FORUMS: list search recent posts

720p 29.97 field dominance

COW Forums : AJA Video Systems

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Oliver Peters
720p 29.97 field dominance
on Dec 27, 2008 at 1:59:56 am

One of my clients shoots P2 as 720p at native 29.97 fps. This creates a timeline match in FCP with a timebase of 29.97 (progressive). 29.97/30fps in the 720p format is technically invalid, since the broadcast and display settings really are only 59.94p. So, the Kona cards play this with a 2:2 cadence (effectively doubling each frame). This has been mentioned in other threads.

It appears, however, that this output on both the HD feed and an SD downconversion to 525i is Field 2 dominant. Therefore, all edit points (normally Field 1 in FCP on other 29.97 formats) become split-field frames in the NTSC downconversion. Anyone know if this can be changed or reset in some way to force Field 1 Dominance?

System: FCP 6.0.3, Kona 2 card on a G5 (OS 10.4.11).

Thanks,
Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Kevin Wild
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 8:02:55 am

Man, this stuff has given us a HUGE headache. I posted a while back on mixing all of these framerates, sizes, etc. Not much feedback...

The best way we've found is to bring it into After Effects, interpret footage to LOWER (SD) and size to SD from HD in AE...and then render out.

We have not found the Final Cut Pro "shift field" OR Kona LHe output or Nattress Plugin in FCP to be good enough at dealing with these frame rate and downconvert issues. We have had all sorts of problems mostly with graphics and text not downconverting well or having that field switch issue (HD=upper, SD=lower).

Hope this helps. The After Effects field convert works awesome.

Kevin



Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 2:31:42 pm

[Kevin Wild] "We have not found the Final Cut Pro "shift field" OR Kona LHe output or Nattress Plugin in FCP to be good enough at dealing with these frame rate and downconvert issues. We have had all sorts of problems mostly with graphics and text not downconverting well or having that field switch issue (HD=upper, SD=lower)."

Kevin,

The issues you are talking about above are Field Order, not Field Dominance. Apple improperly labels this in most of their FCS apps. In this case - Field Dominance - all 29.97 fps edits that start on Field 1 of an FCP 29.97 video stream are shifted to start on Field 2 in the SD downconversion - or to the even frames of a 59.94 fps progressive video stream.

Yes, you are right that AE does a better job of rendering and I have used that in the past. In this case, I'm trying to stick with the real-time benefits of the hardware. And yes, I haven't received much feedback either on several different forums.

Cheers,
Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index


Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 8:35:12 pm

[Oliver Peters] "And yes, I haven't received much feedback either on several different forums. "

Hey Oliver. I have never seen this before and I want to test it out. Have you written or called AJA about this? This is not something that can be 'fixed' in the KOna Control Panel if it is indeed a problem.

If you're working 720p30, you should be getting 30psf SD.

You've tried the obvious uninstall and reinstall the software and make sure you have the latest firmware on the card, right?

JEremy


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 10:38:45 pm

[Jeremy Garchow] "Hey Oliver. I have never seen this before and I want to test it out. Have you written or called AJA about this? This is not something that can be 'fixed' in the KOna Control Panel if it is indeed a problem. "

Jeremy,

Since this isn't my own system and I know AJA monitors this forum, I've started here among other places. I will check directly with AJA to see what they say. Thanks.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 10:45:02 pm

[Oliver Peters] "I've started here among other places. I will check directly with AJA to see what they say. Thanks."

No problem. It's your best bet. PIcking up the phone and calling them will be the fastest as it's very rare that AJA does technical support on this forum, all though they do listen :)

Have you done the software/firmware updates for the Kona? At least a driver uninstall and reinstall?

ALso, what deck are you going to? Have you tried a different deck or made sure the settings are proper on the deck?


Return to posts index


Oliver Peters
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 10:57:06 pm

[Jeremy Garchow] "Have you done the software/firmware updates for the Kona? At least a driver uninstall and reinstall? "

Not my system.

[Jeremy Garchow] "ALso, what deck are you going to? Have you tried a different deck or made sure the settings are proper on the deck?"

