Personal View site logo
Official Low GOP topic, series 5
  • 751 Replies sorted by
  • RRRRRRRR

    @bkmcwd

    On tuesday I was out trying your latest golgop.. I planned to shoot nature "stills", in other words tripod shots with set frame. It turned out there was a lot more movement than I expected in the frame but your settings held up remarkably well! Graded beautifully.

    I hope to be able to show you the results after tomorrow as my show opens then and all my work has been towards that during the last week.

    With a bit of luck I might be able to squeeze in a vimeo upload tomorrow.

  • flaschusflaschus

    @ towi.. WoW! nice work... quick question...

    whats the difference between your HQ fallback matrix and no fallback??

    unaltered screenshot: gop3zilla2.0 w/towi HQ matrix w.fallback 14mm@ 640iso, 3.5.

    sorry double uploadedd

    vlcsnap-2012-03-15-17h51m43s15.png
    1920 x 1080 - 3M
    vlcsnap-2012-03-15-17h51m43s15.png
    1920 x 1080 - 3M
  • bkmcwdbkmcwd

    @flaschus About qp, I am also still studying. And I am sorry, about a matrix, I understand nothing yet.

    @towi Thanks for feedback mate. :-)

    Since it is difficult for me to check all the modes finely, your result helps me very much. I will improve.

    @Bonzai Thanks for your footage, Bob! :-) I think that the result of qp is good as HBR. Noise is unavoidable in case of the dark lens of Pany IMO.

    @RRRR Thanks mate. :-) I am looking forward to seeing it.

    GOLGOP3-13_v4.1_HBR_sample1_ey.JPG
    1193 x 661 - 82K
    GOLGOP3-13_v4.1_HBR_sample2_ey.JPG
    1193 x 661 - 96K
  • bkmcwdbkmcwd

    @towi Although I repeat the test of 25p with natural v2.0 repeatedly, I cannot achieve the same results as you. Would you tell me your test conditions? Are any matrices possibly combined?

  • towitowi

    @flaschus

    "whats the difference between your HQ fallback matrix and no fallback??"

    Here's a brief explanation of fallback mode by LPowell: http://www.personal-view.com/talks/discussion/comment/38868#Comment_38868

    @bkmcwd

    "Although I repeat the test of 25p with natural v2.0 repeatedly, I cannot achieve the same results as you. Would you tell me your test conditions? Are any matrices possibly combined?"

    Yes, matrices were applied. Also FSH BR lowered to 72M (for PAL mode & SanDisk 30MB/s cards).

    Just did a quick test (HBR 25p @ 72M) without matrices and the stream looks healthy. Sorry for the confusion!

    v2_HBR_72M.PNG
    746 x 398 - 44K
  • duartixduartix

    @bmcwd : I'm amazed!!!

    The cadence on this file you published for natural v2.0 is amazingly solid. Even if it would fail in some circumstances on death charts, no evil would come to the world. The final objective of the settings is not filming death charts.

  • towitowi

    @duartix

    "Even if it would fail in some circumstances on death charts, no evil would come to the world. The final objective of the settings is not filming death charts"

    While this is of course true highly detailed landscape shots (especially with slow pans) are much more demanding for the codec. This is why death charts provide a good starting point to test the reliabilty of settings... IMHO.

  • duartixduartix

    @towi : You are right. Death charts are good at stressing FULL INTRA settings. A detailed seascape can provide even more stress through motion to P/B based settings.

  • driftwooddriftwood

    Simple terms: Fallback mode is just very high matrix allowing every Q possibility through and not a particularly good picture.

  • bkmcwdbkmcwd

    @duartix

    Thanks mate! :-)

    @towi

    "While this is of course true highly detailed landscape shots (especially with slow pans) are much more demanding for the codec."

    In my idea, surely, although a motion applies stress to a codec, it does not confuse a codec. It only raises up qp. A codec is not confused by the stress of a motion in my newest settings. However, since stationary highly detailed landscape shots was sometimes confusing the codec, I have solved it. For me, when using a codec normally, the most difficult object is a green tea field.

    One _Of_GOP3ZILLA_Test_setting_Green Tea Field.JPG
    1294 x 608 - 161K
    One_Of_GOP3ZILLA_Test_Setting_Death _Chart.JPG
    1296 x 609 - 171K
  • duartixduartix

    @bkmcwd : That chart has a lot of repeating patterns. I wouldn't be surprised if the motion estimator got confused. I usually use one of these instead.

    However, there's a penalty, it's harder to assess IQ.

    975c857fa81b1adcae5bf884f0c2d0.png
    1920 x 1080 - 5M
    Colour_Texture_06_by_Tackon.jpg
    1920 x 1080 - 4M
  • bkmcwdbkmcwd

    @ALL :-)

    I release the newest "GOP3ZILLA_V2.1" and "GOLGOP3-13_V4.2" setting. These are also TEST settings for using low qp with 3GOP as same as previous versions.

    I am not sure whether these are better than old versions. NEED TEST. However, I improved so that a codec might work more stably in 24p.

    Of course, this is also the setting focused on image quality with LARGE I-FRAME SIZE except 60p and 50p.

    "GOP3ZILLA" is almostly all-round high image quality setting especially in 24p.

