Personal View site logo
GH2 driftwood hack in shootout with Canon C300
  • 471 Replies sorted by
  • Those are Clog? They don't look so much flat... they don't look flat at all, shadows are almost painted
  • edit to my last post: the first images were downscaled to 1280 x 720, it seems.

    @Manu4Vendetta posted the full 1080 frame.

    At a glance, it looks like there is not much difference in terms of resolving power.
  • From FB: With the C300
    AA0002_01.jpg
    1920 x 1080 - 2M
  • If that is a 720p frame grab, it ain't chopped liver. I want!
  • I'd advice everyone to hold on until the actual comparison is presented.
    The c300 landscape looks seemingly similar to the gh2 one. Yet, it is a 720p frame grab (c300) vs a full 1080p framegrab..
  • From FB: Gh2 Driftwood Hack 70mb/s
    Clip-3.jpg
    1920 x 1080 - 703K
  • I'm thinking about what x vs. y tests would be important to me.

    1) Skin tone

    2) People or cars moving around. Hey, still, locked down shots tell very little about motion effect and how codecs deal with lots of change.

    3) A line resolution chart, (same lens) see how the binning or line skipping effects the resolution.

    4) Highlights - I'd like to see how highlight clipping is handled (Red CineX) has that nice feature to rebuild one clipped color channel from the others info.

    5) Shadow detail. How dark can you go and still have useable detail vs. noise.

    6) Dynamic range test. If you carefully matched the look (exposure) on each camera, Then progressively tried to boost the highlight/shadows 1 stop at a time in post. Which holds up for how many stops. So, the C300 should have the most dynamic range. How many stops can you go either way and still have a good image?

    Any other thoughts?

    7) Banding test. Find a subject the the sky or dark shadows and see how well the different bit depths handle banding. What I HATED in my GH-1 tests.
  • Seriously though. Just let the test results come out and then decide if you like the testing method or not. But let's not get so precious about our equipment. We know how good it is, and even if PB's test results show that the GH2 is 80% of the C300, the price differential is still astounding!
  • Philip and his team probably took that shot in 1080i FSH or 720p stretched! Seriously though theyre the first of many pics. But they're the only settings at 70M ;-) Lets wait for the final report.
  • after looking at the stills between the c300 and gh2, it's really obvious that an unfair lighting advantage is given to the c300. I'd imagine the image would be better than gh2 anyway, why help it?

    Nevertheless, I'd give Bloom the benefit of the doubt until the review is in.

    I've never thought he would be a completely unbiased opinion anyway. Anyone who loads their site with paid advertising is suspect as far as I'm concerned.
  • Sorry Vitaliy. I come from a world where total strangers sometimes say bad things about their colleagues. And good things too, of course.
  • but, we can surely be looking at gradations on smooth surfaces, noise, macroblocking and other artifacts from this test.

    even though at the moment at least (it is incomplete) it seems that the color profile was set to favor the saturation on the canon to make it look more atractive

    but i think is worth it still

    ...assuming that he will eventully post the settings used on each camera, like iso, film mode, lens, aperture, etc..
  • I recognize these kinds of tests are just about always lose-lose from an audience complaint point of view, but, like, is it really too much to ask for the images to be KIND OF similarly exposed and KIND OF shot under similar light conditions/time of day/etc?

    But all we're seeing is a couple jpegs at this point, so maybe the full thing will be more enlightening. Let's hope.
  • Very odd.
    I realize it's about dynamic range, but he chose to blow the clouds out on the GH2. Also, looking at the same shot on the C300, it appears he shot this during "magic hour". Of course the highlights look better, not to mention the warmer color temp at that lower sunset. I wish he did the same for the GH2. Also what's with 70mb/s ? Where did this come from? I thought Quantum was 154Mb/s.
    >>>>>>>>>>>>>>

    The prosecution rests...

  • Very odd.
    I realize it's about dynamic range, but he chose to blow the clouds out on the GH2. Also, looking at the same shot on the C300, it appears he shot this during "magic hour". Of course the highlights look better, not to mention the warmer color temp at that lower sunset. I wish he did the same for the GH2. Also what's with 70mb/s ? Where did this come from? I thought Quantum was 154Mb/s.

    I'd like to see real tests where all the cameras see the same subjects under the same lighting.
  • Hmm, the plot thickens then.
  • Bloom's caption of the GH2 screen grab in the updated blog post says 70Mbit GH2 hack. Just wondering
  • @EOSHD, it's @driftwood's latest Quantum v5 (technically still in beta), the ~150Mbit intra version, assuming Bloom shot H mode. If not, even L mode is 100Mbit :)
  • So which patch did you install on the GH2 in the shootout @driftwood? Is it 70Mbit AVCHD i-frame? I'd like to give it a spin as well.

    Currently putting GH2 up against Sony A65 in my own shootout... So far it is winning :)
  • of course, but assuming that those images were rendered at the same quality they are still a valid comparison for image compression, and the gh2 has more artifacts,
  • @lolo... they are jpgs --- keep that in mind...
  • there are some pics up on bloom´s site!
    and the gh2 compared to the c300 has artifacts (very noticeable on the trees)
  • really.
    Sorry I thought that the fact that he is doing the test not you or I, had some merit in the discussion. No need to go any further on this
  • >Do not disuss Philip Bloom?

    Yep.

    >Isn't he the reason for this topic in the first place.

    No. Reason is completely different.
  • DELETE ME PLEASE !

    COde 401!! i CANT DELETE THE PICS...
    Clip #36a.png
    1920 x 1080 - 1M
    Clip #36b.png
    1920 x 1080 - 1M
    Clip #36c.png
    1920 x 1080 - 2M
    Clip #36d.png
    1920 x 1080 - 2M