Tried both Panasonic 1200A and Sony Digibeta. The Panasonic 1200A is getting the 59.94 fps signal over HD-SDI and all edits from the 29.97p timeline occur on only even frames of the 59.94 stream. On the downconverted SD signal from either the 1200A or the Kona card directly, these same 29.97p edit points occur starting on Field 2 of a frame (Field Dominance).

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 11:41:48 pm

[Oliver Peters] "Not my system. "

I don't understand that logic. Not your system which means.....What? You can't touch it or trouble shoot it? If you're driving a friends car and you get a flat, do you leave it at the curb while he comes and fixes it and you hop on the bus?


[Oliver Peters] "Tried both Panasonic 1200A and Sony Digibeta. "

That is very interesting. It's been a good while since I have done a 30p project, but I certainly don't remember this being a problem, although I can't even remember if I checked it in SD.

Have you tried a firewire layoff just to test it?


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 11:50:34 pm

[Jeremy Garchow] "I don't understand that logic. Not your system which means.....What? You can't touch it or trouble shoot it?"

I'm there as a freelance editor. They have staff post supervisor and pay Bob Z to do installs when it exceeds what they can do in-house. So no, I don't proceed to install drivers, plug-ins or other software, unless I'm asked by the system owner. I think there's a pretty long list of disasters here on the COW caused by editors arbitrarily installing software and accepting automatic updates ;-)

Right now my effort extends to seeing whether or not there is even an actual problem first. After all, I'm being paid to edit content, not troubleshoot software, especially when there might not actually be a problem.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index


Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 29, 2008 at 12:06:45 am

Hmm...well it's tough to try and trouble shoot when you can't do anything about it.

I would at the very least unload then reload the drivers and make sure you firmware is up to date with the latest drivers.



Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 29, 2008 at 12:22:30 am

When I hear from AJA, then I'll see what their post supervisor wants to do. At the moment, this is all happening over the holidays and I'm the only one there. Plus it's not holding me up. It's mainly a nuisance when I bounce the footage out to SD and then ingest it back in as SD.

As far as your earlier question - yes, the SD is effectively 30PsF. It's just that now, all progressive frames start on Field 2 of any interlaced frame. So all PsF frames are now F2/F1, etc. as you step through the media file or jog through it on a VTR. This means that if you attempt to make a trim on a previously-edited cut, you have to trim in an extra frame.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 29, 2008 at 12:29:17 am

Yeah, I understand the problem. I just don't remember seeing it is all.

Sounds like a firmware issue to me, if I had to guess.

Jeremy


Return to posts index


Matt Larson
Re: 720p 29.97 field dominance
on Dec 29, 2008 at 9:26:40 pm

[Oliver Peters] "It's mainly a nuisance when I bounce the footage out to SD and then ingest it back in as SD."

I'll be interested to hear what you find out also. I ran into this myself about a year ago, and when I contacted AJA about, they didn't really have an answer for me. Maybe you will have better luck explaining the situation. I saw it on a Kona 2 and 3 on separate machines, so I don't think it's a driver issue.

(In all other cases though, AJA support has been great)




2 x3Ghz Quad MacPro
9 GB RAM
Mac OS X 10.5.5
QT 7.5.5
FCP 6.0.5
AJA Kona 3 (6.0.1 drivers)
G-Speed XL 12 RAID


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 12:36:32 am

[Matt Larson] "I'll be interested to hear what you find out also. I ran into this myself about a year ago, and when I contacted AJA about, they didn't really have an answer for me. "

Matt,

I have been in contact with AJA and have a few things to test. So it will probably be a few days. The preliminary info I'm getting from AJA is that the frame-doubling occurs in FCP. FCP takes the 29.97p timeline, adds the 2:2 cadence and "presents" it to the Kona card as "standard" 720p/59.94 video. BTW - this is apparently also what Avid does in their hardware. So in each case, the NLE software is doing the work. More to come.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

walter biscardi
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 3:54:42 pm

[Oliver Peters] "One of my clients shoots P2 as 720p at native 29.97 fps"

There's no such thing. 720 is 59.94, 30 or 24 unless you're shooting with the Varicam where you can set some other frame rates.



