@mintcheerios I think your numbers of DR is not good at all and exaggerating the DR gap. From last year Zacuto test the Canon DSLR where all in the 11 stop range and the red in the 12 stop one. From all independent test I have seen from zacuto to provideocoalition the red is around the 12 to 12.5 stops and the Canon dslr from 10.5 to 11 stops. The HDR mode is only good where there is no big motion so it cannot be counted.

You can see a test I have done comparing the 7d vs the gh2 http://www.personal-view.com/talks/discussion/1795/dynamic-range-test-between-gh2-and-7d/p2 . As you can see the gh2 is at least as good if not better than a cinestyle 7d because of the retention in detail in the shadows from the hack. So I would say that the gh2 would be 1 to 1.5 stop lower than most of the other camcorders except the Alexa, f65 and slog F3. If it was like 6 stop lower than the red for example, believe me we wouldn't be talking about the gh2 surprise result but how it was the laughing stock of the test.

the whole point of this test was that it's not the camera but you who makes the great/bad image. I think they clearly stated this on their website.

Obviously they made their point: a camera that is without a doubt twice as good as another one comes out worse because of the man who was operating it.

They just want to say that if you have a cheap camera you can still make great images with it if you know what you're doing. Same goes the other way round: if you have an expensive camera but don't know what you're doing you won't make great images.

And in the end it is RED's own fault: if they had sent some one (like the other camera manufacturers) to lit the scene and operate the camera to best extent, it would probably have looked a lot better.

It is RED's own arrogance that's responsible for the bad advertising in the Zacuto test

@bwhitz

You say the hack erases pretty much all artifacts. Well I wish it were true but in underexposed areas there's still a lot of blocking, colored noise and whatnot. I've found that using a film mode that gives up 1 stop in the blacks and at the same time setting the contrast level on the gh2 to 0 lessens the artifacts. On the other had you do loose a stop of DR. But if you know a thing or two about lighting this won't be a problem.

Personally I'd rather loose a stop than to have artifacts

@jrd @pinger007

Ahhh, I stand corrected. That makes sense. In that case, it is an interesting little test. One part to show the absolute abilities of a camera, and another to show the differences when you work within a camera's strengths/weaknesses, with appropriate lighting and color correction.

@jrd The first part of the test sounds like it had the results I would have expected.

@dbp

You can't have different DOPs lighting each scene in their own way, and then call it a camera test.

Yes, you can: it's like "this apple is redder than that orange."

To be honest, I kind of agree with the Red guys. You can't have different DOPs lighting each scene in their own way, and then call it a camera test.

For a proper camera test, you'd have to have a wide range of scenes with consistent lighting, then shoot them all with each camera. Of course, that would take forever.

It sounds like it was a great example of Talent + low end gear surpassing Mediocrity + high end gear. But then again we should already know that, haha.

More proof that the camera is no longer much of an excuse. GH2 is certainly good enough for high level looking content in the right hands.

@bwhitz

To admit that the GH2 looks better to audiences than the Epic is to admit that in terms of image at the final stage (projection in a theater), the camera that cost 66 times more couldn't win. This really is insanity because it goes against so many of our core beliefs. How often do we expect to get worse performance out of any product that we pay 66 times more for? It just can't be real, so we deny that it's even possible without considering otherwise. Even the most zealous GH2 fanboy isn't willing to outright say the image on his camera can beat the Epic (though I don't know after this test).

Every measurable spec of the Epic is objectively better than the GH2. It has R3D RAW vs AVCHD (hacked), RAW vs 4:2:0, 18 stops of dynamic range with HDRx, 5k vs 2k, and 12-bit vs 8-bit color. Yes, the resolution advantage has been compromised due to downscaling, but it should be noted the Epic gets to have true 4:4:4 levels of color because of this. On paper, the Epic slaughters the GH2.

But if we break down each spec, we can see how and why the GH2 was able to compare (and win according to many viewers).

Resolution: This is where the Epic wins hands down, but since Zacuto's test had all the cameras sized to 2k, the Epic didn't get to shine here. The GH2 resolves pretty close to 2k, so resolution was probably comparable in this test. While I realize Red didn't get to flex all its muscles in this test, I agree with Zacuto that the final output should have been 2k since that is what most people see in theaters as of 2012.

RAW vs hacked AVCHD 8-bit color: I think the amount of extra quality gained from RAW (even with grading) is not as significant as many people think. 8-bit doesn't seem like much, but it actually is. That's 256 colors each for red, green, and blue meaning a total of over 16 million color combinations. Out of camera JPEG files (including the highly regarded Olympus JPEGs) are all 8-bit. The BMC stills that people have been so meticulously grading have all been from an 8-bit JPEG. Blu-ray is 8-bit even!

