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.
Official Low GOP topic, series 2
  • 1022 Replies sorted by
  • @elleyyh Here's a tip. Once flashed, switch off, replace with either a different sd card to record on or delete the .bin firmware file off the card (put it back in your pute and delete it!) and then put it back in camera & record a few seconds recording, playit back once to check, then you're good to go. Playback should work from then on.
  • @driftwood "My own tests confirrm 'reAQuainted' is making best use of quality macroblock QP. But I'll be happy to be prroved wrong. :-)"

    Somehow I feel you're going to be right on this ;) thanks man! I'll check it out soon
  • Thanks guys. I did format the card before recording. It says I should get 25 minutes on the 32 gig card, which seams right but it will stop every six minutes to give me 4.2GB files. Any ideas on how to get the files to span? Thanks all. Since joining this board, I feel like I have a whole new camera.
  • I ran some tests today and you couldn't see any difference until you pushed the codec to demand more bit rate. Shaking a plant didn't do it. Playing with little beads showed some obvious clarity difference but the real magic came when I filmed blowing leaves in the trees on this blustery day. The codec in the stock cam began to fall apart and started throwing out data.  Leaves became muddy, the dark woods behind them got blocky and the sky showed gradient lines. The normal stuff we're used to. The hacked version held it all together and showed none of this when the action happened. No muddy leaves, smooth sky, no blocky dark areas. So, I guess that's it. The codec has to be pushed, which probably wont happen at the getting ready or ceremony but will happen at the dark reception. 

    It wouldn't span files. So, it recorded in 4.2GB chunks and then stopped. No crashes, just stopped at that point. I have to figure this out. But it is stable on the class 10 card. If I can get files to span then I will get the fast cards and go for it. 32GB gets me 24minutes. So, over a gig a minute. HUGE files. Vectorscopes are showing more color range and better graduation. It feels like a nice upgrade to your gear.  Just have to get it stable. 

    Clean is green my friend.  Well, it can be. Soon we will all have clean perfect footage and then what the heck are we going to do?  :0)

    Peace. 
  • @mclarenf1
    forget spanning at these rates and AQ3.

    drifwood has set out for what many have been waiting for - better AQ settings optimized for this high bitrate.
    I'm sure he spent a lot time optimizing i-frame size / Q settings / buffers / SD write performance
    what else could we ask for... well, maybe a few brews, women, etc..... !
    ;)
  • Quite gobsmacked over the amount of emails and gratitude regarding INTRA on the GH2 we're getting here at Driftwood! It seems the story has gone global!!! I hope everyone has made a donation/s to Vitaliy so that we can all enjoy the fruits of his labour now and in the future. Without his endeavours we wouldn't be here now.
    Think of the possibilities he could do to other worthy cameras if everyone helped him with finance so that he could employ a crack team of debuggers/programmers. We need a money raising scheme.
  • @driftwood - What would you think about trying the new AQ3 at AQ4 for lowlight, low-ish ISO shots.

    That flood of emails is a sliver of the gratitude we feel for your relentless testing, man.
  • @driftwood you didnt see the 43rumors about you and my test videos? BOTH OF THEM!!!
  • Hi Driftwood and thanks for your hard work, I was wondering I have no problem using QuantmeBaby setting, but when I use 'reAQuainted' file it stop recording at 11 second, is it because my Class 10 card is too slow ? Even through they are the same meg at 176. Thanks again
  • @driftwood cau you make setting to trancend 32gb class 10
    low setting but GREAT QUALITY :D
    stable for all setting
  • @driftwood , can you come help me move my furniture to my new apartment on Sat?
  • @driftwood, white, no sugar, and don't forget the fucking custard creams.
  • LOL, you guys... I'M the one who needs fuckin help round here! The Driftwood team are filming Nuit Blanc Festival Brighton 2011 next Saturday and I need to find the best SDI to X? to laptop to stream live footage to a webcast. So tell me what X? - is it AJA, blackmagic or what? It NEEDS to connect to a mac pro laptop or sony vaio.
    Use this thread: http://www.personal-view.com/talks/discussion/1207/gh2-live-streaming-solutions#Item_31
  • @driftwood
    Thanks for great precise work! :-)

    @someone who needs this
    This is not my usual extreme patch but a safe setting.
    Even if an extreme stress starts, a flat line is almostly not made in cadence of Streamparser.

    The first screen shot is a result of the test done moving and zooming with Stray's chart.
    The second is a result of testing moving with test chart.
    The third is the result of testing moving at 24H80%(30p).
    Although the size of the I frame exceeds 1M in a bright place, under lowlight, it becomes the size of the same grade as Chris's 44M.
    Since the weather is bad today, an outdoor test is not ready yet.
    Since I think lowlight as important, I use the former NINJA patch, but probably, this may be sufficient if a bright place is main.


    It can play in camera satisfactorily.
    Of course, there is cadence issue.
    It is appropriation from Chris's 44M except 24p.

