Nikon D800 Live HDMI Stream

Stephenh

Well-known member
This is an update regarding the Live Output of the D800. All progressive formats whether 24, 25 or 30p or 60 and 50p are wrapped in a 60i stream for TV compatibility even though the footage is progressive. The Atomos Ninja is able to detect the 24, 25, 30p or 60p, 50p inside the 60i wrapper and remove it leaving the progressive image from the sensor untouched. I believe the new 2.1 firmware is required to do this.
 
You neglected to provide any context for your "update". I'm guessing that you are describing the Atomos Ninja product capabilities rather than the Nikon D4/D800 capabilities.

The Live Output of the D4/D800 will be 1080p24/25/30 in 8-bit 4:2:2 YCbCr format (in HDMI terms, this would be 24-bit 4:2:2 YCbCr format) when connected to appropriate devices which support these formats. Note that the D4/D800 does not include an option for 1080p50/60 video acquisition or display (much less embedded in a 1080i50/60 wrapper).

This topic has been well discussed, with a great number of test report contributions, in other forum threads. The Atomos Ninja is mentioned prominently, several times. Give the "Official-HDMI-1080-IS-INTERLACED-ONLY" thread a nice, slow reading (the thread title is misleading).

-- Bob Elkind
 
Last edited:
You neglected to provide any context for your "update". I'm guessing that you are describing the Atomos Ninja product capabilities rather than the Nikon D4/D800 capabilities.

Note that the D4/D800 does not include an option for 1080p50/60 video acquisition or display (much less embedded in a 1080i50/60 wrapper).[/I]


-- Bob Elkind

You misunderstand me and I should have made myself a little clear. I just assumed that everyone would pickup on the fact the 50/60p would relate to 720 format and not 1080p. Are we talking about the same thing here. I am referring to the interlaced wrapper that Nikon put the progressive stream into to make it universally compliant for existing Television.

Steve.
 
What Bob is pointing out is that there has been a lot of discussion on this topic in another thread. When you say there has been an update, its kind of hard to know wether you are aware of the reports that have been posted in this thread.

Take for instance a look at this post http://www.dvxuser.com/V6/showthrea...D-ONLY&p=1986110019&viewfull=1#post1986110019

From this you will see that the D800 is capable of outputting 1080p24 without a 60i wrapper, given the right set of circumstances. The Atmos Ninja seems to want the D800 HDMI stream in a 60i wrapper, but that is a different story more to do with the Atmos Ninjas capabilities. So yes, we are a little confused when you say there has been an update :)
 
Now I get it. :thumbsup: Thanks dfreerider, and appreciation to eteam(Bob). The penny just dropped. it is indeed a Ninja requirement for 60i wrapper. I will follow up both your links for further clarification.

Steve.
 
Hi, Just had a conversation with Jeromy (Atomos) about the live stream output of the D800. It seams that when you take the cards out of the camera, it switches the format of the data coming from the sensor to what ever you set in the Movie Settings menu. I am not going to go in detail as i am not qualified, but I have asked Jeromy to post some info about the process with the Atomos Ninja in relation to the Nikon D4/D800 cameras.

Another bit if info that may be of interest is the Live Stream output data rate is 1.3Gbits/sec before conversion by the recording device.
 
Last edited:
Hi, Just had a conversation with Jeromy (Atomos) about the live stream output of the D800. It seams that when you take the cards out of the camera, it switches the format of the data coming from the sensor to what ever you set in the Movie Settings menu. I am not going to go in detail as i am not qualified, but I have asked Jeromy to post some info about the process with the Atomos Ninja in relation to the Nikon D4/D800 cameras.
You are catching up to what has been discussed in the other D4/D800 threads. It would be good to hear official word from Atomos on their development efforts with respect to the D4/D800. If Atomos relies on a 1080i wrapper for PsF video (using pulldown, essentially), they will do themselves a favour by speaking to the issue of reverse pulldown back to progressive format.

Another bit if info that may be of interest is the Live Stream output data rate is 1.3Gbits/sec before conversion by the recording device.
This is not remarkable for any HDMI interface, either 1080p24/25/30 or 720p50/60.

-- Bob Elkind
 
This is not remarkable for any HDMI interface, either 1080p24/25/30 or 720p50/60.

-- Bob Elkind

Yes! I just thought that I would post it for general information for other forum members.