RAW vs 4:2:0: When you look at explanation of color subsampling, you come away thinking "4:2:0 is awful! How can so much color information be tossed out like that?!". The fact is, all this stuff is at the pixel level, so most layman won't be able to tell a difference. The amount 4:2:0 hinders an image also depends on the kind of image. On some images, it may be impossible to distinguish between 4:2:0 and 4:4:4 without some serious pixel peeping. This would be a problem if movie audiences had a button on their seats that froze the film and magnified the image by 2x. I must admit though that 4:2:0 does suffer a bit for green screening (though be careful not to blame it on that if you are using bad keying techniques).

RAW vs 175mbps hack: In terms of bitrate, the hacked GH2 gives each frame the amount of bitrate a still OOC JPEG gets which is good enough to get rid of most noticeable artifacts. If the hack can render fine sensor grain, I don't think the compression is hurting the image that much. Again, most layman (and apparently some pros) wouldn't see these artifacts on the GH2 footage.

14 (or 18 with HDRx) stops of dynamic range vs 8 stops: This is what the GH2 needs more of in my opinion, but it took me awhile to realize that 8 stops is actually quite good. The Epic has no advantage in this respect if the image doesn't require more than 8 stops, and that's where lighting and NDing windows comes into play. In most cases, you can fit pretty much the entire image within the GH2's range. On really sunny contrasty days, I've been able to underexpose to protect highlights, and lift the shadows up a decent amount thanks to the hack. Apparently, the Zacuto test had a bright window in the shot, and I suspect the cinematographer NDed it good for the GH2. Also note that just because you capture 18 stops of dynamic range doesn't mean you use it all on the shot. Actually, 18 stops (or even 13 stops) is bound to look like garbage before you grade it. It is however a nice get out of jail free card when you botch the exposure. Just be extra mindful of the exposure with the GH2 and you'll get awesome shots. When I shot my feature on the GH1, I didn't need any extra dynamic range on most shots (I've blown out the sky a couple of times though, doh).

With all that said, I think the Epic is probably able to match any of the cameras in the test with the right person behind it. I also think the Epic's chance to shine would be in situations involving high dynamic range (like a forest under bright sunlight) and of course, when cinema moves to 4k+. For aspiring filmmakers, I think the hacked GH2 is better than what we deserve.

"...but RED declined to accept."

Yep. This seems to be the Red way. Decline everything, then say "it's done wrong" when someone else doesn't get acceptable results.

David Mullen made a funny point in the thread that in the first shootout, Redusers complained that it wasn't fair because Red didn't get special treatment and proper work-flow to make it competitive to the Alexa and such. Now when each camera gets special treatment to make it sing... and Red still did come out on top... they're complaining that it wasn't fair that each camera has individual lighting? WTF? These guys are nuts.

I mean, I'd still like a Scarlet or Epic myself (for frame-rates and flexibility)... But this is getting crazy. Maybe the GH2 really does just have a better look to most audiences. Did anyone think of that? Why does the most expensive piece of gear ALWAYS have to be the best or someone is cheating?

@rajmalik Those settings were released in mid-March. I thought I read that the tests were shot in February, but I may have remembered incorrectly.

@dbp @kholi Framerates are an interesting question, and games are a very different question from passive media (that being a label for non-interactive media). For games, increases in framerate are synonomous with a decrease in latency. This is a non-consideration for film framerates, because their is nothing for the audience that requires the audience to do somehting "to" the content and the interaction is either entireley internal or social. So latency is no longer a reason for higher frame rates.

But note that in my comment about animation, I was speaking about framerates relative to 24, 25 or 30P where a lower frame rate would be "shooting in 3s" or "shooting in 4s" etc. as opposed to "shooting in 2s" or even at the frame rate itself. So if we were talking about for 24P, shooting in 3s would result in an 8P look, etc.

Now some of the 90s cartoons had interesting and influential character designs that are more aesthetically pleasing to me than their modern equivalents (X-Men and Batman TAS for instance) but when I watch them with friends that have only seen the newer series, they often frustrated by the animation being low framerate, etc.That is not to say they would prefer framerates exceeding 30P, just that they dislike framerates as low as 8P.

@bressonftw Actually I disagree. The issue with the GH2 is not lowlight performance (which actually holds up quite well to a lot of the similar sensor sizes and somtimes even larger ones) but more specifically, it is the way it handles the lowest part of the histogram. The unpredictability of shadow exposure is troubling, in that certain scenes can result in shadow flash or flicker, even with intra codec settings.

I think a list was already made in the April Fools thread.

@Bressonftw

My suggestion is to create a topic on 'the weaknesses of the GH2

Nah, I'd rather film a bush and marvel at the resolution. Seriously, that's a great idea. How to work within limitations is better than getting drunk on the strong points -- though it is fun to get drunk sometimes.

Good idea @Bressonftw, Looking forward to you starting it. :)

My suggestion is to create a topic on 'the weaknesses of the GH2', since everyone pretty much agrees on 'working around them'. Why should we not make a list with problems/weaknesses with possible solutions to work around it?

For instance: 'bad low light' => possible solutions: fast lens, neatvideo noise remover, etc.

I see this as an analogy to painting. Do you prefer expressionism or photo realism? Is the emotional resonance in art linear to the degree of "realness"?