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.
'Apocalypse Now' Experimental Series 1 Thread - BOOM, Intravenus - cbrandin/driftwood AN Soft/Cinema
  • 1089 Replies sorted by
  • @Driftwood Any idea when the new Intra 444 will be released? Days or weeks? Very curious.

  • Here are unedited 200% crop frame grabs from back-to-back videos taken with a Lumix X 45-175mm f4-5.6 @ 45mm f8, ISO 200. The first was shot with Flow Motion v2.02 and the second with Apocalypse Now 6-GOP 444 Soft.

    The first pair of shots are uncorrected. In the second pair of shots, I boosted the gamma to show more of the details.

    Flow Motion v2.02 @ 24H: image

    Apocalypse Now 6-GOP 444 Soft @ 24H: image

    The FM2 video has an average bitrate of 88Mbps with a DC quantizer of 5, range 5.

    The AN6 soft video has an average bitrate of 75Mbps with DC quantizers of 5-7, range 5.

    FM2-Electrical Towers 24H.jpg
    1024 x 594 - 209K
    AN6-Electrical Towers 24H.jpg
    1024 x 594 - 214K
    FM2-Electrical Towers Gamma.jpg
    1024 x 594 - 262K
    AN6-Electrical Towers Gamma.jpg
    1024 x 594 - 242K
  • @LPowell - I'm curious, how were the screenshots taken? I ask because I've run into issues with screen grabs with Vegas, for example. They offer two formats for screen grabs JPG and PNG. The problem is that they come out somewhat different, so I'm never sure whether what I am getting is accurate. When it matters, I use a different method to produce uncompressed screen grab files. But even with that it's hard to be sure that it is actually what is in the image. Some screen grab utilities, for example, capture what is being sent to the display system - which has been processed for screen gamma, monitor color space, etc...

    It seems to me that a lot of people have to be running into such issues. Can you suggest a method of capturing frames that most people will have access to that doesn't involve recompression, display adjustments, etc...?

  • @Zaven13 I have to agree with @cbrandin, basics, long lens work, images subject to heat rise in the mid ground to the distant. I remember using heat generating tools to get just that effect shoot on a long lens.. or shooting a long road shot and the heat rises from the tarmack. Out of curiosity, what was that lake side store producing? and what lens were you using?

  • @cbrandin To produce frame grabs, I use Adobe After Effects CS5.5, which is one of the few video editors that support color management via ICC profiles. When using AE, it's important to set the Project Properties to 32-bit and Rec. 709 profile before importing your footage. This insures that the color space of the footage will remain unaltered.

  • Oh I feel so chilled out here in France.

    @cbrandin Screen grabs are down to your monitor's pixel resolution size. Here's one on an apple macbook air at its native 1280x800 res.

    @LPowell Remember, Cluster 6 GOP is set up to use less bandwidth under GOP6 than your GOP3. Indeed what you're suggesting is AN is actually more resourceful than FM and produces similar dc values looking at your charts.

    I actually use Elecard's YUV viewer and picture quality estimator. I wouldn't travel through any third party or middleware to analyse footage.

    If its not too much trouble, please could you link me to your mts footage so I can evaluate how each are looking - so I can add it to my database. I have already analysed Cluster v6 against Cluster v5, FM, Sanity - checking dc values and was very happy with the Cluster v6 results. Cluster v7 with 444 matrix was built under similar techniques but is in NO WAY a finished product - its a simulated matrix which does extremely well under certain conditions.

    Alas, Im always intrigued by your choice of subjective tests.

    below: A picture from AN 'Boom!' sharp for GOP1.

    Lo res screen grab from AN on Air.png
    1280 x 800 - 2M
  • @LPowell - I took details from the grabs you posted and lifted the shadows so I could see how noise is handled (i.e. I set levels to 0-128). These also show how detail is rendered differently - look at the rivets. Interesting stuff...

    FM2:

    image

    AN6:

    image

    444.jpg
    268 x 217 - 66K
    FM2.jpg
    266 x 216 - 74K
  • @LPowell @cbrandin Great work guy's from what I see from the results AN 444 Soft is a keeper, The reason I'm saying so it's because it renders the edges much better to my eye it's smoother. Let's not forget about the noise... once again AN 444 Soft is way more pleasing..

