Personal View site logo
Driftwood Settings, Series 5: Cluster v6 ULTRA RELIABLE, Mysteron, Crossfire, Quantum 9B, Sedna...
  • 1029 Replies sorted by
  • Nice work @DeShonDixon. USE that rolling shutter to great effect - fuckit - why not - looks fantastic. :-)

  • @conscius as everyone else has said, excellent work. Really enjoyed it.

  • @deshondixon @driftwood I was thinking the same thing...rolling shutter, fuck it, looks great! Nice edit and grade.

  • What is the procedure when one reaches firmware version 99? Is there a way to go all the way back to v11?

  • @MisterBink You can use the same firmware version forever - e.g. 'GH2__V12.bin' or if you prefer naming your version in an order upto the max of 99, just start again - always from GH2__V12.bin. Just remember the last two digits only change in the filename, and always from 12 upwards to be safe.

  • @mo7ies you are right, my fault. Wronk colorcorrection. End of this week I'm gonne update it. And thank you! thanks driftwood and itimjim :)

  • @driftwood in reference to post by @MisterBink ...if my next update was going to be GH2__V14.bin, then I should be ok renaming it GH2__V12.bin from now on? ...but I should stick with naming it GH2__V14.bin, from this point forward, to be safe?

    Was a new setting included in Ptools to accommodate, so that users wouldn't have to update or increase the file name by 1 increment? ...thought that I read this way back, but was still confused on the matter.

    Looking forward to Cluster v6 ...I'm starting to wonder if you and your crew ever sleeps. Have a good one.

    @Bueller ...nice slow motion on the horses.

  • Reference for Cluster v6 Beta Testers: B&W modes. Please be aware that to equalise exposure to other B&W rec modes B&W Smooth's aperture will need to be adjusted. In this example Smooth was adjusted to f7.1 for perfect exposure, where Standard and Dynamic on the same scene were perfectly exposed at f9.

    Pic1. Standard, Pic2. Smooth, Pic.3 Dynamic. Pics taken from GOP6 version of Cluster 6 at 24p, 50 ss, 640ISO.

    Bitrate - In these near static scenes (each running for 20 secs):-

    Dynamic uses 83Mbps

    Standard uses 81Mbps

    Smooth uses 67.5Mbps

    B-W Standard f9 50ss iso640 - Screen shot 2012-06-27 at 18.57.43.png
    2560 x 1440 - 3M
    B-W Smooth - f7.1 50 shutter 640iso Screen shot 2012-06-27 at 18.56.18.png
    2560 x 1440 - 3M
    B-W Dynamic f9 - 50ss iso 640 - Screen shot 2012-06-27 at 18.58.59.png
    2560 x 1440 - 3M
  • @MartinD Yes. You could save it as v12 if you last installed v14 or whatever higher... the GH2 only/always knows (even after you update with new bin settings) that it has v11 installed. Any new bin I test I always give it the same name - clicking on the sdcard and overwriting the old one everytime... I only ever change a number (for my own memory/notes) so I can recall a major setting change/update later...

  • @Driftwood Haha thanks! Rolling shutter doesn't always have to be a bad thing.

  • @conscius that was very well done the story line was well put together

  • Ok this seems a universal issue across the hacks, please tell me what you think.

    Scenario: landscape video, heavy fog covers NYC skyline making it very low-contrast and 2 stops underexposed (according to GH2's meter) so to keep at least some features visible, instead of the bland white screen at normal exposure.

    What I get is mad dancing artifacts (macroblocks?) and uneven arches/bands snaking horizontally throughout the screen.

    Unusable image, really. I tried Cluster v4 and Sanity v5, same thing. Clearly I'm hitting limitation of GH2. I'd like to better understand this, so to avoid unusable shots in the future.

    I also noticed that sometimes parts of a perfectly lit image are also very noisy. Example: beautifully made video here, but look at the lower half of the marble column starting at 0:34:

    Very very distracting noise... this one with Flow Motion... so the issue seems universal across all hacks.

    What is it? GH2 hates shades of gray of certain luminance??

