ATSC broadcast may push over the air 1080p broadcast.

The new digital TV broadcasting standard ATSC that is set to go into effect next year could eventully be broadcasting over the air 1080p. We went from Black and white, to color, to cable, to sattlelight, to HD.

Right now the maximum output signal for HD broadcasting is 1080i, meaning that if you own a 1080i set then you don't need anything else because your recieving the signal at it's fullest quality. Right now the only way you can get a 1080p broadcast is to use the PSF(Pogressive segemented frame) metiod. It allows 1080p on interlaced media that can only handle 1080p. That means if you own a 1080i set, you'll be viewing 1080p regardless. For those who own 1080p sets, you won't to rely on anymore de-interlacing.

If the PSF method can be used, why don't broadcasters use it?

Where did you get this? Got Link?

Since the ATSC standards don't support 1080p/60, they presumably would be talking about broadcasting 1080p/24 or 1080p/30, which would fit within the existing bandwidth of 1080i/30.

1080p/24 would be great for films and would even free up some bandwidth for the broadcasters. I can't imagine them producing programming in 1080p/30, as it's motion response would be inferior to either 720p/60 or 1080i/30.

Hopefully they wouldn't consider de-interlacing content sourced at 1080i/30 to 1080p/30--that would be ugly...

Films could be sent at 1080p/30 with same motion response as 1080i/30, but why?

Assuming the TV tuners actually would handle 1080p/24 my question would be how would it be displayed? If it still used a 60 hz refresh, there would be no gain.

Except the broadcasters would be left with extra bandwidth to make money on :D

The same caveat could be said about Blu-Ray and HD DVD players--if your display doesn't support 1080p/24, or it supports 1080p/24 but not a refresh rate that is a multiple of 24Hz, it doesn't get you anything. But more and more new HDTV's are going to be supporting that feature in the future.

Signals captured at 1080p/30 have half the motion update rate of signals captured at 1080i/30. A 1080p/30 signal updates frames 30 times per second, while a 1080i/30 signal updates fields 60 times per second--temporal resolution (i.e. motion response) is twice as good although spatial resolution at that rate is half as good.

Yes a gain for broadcasters, but none for the user which I was saying to the OP. That post implied there would be some benifit to the viewer and I don't see it. That's why I asked for a link to try to see what was behind the post.

It is quite possible, even though ATSC tuners are supposed to support 1080/24p and 1080/30p, that many in use today do not. I don't think broadcasters will take that chance.

Personally, I doubt that *any* of the ATSC tuners currently out there (at least in conjunction with their built-in MPEG2 decoders) can decode 1080p/24 or 1080p/anything. I really can't imagine a penny-pinching CE company bothering to provide compatibility with signals that aren't in use and weren't planned to be in use, even if they are in the ATSC spec.

I could be wrong, but I doubt it...

PsF... only means that it is recorded on tape progressive but the output is still interlaced. That is why the CRT monitors can handle it, if it were true progressive the CRT monitors would freak out. So there would be no difference to a signal that is film based, but if it were video based it would have jaggies.

The motion response of 1080PsF would be half that of 1080i in the same bandwidth--1080PsF/30 would capture 30 Frames-per-second, while 1080i/30 would capture 60 fields-per-second.

There is not enough bandwidth to transmit 1080p/60 or 1080PsF/60, so any signal captured at 1080PsF they could transmit would have inferior motion response compared to either 1080i/30 or 720p/60.

Films are already essentially transmitted as 1080PsF/30, as the modern, digital Telecine process is based on always extracting fields from the same film frame and repeating whole frames as necessary to pull down the film from 24Fps to 30Fps (i.e. 60 fields-per-second).

The motion response of 1080PsF would be half that of 1080i in the same bandwidth--1080PsF/30 would capture 30 Frames-per-second, while 1080i/30 would capture 60 fields-per-second.

There is not enough bandwidth to transmit 1080p/60 or 1080PsF/60, so any signal captured at 1080PsF they could transmit would have inferior motion response compared to either 1080i/30 or 720p/60.

