Personal View site logo
Driftwood VY Canis Majoris: This thread will now be integrated into 'Driftwood Series 6 Settings'
  • 901 Replies sorted by
  • @ishvar what post did you do? That's quite a streetshot!

  • @jhero Youre totally correct. I do all my testing on Standard profile as of Canis. Hey, and none of those pictures I put up were graded yet. Theyre looking exceptional under devinci and in fcpx

  • @ishvar Forget the settings for a moment. Those kids were brilliant, wonderful scenes of joy.

  • I couldn't agree more. Those kids brought a big smile to my groggy early morning face. Thanks ishvar for posting. Great stuff!

  • @driftwood amazing work! i used the CM night as directed...in performance/installation space - object lighted properly and the rest dark...well i am blown away by the results...i don't think i have ever seen anything like this so far from this little camera! to the people that don't see the differences from stock to (latest)patches here - you need to check you vision...i had relatively fast zoom f2(old) lens but it normally gives a little too soft image at f2 with CM night it looked sharper then usual...maybe just the colors stick out more...a big THANK YOU Nick!!! p.s.i am broke as hell now...but the moment i get a pay check will be donating. i think all the professional(people that make a living out of film/video, i am not one...) and use this site, patches information, etc...should be donating big time to Vitaly, or its bad karma! ; )

  • And still, just as with every other single patch I've tested, horrible blocky splotching in low IRE skintones. It's readily apparent in the first picture posted on the bottom of page 3. Look at the right side of his face, A bit over middle gray exposure wise, perfectly fine. The let side of his face, low IRE, let the fun begin. There's more or less just 3 giant blurry gobs of color there. I love this camera, but maybe I've been spoiled by my ex1 for years. I have never, ever had this issue with it, and I have shot probably in the hundreds of underexposed skin. An that was mpeg-2 35mb/s. Really wish someone would address this and stop blaming it on 8-bit space. (protip - it has nothing to do with bit depth. It is not banding on a smooth noiseless luma gradient. All those beautiful, smooth, barely lit skintones you see on your favorite bluray or dvd? Those are in 8-bit space, and at a much lower bitrate than these patches)

    What the eff is going on here! It is so annoying! I've been using my ex1 to shoot pick up shots when I know a subjects face is going to be in low light. Less than Ideal

    here's a cap from sedna aq1 - literally the ugliest skintone compression I've ever seen - http://img848.imageshack.us/img848/8484/tfff.png

  • @Fohdeesha i think this ex1 has one more digit added to the price...is this a possible difference?

  • Yes, a financial difference. I wasn't aware the amount of digits in a financial indicator had anything to do with compression coefficients, but I might be wrong!

  • red is not alexa no matter what and there is a price difference in this (obvious) sense...

  • and btw is stock any better on the gh2? and if you know a thing or two on the subject of compression coefficients why not get a better setting here...vitaly has made it possible for anybody with knowledge and time to make it better...

  • @fodeesha I think when we look at the video performance potential for the GH2, a better reference point than other cameras is the camera's own JPEG mode. Why not take two identical shots (one a JPEG shot in stills mode, the other a frame from the video mode of your choice) and compare them to illustrate the presence or absence of the performance delta you perceive?

    Clarification (in light of the comment below): I am not saying that we should not compare the camera to the EX1 or any other camera. I think that it's good to get concrete information about any such comparison. But in terms of looking at developing settings, it seems that the best starting point is "what is left to unlock on the camera?"

  • @Fohdeesha Are you talking about comparing the Sony EX1 professional camcorder that retailed for nearly $8,000 US to the $700 US GH2 camera?

  • Why bother? That post was meant to stir up trouble. I suppose if anyone says anything negative about his post that we'll be labeled in the same category as Red fanatics. Honestly though there is nothing constructive about it.

    Plus...I personally dont see anything wrong with his still shot. What are we supposed to be looking for?

  • @Fohdeesha is quite correct yet your 'and still' announcement disregards attempts in trying to improve things. Its fiddly exposing skintones correctly on the GH2. Achieving around 50% to 70% (variable skin colour) IRE scale is easier on my EX3 and other dedicated video cams/decent monitors with zebras. The light on the subject matter and surroundings can also change in whatever conditions so you have to adjust. This in turn makes it harder to test settings on a gh2. Its not perfect but its getting better. The picture profiles on the GH2 have yet to be cracked and so we are left with a fight with the few ptools settings we have to make the best of things. Incidently, the picture you refer to was under the 'standard' profile. Filters can help and there are plenty of topics on the case. Indeed the EX1/3 is not perfect.

    Adjusting red saturation in the matrix is one of the things Im continuing to delve into. This takes many hours/days to test for the few testers doing settings; please by all means suggest chroma and luma frequency adjustments to me at any time - and try and help!

    Provide detailed analysis and send identical images where you have had great success with the EX1 in similar conditions with the GH2.

    I suggest you test Orion. This is using the official Panasonic AVC Intra matrix. Now you will realise how difficult it is.

