Any GH2 news on the HDMI output front?

Another Nanoflash post. lol. But it can remove 3:2 pulldown on the fly, for those who may be curious.
 
NanoFlash? I'm thinking KiPro Mini myself ($1K less expensive than NF & has 2 XLR inputs), but it's all good.

Wow. Just wow!

Thanks guys. This is looking very promising indeed. Well done!
 
Uh oh.. a new development... I did some more tests and noticed that Final Cut was choking on some of the files (stuttering). I opened them in Quicktime and stepped through frame-by-frame to see the pattern. It changed! This one goes like this (I know this isn't technically how you note cadence, but this is just my way of indication Progressive and Interlaced): PPIIPPPIIPPPIPPPPIIPPIIIPPPIPPPPIIPPIIIPPII and so on... this seems random to me, but maybe there is something to it? Here is the file: http://www.farfromsubtle.com/gh2/pulldowntest/pulldown_weird.mov
 
More tests... it looks like the feed jumps from standard 3:2 pulldown, to whatever it is I posted previously. I did a longer 30 second test, and it starts out with the weird (random to me) pattern, but eventually goes into the very recognizable "PPPIIPPPII". After (I believe exactly 15 seconds) it goes back into the random pattern. I am posting the file so that someone smarter can figure this out. http://www.farfromsubtle.com/gh2/pulldowntest/pulldown_weirdlong.mov (still uploading- done in 13 minutes)
 
Last edited:
That's not a problem.
The pattern changes based on when your digitizing began.
The stream from the camera comes into the Intensity card at a constant pattern, as you see, of PPIIP-PPIIP etc. But, if you think about it, how different is that from PIIPP-PIIPP or IIPPP-IIPPP. Depending on when you start recording, your first frame could be anywhere in that pattern of 3Ps and 2Is. That's why there are 5 letters to choose from when removing pull-down. Until you look at each clip, there's no way to know where the clip began on that cadence. So, the clips will always change on you.
Final Cut is choking because it's not the fastest application for handling changes in frame-rate on the fly. So nothing going wrong here. Just new horizons open up new problems and solutions.
Do you have After Effects or Shake?
I'd be happy to help you remove pull-down with those apps.
 
What about the single and triple interlaced frames in that pattern? That sort of seems out of place... check out the most recent clip I have uploaded.* (sorry I don't have either of those apps, I'm running all apple suite at the moment)

*6 minutes remaining
 
Half way there on download. If you want to try a reverse telecine in Cinema Tools for your pulldowntest_intensity file, use Capture Mode set to Field1- Field 2, and Fields set to BB. On Frames, Conform to 24, and I unticked "Standard Upper/Lower." I haven't used this software in the past, and it doesn't impress me, but at least you can see a solution for that first file you posted.
There is a tutorial on an "automatic" method to this on Ken Stone's website, but from what I can see, they're just saying to conform the footage, which didn't really reverse tele for me at all. Maybe I need to play around with it some more, but for now, I'm not impressed. With AE and other compositing packages, you can scrub through the footage, while selecting from the 5 possible patterns, and very quickly figure out what the right cadence for reverse tele needs to be.

Getting to be a long night :)
 
That worked on the original test file! But it certainly isn't a magic bullet. I tried this on my other tests... some worked, some didn't. :S

Also, on the file you are currently downloading, the pulldown removal starts working at the 17 second mark, then stops working at the 26 second mark. The same spots I found the change in cadence to being with.
 
Last edited:
Just playing with cineform... it seems to be better at detecting the proper pattern... most files start with a few interlaced frames, but then smooth sailing.
 
Yeah, I was looking at it in Shake and noticed several things.

First, you start recording with your menu items on, and when they turned off, the sync was lost for a moment and then recovered.
So I'd like to see another test where your menu items are off before you even start recording.

Second, I noticed that the cadence changed around every 200 frames. Not to the frame, but certainly in that ballpark. And to solve the temporary adjustments in the video, I found myself moving forward to the next letter, so to speak. Also, the break in cadence happens in short pairs.
Clean for 180-200 frames, then break and 4 or 5 more frames, then break again, followed by new correction for cadence and another 180-200 frames. Then break 4-6 frames, then break, then 180-200 frames again.

Now, I guess I should put on a WACKO alert, because it got me thinking weird thoughts that have more to do with me not getting any sleep than anything, but here's my wacko idea of what I HOPE is not going on.

Do you remember this comment pasted on Phil Bloom's blog by a guy who claimed he'd spoken to Panasonic about this?
Here's what he said, "The HDMI live-view output is not suitable for recording use, using nanoFLASH or Ki pro, Ki pro mini, etc..
You will get one freezing image every 17 seconds.
It happens because of the difference between 60 fps, which is the frame rate at which the sensor reads the light, and 59.94 fps, which is the frame rate of the HDMI output.
This is the explanation I had from Panasonic."
The link is here http://philipbloom.net/2010/11/21/gh2/#comments and the commenter was Makoto. (About 3/7 the way down the page).

W-e-e-e-e-ll, 200 frames is around 8 seconds (in 24 fps), half of the 17 he was speaking of. And since we're dealing with interlaced footage, could the GH2 feed be adjusting itself every 8 1/2 seconds, or so, causing the Intensity card a little trouble every 8 seconds or so? We are seeing an added I frame in there aren't we?

See, I told you it was stupid. Just a little too punch drunk-with diet Dr. Pepper.

But, the pattern is definitely not random (like a loose cable would be).

I'm tired - going to bed. But this isn't good. Certainly, this footage could be sewn together by hand - find each interruption and conform just for the range that it works. And I guess that's cool. I mean, how many shots last 8 seconds in a final project? Aside from those that enjoy those long landscape panning panoramas with world-class violinists playing in the background LOL.

Time for someone smarter to jump in and figure out what's going on. I'm beat :P

PS A lot of people discredited what Makoto wrote, and he never responded. It's gonna be funny as hell if he was right!!
 
I am uploading a 5 minute video of my ceiling fan. Hopefully this will be enough reference material for everyone to work with and figure out what is going on. This will take a bit (3.56GB upload) but it should make for an end all test file for exploring the cadence issues.

If there is something funky going on with this, perhaps it will be workable by building a tool specifically for the GH2's unique wrapper? I am sure Cineform or the ClipWrap guys would be up to it, and I am sure we would be up to buying it.
 
Here it is, the 5 minute fan test: http://bit.ly/gQdT2T

I am headed to bed, but this will be complete in about an hour from the time of this post.

(Sorry for the boring subject matter, not much to work with at this time of night. Also, this is shot at 3200 iso, so it's a good stress test on the noise factor!)
 
Last edited:
One last note, I sent this through Cineform then played it back counting to 8... without fail, interlaced frames for about 1 second every 8 seconds. Other than that, Cineform did a great job... there is just something that it trips over, and it certainly is NOT random. I am fairly confident that someone will figure this out and/or sell us a one-click solution.
 
Well, I certainly do appreciate the test footage. There certainly seems to be an issue with the 3:2 pulldown. I wont hold my breath waiting for a firmware update fix.
 
Just playing with cineform... it seems to be better at detecting the proper pattern... most files start with a few interlaced frames, but then smooth sailing.
That's exactly how Cineform does it on all telecined clips. It takes about a second before it starts to remove the puldown...so it's not unusual to see some interlacing in the first second or so.
 
Back
Top