Films are already essentially transmitted as 1080PsF/30, as the modern, digital Telecine process is based on always extracting fields from the same film frame and repeating whole frames as necessary to pull down the film from 24Fps to 30Fps (i.e. 60 fields-per-second).The Telecine 3:2 pulldown is not done in frames it is done in fields so there are still mixed frames.
There is no 1080PsF/60, it is 23.98, 24, 25, 29.97, & 30.

Yes, Telecine is still done in fields, but with current digital Telecine, the two video fields (which essentially create a video "frame") always come from the same frame of film, not like the mechanical Telecine process which periodically created new video "frames" by combining fields from different film frames--very bad if there was a cut or scene change in the film between the scan of the first field and the scan of the second.

That was all the result of moving the film through the scanner gate synchronized to the video scanner vertical sync rate. Since there was no real-time storage (the film itself was the "storage"), the fields came from whatever film frame happened to be in the gate at the time. Now they can keep the film frame in memory and derive the two fields from that to avoid the problem of scene changes between the first and second field.

At that point, the interlaced result is effectively PsF, as both fields of each video "frame" come from the same instant in time even though the are transmitted as two interlaced fields.

The telecine 3:2 pulldown takes a film frame and makes two fields of video from it, then the next frame of film makes three video fields, then the next frame of film makes two fields of video, and then the next frame of film makes three fields of video and so on and so on. Thus the third frame of video is a mixed frame from the second and third film frames and the fourth video frame is also a mixed frame from the third and fourth film frames. That is also how the video tape machines that record 23.98 make an NTSC signal.

So I don't understand what you are saying about no mixed frames.

That's the old system. I'm not talking about video tape.

AFAIK, Pull-down for HD is done digitally, so it never takes the odd video field from one film frame and the even video field from another film frame--it goes back to the film frame in the buffer memory and always takes both the odd and even fields from the same film frame.

It's not a real-time process, there is always several frames of delay between the film frame currently being scanned and the frame (fields) being output, or it's completely done as a post process after all the frames are in memory.

That's the old system. I'm not talking about video tape.

AFAIK, Pull-down for HD is done digitally, so it never takes the odd video field from one film frame and the even video field from another film frame--it goes back to the film frame in the buffer memory and always takes both the odd and even fields from the same film frame.

It's not a real-time process, there is always several frames of delay between the film frame currently being scanned and the frame (fields) being output, or it's completely done as a post process after all the frames are in memory.Every HD tape that I have sent out for broadcast here in the U.S. has been a 1080i/59.94 and it has a 3:2 pull-down in it (if it was film based) and it is done just like I wrote earlier. The 3:2 pull-down has been digital for the last 20 years, Rank Cintel made the digi-scan and that produced the 3:2 pull-down for telecine.

BTW it is real time. Both from telecine or from a video tape machine.

I've been running video for about 20 years and I have never seen the 3:2 done any other way than how I explained it (except for the new pro-sumor half-ass HD cameras). In those 20+ years I have been an on-line editor, telecine colorist, tape operator, and the manager of technical services. So if you could explain a little more about how the new digital 3:2 works or give me a link on where I could learn about I would really appreciate it.

How is it done for 720p/60 broadcasts?

Every HD tape that I have sent out for broadcast here in the U.S. has been a 1080i/59.94 and it has a 3:2 pull-down in it (if it was film based) and it is done just like I wrote earlier. The 3:2 pull-down has been digital for the last 20 years, Rank Cintel made the digi-scan and that produced the 3:2 pull-down for telecine.

BTW it is real time. Both from telecine or from a video tape machine.

I've been running video for about 20 years and I have never seen the 3:2 done any other way than how I explained it (except for the new pro-sumor half-ass HD cameras). In those 20+ years I have been an on-line editor, telecine colorist, tape operator, and the manager of technical services. So if you could explain a little more about how the new digital 3:2 works or give me a link on where I could learn about I would really appreciate it.

You may want to look at the specs. of the equipment movie studios are using for digital film conversion and for DTV production.

There are multiple non-real-time/semi-real-time methods including virtual pull-down and soft pull-down. When working with a digital intermediate there is no problem at all adding pull-down without creating mixed frames.

