GH2 lets me down!

[filipeG]

On the last couple of days I was shooting a short movie, with 2 GH2. One of them was in NTSC and another in PAL. Although the hacks were different. Because there was a one minute sequence that needed to be filmed in slow motion, and we needed the 720p60 to do that. But when we were going to do the shot the it only filmed for 6 or 4 seconds, and the error that the writing speed was to high for the SD card showed up. I own a 16GB Sandisk 95mB/s!! I went on the computer to do a different hack while all the team and the actors were waiting, put the bitrate lower at 88, and 44 mbit, and the same error appear.. We had to do with the PAL GH2 at 720p50.

It's the first time that the gh2 lets me down! Maybe it was the patch that I used, but all the settings were good.. don't know.

Any ideas?

[Sara]

Your memory card is probably a fake or damaged.

Where did you get the card from?

[filipeG]

NO!! It's one original Sandisk, I bought it from amazon! ExtremePro 95mB/s 16GB!

Probably it was the patch I used..

Edit: also I  normally use 24H at 176mbit and works fine..

[Stephen de Vere]

I have the same sort of problem. Just loaded the EOSHD 88Intra and the camera freezes when the scene gets detailed, like a lots of foliage filling frame, and when using the Panasonic 14-40 lens when there's incamera sharpening going on. This is with HBR (1080/25p). I haven't tried other modes.

I went to some trouble and expense to get the card from a registered Sandisk dealer.

I was previously using Driftwood100Mbps Flow Motion Patch v1.11 which was fine (with the same camera/card) but I wanted to use the EOSHD 88Intra to keep file sizes and data rate down a bit.

Apparently doing a full format of the card in Windows can help (defragment) but it didn't for me.

[Axel]

In the popular german site [i]slashCam[/i] someone finally set up the ultimate test in a thread called [i]The Naked Truth About The GH2 Hacks[/i]. He mounted two identical GH2s with identical lenses (Nokton 25mm) and identical settings on a rig. There were only two differences: 1. The slight parallax, comparable to that of the two human eyes. He could have used a 3D rig to avoid this, but he didn't have one. 2. One GH2 had standard FW 1.1, one had Quantum v9b.

He had a hard time comparing cropped parts of the test shots. [u]There is no more detail[/u]. If there could be a small difference, it was dubious if for the worse or the better.

Only at extreme lowlight, the shadows of the unhacked camera were filled with more colored noise that looked worse in motion and made grading difficult. This noise didn't disappear in the Quantum clips though, it just became finer and looked better. The screenshots were ambiguous, but he promised to upload original files for everyone to compare.

> Unless you don't use to shoot at night: Unhack! It is ridiculous to state that all people who experience problems with hacked GH2s (are "let down") are all in the wrong because they used the wrong cards or didn't swipe them properly. If Magic Lantern had so many failure reports as the GH2 hacks ...

> Use a hack only, if you make short, instantly controllable takes. Don't rely on a hacked Lumix if you are in the situation. Get yourself a second body to switch faster.

> I have lengthy recordings of a two hour concert, filmed with only the 29'59" restriction removed (for the same reason, reliability). It looks phantastic, and if I told anyone of the fundamentalists at personal view that it was an Intra-hack, they had believed it.

> Next step in the test series is, with what patch and at which bitrate the postive effect described above will be distinguishable. I am curious for the results and will link the thread then (you can translate it with Google, funny read), stay tuned!

[HurtinMinorKey]

Were they static test shots? Unless I am mistaken, the extra detail should only be noticeable when you have a of elements changing frame to frame.

[Axel]

[quote author=HurtinMinorKey link=topic=800.msg5863#msg5863 date=1338919234]
Were they static test shots? Unless I am mistaken, the extra detail should only be noticeable when you have a of elements changing frame to frame.[/quote]

This is what the video engineers tell us. You need to "stress the encoder". A lot of random detail, movement, wind in the foliage, raindrops, pans where no pixel stays in place. This was demanded of the set-up.

But there is a second theory, connected directly to the situations where the patches shine. What if an image seems so [u]simple[/u] that the original firmware's bitrate gets too low to describe very fine differences in shadows? The engineers tell us, no, this is not the case. However, to me it seems obvious that when I find compression artifacts in a clip that uses only 16 mbps of 24 mbps, this could after all be a point. It will also be tested.

[Stephen de Vere]

I believe it is precisely these very slight differences in quality, difficult to notice in the straight-from-camera result, that are important to professional producers. It is during post-production operations and final re-encoding for distribution/broadcast that any small problems get multiplied into serious problems.

Hence the BBC insist on data rate minimum of 50 mbps (long-gop) or 100 mbps Intra, amongst other things. The BBC have some of the highest standards in the world so has become an industry standard but if, as most GH2 users don't, you don't need this level of quality then hacking is probably not worth it.

I do work for the BBC so even the smallest improvement in eg. shadows is the BIG issue for me with the GH2 - I cannot use it as a 1st camera but I want the best quality for the few shots (remote, risky etc) I am 'allowed' to use it for as 2nd specialist situation camera.