Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Zacuto 'Revenge of Great Camera Shootout,' featuring GH2
  • 261 Replies sorted by
  • sam_sticklandsam_stickland

    @bwhitz You mean being a film maker is like Renaissance art, where it's all done on commission, as part of a guild, with a whole team of helpers and apprentices? ;)

  • dbpdbp

    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.

  • pinger007pinger007

    @hemp22

    At the shootout in chicago I questioned the moderator regarding the use of nd filtration on the window. He said that none of the cameras used nd on the windows. When comparing the original footage with the corrected, it appears the dp exposed primarily for the window and brought up the shadows in post.

    @Dbp

    There were two shoots: one with lighting set up by a single dp for ALL cameras, and then a second shoot where each dp was given an additional 30 minutes to tweak the lighting setup without removing any of the original setup. The gh2 footage looked good in both.

  • RobertoRoberto

    @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."

  • jrdjrd

    @dbp

    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.

    Zacuto came reasonably close, within practical limits. What nobody mentions here is the first round of the test, which consisted of shooting two scenes with each camera (same lighting in all cases) and giving everything a one-light grade.

    I swear this will be the last time I point this fact out -- apparently nobody wants to hear it -- but this first part of the test did not cause anyone to swoon over the GH2, because the GH2's limitations were obvious. Its highlights and shadows had noticeably less detail than everything but the iphone4 and the Canon 7d.

    Note also that, in the second portion of the test, the GH2 team lit to significantly reduce the contrast range and graded to increase color saturation. Except for the iphone 4 stuff, the other teams didn't. This may account for why the GH2 footage was preferred by many in the audience, particularly in scenes (like the Zacuto test) where resolution differences wouldn't be obvious.

  • dbpdbp

    @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.

  • ButtButt

    GH2 Vitaliy/Driftwodd is still the No. 1 for 800 Dollar . . . Heeeeeeeeeee!!!!

  • 2many2many

    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

  • thepalaliasthepalalias

    @2many I normally overexpose as much as I can without missing the most important details in the scene to deal with shadow issues and then darken in post. I do not mind the shadows so much in other things (unless the shot happens to be having the sporadic "flickering shadow" issue) but it can be rough on skin.

    Honestly, while there is always room for improvement, I think that the codec (within the limitations it works with) does a really good job in some of those areas with the right settings and takes it further than I have seen other AVCHD/H.264 cameras do. Have you used CM Night? It's at a whole difference performance level in terms of shadow macroblocking from Quantum v9c (though I have not compared to V9b, which they used for that test). There is still a noticeable gap vs JPEG, but much better.

    http://www.personal-view.com/talks/discussion/comment/60506#Comment_60506 http://www.personal-view.com/talks/discussion/comment/60002#Comment_60002

    In the second link, CM Night is in the upper right and Quantum V9c is in the lower left.

  • danyyyeldanyyyel

    @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.

  • thepalaliasthepalalias

    @danyyel The GH2 has a much lower dynamic range than the RED, not just 1 to 1.5 stops. My testing suggested at least 2 stops at the bare minimum and scientific comparison may show even more. I can tell you this from shooting scenes with dynamic range that exceeded even the RED without HDRX with both cameras (though I have no comparison footage I can show at this time). And that's before we even deal with the GH2's comparative difficulty in dealing with the shadows under certain conditions.

  • kholikholi

    GH2's useable DR is around 9 to 9.5 stops.

    Epic's useable DR is around 11.5 stops @ 5K FF

    So yeah, it's a good two up.

    HDRx is a rarity, no sense in counting it.

  • thepalaliasthepalalias

    @kholi For most scenes, it doesn't make a lot of sense to use HDRX. But it can be helpful on some of them - I used it to keep detail in the sky while I exposed for a backlit performer in shadow in a scene that used only natural light once. :)

  • shianshian

    well, because I have tested it, and am still testing it across film modes and hacks, let me tell you. The GH2 has a full range of 8 and 1/3 stops, only 6 of them are effective stops, and only 4 stops of detail between 80IRE and 20IRE regardless of film mode. Not even close to the RED.

    No DSLR at the moment. not even the 5D has a larger DR... they are all about the same. which is a 5.6 to 6 stop effective range. It is a myth that DSLRs have a significantly greater DR than any other standard HD video camera. Don't believe me, take an F900, XD-cam or an HVX and chip chart it with a 5D and a GH2... they are all pretty much the same.

    And to be perfectly clear, since some people don't understand what dynamic range actually is - Dynamic Range is the exposure range (in stops) across a single exposure from pure white 100IRE to pure black 0IRE. There is a REASON your in-cam EV meter only goes to +3 and -3.

    EDIT - the 6 stops of effective range refer to the 6 stops between 95IRE and 7.5IRE (broadcast)

  • thepalaliasthepalalias

    @shian Well said, Shian. The difference I notice between the some of the other older cameras and GH2 is not the dynamic range, but the detail contained within that dynamic range. The compression artifacts and noise do a lot more to compromise it in some of the older cameras and is not uncommon for the camera to "blend away" some of the detail close to the highlights or shadows in certain shots, even if the camera can sometimes capture them (or for it to be so noisy that the detail gets lost in noise reduction anyway).

    While the GH2 improved on that, the RED cameras provided consistently usable visual information within the dynamic range the sensor offered and I did not have to keep a close eye on it to watch for errata the way I do with the GH2 under very specific conditions.

    As I said, I unfortunately did not have time to do the scientific tests I had planned, but my subjective reaction to the dynamic range gap was "a lot". :)

  • danyyyeldanyyyel

    @thepalalias and @kholi, lets say it is 2 stops compared to the 1.5 from the zacuto test, but not the 6 stop I made reference from. I base it from some test I have seen from provideocoalition and now the zacuto 2011 test.

    http://www.personal-view.com/talks/uploads/FileUpload/ee/1f710f6d505171964b288d498ed158.jpg

  • danyyyeldanyyyel

    I don't think there will be chart in this years test, so we will have to see when it comes out.