Except for the rare commercial DVD in which films have been Telecined before encoding (which may be done if the master is only available in Telecined form), the DVD player itself repeats the fields on playback to add pull-down to 24Fps content on the disc. It's very wasteful of disc space to store 24Fps content in Telecined form since the majority of stored fields would be repeats. The DVD player uses repeat flags encoded on the disc to know what frames to repeat and how often. Since each odd/even field pair on the disc comes from a single frame of film, DVD's of most films are essentially encoded as 480PsF/24.

Not that Wikipedia is the end all of accurate information, but you may want to look at the information on MPEG2 encoding used for HD broadcasts. Since the TV station is transmitting MPEG2 compressed DTV signals, they can even transmit a 1080p/24 signal and have the MPEG2 decoder in the ATSC tuner add pull-down:

http://en.wikipedia.org/wiki/MPEG-2

Anyhow, I think we debated this several years ago to no conclusion. As I recall, you tried to convince me nobody could see judder resulting from the uneven repeat cadence of 3:2 pull-down, which I, of course, disagreed with as I see it all the time. Now that everybody and his brother in the CE business is talking about 1080p/24 and displays with 72Hz or 120Hz refresh to eliminate judder, I feel pretty good about my position.

Let's just say your experience doesn't match up with mine and agree to disagree. As part of my consulting, I design, set up and help debug hardware and know how it works. Is it your contention that all 24Fps film transfers pulled-down to 60 fields-per-second periodically have bogus frames made from two different film frames?

How is it done for 720p/60 broadcasts?In the case of 720p/59.94, each frame of the film is repeated either 2 or 3 times depending on the 3:2 cadence. In the 10+ years of running HD I have only made a handful of 720p/59.94 tapes.

You may want to look at the specs. of the equipment movie studios are using for digital film conversion and for DTV production.First off I deal with the studios (Fox, Sony, Universal, & Miramax just to name a few) everyday I work so I know their specs pretty well.
Not that Wikipedia is the end all of accurate information, but you may want to look at the information on MPEG2 encoding used for HD broadcasts. Since the TV station is transmitting MPEG2 compressed DTV signals, they can even transmit a 1080p/24 signal and have the MPEG2 decoder in the ATSC tuner add pull-down:

http://en.wikipedia.org/wiki/MPEG-2

Anyhow, I think we debated this several years ago to no conclusion. As I recall, you tried to convince me nobody could see judder resulting from the uneven repeat cadence of 3:2 pull-down, which I, of course, disagreed with as I see it all the time. Now that everybody and his brother in the CE business is talking about 1080p/24 and displays with 72Hz or 120Hz refresh to eliminate judder, I feel pretty good about my position.Thanks for the heads up on wiki... that explains a lot of how the broadcast side works and now I understand more of why we always had to make sure the 3:2 cadence was correct. Because they stripped it out for transmission and then the decoder added it back in, and now I also see why you would see more judder that way because they are repeating a frame not a field.



Let's just say your experience doesn't match up with mine and agree to disagree. As part of my consulting, I design, set up and help debug hardware and know how it works. Is it your contention that all 24Fps film transfers pulled-down to 60 fields-per-second periodically have bogus frames made from two different film frames?And for experience... I also designed the post house that I work for now, I work with Sony, Panasonic, and TeraNex to de-bug and test their products all the time and I have been involved with a few beta sites for a range of products.

If you would ever like to see a mixed video frame I would be more than happy to have you come by my work and I would show you that in any 1080i/59.94 tape that was film based. We have everything from VHS to a new DI center. I would love to show you around the plant. Would anyone else like a tour? If so send me a PM.

In the case of 720p/59.94, each frame of the film is repeated either 2 or 3 times depending on the 3:2 cadence. In the 10+ years of running HD I have only made a handful of 720p/59.94 tapes.Hmmm. That's why I tend to notice jerky video on ABC (Boston Legal was particularly bad last night) more than some of the other networks, I guess. CSI Miami has great video, but some very jerky pan shots and it is 1080i, but I can't recall too many on CBS or NBC.