Personal View site logo
Driftwood Settings, Series 5: Cluster v6 ULTRA RELIABLE, Mysteron, Crossfire, Quantum 9B, Sedna...
  • 1029 Replies sorted by
  • @driftwood what we can expect from new ptool Ort modified quantum v9b ?? is it gonna be a quantum v9b with aq4 @ 176mbit???

  • I spent a lot of time with Cluster V2A and Flow Motion V2,1 last night and more of today. I need to work on it a lot more..... then do the same to Cluste 2e.

    Image wise in perfect conditions they both seem about the same, Flow Motion seems to have deeper colors for the same settings. Since I do nearly 100% run and gun, performance in various conditions are key factors. At night, Flow Motion seems a lot cleaner but I think it's because it's crushing the blacks. Today while grabbing a few quick shots under a shelter, there was a lot of noise with Flow Motion and there shouldn't have been.

    So..I need to sit down and strip some channels in order to figure out what's going on. The daytime footage of Cluster 2A in the shelter is very clean.

  • Awesome - cheers for the reply. I've followed and used your hacks keenly from SMBU, Orion and now Cluster. Just don't quite understand all the finer tech-speak about AQ Q20 etc etc. I'm flashing with v2f now! :)

  • v2f will just turbocharge v2e with its auto quant set to 4 (more to details) since GOP6 as a GOP is quite underpowered... Its an ages old aid developed by cbrandin to maintain low Q attempts on the fly.

  • Cluster v2f released (same as v2e but with AQ4 on). Sorry I'm dumb. What will AQ4 do for me? (Currently using v2e and quite impressed.

  • @driftwood, I think i'm going to print a t-shirt with the beginning of this page. Congratulations. Thank you for everybody's work. Your are really generous.

  • Please explain the "2K HD resolution 2048 x 1152 pixels - Highest quality MJPEG" mode! Is it 2K straight from the camera at MJPEG mode? Curious that's possible at GH2.

  • I don't think there would be any advantage to having a quantization value of less that 4, after factoring in the scaling tables. A quantization value of 4 produces 255 possible levels, which is the most an 8 bit codec can represent. Quantization values lower than 4 really only offer an advantage with 10-bit codecs.

  • What would the advantage of shooting Mjpeg 2K over Sedna A AQ1 24p?

  • @driftwood. Nick, I have always been a fan of Sedna A Q20 setting for 24P and I agree that nothing beats GOP1 for 24p. I have been using a Sedna A Q20-Sanity hybrid for a while but I like what I see in Cluster v2e for 1080i/HBR/720p. I was contemplating to do a Sedna A Q20-Cluster V2 and a Sedna A Q20-FM 2.01 hybrid which you call it Crossfire now. I think you can do a much better and thorough job of it than I can. When can I expect the Crossfire setting for Sedna-Clv2? I am willing to be an early adopter and test it.

  • @Sharpixels I was consistently getting 90MB/s with Cluster V2e. @driftwood Thanks for Cluster V2. Some initial tests look good. Works fine on a SanDisc Extreme 30MB/s SD card.

  • @Driftwood Thanks for organizing your settings/patches so neatly. That's a lot of patches and work! Thanks, very much!

  • For 24p Try the Quantum 9b which is constant and some of the other cbr settings.

  • Try v2e for PAL.

  • Thank you for your work! I like bitrate 100-50Mb/s at Cluster v2a - but required PAL version. Is v2a useless at PAL?

  • Thanks @peternap Try and test v2e too. Version 3 is in progress to try and balance the memory which gets allocated to P/B frames - which incidentally I have found the more you power them, the more the i frames drop... balancing the available memory gets everymore tricky... :-)

  • I took Cluster 2Va to the water today. I also switched to a 32 gig 45mb card (I'm not buying anymore 95mb'ers). Nearly everything was in 720-60 and it did fine....until I got to the killer. Wake crashing on rocks. In 720 it locked up tight giving a card speed error and requiring me to pull the battery. After a couple of times doing that, I switched to 24P and no other errors.

    In last nights test, I got a lot more noise than I expected so tonight I'm taking the same lens to the same place I tested flow motion v2.1, and see how it stacks up.

  • @Frame Shoot the same stuff against stock and try v2e too. You PAL land ?

  • Maybe that is a beginner´s question, but can GOPs be allocated to bitrate settings? Let´s say, 24H is Intra, whereas 24L switches to longer GOPs? But I guess the GH2 will not allow this. Made a short test beween Canis Night an Clusterv2a. With cluster, I see some mosquito-like noise (hope mosquito-like is correct, looks a bit like haloing or ghosting) on the edges of highly detailed opjects. But very, very subtle

  • @karl On a lot of the newer GOP1 settings the only way you're going to get crashes is if the sd card is either not good enough or fragmented. Seriously, youre missing out. :-)

    PS How did you get on with Cluster v2? That will understress most decent type 10 cards and still retain good quality at 24p, 30p and 60p... etc...

  • @karl Try SpanMyBitchUp v2b. I's Gop 1 and spans and the quality is great.

  • @driftwood: Regarding your statement "if you want the best, stick to GOP 1 on 24p": Is there any GOP1 setting that will not abort recordings or crash the camera while still delivering quality better than the best current GOP3 settings?

    I wonder because the only thing I do is documentary stuff, and unlike at a feature film set, I cannot ask the "actors" to do a re-take when something goes wrong. Also, when shooting underwater, unpowering the camera is simply not possible.

    Therefore, a GOP1 setting would be useful for me only if it was known not to abort or even crash.

  • GOP1 on 24p/HBR/1080i50, cluster style settings on the rest. New matrix.

  • @driftwood What can we expect from the crossfire settings?

  • The 'hack' is applied throughout. Settings for all modes.

This topic is closed.
← All Discussions