Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Please, support PV!
It allows to keep PV going, with more focus towards AI, but keeping be one of the few truly independent places.
Official Panasonic GH3 topic
  • 1288 Replies sorted by
  • thepalaliasthepalalias

    @Svart This data is extracted from the specifications available at the Panasonic website.

    GH2 Active Sensor: 17.3x13.0 (in 4:3 mode) Horizontal Resolution: 4608 (in 4:3 mode)

    GH3 Active Sensor: 17.3x13.0 (in 4:3 mode) Horizontal Resolution: 4608 (in 4:3 mode)

    The difference is that the GH3 maintains that horizontal resolution (and horizontal active sensor dimension) while the GH2 expands both (and the horizontal resolution in JPEGs hits 4976).

    The two differ in the total sensor size (ca 18MP GH2 vs ca 17MP GH3) but the active sensor area in 4:3 mode (and resolution in that mode) are identical, at least as reported in the Panasonic official specifications on their site.

  • jonbeeotchjonbeeotch

    Hi, im new here, but i've been following this site since the gh1.

    i finally downloaded "genesis"... saw some weird frames. im not super "tech", so i apologize if this is totally obvious. anyone know whats going on in these stills? looks like twixtor weirdness. would it be the camera, compression, plug-in artifacts???

    GH3_weird_1.jpg
    1926 x 1063 - 675K
    GH3_weird_2.jpg
    1923 x 1065 - 741K
    GH3_weird_3.jpg
    1923 x 1069 - 667K
  • mpgxsvcdmpgxsvcd

    @svart

    I think that there is a general misunderstanding of what the "Sensor Size" spec means. For the Panasonic GH3 it means that the size of the area of pixels used in 4:3 mode is 17.3 x 13.0 mm. Panasonic even puts a special note on there that this size is only with respect to the 4:3 aspect ratio mode.

    This is NOT. and I repeat NOT the total size of the entire sensor in the camera. For multi aspect ratio cameras like the GH2(and yes the G5) the total sensor size is bigger than 17.3x13.0mm because it has to accommodate the different aspect ratios with the same field of view.

    For non-MAR cameras the total sensor size is probably greater than 17.3x13.0mm because they can't use the pixels edge to edge. However, it is probably not as big as those cameras that have MAR capabilities.

    There really is no way to convince you without showing you the exact same picture(Same Resolution and Same Field of View) taken with the GH2 and the GH3 in 4:3 mode. So until I have my GH3 you will just have to trust the specs that are listed on the Panasonic pages.

    The sensor dimensions and resolutions in 4:3 mode are identical for all three of these cameras(GH2, G5, and GH3).

    http://panasonic.net/avc/lumix/systemcamera/gms/gh3/specifications.html

  • svartsvart

    @mpgxsvd @thepalalias never said it was the "total size", I said "usable size". Please re-read my post before you preach! But I think you should also be thinking in "effective pixels" as pixel binning and line skipping can yield the same effective resolution with vastly different numbers of total pixels. My point is that people, including yourselves, are using hearsay and unclear marketing terminology to derive a line of thought that doesn't have much fact to back it. We know that the total pixel counts for the GH2 and GH3 are both far above what is needed for HD video, from there panny could be using any number of different types of math to create various aspect ratios. They could also be using something we've never seen before too. Just because it isn't using some marketing jargon, I.E., MAR, that doesn't mean that it's any less effective. For all we know, they have found something superior. I implore you to keep down the supposition for a while until we get more facts. We've already had dozens of posts where people are poo-pooing the GH3 and swearing never to buy one based on only the thoughts of others in these threads and the damn camera isn't even released yet!

  • GH2UWGH2UW

    @thepalalias

    Yes...I am aware of that. But as you say depends on how the GH3's are set up. The point I was making is this sort of material does not show performance compared to hacked GH2 (exactly same scene with cameras side-by-side). That is the only test that is relevant to me for making decision about upgrade or not.

    @Vitaliy_Kiselev

    Your enlightened "bullshit" comment the other day about what I said regarding manufacturers cynically "unlocking" new features in the same camera hardware with only a firmware upgrade is itself "bullshit" in view of recent comment from Canon that:

    "Canon confirmed ... the 1D C 4K DSLR is a 1D X with a firmware update. Identical hardware." $6,000 firmware that is. (canonwatch.com)

    So, as I said, maybe the GH3 hardware is actually capable of much more than Panasonic would like us believe.

    @svart

    But FOV of lens will still be adversely affected. That is the problem using less of the sensor. Down scaling algorithm irrelevant for this.