    NOTE:
    It is not what is called cadence issue that has appeared in this first chart, and it is only based on a motion of my camera.
    If camera moves with zooming from a top to the bottom on a chart, while a terrible stress will start, when moving upwards from the bottom, stress does not start so much.
    It only seems that cadence has occurred according to the difference of the strength of the stress.
    When a fixed stress continues being applied, cadence does not occur in this setting as like the second chart.
    All tests are with ISO3200 SS1/800.

    ADDITION: The result of 30p(24H80%) outdoor in 4th screen shot
    setf.zip
    556B
    176mq18fb2.7fl1.05tb6.6vb3.6.JPG
    1297 x 634 - 187K
    176mq18fb2.7fl1.05tb6.6vb3.6_4.JPG
    1295 x 632 - 189K
    176mq18fb2.7fl1.05tb6.6vb3.6_30p.JPG
    1296 x 633 - 199K
    bksleaves_30p2.JPG
    1295 x 632 - 200K
  • @bkmcwd Nice work this is very close but you need to sort cadence issue. Remember, with the Driftwood INTRA patches average frame size defeats virtually any other patch. However, this for top quality i frames is looking good. Get Ralph_B to check for smudging at odd frames too.
  • @driftwood
    Although there is no telling whether I am well told to you in my English, I understand your opinion. :-)
    It is not what is called cadence issue that has appeared in this first chart, and it is only based on a motion of my camera.
    If camera moves with zooming from a top to the bottom on a chart, while a terrible stress will start, when moving upwards from the bottom, stress does not start so much.
    It only seems that cadence has occurred according to the difference of the strength of the stress.
    When a fixed stress continues being applied, cadence does not occur in this setting as like the second chart.
    Thank you for always pointing out the noticed place!
  • Some more testing w/ reAQuainted.

    I tested Stray's chart:
    - Transcend Class 10 SDHC 16GB - stopped around 14sec.
    - Transcend SDXC 64GB - stopped around 20 sec.
    - SanDisk Extreme Class 10 SDHC 32BG works fine. It had been stopped by the 4GB limit of file-size.

    I tested low light shooting:
    - w/ ISO1600&3200 it seems a bit noisy but it looks like film grains, there could not be found any artifacts (macro blocking, I meant) in dark area, noise and gradient are much better than former patch(v3).
    - w/ ISO800 I just like to say the very word of 'awesome'.


    I think this patch is really nice if you could have got Sandisk Extreme SD card. (I haven't got 'ExtremePro' and not sure how it works).

    But 4GB file could capture only around 4-5min, and it seems the battery was quickly dried out.

    Thanks again @driftwood :)
    GOP1_reAQuainted_ISO1800.png
    1920 x 1080 - 5M
  • @hiroshisaito - "noise and gradient are much better than former patch(v3)." do you mean better than the gopstoppa patch? Thanks in advance.
  • @cosimo_bullo I mean Driftwood 176M Intra GOP1 ver.3 QuantMeBaby.
  • @driftwood - Are you being coy, man, or do you really not know how good the new AQ3 patch is?

    It's amazing. In low light, it's now passing Henry's admittedly flawed patch in bitrate. Totally reliable so far. We just tested gopstoppa, this new AQ3, Henry's, the old 220 gop1 at 160, 400, 800, and 1600. The AQ3 seems to win at all ISOs. Just got the stuff into the computer so will mess around more tonight if time allows.. but MAN!!! Thank you again is all I can come up with. This new one is a gift, really the filmmaker's patch.
  • IDK maybe its me but I dont see much difference in the new patch exept its smoother... less pixelation and higher bitrate in low light... both of these shots were done at the same settings. Quantmebaby had a bit rate of 66mbit/s and reAQuainted had a bitrate of 123mbit/s
    Test.jpg
    1920 x 2160 - 1M
  • Great example Perry. You see it in the grain in the bokeh. The way I see it is that these are now both great choices. If you want to sip a bit more with your cards, use quantmebaby, if you're going for max resolution use reAQuanted. --- unless driftwood is able to unlock more down the road. I assume Panasonic or Nikon will be begging him to come work for them sooner or later (this assumes they're paying ANY attention to what excited the pro/semipro customer).
  • Oh they are paying attention you didn't seenthe post about them supporting the hacking of their cameras they love ustheae guys... Watch out for my cage im starting the kickstarter this weekend t everyone that like the rewo cage won't want to miss this 280$ replica im making
  • Wow that English sucked ass sorry guys im at world of beer celebrating my baby with my brother
  • @driftwood, I was in the process of uploading your latest AQ patch video test sample to Youtube, but I was surprised to see lines. The lens was an Oly 45mm wide open @ iso 3200. Lighting was day light from the window. I didn't have a light meter with me at that time. I think I will stick with your last version. Any suggestions appreciated. thanks.
This topic is closed.
← All Discussions