[Oliver Peters] "It appears, however, that this output on both the HD feed and an SD downconversion to 525i is Field 2 dominant. Therefore, all edit points (normally Field 1 in FCP on other 29.97 formats) become split-field frames in the NTSC downconversion. Anyone know if this can be changed or reset in some way to force Field 1 Dominance? "

Lower Field first is correct for the NTSC Downconversion. If it's upper field first you'll have a reverse field interlacing issue.



Walter Biscardi, Jr.
Biscardi Creative Media
HD and SD Production for Broadcast and Independent Productions.

Read my Blog!

STOP STARING AND START GRADING WITH APPLE COLOR Apple Color Training DVD available now!


Return to posts index


Oliver Peters
Re: 720p 29.97 field dominance
on Dec 28, 2008 at 10:52:38 pm

[walter biscardi] "There's no such thing. 720 is 59.94, 30 or 24 unless you're shooting with the Varicam where you can set some other frame rates. "

Walter,

The P2 cameras can record in native and non-native frame rates. Check the record menus for both "P" and "PN" frame rates. So you can record 29.97fps or 23.98fps over 59.94fps, just like in VariCam. That would be "P". However, you can also shoot native frame rates. This means the card is ONLY capturing 23.98 or 29.97 progressive frames. That would be a "PN" setting. When you drop this media into an FCP timeline, the timeline will set itself up with a timebase of 29.97 (if you set the camera to 720 29.97 PN) and "none" for its field setting. This is valid inside the software only. You are right that the format is 720p 59.94, which is why the Kona card doubles the frames. That's not the issue. The problem is that it starts the frames on the even frame of a 60 frame stream or Field 2 of the downconverted, interlaced 525i signal.

[walter biscardi] "Lower Field first is correct for the NTSC Downconversion. If it's upper field first you'll have a reverse field interlacing issue. "

That's not what I'm talking about. Apple is totally wrong in how they label this in the FCS apps. They are referring to Field ORDER as Field Dominance in their menu settings. So yes, Field ORDER is lower field first in NTSC, but Field Dominance of the edits can be either Field 1 or Field 2 in an interlaced signal. Most NLEs and VTRs are Field 1 dominant, but you will often find live video switcher feeds that are Field 2 dominant.

Thanks. I'll check with AJA and try to post the info when I find out.

- Oliver


Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Kevin Wild
Re: 720p 29.97 field dominance
on Dec 29, 2008 at 4:38:04 am

What is the difference between field order vs field dominance? I guess I don't understand this myself, but I have had the exact issues you are having (We were mixing ftg from 59.94, 23.98 and SD 29.97). I listed them below in another post. It was a pulldown issue, I'm sure. The way we fixed it was to trim an additional frame to get rid of a very sleight flash type frame that was happening due to an edit starting on the wrong field, I suppose.

Please let us know what AJA says if/when you find something out.

Thanks.

Kevin



Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 12:31:18 am

[Kevin Wild] "What is the difference between field order vs field dominance?"

Kevin,

Field order and field dominance relate to interlaced frames. Field order is whether the scan lines of a complete frame start on the odd (Upper) numbered or even (Lower) numbered lines when displaying an interlaced frame. Or from Adobe:

"The field order for an interlaced video footage item determines the order in which the two video fields (upper and lower) are displayed. A system that draws the upper lines before the lower lines is called upper-field first; one that draws the lower lines before the upper lines is called lower-field first. Many standard-definition formats (such as DV NTSC) are lower-field first, whereas many high-definition formats (such as 1080i DVCPRO HD) are upper-field first."

Field dominance is whether an edit point or a cut occurs on Field 2 or Field 1 of a frame. From Wikipedia:

http://en.wikipedia.org/wiki/Field_dominance

Basically Field Order is a dislay issue whereas Field Dominance is an edit issue.

