GH2 HDMI Discoveries
- Thread starter Ralph B
- Start date
Barry_Green
Moderator
I searched through the manual and didn't find any reference to HDMI becoming disabled in 720p mode or in ETC mode. Maybe it's entirely due to the device on the other end? But even when I forced HDMI to be only 1080i, or only 720p, it still disabled it when connected to the Nano. Which is odd, because the Nano can record 720p, so it should be communicating back to the camera to tell it it's a 720p-capable device.
Oh well, not too interested in spending much more time on this, just throwing it out there for those who think they may have a way to work around the 24p-within-60i-converting-to-59.94i-converting-down-to-23.976p issue.
Oh well, not too interested in spending much more time on this, just throwing it out there for those who think they may have a way to work around the 24p-within-60i-converting-to-59.94i-converting-down-to-23.976p issue.
Barry_Green
Moderator
Why not make it 55.000, that way we can finally meet halfway with PAL and get one overall worldwide standard once and for all? Of course, that means problems for film, but ... eh, we'll worry about that later.
farfromsubtle
Well-known member
A little taken aback to see Barry say flat out "DON'T" to HDMI capture aspirations after all that work in the other thread to prove that 24p IS recoverable from the weirdness in the signal. If the subsampling is 4:2:2 why would this not be worth the effort? And by 'effort' I mean the development of an app/script or addition to an existing app that would deal with it automatically.
What I meant was the digitized footage is designating itself at a slower fps than is real.
You can drop a 24 P clip that's 24 frames long on a 29.97 timeline and it should take up a full second. And if you drop a 60p clip that's 60 frames long, on a 29.97 timeline, it too will only take up one second. An editing system should interpolate based on FPS designation for each clip (compositing packages are different, so I should have asked what you were using, I guess.)
So two clips of different frame-rates should be exactly the same length on a TIME line - assuming their fps designation is accurate.
The fact that the digitized piece ends up longer means it's designated FPS is wrong, in this case, slower, which I found interesting. It confirms the idea of a 60 hz signal being jammed into a 59.94 signal.
Umm - where's the greenscreen footage? Heh.
I don't know about the others, but I think the GH2 color over HDMI is a mess. The AF100 has no problems sending out a legitimate 4:2:2 signal - GH2 looks nothing like it.
Would still love to try pulling a key with both types of footage, though - if nothing else than to put it all to bed.
You going to have time to give it a shot soon? I'm about to move over to that Canon 550D thread (JK)
rogar-Dum(DBG):
Now you're starting to sound like Sony.
You have to wonder: Some person decided on 60fps vs 23,976x2 and most likely for a reason.
I just feel reminded of those newer LCDs that have "True motion" modes or however they call it, where frames are being doubled to make footage look like "reality" meaning what we try to avoid: video look... WTF?
To me it doesn´t sound logical, that basically every demand from indie filmers gets met by Panasonic but they stop at this last feature by accident. Sad though that they really seem to do stop right there, where the GH2 could become the next RED One-ish piece.
I'm still confused as to whether the GH2 HDMI outputs at 60i or 59.94i. I keep reading reports indicating one or the other. And if the GH2 outputs exactly 60i, how does that work? I thought 29.97 and 59.94 were originally derived and required to produce color TV. Or are we only talking broadcast?
In any case, I wonder if the external recorder or FCP would be capable of converting a 60i into 59.94? I would think FCP could.
I hold in my hands three great BBC Blu-ray videos: Galapagos, Ganges, & Wild China all sporting "1080i." Not to say they were shot that way. Governments like Canada insist upon 1080i. HDCAM mainly shot 1080i 59.94. But your point may be that no one literally shoots 60i, not that anything other than 24P is trash.
If you are using an external HDMI recorder, one would also use its superior audio recording capablilities.
RBG
In any case, I wonder if the external recorder or FCP would be capable of converting a 60i into 59.94? I would think FCP could.
I hold in my hands three great BBC Blu-ray videos: Galapagos, Ganges, & Wild China all sporting "1080i." Not to say they were shot that way. Governments like Canada insist upon 1080i. HDCAM mainly shot 1080i 59.94. But your point may be that no one literally shoots 60i, not that anything other than 24P is trash.
If you are using an external HDMI recorder, one would also use its superior audio recording capablilities.
RBG
Barry_Green
Moderator
According to the Nano, it's 60.000. And my Dell 2405 won't even display it live. It'll show playback, but it won't show the live stream.
Well that's the point. I'm saying the only thing you can effectively record now, and use, is shooting live 1080i. And even then you'll have to convert it to 59.94i for distribution. Which means duplicated fields that will have to be removed.
And which superior audio recording capabilities would those be, on the Nano? The 3.5" stereo mic jack with no volume controls?
Peter J. DeCrescenzo
Veteran
Happy new year, Barry!
RGB may have been referring to the KiPro Mini's XLR inputs, but as you've said, external HDMI recording is currently probably not worth the trouble, unfortunately. As we know the GH2 doesn't output audio in its HDMI stream. Cheers.
Heaven knows we should all aspire to shoot Panaflex Hollywood features. Well then we can only hope the ancient and hallowed 24P will submit to a dignified death (as 25P seems to have on the GH2) as technical capabilities permit higher frame rates with 1080P.
Convergence Presser:
"The new firmware brings 8-channel audio support for 5.1 and 7.1 audio programs as well as Holophone (5.1) recording."