Continuing on from my previous post it appears that Nikon are wrapping the progressive stream in a 60i stream. Now don't everyone jump on me at once. To give an example 1080/24p, I am told that Nikon are duplication every fourth frame (e.g. 1,2,3,4,4A) which when multiplied out becomes 30 and it is doubled to give the 60i stream. I am not technical enough to get into a discussion with my more learned forum colleagues. Jeromy (Atomos) has promised to post a much better detailed explanation as time permits. I do know that Atomos has been in close communication with Nikon with regard to the external HDMI protocol.

Hopefully we will not have to wait to long for this update.
 
Continuing on from my previous post it appears that Nikon are wrapping the progressive stream in a 60i stream.

As far as I understand it, ATOMOS requires you to set the D800 to wrap the progressive stream in a 60i stream,
so that they can record it. At least that is what they say in firmware release notes. It is unknown so far why they can not record native progressive signal from the D800, which it seems to be capable of putting out when you remove both flashcards from it. I have asked ATOMOS support about the reason for this, but I am still waiting for an answer.

Martin
 
As far as I understand it, ATOMOS requires you to set the D800 to wrap the progressive stream in a 60i stream,
so that they can record it. At least that is what they say in firmware release notes. It is unknown so far why they can not record native progressive signal from the D800, which it seems to be capable of putting out when you remove both flashcards from it. I have asked ATOMOS support about the reason for this, but I am still waiting for an answer.

Martin

The HDMI settings menu (or as far as I can tell the manual) does not mention support for 1080p, so the 1080p output is kind of undocumented. For those jumping into this thread, please read http://www.dvxuser.com/V6/showthread...post1986110019.

There are just so many claims that it does not support 1080p24 that something in the testing methods must be a little off. Or the handshake between the capture device and camera is just not working or something. Why Atmos would update their firmware to support a Nikon 60i wrapper format, instead of focusing on getting the 1080p24 output to record directly is confusing.

Hope they get back to you with an answer soon.
 
Last edited:
Continuing on from my previous post it appears that Nikon are wrapping the progressive stream in a 60i stream.
Freerider's post is an excellent followup on this point.

I am told that Nikon are duplication every fourth frame (e.g. 1,2,3,4,4A) which when multiplied out becomes 30 and it is doubled to give the 60i stream.
In my posts, I refer to this sequence as "2:2:2:4 pulldown" -- which should be distinguished from "3:2 pulldown", a popular alternate implementation of much the same thing.

... Jeromy (Atomos) has promised to post a much better detailed explanation as time permits. I do know that Atomos has been in close communication with Nikon with regard to the external HDMI protocol.
I cannot imagine that 1080i plus pulldown is anything more than a self-serving interim explanation -- a placeholder -- until a proper solution is completed and tested. If Atomos and Nikon do their work correctly and completely, this business with 1080i wrappers and pulldown will go away. This is the work that AJA has already undertaken and successfully completed, by their (Jon Thorn) reports in this forum.

To be honest, all Atomos (or anyone else, for that matter) needs to do is: buy an AJA recorder with the updated D4/D800-friendly firmware, and reverse engineer its EDID table. This is simple to do with freely available software tools.

-- Bob Elkind
 
Last edited:
Hi, Just had a conversation with Jeromy (Atomos) about the live stream output of the D800. It seams that when you take the cards out of the camera, it switches the format of the data coming from the sensor to what ever you set in the Movie Settings menu. I am not going to go in detail as i am not qualified, but I have asked Jeromy to post some info about the process with the Atomos Ninja in relation to the Nikon D4/D800 cameras.

You are catching up to what has been discussed in the other D4/D800 threads. It would be good to hear official word from Atomos on their development efforts with respect to the D4/D800. If Atomos relies on a 1080i wrapper for PsF video (using pulldown, essentially), they will do themselves a favour by speaking to the issue of reverse pulldown back to progressive format.

It's my understanding that "PsF" means the fields are just separated, so basically it'll turn 24P sensor data to 48i, 25p sensor data to 50i, and 30p sensor data into 60i. This is how it worked on my AF100 thru SDI anyway. It's also my understanding that the correct definition of "Pulldown" means adding fields. If this is that accurate description for "PsF" and "Pulldown" then the way Bob writes the quote above is confusing.
 
"PsF" means that a single progressive frame is separated across two fields. This is different than interlaced video b/c w/ interlaced each field is from a different instant in time. E.g. each field in 60i is one 1/60th of a second later than the previous field. If you know the cadence used to map the full resolution frames to across the half resolution fields for the entire length of the footabe, then a completely accurate reconstruction of the progressive frames can be done.
 