  • bwhitzbwhitz

    @Astro

    In 3d development, the texture and light is the difference between sterile and realistic, I spend hours tweaking and aging surfaces, developing subsurface scattering on skin to get a real world look, then theres light...

    haha, I'm also in 3D development... interesting. Are we just insane detail freaks? :)

  • meemee

    Pure speculation on my behalf, and I don't reckon it will cool any of the catfights stirring in here.. but.. Is it not possible, neigh probable that Panasonic knows the true professionals will be right here on Personal-view.com waiting for the VGH3K to roll out, and thus they just make a marketable camera for the mainstream, and knowing doors they locked will soon be jimmied open by our beloved VK, they also know that the Pro's and Pro wannabes (mee) will buy their camera soon enough. Again pure speculation, but it seems probably obvious to me. At least if them boyz in the panahood have smarts.

    What I am really trying to say though is... hush thy pussy, and relax till we know more =)

  • thepalaliasthepalalias

    @svart The "active sensor area" is not an ambiguous term. Panasonic can misrepresent it if they want to, but what the put up as the official specifications is specific and can only accurately be interpeted as meaning one thing: what you seem to refer to as usable size.

    They have said that it takes 17.3mm on the GH2 sensor to get those 4608 pixels on each sensor and that it takes 17.3 mm to get 4608 pixels on the GH3 in 4:3 mode. That is is the "usable area" that is employed in that shooting ratio.

    Since the GH2 has an MAR sensor, the active area changes in different shooting modes and that is reflected in the other aspect ratios So 4:3 gets 4608 wide and 16:9 gets 4976.

    So the active and usable area is very explicitly defined for 4:3, without ambiguity, for both the GH2 and GH3. The MAR aspect is inferred from the resolution: no MAR sensor released in a mass produced device has used differing pixel sizes in different aspect ratios. The pixels are a certain size for the sensor so either it adds more of them when it goes wider (and therefore employs a greater active or "usable" area in that mode) or it crops. The version implemented in the GH3 is very clearly a crop approach in the specifcations given.

    The only way I can try to interpret your comment as making sense is that you might be trying to say that Panasonic could setup the GH3 to use more of the sensor in the future in certain shooting modes and thus make it an MAR. Is that what you meant or do I misunderstand you yet again?

  • svartsvart

    @thepalalias yes, to some degree that is what I meant but I also wonder about how they downscale the horizontal or vertical pixels to 1920x1080 resolution. Since we know that the GH2 binned pixels, the MAR action meant that the binning probably used the same number of binned physical pixels to equal a single summed "effective" pixel. In effect, the X/Y ratio stayed the same and necessitated a MAR sensor to do this. Now who is to say that they didn't figure out a new way to do this where the binned number changes or even switches to line skipping in the horizontal/vertical axis to change the X/Y ratio and utilize fewer pixels in either axis in order to optimize something like pixels size, ease of manufacture or simply cost of the sensor? At this point, I'm also just pondering different approaches based on limited data, much like most of the others in this thread. I'm basing all this on the fact that people are seeing aliasing and moire, which isn't a side effect of a non/MAR sensor but an effect from how the data is treated/processed from the sensor, I.E., line skipping/binning. I think once someone get's their hands on a GH3 we'll know a lot more.

  • TronTron

    @jonbeeotch

    Looks like sloppy timewarp in AE ... the editor might have had pixelmotion enabled while trying to increase the speed of the car's launch, then a few frames ended up with interpolated pixels rather than cleanly dropped frames. There was definitely something less-than-smooth in that diner scene and follow-on whip pan. Not sure about the taxi though.

  • doczdocz

    Why are people whining about who made the sensor in the GH3? Does it really matter if it is made by Sony, Panasonic, or Toyota? These electronic products contain a gazillion parts from a gazillion different manufacturers. Can't we just wait and see if the GH3 has got the stuff we need?

    Anyways, by the looks of the P.Bloom short, this thing really can produce some stunning footage, and most definately is a tool to be reckoned with.

    I'm really excited about this, I skipped on the GH2 because I was happy with my GH1, now maybe I'll take a look at that GH3, because all of the new features are really looking promising to me.

    DocZ

  • TheNewDealTheNewDeal

    It's $1295. And most of us will not get our hands on one until 2013.

    Here is my advice: Call a girl and take her to dinner tonight.

