Personal View site logo
Driftwood VY Canis Majoris: This thread will now be integrated into 'Driftwood Series 6 Settings'
  • 901 Replies sorted by
  • @driftwood. Nick, do all of the variations of Canis Majoris use VBR and the only difference is the matrix values that more closely match the time of day?

  • @HillTop1 No settings will help you to eliminate RS, it's inherent to current CMOS technology. If you are concerned about fast pans only, the filter from The Foundry or the one in the new Adobe version can fix it pretty well. Different motions in the frame can't be fixed perfectly.

    And if you have further questions about RS, please don't derail this thread, there have been long discussions on the subject.

  • @thepalalias,

    "If you really want to use lower shutter speeds, you can use VMM to compensate. If I use a 1/15 shutter, I use VMM 300% to maintain 180 degree rule."

    I've just checked that this line dissappers starting from the shutter of 30 and below. If you start getting the shutter upper during recording then you'll notice this line goes upper on your screen and it doesn't go away even on the shutter of 100. I've checked it using ISOs 1000-12800.

    @thepalalias, could you explain what the VMM is? Is it some sort of a video filter?

  • VMM - Variable Movie Mode

  • @matthere, thanks a lot.

  • Guys,

    Regarding this LINE issue....

    This issue has been discussed already (since August 2011) and the thread is "GH2 showing fixed darker band/stripe 20 pixels high across footage". This is a hardware problem and it cannot be fixed...

    @driftwood, VK, sorry for doubling this line issue in the current thread.

  • Well we gonna have learned a lot about this horizontal line and we are not careful at the beginning: D thank you for the explanation and use of VMM

  • Canis Majoris 'Night' Smooth -2 -2 -2 -2 24P, 1/50 14-140mm kit lens. Performed well in the low light areas ...ISO from 160 to 2500.

  • Yes all Canis are VBR so if theres not much in the image there's going to be bitrate reductions. You can always set it to AQ1 if you feel a must for permanent high bitrate which is unnecesary.

  • As promised I've been doing some side by side testing with two GH2 bodies, one hacked with CM Day PM, and one left completely stock, unhacked. Settings were ISO160, standard mode, and sorry I forgot to play with the picture profile so it's all 0,0,0,0. Lens was a Zeiss Planar 45mm f2 (set at f 2.8 and focused on the rock in the middle of these crops) from a Contax G2. Exposure was as recommended by the camera, which produced a good spread over the histogram.

    It's a grey, overcast day in London, and it just started to rain as I changed the lens over to the other body, but conditions were otherwise reasonably consistent over the course of the 4 minutes of test shooting. I tried a static shot and some pans of various speeds. These are 200% crops, overlaid, rendered, and screen grabbed directly from FCP7.

    First up, 720p50 - Not much to choose on the static, but the low bit rate stock codec clearly breaks down on the pan. It's even more obvious going through frame by frame, there is a lot of 'mud', big nasty blocky areas. The hacked Canis Mojoris Day PM footage has no such problems anywhere. Clear win the CM Day PM.

    1080p24 and HBR 1080p25 - There's no question that in motion, there's more grain on the CM footage, though it's hard to see on these stills. But from here, the additional grain does not seem to equate to extra detail. Beyond that....... well, you decide for yourself, opinions would be appreciated.

    At some point I will try a scenario where the codec really fails, which in my experience is something like a soft grad on an underexposed monochrome wall in the background of an interview setup, which picks up unpleasantly digital noise.

    720p static.png
    1322 x 753 - 1M
    720p pan.png
    1325 x 754 - 708K
    1080p24 static.png
    1322 x 753 - 1M
    1080p24 pan.png
    1322 x 752 - 731K
    HBR 1080p25 static.png
    1323 x 752 - 944K
    HBR 1080p25 pan.png
    1322 x 752 - 624K
  • Superb work @Driftwood (and of course @Vitaliy_Kiselev - Thank you guys). I see most of the tests are at 60 fps, but what is the general reaction with 24p? I'm going to test now!

  • @Driftwood Do you have any preferred settings for VYCM Day PM? Standard/smooth etc?

  • @Itsguy Look at the rock on the right between the two in 24p! Much more detail. Besides that do some grading.

  • @istguy, I am rocking a 720p monitor, so the 1080 stuff likely has detail I am not seeing, but I also see more detail on the rock on the right, and on the left rock near the lower right corner in the dark area where there are horizontal slightly curved upward dark lines it has more detail creating more of a shadow effect. On top of that it seems to have a bit more depth in the area just above this. It is closer than I expected, but that's what I can see on the 24p. On the 720p, I see more detail and depth all around on both rocks, but it is most obvious with the rock on the right.