    Finally, the picture from Sedna you show us isn't brilliant is it?! Moving vehicle, moving hands, very moving... OIS on the EX is certainly better for this kind of shot. Just realised you 'joined' this site very recently so its about time you changed your name again.

  • @thepalalias this is a very good point. i meant to say this for a while, the latest patch developments have narrowed this jpeg - video difference. things are not perfect but we are in uncharted territory here...in my experience light conditions,film mode settings, lens used etc do make (sometimes very big) difference in outcome...but like any "new tool"(patched gh2) it needs its time to be improved...testing, testing...

  • Shot some live music under very challenging light last night using CM Night - looks brilliant! Even ETC mode looks very strong, yes it's grainy, but to me it has a very pleasing look like the grain you see on blu-rays of 60s music docs, like Monterey Pop.

    I was mostly using ISO640, but once I had my primary material I think I shot a little more at 1250 and above just to experiment. Again, all looks awesome to my eye and taste at least. Sure it's not clean, but it is definitely organic!

    It's for an ongoing music video project so I can't post anything yet but I'm super happy with it so far.

    The only downside was that I did have a couple minutes of footage flat-out go missing - and I noticed several times recording stopped without an error, just stopped, at various points in clips. Nowhere near the 4 minute point either. However when the recording stopped in one instance it did seem to have taken my footage with it, sadly. Sandisk Extreme 32gb 30mb/s card.

  • Here are frames from one of the many comparisons from the last couple days that I have deemed a failure (read: almost every comparison involving natural light so far :) But they still serve a use in regards to resolution.

    The reason I include them is because in looking at small text in each picture, you can the resolution gap that remains between video mode and JPEG in the finest detail.

    The second set is labeled by when they were shot, and has had the contrast and brightness pushed strongly.

    Let me be clear: the light changed too much during testing in each of these tests. The JPEGs shot at the start and end of each test did not match up sufficiently. Do NOT use these as a reference for other things.

    Lens: Rokkor 58mm f/1.4

    24H (except for the still)

    1/50, f/5.6, ISO3200

    Smooth -2, -2, -2, -2

    5000K WB - illuminated by outdoor sunlight through visible window only.

    P1050125.JPG
    1920 x 1080 - 1M
    3 CM Soft Skin Tone0000.png
    1920 x 1080 - 2M
    4 CM Night0000.png
    1920 x 1080 - 2M
    2 CM Day PM0000.png
    1920 x 1080 - 2M
    1 GOLGOP0000.png
    1920 x 1080 - 2M
    11-08AM GOLGOP3-13_SednaAQ1C_V2.1_BETA4.png
    1920 x 1080 - 3M
    11-16AM CM Day PM.png
    1920 x 1080 - 3M
    11-23AM CM Soft Skin Tone.png
    1920 x 1080 - 3M
    11-31AM JPEG Still.png
    1920 x 1080 - 2M
    11-29AM CM Night.png
    1920 x 1080 - 3M
  • @Oedipax ETC in low light and a few other conditions is big no go zone for me...stock,patches, whatever, it only works fine in bright daylight .

  • @luxis, I'm inclined to agree except CM Night really surprised me. But the look we're going for in this video is definitely a 16mm kind of feel, fuzzy lenses, haze filters, all that to evoke a certain mood. So for that it's perfect. In a narrative context, unless deliberately and carefully chosen, I'd still say it's dangerous territory :)

  • @Odeipax I think @luxis was referring to the ongoing stability issues he's suffered with ETC. I've been lucky enough not to have them to the same extent, but it sounds rather frustrating.

  • it'd be cool to have multiple matrix/settings in a way akin to colour/style profiles in the GH2, rather than having to flash the firmware multiple times. :)

  • @driftwood thank you so much!

  • @rajamalik In response to your e-mail, here are some more frames from the daylight footage with the different settings. Note from the two JPEGs at the start, just how much light had changed between the start and end of the scene.

    You have both the JPEG still and "GOLGOP3-13_SednaAQ1C_V2.1_BETA4" using the Sedna matrix to reference against (and the latter was shot as close after CM Night as I could). As I said in the e-mail, not close enough in lighting conditions to judge which is best, but you can see the CM Night is not limited to high ISO or low light shooting.

    ISO 200, f/4.0, 1/50

    Also, in the very limited skin tone testing I did using photos contained in the Canon EF Lens workbook, CM Night did not appear any less capable than the other settings shown here. However, the light was changing even more rapidly by that point and I do not own the rights to the image, so I cannot post any images from the test.

    P1050118.JPG
    1920 x 1080 - 618K
    P1050119.JPG
    1920 x 1080 - 593K
    1 CM Day PM0000.png
    1920 x 1080 - 1M
    2 CM Skin Tone Soft0000.png
    1920 x 1080 - 1M
    3 CM Night0000.png
    1920 x 1080 - 1M
    4 GOLGOP3-13_SednaAQ1C_V2.1_BETA4.png
    1920 x 1080 - 1M
  • Canis Majoris 'Night'.

    Hope to get some better shots this weekend.

  • Canis majoris FSH 720p mode does not work ... error writing :/ Sandisk 64GB 95MB / S

This topic is closed.
← All Discussions