  • jleojleo

    Hasselblad is also using a Sony Sensor

    http://photorumors.com/2012/09/18/breaking-this-is-hasselblads-mirrorless-camera/

    http://www.bjp-online.com/british-journal-of-photography/news/2206781/hasselblad-were-not-robbing-people-off-with-lunar-camera

    Yup. Forget the GH3. At least until the Panasonic $300 rebates...meanwhile:

    Dear Arri:

    Alexa 2 for $899. please. With optional Rolls Royce hood ornament. -Thanks

  • GH2UWGH2UW

    @docz

    "I'm really excited about this, I skipped on the GH2 because I was happy with my GH1, now maybe I'll take a look at that GH3"

    LOL...Your decision making process is definitely easier than for a GH2 user!

  • amateuramateur

  • AriaAria

    I'm actually quite impressed with this footage. I think given the newness of the camera and it not having finished firmware, it looks great to me. LOVE the skin tones and highlight handling. It looks a lot closer to what I liked about D800 footage in terms of color. There looks to be PLENTY of sharpness for me.

    All of this from a pre-production model that isn't even fully implemented and the user not having any idea yet of just how to get the absolute best out of the camera, is pretty impressive to me. This just makes me want to see more.

  • bwhitzbwhitz

    Crap. Well, Voldemort video is just another nail in the coffin. He usually absolutely do not know what he's doing with cameras and I'm still only seeing the detail level of a 7D. Texture is abysmal. Everything looks like plastic or clay. Noise and patterns still don't look any better than the GH2 with the right patch (although I'm sure you'll be able to go higher overall). This is a horrible step backwards. I'm be more angry... but the thought of the BMCC soothes my rage. I was really looking forward to the 1080p 60fps stuff, but honestly, the GH2 still seems to have more detail in 720p mode.

    I'm wondering if they just made a compromise, hoping people wouldn't notice the terrible resolution... or if it was done intentionally? I guess it doesn't really matter either way. There's no way I'm buying another fake-1080p camera.

  • 5thwall5thwall

    Anyone know if ETC mode is enabled through HDMI out with the GH3?

  • StrayStray

    @bwhitz Are we sure thats raw footage ? It seems to be half the bitrate.

  • AriaAria

    From what I could tell the lighting in that place didn't look too flattering for filming so I just don't get the negative aspects of what some are saying they see here. Not to mention that this is a PRE-PRODUCTION camera and we know that for a new camera with a new sensor it's quite possible we haven't seen the best this camera can do. It's not as bad as a 7D in terms of detail!!! I mean come on.

    Andrew even said he wasn't able to turn down settings so it's quite possible that NR is set high as are most of the other settings we might choose to tweak if we had a fully functioning firmware. It just doesn't look as bad as some are making it sound.

  • bwhitzbwhitz

    @Stray

    Are we sure thats raw footage ? It seems to be half the bitrate.

    Well, I downloaded the original. Maybe there's a mistake, but even if compressed, a real 1080p image is usually detectable (see compressed BMCC downloads) and I'm just not seeing it here.

    Like Astro said... I'd love to eat my words if this is just some "rough down-sample" in the beta firmware. But as it stands, the "1080p" looks terrible... no better than my 7D, and nowhere close to the GH2.

  • thepalaliasthepalalias

    @svart Ah, I see. The MAR sensor does not play a role in the downsampling/downscaling/line-skpping choice. Since the aspect ratio is the same either way, and the resolution has [TYPO: should be "has not"] been changed by doubling (or similar) it would be unlikely for it to affect the approach.

    However, I am also every bit as interested as you to find out precisely what approach they are using and how it compares to the GH2. Repeated tests have shown that line-skipping results in a loss of detail compared to decent downsampling methods, but my favorites are often considered too slow for real-time use (Premiere CS6 being one exception when GPU accelerated).

    I am no expert on the algorithims, but I have repeatedly studied the varying effects of dozens of them on still images and Lanczos3 is one of my favorites in post because of how well it maintains detail while combatting moire.

  • AriaAria

    Too bad no one tried to shoot the same location that day with a Canon and GH2 just to compare directly with the GH3.

  • TronTron

    @Aria Agree completely. That 7D comment a few posts back is just f*cking retarded... post that super-sharp 7D footage for the world to see!! Are you watching the vid on an iPhone, bwhitz?

    My main concern at this point is not resolution, sharpness or codec... it is mystery moire. Entirely valid concern. I hope Mosaic can cram something in there if it's not a lens correction anomaly.

  • bwhitzbwhitz

    @Tron

    Agree completely. That 7D comment a few posts back is just f*cking retarded... post that super-sharp 7D footage for the world to see!! Are you watching the vid on an iPhone, bwhitz?

    I don't think so. I think your eye's are blurry. 7D footage is not sharp by a long shot... and neither is the GH3. I'm watching on a 24" Apple Cinema display...

This topic is closed.
← All Discussions