    "GOLGOP3-13" is also all-round but especially good at indoor or night shooting in 24p.

    In some case, the old versions may be the more nearly better. Please try them.

    NOTE:

    *These settings are focused on especially 24p mode.

    *24H mode is high quality 3GOP setting with mega size I-frame.

    *24L mode is for 24p80%(30p).

    *In camera playback is available.

    *EX Tele works.(Not tested yet, but probably)

    *Spanning in 24H, 24L, maybe in FSH, HBR, and SH with SanDisk 95MB/s 64GB SDXC Card.(Not tested yet, but probably)

    bkmcwd GOP3ZILLA_V2.1.zip
    726B
    bkmcwd GOLGOP3-13_V4.2.zip
    721B
  • bkmcwdbkmcwd

    @duartix

    Thanks mate. Although I also have them, I have seldom used. From now on, Ill use. :-)

  • flaschusflaschus

    @driftwood @towi

    thx for the explanation on fallback!

    @bkmcwd

    yeaaa! a new gop3zilla monsterr!
    many thanks.

  • towitowi

    re @bkmcwd Natural V2.0 ...

    testing goes on… some notes on 24H and HBR in PAL mode (25p):

    • the settings seem to work fine even with SanDisk 30MB/s cards (maybe not for the most demanding scenes, but generally 30MB/s cards will survive most shooting scenarios)

    • HBR 25p spans on SanDisk 30MB/s cards (as long as the scene does not require to utilize max bitrate)

    • settings span on SanDisk 95MB/s 32GB in HBR mode (25p)

    • FSH (50i) may show an uneven cadence of the streams. Not noticeable in the actual footage

    I've compared bkmcwd Natural V2.0 with different matrices (my HQ-matrix, Driftwood's Sedna A, Driftwood's Orion, LPowells Flowmotion):

    • my HQ matrix and Driftwood's Sedna A look good (the other matrices do not)

    • "HQ" shows somewhat less flickring at edges, however the streams may show some (minor) irregularities (while the stock matrix always shows healthy streams so far)

    • "Sedna A" shows the most gentle fall off in the blacks of all matrices, however also with Sedna A the streams show some irregularities

    • All in all: for the "bkmcwd Natural V2.0" setting I'd stick to the stock matrix...

    • bkmcwd rocks!

  • KihlianKihlian

    Here a comparison shot with FD 50 at F 5.6 on 25P

    Gop3zilla with towi matrix 44830 Kbt/s

    Sedna A Q20 92265 Kbt/s

    Gop3zilla with Sedna A matrix 61808 Kbt/s

    My impressions are that the Gop3zilla featuring towi is much more sharper on 25p that Sedna, the differences between Gop3zilla with and without sedna matrix are not significant, but when I playback the video, with the sedna matrix there is less noise. In attach the gop3zilla with the sedna matrix

    Gop3zillaTowi.png
    1920 x 1080 - 3M
    Gop3Sedna.png
    1920 x 1080 - 3M
    SednaAQ20.png
    1920 x 1080 - 3M
    setd.ini
    6K
    setd.zip
    992B
  • StrayStray

    @kihlian Have you tried these settings much with 24H at all ?

  • KihlianKihlian

    @Stray Ops! forget to mention... I shoot it all at 25p...

  • fehafeha

    @Kihlian

    what is the difference between your two sets above ?

  • KihlianKihlian

    Are the same, the second is a ZIP file so it could be downloaded, was my mistake..

  • fehafeha

    Ah Ok, i tested i like it ... btw where can i download : bkmcwd Natural V2.0 ? also does bkmcwd Natural V2.0 work in HBR 25p ?

  • KihlianKihlian

    @feha You can download it in the pages back of this topic, works in 25p but how @bkmcwd says it could not archive so good results as in 24p, or like @towi settings. I think @towi is working on NaturalV2.0 for optimize it in 25p, hope to see a setting from him soon.

    @towi (P.S. sorry to speak or interpret for you)

  • bkmcwdbkmcwd

    @flaschus

    Please try them, mate. :-) I am not sure which is better, new or old...

    @towi

    Thanks for many tests, mate. :-) It is natural supposing the stock matrix is most stable with my settings. It is because I have set so that it may be stabilized with a stock matrix.

    @Kihlian

    Thanks for interesting comparison.

    Why is it that a horizontal striped noise is visible only to screenshot of Sedna A Q20?

  • bkmcwdbkmcwd

    @feha

    Thanks!

    My newest settings in which 25p probably works are GOLGOP3-13_V4.11, GOLGOP3-13_V4.2, GOP3ZILLA_V2.0, GOP3ZILLA_V2.1 and natural_V2.0. http://personal-view.com/talks/discussion/comment/42201#Comment_42201

    These are almost same in 25p.

  • thougts2ukthougts2uk

    I have a couple of questions...

    How does the GOP3ZILLA compare to Sedna A AQ1 (which is supposedly the highest quality GOP1 Intraframe).

    Also would I use GOP3 instead of GOP1? What advantages does GOP3 have?

    Which is the best GOP3ZILLA patch at the moment? I am so confused with v2.0, v4.2, v4.11, v2.1??

Add a Comment