One thing to remember when doing thee test is how the GH2 MTS file is converted.
It makes a huge difference.
For example using my favourite method with 5DtoRGB you can get a lot of range using the 1.8 gamma flagging. I'm not sure it was used in the Philip Bloom shootout and that might be why the GH2 screen shot was so contrasty.
Unfortunately 5DtoRGB has left out the 1.8 Gamma option on the new paid for "batch" app.
They offer a way to do it using using GLSL Fragment Shaders, but I have not been ably to figure that out.
Anyway, dont forget when comparing cameras to export the flattest image possible if you want to see the range.

Yes, the usable DR is another thing. Two things that chart have a lot of problem dealing with are at the highlight level channel clipping and in the shadows not only the noise level but its characteristic. Like in my test the shadow noise of the hacked gh2 is much finer and looking more like grain, that it is much more acceptable/usable to dig into them than in the Canons ones.

@cowpunk52 is the "backlit stoufer chart" the same device with little holes giving more (twice) and more light, the same as zacuto test?

@nomad:thx for information :-)
How do you measure information of luminance of your chart? From a still grab of your footage do you?
Then from a software (maybe photoshop?) you take measure of luminance from each point of light, looking for non clipped data? How?
What I dont understand is that each spot of light is double lighted from the last. So considering the grey scale of your photo, have you got enough information do measure 14 stops!?

Thanks nomad.
Good to know.

@magnifico - i'm going to try hard to get a hold of a backlit stoufer chart before i start shooting a feature in January. I would very much like to know this information as well before i get neck deep into production.

DR for video of course.

@Ian_T yeah, the trick is the downscale. A bayer chip can usually resolve at max about 80% of it's native resolution. If the downscale is done with the proper processing from a suitably oversampled sensor, then it should be able to resolve as much as a native 1080p 3-chip camcorder. But it's very difficult to resolve more than 1000 TVL in the HD format, because there are only 1080 vertical pixels in the frame. Also, downscaling is no simple task - just try and do a straight conversion from HD to SD and see how much aliasing you get! I'm thinking a fair amount of anti-aliasing or blur is applied during the encode process of the GH2's video which results in it's softer than native resolution, but that's pure speculation. In a perfect world we could get a 3-chip S35 sensor block at HD or higher resolution, but that would probably result in a very large camera!

@Ian_T - actually, i am explaining this correctly. All of those cameras have three sensors, and each of those sensors have 1920x1080 photosites. Which is why they are known as 3 chip camcorders with 3 chip sensor blocks. It's really super easy to confirm this: just go to the B&H product page for the Canon XF300. Notice the part in the description where it says that the camera has three native 1920x1080 sensors? Yeah, there ya go. But how can this be? How does such voodoo work? Very simply - In 3CCD or 3MOS camcorders, a beam splitter separates a signal into three different versions of the same image - one showing the level of red light, one showing the level of green light and one showing the level of blue light. Each of these images is captured by its own chip - but each measures the intensity of only one color of light. The camera then overlays these three images and the intensities of the different primary colors blend to produce a full-color image. A camcorder that uses this method is referred to as a three-chip camcorder, and because it doesn't need to do a bayer interpolation like a single sensor has to do, it's typically able to resolve extremely close to it's native resolution. The RED and F3 have aliasing and moire because they are single cmos sensors that have to do a bayer interpolation of their red, green and blue pixels on the sensor. It's impossible for those cameras to produce an image that resolves the sensors native resolution, and so some information is made up, and that's why aliasing is sometimes referred to as 'false detail.' The GH2 has to do a bayer interpolation too (which is why it will also alias and moire given the right detail frequency), then downsample that to a 1080 video signal. However, perhaps due to processor horsepower or encoding algorithm or some other reason, the video resolution is not as good as a proper downsample from the "5k sensor." Anyone can test this very very easily: take a 16MP raw still and a 1080p video of the same high detail subject or scene. Downrez your GH2 raw still image to 1920x1080, then compare it to a video frame grab from the same scene. Post your results.

Also, we ended up here because a question was raised regarding the validity of a resolution chart shot with a GH2 and posted on another forum. One poster here stated that it was not accurate because the chart poster was just trying to protect his $40,000 2/3" camcorder investment. I stated that it was accurate given my similar findings with the GH2's ability to resolve detail in HD video.

"a camera with three 1920x1080 sensors can easily resolve a 1000 tvl picture when shooting HD format. That's why the XF300 can do it, as well as the EX3, HPX250, PMW350, HPX3100, and a bunch of other full-raster 1080p camcorders. "

@cowpunk52 Um....I don't think you are explaining this correctly. None of these cameras have 3 sensors. They have one 3CCD sensor that takes separate readings of red, green, and blue values for each pixel as opposed to a single CCD camera. And...just because a camera is 3CCD does not necessarily mean it can resolve 1000 lines. But...I do believe the EX-1 is rated at 1000 lines. I've read some testemonials in the past (I forgot where) that it resolves better than the EX-3.

As far as the GH-2's resolving power.....I believe it's very close to the Sonys (on paper) but not quite as much. BUT in real world shooting it shows a bit different. I think it all depends on the shooting situation.

I agree with @bwhitz about viewing these DSLRs as like different film stocks. I mean...film is film....but not all film stocks resolve the same...or even look the same. So, why is it that we all get so worked up about DSLRs vs....whatever?

Also, I've seen examples of EX-1/3 footage showing some form of aliasing in the past. So it's not just a CMOS thing (though I understand where it comes from with CMOS cameras). And yeah....even the more expensive CMOS cameras like the Red, F3 etc. all have aliasing. I only bring this up because it seems like people who like to put DLSRs down always use that issue as a way to seperate DSLRs from the "big boys." I also found it funny that those same people never use the GH-Xs in their argument. Fact is these Panasonic cameras do an excellent job in controling that issue.

......what were we talking about again..... Oh yeah....Dynamic Range between GH-2 and 7D. How did we end up here?

@bwhitz I think you just need to do some side by side test by just shooting some high detail scene, better in low iso because we want to see them at their best (if else someone wouldsay that the iso noise is higher in one or the other. post the result in full 1080p png files. As the chinesse say a picture says more than a thousand words.

bla bla bla

Also, was the landscape shot filmed at 1080p on Philip Bloom's page? Because in that instance, the C300 is clearly resolving more detail.