It just seems Stephenh said Atomos is saying the D800 streams PsF when no card is inserted, and it seems eteam (Bob) is wanting Atomos to explain how they do the pulldown removal, when there is no pulldown added in the first place. Just seems confusing.
 
Last edited:
23.976 embedded in to 60i splits frames into fields and *also* has pulldown. It has to have a pulldown b/c 59.94 is not multiple of 23.976. Bob is talking about what cadence Nikon uses for the pulldown. There is the standard 3:2 pulldown, but others can work as well. This gets rather tedious. But if the wrong pulldown cadence is assumed, losslessly reconstructing the progressive frames from fields is not possible.
 
24PsF in my AF100 was 48i out the SDI stream, no pulldown added. Stephenh seems to say Atomos is saying the D800 is PsF. "it switches the format of the data coming from the sensor to what ever you set in the Movie Settings menu". To me that means PsF, and 24PsF would be 48i.
 
Last edited:
I think I know what the confusion is. I was assuming the menu settings were 24p, 25p or 30p. It is probably 1080i or 720p. Assume make an ass out of you and me :)

Sorry guys.
 
To Brian and BobbyMurcer Fan (note: you don't specify whether you are a fan of the player Bobby Murcer or the team executive Bobby Murcer -- [I'm a recovering New Yorker, born and bred] ):

Yes, indeed I am conflating pulldown, PsF, and interlaced -- because in this context it is entirely appropriate (if imprecise).

I do not have a D4/D80 camera to test, so this is a disclaimer. On to the nitty gritty...

Has anyone seen reports of 1080i48 output from the D4/D800 (I'm directing this question specifically to Brian, in reference to post #12)?
Neither have I.

What we have seen (over and over and over) are reports of

  • 1080p24 in the camera => 1080p24 on the HDMI interface (with fully compatible device attached)
  • 1080p24 in the camera => 1080i60 on the HDMI interface (with a "less" compatible device attached)

We do not see 1080p60 on the HDMI interface, but we can see 1080i60 on the HDMI interface. 1080i60 is interlaced. This is where interlaced enters the discussion.

Question: How is 24 FPS progressive video mapped to 30FPS 1080i60?
Answer: With repeated frames.

Question: What is the difference between 24FPS progressive video converted to 30FPS 1080i60 signal
with PsF and without PsF ?
Answer: None. PsF is a distinction without a difference, in this context.

Question: What do you call the process of converting 24FPS progressive video to 30FPS 1080i60, using repeated fields or frames?

Answer: Pulldown.

Based on anecdotal (sketchy) reports in these forums, it appears that the D4/D800 will output 1080p24 source video as 1080i60 with 2:2:2:4 cadence (every 4th frame is repeated once). Without any special markers or flags or special user intervention (e.g. select from a drop-down menu that the video source is a D4/D800 in LIVE Movie mode), this signal is indistinguishable from:

  • straight 1080i video from a set-top box (or a D4/D800 in video playback mode or picture review mode or recording on media mode)
  • 1080p24 video from a telecine or DVD player applying 3:2 pulldown
  • 1080p24 video from a D4/D800 in Live Movie mode applying 2:2:2:4 pulldown for "non-native" signaling to "less" compatible devices
  • 1080p30 video from a D4/D800 in Live Movie mode applying 2:2 pulldown for "non-native" signaling to "less" compatible devices

What's the point of all this minutiae and syntax-nerding? I'm glad you asked:

Settling for 1080i60 HDMI output from the D4/D800 involves risk, because the signal can be confused with any number of original source formats (and more than one pulldown scheme to reverse). If someone (AJA, Atomos, Sound Devices, et al) tells you:
"Don't worry, we have the smarts to take care of everything for you"
​

your response should be:
"Oh yeah? How? I want specifics, and I want you to tell me how this can't possibly be screwed up".
​

These potential issues and risks are simply swept aside once the attached device is entirely compatible with the D4/D800's native HDMI output formats.

Does this make sense to you?

-- Bob Elkind
 
Last edited:
But if the wrong pulldown cadence is assumed, losslessly reconstructing the progressive frames from fields is not possible.
BINGO! Phew! And even the use of pulldown is an assumption which must be made.

-- Bob Elkind
 
Last edited:
Back
Top