Personal View site logo
Make sure to join PV Telegram channel! Perfect to keep up with community on your smartphone.
Driftwood Settings, Series 5: Cluster v6 ULTRA RELIABLE, Mysteron, Crossfire, Quantum 9B, Sedna...
  • 1029 Replies sorted by
  • I'd say this answers my question about noise with Cluster 3. This was taken with a Canon 50mm Nifty 50 lens, The full moon was up and there was some shore lighting. Smooth -2, 24P.

  • Driftwood, would it be possible to create a high res high quality MJPEG at 4:3 so people with 2x anamorphic lenses can get 2.39:1. From what I understand cluster v3 creates 2K MJPEG but it's aspect ratio is 1.778. I'd love a high quality 4x3 MJPEG.

    What's the feasibility of such a thing?

    Thanks for all you do.

  • In MJPEG for Cluster v3 does anyone else have an issue with the GH2 not correctly interpreting remaining time on the card? I have the 32gb 95mbs card and it tells me I only have 2 minutes of footage. I can squeeze about 2 and half minutes before it stops recording and when it finishes writing to the card it tells me I have 2 minutes left.

    This is a great patch, this is the only problem from keeping me using it all the time.

  • There are only two patches worth looking at right now for HBR: Cluster V3 and FlowMotion V2.1 Both render movement and colours well at reasonable bit rate to storage ratio.

  • Sir Driftwood, could you please clarify what does it mean "5 seconds high detail, longer less detail". I am talking about the Mysteron Burst Mode specifically. Thank you.

  • @kazuo I'm not sure I'd completely go along with that Kazuo. There are a lot of unbelievable patches. Nick is a wizard.

    Many are specialty though and as far as all around use...I think I'd have to agree that those two and Sanity are the top choices.

  • Today I had an opportunity to try the "cluster v3" with my usual "moving tree leaves in the sun" setting (though some clouds blocked the sun every other minute).

    Alas, cluster v3 aborted recording not only in 720p60 after a few seconds (with camera requiring power-cycle, not just write abort), even recoding in 24H aborted in that scenery after only a few seconds. (Therefore, I cannot comment on the image quality of 720p60 or 24H.)

    HBR (30p) did not abort, but (at ~50MBit/s) doesn't look too well - see attached screen shot, where many of the moving leaves lack detail.

    (All tests done on a freshly formatted SanDisk 64GB 95MB/s.)

    cluster_v3_hbr_moving_leaves.jpg
    1920 x 1080 - 1M
  • @Karl Thanks Karl! That may save me a lot of testing time. The noise at night is very tight grain and light. Motion testing was scheduled for today. I have to have 720-60.

  • I have more of a general question as i have noticed this on other patches too (24p and HBR). You see a shift in gradients (?) when i move my nose to the edge of the frame, and back. What is causing this?

    Check out the clip on Vimeo (password: nose):

  • @gorgal It's a banding issue. (old problem in 8 bits pictures)

  • I see this a lot in my footage. Makes the material unusable. Unless i go black and white.

  • @karl Really? Can you provide a 2-3 sec recording in 24p and 720p (ie before it stops, stop it manually after 2 secs) on the same subject in your jpeg and get a copy to me to download. I need to analyse it. Thanks. Also experimenting with diifferent matrices for Cluster v3.1

  • @karl, @driftwood. That is strange because I have tested Cluster v3 using Sandisk 32GB 30mbs and demanding panasonic 14-140mm lens shooting 24H with success fully zooming to 140mm and good quality. As for 720p60, I have had success with SH mode as long as I don't zoom but H mode has been working well with good quality also. Nick, any news of Sedna GOP1-Cluster Crossfire soon?

  • Id tested on my forest of death at the weekend without problems. Im certainly going to be trying to verify what Karl has come up with.

  • I'm with Zaven13...except I recorded to a Transcend 32GB Compact Flash (CF) 400X Flash Card Model TS32GCF400. Tested 3200 ISO low light setting with 24P Cinema, 30P HBR an 24P Manual at 80%. Also tested daylight setting (the same as I always use with blowing leaves, zoom, waving flag) using 24P Cinema and 30P HBR, All clips were 1minute each. Surprised I had no write error issues as Cluster v2e was just too much for the Transend card.

  • Oh yeah..and the lens was a Panny 14-140 with OIS enabled and AF on. Low light clips were shot at 30 & 24 SS while daylight was 60 SS and ISO 160.