  • @mo7ies it's a misunderstanding of the camera's optimum usable dynamic range, which is about 4 stops. If you don't squeeze your image into this range you will get noise in skin tone and gray images.

    http://www.personal-view.com/talks/discussion/comment/67029#Comment_67029

  • @mo7ies Without looking at your images (I am checking on a mobile) I would say that 1) it sounds like you are running into issues with the issues in the lower exposure range that the GH2 sometimes demonstrates 2) Shian has got the right idea and has tested this a lot, but in my own experience it is not guaranteed that the GH2 will perform poorly in that range - it is a range where it performs inconsistently. 3) I usually deal with the conditions you mentioned by shooting it properly exposed with a high bitrate setting (not CM Night, though, because it has slight issues dealing with haze) and then compensating in post.

    To get back to what @shian said, the GH2 has an optimal exposure range (I will not comment on how many stops) where it performs very consistently. If you shoot your material in that range, you can generally get the results you want in post by modifying exposure, colors, sharpness and contrast. But if you shoot in the less predictable range, you may find the camera behaves less predictably - on top of that, underexposure can lead to elevated noise levels and the camera may have to compress the color information into a smaller range.

    Try re-shooting that kind of scene with a very thoroughly tested intra setting meant for high detail and equally balanced for different conditions. I usually rely on Sedna for that sort of testing. Then shoot it properly exposed and see how the results fair in post. You may be surprised how much detail you can extract.

    Other than that, all the usual shooting advice applies: make sure to get sharp focus through the haze and use focus assist to make sure - also make sure to get deep DOF and shoot at an aperture that gives a good combination of that with close to the optimal resolving power of the lens. For Cannon zooom lenses, f5.6 to f8.0 tends to give optimal resolving power and f11 trades a little bit of resolving power for depper DOF. I rarely stop down further than f11 with them.

    Oh yeah, and make sure that the lens can get enough contrast and resolution to do what you want in the haze. Shoot controlled tests with RAw and JPEG at the same time to reference against. If the lens is not up to job, switch lenses. The 20mm f1.7 does a surprisngly good job at dealing with resolving throustant haze.

  • @thepalalias yes "inconsistent" is a better term.

  • @davidhjlindberg

    Wonderful work. You say you used the Samyang 35mm 1.4 the whole time? Wonderful!!! Quick question: for the AMAZING shot of her by the water, what f-stop were you shooting at?

    Thanks.

  • @shian said:

    camera's optimum usable dynamic range, which is about 4 stops.

    Well, I was still within that range at -2, yes? Isn't the "consistent quality" range -2 to +2 stops?

  • @mo7ies it's really difficult to diagnose without seeing the footage. There are a ton of factors involved, one being the film mode, the other: the specific hack (also quality 24H, 24L, 60p, 25, HBR...??) and you'll have to bring @driftwood into the conversation to tell you how the matrix for that hack is interpreting the frame. And depending on which zone that color is supposed to be exposed in, versus which zone it was actually exposed in.

    And as for the vid you posted: You can't use a vimeo vid as a true example. I've shot stuff that was completely clean, and picked up that kind of noise during mp4 export and/or vimeo conversion, but it wasn't there in the raw or graded footage.

  • @mo7ies It depends on where you started. If you are basing it on in-camera metering, then the metering mode comes into question to start with. If it was center spot metering, then it depends on what area you pointed it at, etc.

    In a scene with a wide dynamic range, the metering can get tricky and inconsistent. Ever watched a camera try to deal with metering a wide shot of a sunset? It gets pretty amusing. But if you ever want a more consistent indication, you can use the flashing clipping indicator to establish the upper exposure limit for the area and then start counting down stops from there. @Shian could give you better advice on how many stops down to go, and I bet @Driftwood could to. But it is a much more consistent way of checking than standard metering, at least in my experience.

    Anyway, try shooting the same area with Sedna not underexposed and then bringing the detail out in post to see if you like the results better. I prefer the way GOP1 settings handle haze and shadows, because you do not find tiny detail getting "stuck" for a frame or two, which sometimes happens with Sanity for instance (though I like Sanity a lot).

  • @conscius Great story, flows well.. just damn good.

  • Quantum X v4b, will have to test out Sedna AQ1 soon! :)

  • @retrospective real nice! A couple of shots (:15, 1:05 for example) look like a slider, if so - which one did you use?

  • Cluster v5.1

    00030.MTS - 1.png
    1440 x 900 - 1M
    Subrail Dubs Spine Flat.MTS - 1.png
    1440 x 900 - 1M
  • @mo7ies thanks! It's an indiSLIDERmini, not the smoothest slider out there.

This topic is closed.
← All Discussions