half the resolution, because the fields don't line up anymore.

So, as you can see, a theoretical conversion of 1080p to 1080i will result in up to a 50% loss of potential resolution.

+

Wrong, PsF means it's recorded as 1080i, i.e. interlace.

My reply.

I think you are incorrect in saying that a 1080/50p recording converted to 'i' is a loss of 50%. If I was to record a 1080/25p tape, I can play that tape back (without any change) as 1080/50i. The resolution is the same, its the 'presentation of the frames that is different. This will result in a difference in the apparent 'flicker' on Plasmas/LCDs. Most CTRs don't even do 'P' mode, only 'i' mode.

However if you take 1080/50i and convert it to 1080/50p there can be a loss in resolution.

PsF means 'Segmented' as opposed to 'Progressive' or 'Interlace'. You are close in that PsF captures the image for both fields at the same time but records them as 2 fields (line 1,3,5 etc then line 2,4,6) like 'i' mode. Progressive captures a frame and then records it as line 1,2,3 etc. Interlace captures 2 fields at different times and records them as 'i' mode, line 1,3,5 etc then line 2,4,6.

This is all from the Sony HDW-F500 manual and the Miranda format standards chart.

Link to comment
Share on other sites

The 24PsF format enables scenes to be added due to the 24-frame-per-second video rate. Prior HD formats were 30-frame based which meant the film images were recorded with a 3:2 sequence in order to fir 24-frame-per-second film on a 30-frame-per-second video format. The 24PsF format allows seamless editing and post production to take place without the 3:2 sequence, implenting this only after scenes are added. The produces a perfect 3:2 sequence integrity which is critical for the MPEG-2 compression forDVD

I think I get what they're on about now.

In effect, what they are saying is that last time they transferred the movie and cut scenes into it, (once it was in the video domain), they stuffed up the 3:2 cadence. As one frame in every 3 consists of two fields from different film frames it's essential that edits take this into account, otherwise it can confuse the deinterlacer and cause jaggies.

Happily for us we don't have to worry about 3:2 because the movie just gets sped up to 25 fps, i.e. 2:2.

Link to comment
Share on other sites

I think you are incorrect in saying that a 1080/50p recording converted to 'i' is a loss of 50%. If I was to record a 1080/25p tape, I can play that tape back (without any change) as 1080/50i. The resolution is the same, its the 'presentation of the frames that is different.

Actually Chris is spot on. It all comes down to the original format. If shot at 24p or 25p (on either HD tape or film) then yes there is no loss in resolution when converting 1080/24p to 1080/50i. The reason being each field is from the "same moment in time" and these fields can be recombined (via simple weave de-interlacing) to form a true 1080/25p format (with each frame shown twice to display at 1080/50p). This is why 1080i usually looks best for film.

However if something was shot on a native 1080/50p or 1080/60p format (not possible at this stage but no doubt it will be in the future) then you are talking 1080 full lines of resolution per frame (1080 lines per 50th of a second). The only way to convert this to 1080/50i is through the normal interlaced conversion process of splitting frames into fields (even lines first field, odd lines the next). This means each motion update, you are only getting 1920x540 resolution, rather than the full 1920x1080 of the original 50p material (or another way to think of is 1,036,800 pixels per 50th of a second vs 2,073,600).

The only way to then de-interlace this material back to 1080/50p is to upscale each field to 1080p through line interpolation (bob de-interlacing) or at best per pixel motion adaptive de-interlacing, as you cant recombine the fields to progressive frames when they don’t line up (as they are from different moments in time). This means with de-interlaced 1080/50p from progressive video based sources you only end up with 540 unique vertical lines per frame.

This is the reason that 720p outshines 1080i for video based material (anything shot higher than 24 frames a second) as you are always getting the full 720 lines per motion update as opposed to 540 of 1080/50 or 60i.

Put simply, if the material is 24p based - 1080i can be the best looking format due to the ability to reclaim the full 1080 lines of vertical resolution. However, anything shot on video at 50 or 60 motion updates a second is far better suited to a native progressive format and can look far superior (particularly when there is a high level of movement such as sport).

This page has come nice clear diagrams illustrating the superiority of a true progressive format when dealing with motion (and is the main reason the American Department of Defence went with 720p over 1080i) -> http://www.atd.net/HDTV_faq.html

Link to comment
Share on other sites

Nice summary Darklord, why can't material recorded at 1080p at 50 or 60 Hz be encoded as 1080i at 100 or 120 Hz, I suppose it takes up the same bandwidth as 1080p 50 or 60 Hz, but DVB does not support 1080p.

What I want to know is how will the studios encode their films on Blu-Ray/HD-DVD: 1080PsF24/25, 1080p24/25, 1080i48/50, ...?

Link to comment
Share on other sites



  • Recently Browsing   0 members

    • No registered users viewing this page.




×
×
  • Create New...
To Top