  • I sincerely hope that Karl isn't the German alter ego of Lee Powell! ;-)

    I still highly rate Quantum 9b 'updated' HBR - constant 80M and Intra.

  • @driftwood Death watch testing FM2 at 720p60 is more than enough self-abuse for me. No time for sock puppets here!

  • @driftwood: If I was LPowell's alter ego, would I have reported 720p60 SH recording issues in the FM2 thread? ;-)

    But seriously, I'm happy for everyone who does not experience any involuntary stops with anyones settings. I am just trying to find out which of the many settings is stable for me under all circumstances, and the "moving leaves in the sun" test is the one that has caused most issues for me so far.

    Among the high bit rate 24H settings that I have tested so far, only GOP3ZILLA, GOLGOP and FM2 were stable for me, and only the first two also survived all tests with 720p60. I'm sure some of the intentionally lower-bit-rate settings (like Sanity) would also be stable, but since I don't mind high capacity usage on the SD card, I try to find "maximum quality without stability risk" settings for use with SanDisk 64GB 95MB/s cards.

  • @lpowell LOL! I completely agree. Just waiting for Karl to offer some reports in all seriousness so we can improve things. Its all in good sport. dw... Speak of the devil. @Karl, now youre back could you send me those test files?

    Also Karl, are you suggesting none of the Intra 24p settings are stable?!

  • @driftwood: Regarding: "Can you provide a 2-3 sec recording in 24p and 720p (ie before it stops, stop it manually after 2 secs) on the same subject in your jpeg and get a copy to me to download." - I will remember to create such a test recording next time I can try (when sun, wind and free time converge at some noon). (BTW: Can you already name a place where I could upload those files? Getting one < 10MB for uploading here could be difficult...)

    Regarding use of different SD-cards (like "30MB/s" vs. "95 MB/s"): I get a feeling that the influence of those different card types is not as large as we may think. I use a much slower card (actually an older 8GB microSDHC class 6 card in an SD adapter) to transfer the firmware .bin files to the GH2. It happened a few times to me that I at first forgot to exchange this card with the 64GB 95MB/s SDXC card before starting the test-shot, and curiously, the recording on this smaller, inferiour card never yielded a different result regarding stability then after I noticed my mistake and changed for the fast SDXC card.

    Also, if you think about the bit rate of 720p60 recordings, they are relatively low, so it is not very plausible that recording stability depends on the card performance in that mode. I've got the feeling that the CPU power of the camera and maybe some memory buffers it requires is the more relevant factor for recording stability.

  • @driftwood: Am I suggesting none of the Intra 24p settings are stable? So far I tried the 24H settings of Mysteron, Quantum 9b and Canis Majoris, and indeed, they all made my GH2 crash sooner or later. (MJPEG recording was stable, though, even at high bit rates, but does not look as good as 24H.)

    Am I assuming correctly that "span my bitch up" or the unified EOSHD settings would be the 24p Intra settings with the least stability risk? (Then maybe I should try them once, just to get a chance to rate their IQ against the GOP3 settings I've worked with so far.)

  • @karl Dropbox me. See my email address on personal-view on my name - use that to invite me. Dropbox is free for 2gbs. The Intra 24p are top notch - Im not suggesting that they will span forever like some people need, but if youre filming less lengthy stuff like youve put up so far you'll have plenty of time to adapt (and adjust) to the excellent recordings of the Intra setings. Give 'em a whirl again, go on, be a devil!

    And please attempt to describe the sort of crash you experience when using them.

  • BY the way Quantum 9b followers - here's the details of the 'The Revenge of the Great Camera Shootout 2012' in their pdf for those technically minded.

    http://www.zacuto.com/zvideo/RGCSO-TechnicalCameraDocumentation-V1.pdf

    I've been invited to the forthcoming London event, I'll keep you posted as to where its going to be, and it will be a great chance for us fellow GH2 users to meetup and have a chat.

  • @karl While I can't speak for Driftwood's patches, my experience with Flow Motion v2 and a variety of Class 10 SD cards rated at 30MB/sec, 45MB/sec, and 95MB/sec is that the rated speed only has an effect on 4GB file-spanning reliability. In every case where I've observed FM2 recording failures on 30MB/sec SD cards, I saw the same failures occur on 45MB/sec and 95MB/sec cards as well. OTOH, the only cards that I have verified 24H and HBR file-spanning on are the 95MB/sec SD cards.

This topic is closed.
← All Discussions