Although I point out the FCP discrepancy, I should add that many manufacturers use these two terms interchangeably. That's wrong, but they all do it - hence, all the confusion. In the case of all NLEs that I know of, the edits are all Field 1 dominant, regardless of whether the Field Order is upper, lower or none (progressive). This means when you park on an edit point, you are on the complete frame of that timecode value. If you re-edit material that was already edited and is Field 2 dominant, any given cut with be a split-field frame showing one field of the outgoing frame mixed with a field on the incoming frame.

Sorry for the lengthy explanation. Hope that helps.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Tom Brooks
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 11:36:46 am

How is field dominance marked or flagged in a digital video signal? And obviously it has no meaning in progressive formats, so in that case, is there a marker that identifies a frame as the first frame in a pair when you are in 720p29.97? In other words, what keeps 720p29.97 from being plain old 720p59.94? If there is such a marker, wouldn't it be in a sort of conflict with the 720p59.94 time code, which would also define the start and end of each frame?


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 1:49:31 pm

[Tom Brooks] "How is field dominance marked or flagged in a digital video signal?"

Through TC and the inherent nature of the device itself. Jog through a VTR and you can see where the cuts occur as compared to when TC changes. All devices have a field dominance setting. VTRs and switchers have actual menu settings to set them to be F1, F2 or F1/F2 (driven by external commands). I don't know on NLEs, because AFAIK, you cannot change it on them. In a VTR, if the switch is set to F1 Dom, then an insert/assemble edit made on the VTR itself cannot occur on F2.

[Tom Brooks] "In other words, what keeps 720p29.97 from being plain old 720p59.94?"

720p/29.97 as a format technically doesn't exist, as Walter has pointed out. It DOES exist, however, in P2 media (and JVC's HDV) and various software apps, including FCP. When you drop a 29.97PN (native) clip on a blank FCP timeline, the settings are changed to a 29.97 - NOT 59.94 - timebase. All edits and TC are based only on 30 intervals/sec. When you play that through hardware - like the Kona card - that expects to get 720p/59.94, the software doubles the frames in the stream. An exported QT would only be 29.97.

This issue I'm hitting is that in order to turn 30 frames into 60 frames or 60 fields, something has to determine whether the first frame starts on timecode Frame 1 at Field 1 or 2 (interlaced). Since frames are doubled in progressive, the 30 frames of a 29.97 timeline become 60 frames of a 59.94p timeline. In effect, each timecode value in the 29.97 timeline corresponds to every-other timecode value in the resulting 59.94p video stream. As with interlaced, the software must determine whether this starts on even or odd frames. In other words, does Frame 1, 2, 3, etc. (29.97) = Frame 1, 3, 5, etc. - OR - Frame 0, 2, 4, etc. (59.94)? And can this be changed or adjusted?

Does you head hurt yet? ;-)

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Tom Brooks
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 1:57:16 pm

[Oliver Peters] "Does you head hurt yet? ;-) "

Yes it does, but I look for things like this. Thx.


Return to posts index

Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 4:57:39 pm

Oliver's explanation is very good. I think that the use of 30 fps based tc has to be with dragging around the NTSC legacy for broadcast. If all of a sudden 60fps tc was unleashed on current broadcast infrastructures a lot of devices that would be totally confused so sending 30fps tc was probably a no brainer. I'm not even sure if there's a SMPTE standard for 60fps based tc. My guess is probably not, and if there is, it's probably not used too much today.

In FCP, you can change the tc from base 30 to base 60 fps. That way you can see exactly which frames get removed when working with 60fps tape base captures and removing VFR data when doing so.

If you have an ioHD you will notice that when playing out 60fps material and you have the front panel set to time code, the ioHD will display a frame number, then the next frame will have an asterisk indicating that it's a different frame (but same tc number). So if you start on frame 01, the next frame according to the ioHD will be :01*. Basically, every frame number has two frames assigned to it.


The beauty of 60fps workflows is that it fits exactly in to 30fps based material. So if you're shooting 30pN material, this gets translated to 30psf in 60i formats (being 1080i or SD). Shooting 24pN results in 24p material with 3:2 pulldown once going to a 60i based format. Shooting 60fps will turn in to plain old 60i material in 1080 and Sd formats. I wrote something a while ago that explains that concept in more detail.

http://forums.creativecow.net/readpost/98/869490

Hope this helps further your understanding of 720p.