  • @driftwood That look great man...

  • @driftwood, yeah, great looking shot there!

  • @cbrandin Yes, the comparison looks pretty much like we'd expect. AN6: softer edges and smoothed details. FM2: Sharper edges and textured details.

  • @LPowell - I think @bkmcwd is about to release settings that I know for sure are using the soft pseudo 444 matrices in their pure, unmodified form for 1080p24. It would be interesting to see your test shots with that. Some of these other settings are modified variations of my original matrices which makes them a little difficult to evaluate, at least from a matrix standpoint.

  • Is this macroblocking?

    atom.png
    270 x 305 - 24K
  • The one thing I haven't looked at is how AN Drewnet Soft handles 720p I already know FM works pretty well in 720

    This is a quickie in AN. Without pixel peeking, it look fine. I'll try it on a moving subject later.

    460row.png
    1280 x 720 - 771K
  • @peternap: LOL, no kidding about the peeping.

  • it works fine in 720p.
    its not the original AN setting, but a gop2 720 setting.



    but there is a change in the profiles, shots are done in cinema -2 -2 0 -2 WB 5400k
    Cinema looks like Vibrant, ( i dont have vibrant examples but the collors are to much, specially green)..

    Again in cinema.
    First part ungraded, second part graded and overdone.
    those 2 clips are the first test i have done
    I have still a long way to go...


    Upscaled to 1080p

  • @I have still a long way to go...

    When you get done post it please. Those are great!

  • @LPowell @cbrandin What did you expect with a FM / AN comparison? That FM would be softer!? I am totally bemused as to why any setting let alone FM should even be used as a comparison. This is a totally different matrix.

    If you were that unsure that AN wasnt softer try Cluster Nebula v6 as its rate control and GOP structure is not too dissimilar. Ditto cbrandin 66 or DREWnet 12/15 v6 against AN 12/15.

    *** NEW EXPERIMENTAL SETTING ***

    The new AN 'Boom' Flat 4s setting is now up on page one intended for GOP1.

  • @Lambo @johnnym @botha @oscillian @retrospective @goanna @Mariella @DeShonDixon Thank you all, it really helps me to see that my hard work is liked !! :)

    @DeShonDixon Thank you for your interest in my research, I am tweaking the preset for now, when it is ready I'll send it to you. I think it will be a life changer preset, I just need to try making it perfect ^_^

    Here is a new video I have made with a couple of gopros and gh2s also with Cluster v6 'DREWnet' 12-15GOP (Try to guess how the lights where made)

    NOISE VS DYNAMIC RANGE / LATITUDE TEST

  • @Driftwood thanks for releasing the boom in AN Will definitely test it on a Zeiss 50mm 1.4 tomorrow when I have some time.

  • @mozes how did you do the slow motion shot in your third video at 1:36?

  • @driftwood @cbrandin @all

    bkmcwd Valkyrie 444 TYPE-ZERO1 'Apocalypse Now' with cbrandin 'Pure Original 444 Matrix' Released!

    Warning! : in this setting, all the modes of AVCHD other than 24p are experimental.

    Please keep in mind that artifacts often appears in HBR, 1080i and 720p in spite of being very stable.

    These will be improved from now on.

    http://www.personal-view.com/talks/discussion/comment/81033#Comment_81033

    @driftwood I congratulate you on your first edition of AN-Intra released! :-)

  • @lpowell I see in a general way what you mean about smooth/detail, etc, but the first thing I noticed was the difference in the jaggies. On my monitor, FM has straighter lines (going by your very helpful screen grabs) which for the material I shoot is critical; other ppl may have material that is less full of lines.

  • @bangkng twixtor.

    @Driftwood nice video, i did hold my finger on the hold button
    That was a wild party, and the up-tempo editing was just.... i really liked it.

    btw, did i see some people from the jackass tv-series?

  • @bkmcwd how does this new valkyrie 444 type zero1 differ from type zero? they both state to have original cbrandin matrix , it this one jus an even purer form ? what can we expect to see ?

This topic is closed.
← All Discussions