  • GF2 test results -1080p30 (daylight 14mm)

    Canis Majoris DayPM - 79.6 Mbps

    Canis Majoris Night - 81.8 Mbps

    Canis Majoris SkinTone Soft - 83.5 Mbps

    CanisMajoris-DayPM-gf2.PNG
    746 x 412 - 54K
    CanisMajoris-Night-gf2.PNG
    746 x 412 - 55K
    CanisMajoris-SkinTonesSoft-gf2.PNG
    746 x 412 - 55K
  • Canis Majoris gives a really good result, for I am against how if I have to go out all day: D, between the Canis, Night, Day AM / install AM...je sees me wrong every time, the version for AM morning, PM for after noon and night for dinner ... in this case as well stay on Sedna AQ1 not? or take PM Canis day, just to do both :/, unless a version of morning, noon and night, night planned ... well thank you for the work you have accomplished.

  • @Itsguy Thank you for your useful tests. For me. they confirmed once more that the hacked gh2 is a far superior machine in every respect. Also, if you look a frame at a time, you will see very clearly the difference among I frames and not I frames, which make me to use only 1 gop for my narrative works. In particularly I must say I did a shot last night under the rain with CANIS NIGHT and... well, I can't believe my eyes, it almost looks like film. (Nothing looks exactly like film IMOH these days, not even RED or ALEXA, just compare The girl with a dragon tattoo with its former film version, for example. But the GH2 with CANIS is very very close!!!!)

  • I made ​​a small donation, it's not much, but it is mostly to thank you for your excellent work. :)

    Good luck and thank you, thanks to you the GH2 takes each day, a small facelift

    Jimmy

  • Sorry, I just need please the Spanmybitchup V2b but impossible to open here the file (with my Macbook). Someone can post me parameters. MERCI BEAUCOUP.

  • @Itsguy two different rocks? ; )

    @driftwood any cluster v2 in the projects? thanks!

  • [Information] Comment=Quantum X 'SpanMyBitchUp v2b' For v1.1
    As AVCHD-I settings except
    1080i/HBR on 100M Intra
    720p on GOP6 100M max
    Note: 24L, 720/1080i FH/H settings on Trick mode (encoder
    calcs bitrate) SD_Card=type 10 cards Camera=GH2 v1.0E Driftwood [Settings] Version increment=1 All Interface Languages=Checked 30min limit removal=Checked Maximum ISO limit removal=Checked PAL<->NTSC Menu=Checked 720p30 width=1920 720p30 height=1080 480p30 width=1280 480p30 height=720 E1 Quality=512 E1 Table=4 E2 Quality=422 E2 Table=26 E3 Quality=422 E3 Table=26 E4 Quality=486 E4 Table=6 F1 Quality=454 F1 Table=2 F2 Quality=224 F2 Table=36 F3 Quality=224 F3 Table=36 F4 Quality=454 F4 Table=8 Video Bitrate 24H=100000000 Video Bitrate 24L=5000000 Video Bitrate FSH/SH=100000000 Video Bitrate FH/H=5000000 720p50 GOP Size=6 720p60 GOP Size=6 1080i50 and 1080p24 GOP Size=1 1080i60 GOP Size=1 Encoder setting 1 720p=3 Encoder setting 2 720p=1 Encoder setting 3 720p=2 Encoder setting 4 720p=4 Encoder setting 1 1080i/p=1 Encoder setting 2 1080p=1 Encoder setting 3 1080p=2 Encoder setting 4 1080p=4 Encoder setting 2 1080i=1 Encoder setting 3 1080i=2 Encoder setting 4 1080i=4 Initial quantizer=16 Quantizer for 1080 modes=16 Quantizer for 720 modes=16 1080p24 FB2=520834 60fps FB2=472000 50fps FB2=472000 1080p24 Frame Limit=4166666 60fps Frame Limit=1637165 50fps Frame Limit=1997128 1080p24 High Top Setting=95830 1080p24 High Bottom Setting=67081 1080p24 Low Top Setting=47913 1080p24 Low Bottom Setting=33539 1080i Top Setting=95825 1080i Bottom Setting=67077 720p Top Setting=95825 720p Bottom Setting=67077

    Je ne sais pas si c'est ca que tu cherches, jt fait un copier coller du fichier en question :)

  • @gokuan

    YES.

    THANKS YOU!

  • Just a question off topic ... I bought a mini HDMI cable, a small price, quality average, less than $ 10, do you think the image quality suffers on a full HD tv or is it I think a better buy me a better quality cable such as "Panasonic Mini HDMI Cable - RP-CHEM30E-K"

    last quick question, what are the glaring defects that could perceive, the cable has a low price?

    If I can enjoy the real potential of the Canis Majoris ... with a cable that would do credit to hack in question: D

    Thank you.

  • @gokuan003 I have never had a cheap hdmi cable be anything less than great. As I've come to understand it either works or it doesn't.

  • Some cheaper HDMI cables have poorly made jacks which can split if bent slightly, so the pins can short out. But I picked up a HDMI cable at a dollar store and it seems OK.

This topic is closed.
← All Discussions