In Oliver's case it seems that there SHOULD be a way to change the field dom, but perhaps there's not. Oliver, have you tried a firewire layoff to the 1400 yet as a test? What happens if you shot your 30pN timeline in a 60p timeline, render in the pulldown and layoff (FCP actually handles the proper pulldown in a 720p60 timeline). Does it happen then? What happens if you delay your in point by one frame?

Jeremy


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Dec 30, 2008 at 11:55:04 pm

[Jeremy Garchow] "have you tried a firewire layoff to the 1400 yet as a test? What happens if you shot your 30pN timeline in a 60p timeline, render in the pulldown and layoff (FCP actually handles the proper pulldown in a 720p60 timeline). Does it happen then? What happens if you delay your in point by one frame? "

Jeremy,

It will be next week before I get back to that shop. The director/DP is going to shoot some 59.94 footage with his P2 camera for me to run a few tests. I am going to test 59.94 behavior with 59.94 media as well as 29.97PN media if rendered in a 59.94 timeline. AFAIK, the 1200A that I have access to only has SDI/HD-SDI, not FW. After I run those tests, I'll be back and post the results. Thanks.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Jeremy Garchow
Re: 720p 29.97 field dominance
on Dec 31, 2008 at 3:39:50 am

[Oliver Peters] "It will be next week before I get back to that shop."

Me too. I would like to check it out for myself as well.

[Oliver Peters] "AFAIK, the 1200A that I have access to only has SDI/HD-SDI, not FW. "

Got ya. I thought you said 1400 by accident. It's been a while, but I thought the 1200 came with firewire only and you had to outfit it with all the other ports (HD SDI, etc), but maybe I have that backwards and the firewire was an add-on. I forget. At any rate, if it does have firewire, I'd be curious to see if that will fix your problem.

[Oliver Peters] "After I run those tests, I'll be back and post the results."

Please do. I am curious as to what you'll find.


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Jan 9, 2009 at 1:03:37 am

An update... Drivers were updated and I ran a series of tests before and after this update, as well as on two different machines. I ran tests with 29.97pn media in a 29.97p timeline (FCP adds 2:2 cadence for a 59.94p signal), as well as 29.97pn media rendered in a 59.94p timeline. These various tests were output to Digibeta. All results have been the same.

Field dominance of cuts recorded onto tape appears to be random. You can insert the same sequence multiple times and get different results. It appears that FCP "buffers" the first clip of video by between 0 and 2 additional interlaced fields, so the timeline on tape can be either accurate (F1 dominant), F2 dominant or 1 full frame delayed. This appears to be a result of the additional "length" on the first video clip. All subsequent clips are the proper length, so if the first clip is right, all cuts will fall on Field 1. If the first clip is 1 field longer, all subsequent cuts will fall on Field 2.

AJA support is aware of the issue and looking into it, but I'd bet the problem is with FCP or QT.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

Matt Larson
Re: 720p 29.97 field dominance
on Jan 9, 2009 at 3:43:26 pm

So, when I lay off a 720p 29.97 sequence (which goes to tape at 59.94) to the 1400 deck, then re-capture that tape back in FCP with an NTSC easy Setup, this is the same thing I was seeing. (The randomness made me think I was doing something wrong in my tests.)

If I do this and I determine the cuts are on the wrong field, is there a way to switch that in FCP so the cuts are on Field 1?




2 x3Ghz Quad MacPro
9 GB RAM
Mac OS X 10.5.5
QT 7.5.5
FCP 6.0.5
AJA Kona 3 (6.0.1 drivers)
G-Speed XL 12 RAID


Return to posts index

Oliver Peters
Re: 720p 29.97 field dominance
on Jan 11, 2009 at 2:36:47 am

[Matt Larson] "If I do this and I determine the cuts are on the wrong field, is there a way to switch that in FCP so the cuts are on Field 1?"

No.

- Oliver

Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]