  • thepalaliasthepalalias

    @danyell Yeah, I remember seeing that and wondering how they measured. It just did not reflect my experience with the cameras that the gap between the RED and 7D was only 1 stop.

  • soundgh2soundgh2

    Shouldn't this community be happy that it's even considered? If we all bothered what guitars and amps our music hero's used and argued about it ... hats off to taking a codec and camera, by you, far beyond it's original use.

  • TrackZillasTrackZillas

    @mintcheerios Well said, well said I made those same observations long time ago. We have plenty to work with using the GH2 8bit 4:2:0 is good enough, just have to learn how to make it shine..

  • bwhitzbwhitz

    @Shain "Don't believe me, take an F900, XD-cam or an HVX and chip chart it with a 5D and a GH2... they are all pretty much the same."

    well I don't know what charts will tell you. But I just did a shoot yesterday with 7Ds and XF300's. The XF footage was almost unusable at how horrible the latitude was. Blows-out at about half the threshold of a 7D. You couldn't even see the sky on the XF footage, while the 7D rendered it perfect, clouds and all.

    So maybe when measured, the DSLR images "technically" only has 6-8 stops of information. But the 7D really exposes like 11-12 stops.

    "It is a myth that DSLRs have a significantly greater DR than any other standard HD video camera."

    No, I think that this is a myth. DSLR have much higher exposing latitude. Again, when looking at charts and "recorded information", maybe not. But in terms of how much "light" you can expose for while shooting... DSLRs really have the advantage. HVX's, EX's, XF's all blow out so much faster, and have much worse shadow detail, than any DSLR I've shot with.

    We may be taking about two different things here. DR and Exposure Range are two different things. If you're saying that DSLRs only have 5-6 stops in terms of how much manipulation and recorded information you have in post, then yea, that sounds about right. But if you're talking about how many stops you can actually expose for when shooting, then 5-6 is worse than a iphone. DSLRs are all around 10-11 stops here. Not 10-11 stops of workable range, but 10-11 stops of exposing range.

    Say what you will about Zacuto, but I have no reason to dispute them. And this chart really seems the most legit to me out of anything else...

    http://www.academyart.edu/export/sites/marketing/webcheckout/images/latitude.png

  • shianshian

    Dude, the XF must be garbage. I've never tested it. But I have tested the HVX200 vs F900 and the only real difference is detail, not DR. And The same vs. 5D and GH2.. The detail in the shadows is at first blush, way better with the DSLRs, but in terms of pure DR, not a lot of difference. I've shot on the 7D a few times, and didn't see anything close to 11 usable stops... not even close. The original RED ONE had 11, and we could not fit the DR of stuff shot on the RED ONE onto the 7D, we had to squeeze it into 8.

    The GH2 (on the average) has 3 stops above 0 and 5 and 1/3 below. This range moves around as different film modes are selected, but it never really stretches wider. Everything is a compromise. Better shadow response on NOS equals worse highlight response... and so on.

    Now I've shot stuff with the GH2 in Smooth and Nostalgic modes, that seemed to be handling a wider DR than I would normally think capable, but when I went to grade the footage, that "extra" range was complete crap...just falls apart immediately on both the high and low ends when trying to grade it, so I have given up trying to use those outer edges, it never pays off the way you want it to. I'm resigned to 6 that I can work with, readily. And 4, if I wanna do any serious grading.

    Fit everything into 4, and you can do some amazing shit in post.

  • shianshian

    7D vs Alexa

    http://nofilmschool.com/2011/01/hdslr-missing-dynamic-range/

    Right around 8 stops, the sales pitch would call this 10 stops...but 9 and 10 are completely unusable. Which is why I asked the BMD folks how they were calibrating 13.5. How many stops between 80-20? Where does 18% gray hit white and black, they couldn't tell me. So how much is marketing, and how much of it can you use?

    The GH2 has 8 and 1/3 stops - If we count those super noisy, and partially blown stops (but not fully), then the GH2 has 10.5 where you can still see a tiny bit of something there before pure black and pure white...but nothing usable. 18% gray (skin) blows out pure white at 3 to 3.3 stops over, and hits pure black at 5 - 5.3 under. (the variance is due to film modes - depending on which way the image is being pushed)

    My tests and the number of "8.3 stops" is stuff you can use between 0-100IRE. 6 stops between 7.5 and 95, and 4 between 20 and 80.

    You may say, well that just pertains to skin tone. I'm a narrative filmmaker. EVERYTHING in my world REVOLVES around skin tone. All my exposures are based on skin tone.

  • mintcheeriosmintcheerios

    I've heard a lot of different numbers for the dynamic range of all the various cameras, so I just used my own figure for the GH2 and I admit, I used the Red's marketed figures for the Epic since I don't have an Epic. I ended up with 8 and 1/3 stops of range for the GH2 with hack using this method (http://www.luminous-landscape.com/columns/determining-exposure.shtml). I counted what I thought were usable stops, so I'm sure you can get a higher figure with different criteria. If the Epic's usable DR is really around 11.5, then that is kind of disappointing.

    Either way, I think the GH2 has enough dynamic range for most applications if you plan on doing post correction. If you notice in the link where the Alexa and 7D's dynamic range is compared, you'll notice how fast highlights blow and how much detail the shadows retain. Protect your highlights and dig a little in the shadows if needed. The highlight blowout indicator in the GH2 + histogram is tremendously helpful for this. I think the key is to work around the GH2's shortcomings which seems to be the lesson of this shootout.

  • hankhank

    Lol sure

Add a Comment