"http://tinyurl.com/2838zjl
Holophone 5.1 surround sound:
http://tinyurl.com/2gyqt6v
I see the inconspicuous Ninja will have analog gain adjustments.
The KiPro Mini has 24-bit A/D, 2 channel balanced (2 XLR), 48kHz
Input level: Line, Mic, Mic + phantom 48Vdc
These devices must have better metering and electronics than the GH2 that primarily relies on an in-camera microphone. Certainly they all have headphone monitoring. But then again, I think I can recall being wrong once before.
RBG
PROGRESS REPORT
I have successfully identified the cadence of duplicate fields in the HDMI 60i output. (and it is true 60i, not 59.94i)
It goes like this: duplicate field - 1001 clean fields - duplicate field - 1002 clean fields - duplicate field - 1003 clean fields, and then it repeats.
I hope others will verify this. One thing I have to stress is that your capture device/software MUST be set to record at 30P, not 29.97. I tried it both ways. At 29.97 the duplicate fields are scattered anywhere between 1001 and 1006 clean fields and with no repeating pattern.
For those that are interested, here's how I did it. In After Effects, I created a movie of a vertical bar moving across the screen from left to right, at a medium speed. The movie loops and run for three minutes. I rendered it at 60P. I played it on the computer monitor and shot it with the GH2 going into an Intensity Pro board. The GH2 was set to 1080 60i. After capture, I ran the Intensity movie through AVIsynth to extract the fields and thereby creating a 60P movie.
Here is the AVIsynth script:
aviSource("Your_movie.avi")
AssumeTFF()
SeparateFields()
I then took the 60P movie into Premiere to examine it. Having it in an editing program made it easier to navigate.
I have successfully identified the cadence of duplicate fields in the HDMI 60i output. (and it is true 60i, not 59.94i)
It goes like this: duplicate field - 1001 clean fields - duplicate field - 1002 clean fields - duplicate field - 1003 clean fields, and then it repeats.
I hope others will verify this. One thing I have to stress is that your capture device/software MUST be set to record at 30P, not 29.97. I tried it both ways. At 29.97 the duplicate fields are scattered anywhere between 1001 and 1006 clean fields and with no repeating pattern.
For those that are interested, here's how I did it. In After Effects, I created a movie of a vertical bar moving across the screen from left to right, at a medium speed. The movie loops and run for three minutes. I rendered it at 60P. I played it on the computer monitor and shot it with the GH2 going into an Intensity Pro board. The GH2 was set to 1080 60i. After capture, I ran the Intensity movie through AVIsynth to extract the fields and thereby creating a 60P movie.
Here is the AVIsynth script:
aviSource("Your_movie.avi")
AssumeTFF()
SeparateFields()
I then took the 60P movie into Premiere to examine it. Having it in an editing program made it easier to navigate.
Vitaliy Kiselev
Carbonite Member
60/59,94=1.001
So, your report suggest that real sensor framerate is 59,94 and HDMI framerate is 60.00.
I suggest you to not to use computer monitor, btw. Use some real moving vertical bar
Other useful suggestion for tests - record clip and play it back through camera.
May be in this case it'll use proper 59,94 mode.
So, your report suggest that real sensor framerate is 59,94 and HDMI framerate is 60.00.
I suggest you to not to use computer monitor, btw. Use some real moving vertical bar
Other useful suggestion for tests - record clip and play it back through camera.
May be in this case it'll use proper 59,94 mode.
Until proven wrong, I still want to be optimistic here
- When using 1080i, does the PAL version of GH2 output PAL 50i over HDMI? In that case, my PAL version shouldn't be affected by the 59.94i vs 60i issue, right?
- Regarding clipping issues, tbruegg posted his findings in this thread when capturing with Blackmagic decklink intensity. It seemes that when using a smallhd dp6 monitor, he found that GH2 outputs RGB while not recording, and YUV while recording. And if I understood correctly, he also found that the Blackmagic mjpg capture codec didn't work correctly with the gh2 signal, but changing it to something else could solve the clipping issue for capturing (hopefully tbruegg can confirm).
Note: I might have misunderstood some technical aspects here, sorry if I'm reporting wrong.
- When using 1080i, does the PAL version of GH2 output PAL 50i over HDMI? In that case, my PAL version shouldn't be affected by the 59.94i vs 60i issue, right?
- Regarding clipping issues, tbruegg posted his findings in this thread when capturing with Blackmagic decklink intensity. It seemes that when using a smallhd dp6 monitor, he found that GH2 outputs RGB while not recording, and YUV while recording. And if I understood correctly, he also found that the Blackmagic mjpg capture codec didn't work correctly with the gh2 signal, but changing it to something else could solve the clipping issue for capturing (hopefully tbruegg can confirm).
Note: I might have misunderstood some technical aspects here, sorry if I'm reporting wrong.
Last edited:
Wow, my heads bogged down with all the variables with this GH2 HDMI capture stuff. If someone would save me from researching through all the posts again... whats the best GH2 NTSC 24p Mode (60i capture) native HDMI captured sample footage available for us to play around with? I.E. footage that was captured and encoded most correctly to the best of our current knowledge. I'm interesting in doing some AVISynth fiddling myself. Not upgrading to a GH2 myself just yet, though I'm iching to make us of my Intensity Pro card.